It has become a national embarrassment — and tragedy. Millions of Americans go deeply into debt to pay for higher education, only to be disappointed in the job market and left to struggle with loan repayment. Of student borrowers who left college in 2009, more than one in eight had defaulted by 2011. All told, about 12 percent of the $604 billion of direct loans made and held by the Department of Education are either delinquent or in default in 2018. To put it another way: a larger share of student borrowers default each year than home borrowers did at the peak of the recent financial crisis.

The consequences ripple through the economy and society. Borrowers who struggle with their student debt may be unable to buy a home or start a business. And since student debts cannot be discharged in bankruptcy, default can result in garnishment of wages and seizures of tax refunds.

Sometimes, loan hardship is the result of bad luck, like catastrophic illness or losing a job with a firm that’s downsizing. But, all too often, disappointing student outcomes aren’t the result of random misfortune. Poor outcomes are systematically concentrated among former students at institutions where few borrowers complete their degrees or where the completed degrees don’t lead to good jobs.

Students who attend these institutions, regardless of whether they are from higher- or lower-income families, struggle to succeed in the labor market and to repay their loans.

There’s solid evidence that institutional quality, not just students’ background, drives the economic success of borrowers. That’s good news, since it means we could improve outcomes and the pace of loan repayment by improving the quality of schools that borrowers attend. The bad news is that the student loan program does little toward this end. Indeed, failure is a predictable consequence of incentives created by the way Washington subsidizes educational programs with little regard to their quality.

Thinking About a Fix

The most important way that the U.S. Department of Education can affect student outcomes and reduce defaults is through accountability measures that limit the institutions allowed to offer federal loans and grants. This conclusion is based on experience as well as common sense: over the history of the loan program, default rates have surged and declined repeatedly as the federal government relaxed and tightened its standards for which institutions could participate.

Current federal accountability systems are ineffective for promoting quality because they use only the bluntest of tools, and the penalties and incentives affect only a tiny sliver of institutions. For example, the cohort default rate rule bars institutions where more than 30 percent of borrowers default within three years. In practice, though, the rule is rarely invoked, both because the threshold is set so high that it affects only a handful of institutions and because institutions have figured out ways to help students postpone default without making progress in paying down their loans.

The key, then, to improving outcomes for student borrowers is to toughen institutional accountability. An ideal system would align the incentives of schools with those of policymakers and students. One promising approach — and the focus of this analysis — is risk sharing, in which institutions bear a portion of the financial consequences that students (and taxpayers) face from poor outcomes. Risk-sharing approaches differ in details, but all seek to link the financial success of schools to the long-term financial success of their former students.

In broad terms, risk-sharing proposals identify socially valuable outcomes — as measured by degree completion, loan repayment or post-college employment — and set targets for schools. If an institution’s students fall below target, financial penalties proportional to the failure would apply.

Some plans would utilize carrots as well as sticks: revenues collected from failing schools would be used to finance bonuses for institutions that exceeded the target. For instance, the funds could be used to provide extra grant support to schools that have a superior record of outcomes for low-income students.

Origins of the Mess

We get ahead of ourselves. First, some background on the origins of the student loan mess. The federal student loan program has played a central role in financing higher education, especially education for low- and middle-income families, for some time. Today, most people who borrow to attend college do so through the Direct Loan program administered by the Department of Education. The department offers loans to cover the full cost of attendance (including living expenses) at almost any accredited institution, up to annual and aggregate caps. For example, the total borrowing for an undergraduate degree is capped at $31,000 for a dependent student. The program is meant to be largely self-financing, but some loans for low-income students are subsidized.

As of the spring of 2018, federal student debt outstanding was a whopping $1.4 trillion, almost triple the figure in 2007. For the group of students leaving school in 2009 — the cohort for which we base our analysis below — approximately 3.8 million student-borrowers reached the repayment phase with $57.4 billion in aggregate initial loan balances, which amounted to $15,100 per borrower.

In themselves, those numbers aren’t evidence of program failure. On average, college remains a good investment, and many high school grads would be unable to attend in the absence of federal loans. However, with funds available for use at virtually every school, students are free to make uninformed and even bad choices — emboldened with the tacit approval that comes when the federal government is willing to provide the funds. And, unlike in other markets where consumers shell out cash upfront, students may have less incentive to seek out quality and value because federal aid covers much of the immediate cost.

Schools, on the other hand, often have better information than students do about the likely economic return to a particular program for a given student. Institutions also control the quality of advising and job placement services, which can make a big difference in matching a graduate with a good job.

Hence, schools often face a conflict of interest. They benefit from encouraging cash-strapped applicants to borrow and enroll, even when the expected returns are low and the risks of personal financial debacle are high. Of course, it is not as simple as a car dealer eager to sell vehicles to anyone who can borrow the cost. Schools presumably weigh the impact of enrolling student-borrowers who are likely to fail on their reputations. But misaligned incentives and asymmetry of information can lead students to make poor choices, especially when institutions rely more on marketing than reputation to attract students.

Indeed, some institutions stand out on the dishonor roll, exhibiting consistently poor outcomes for their student-borrowers. Consider borrowers who graduated or dropped out of schools in 2009. Under the standard 10-year loan contract, they are expected to have paid off roughly 40 percent of what they borrowed. But at the 255 schools with the most problematic outcomes, ex-students collectively owed more after five years than they owed when they left.

To be sure, 2009 was an especially rough year to enter the labor force. But there are hundreds of institutions for which poor loan repayment is a common and predictable outcome even in recent, more prosperous years.

The figure below shows the distribution of repayment rates by institution for undergraduate loans five years after entering the repayment phase in 2009. The typical school had a cohort repayment rate of 22 percent, meaning that their undergraduate-borrowers in aggregate had repaid 22 percent of the initial principal after five years. This repayment rate is below the rate of amortization schedule on a 10-year loan, implying that some debtors were behind on payments. But it does correspond to amortization over 15 years — not so bad for borrowers who struggled to make a living in the headwinds of the Great Recession.

The bars on the figure are color-coded to reflect the repayment term associated with that five-year repayment rate. Five percent of institutions had cohort repayment rates below zero — signified by blue — meaning that the debtors collectively actually owed more after five years than they did when leaving school. These bars, combined with those in dark purple, show that 32 percent of institutions have repayment rates below 15 percent, meaning that their borrowers are paying down less in aggregate than would be expected on a 20-year repayment schedule. Half of the bars are green, implying that these institutions have repayment rates of 20 percent or higher, which would be consistent with a repayment term of 15 years or less.

While the repayment rate provides only one measure of education-loan outcomes, it turns out to be an excellent predictor of a wide variety of other important outcomes, including student loan default, post-college earnings, and the upward earnings mobility of low-income students. Hence, repayment serves as a rough-and-ready summary of the success of the student loan program — including the likelihood that taxpayers will recoup their upfront investment.

As shown in the figure below, there are some systematic patterns in the sorts of schools that are more likely or less likely to have low repayment rates. In particular, less-than-four-year institutions are generally behind four-year institutions. Note, though, that a substantial number of less-than-four-year schools have reasonably high repayment rates, suggesting that there are good options out there for students seeking associates degrees or certificates.

It is also the case that for-profit schools are overrepresented among institutions with low repayment rates. Almost no students at schools in the top half of repayment rate performance attended an institution in the for-profit sector, but a clear majority of those attending schools in the bottom decile of repayment rates did attend for-profit institutions. The disappointing loan performance of for-profit schools is part of the reason that these schools have periodically been targeted for additional scrutiny and regulation.

Not surprisingly, loan repayment problems are higher for students from low-income families. Low-income students rely more on loans to finance their educations, disproportionately attend institutions with poor repayment rates (for all income groups), and are less able to rely on their families for help with loan repayment when those institutions fail to deliver an education that yields good jobs. Disadvantaged students consequently face the most significant hardship as a result of inadequate quality control.

It is certainly the case that the loan program is critical to promoting educational opportunity, and any attempt at reform should be careful to preserve access to quality education for low-income students. At the same time, however, some institutions consistently ask students to take on more debt than they are likely to be able to repay, harming both taxpayers and the students themselves.

Accountability and Insurance

Over the past decades, policymakers have taken a two-pronged approach to address the problem of poor loan outcomes. One is to give borrowers more time to repay their loans (with interest) to minimize defaults arising from temporary hardships, like a spell of unemployment. Borrowers may go into deferment or forbearance on their loans, both of which halt loan payments temporarily.

To the same end, the loan program offers an income-driven repayment (IDR) option, in which monthly payments are tied to earnings. Borrowers in IDR programs pay a fixed percent of discretionary earnings — often 10 percent if that amount is less than the 10-year standard payment would be — and any remaining balance is forgiven after 20 or 25 years. With the IDR programs students may pay more in total interest because their loans are amortized more slowly, but they are protected against the downside risk of a poor education investment — or, for that matter, any misfortune that makes them poor earners.

A downside to IDR plans is what economists call “moral hazard”: income-driven repayment weakens incentives for students to shop around for programs with lower costs or better economic value, or for institutions to make a real effort to improve students’ financial outcomes. In addition, from an accountability point of view, IDR masks poor student labor market outcomes without fixing them. By design, students enrolled in IDR programs rarely default, even when their loan burdens exceed their ability to pay. Thus, while IDR partially protects students from hardship, it worsens the institutional incentive problem.

The other approach used to address poor student loan outcomes is to strengthen the accountability systems that govern the institutions that can participate in the federal loan program and under what circumstances. After sky-high default rates among student borrowers in the late 1980s, Congress:

• enacted new rules restricting the share of total revenues that for-profit schools could obtain from federal grants and loans to a still remarkably high 85 percent (then called the 85/15 rule)
• imposed “cohort default rate” targets to exclude institutions with very high student default rates from the loan program
• restricted loans for distance-learning programs
• prohibited abusive student-recruiting practices
• required mandatory garnishment of the wages of defaulted student loan borrowers.

However, the bite of these accountability measures has dulled over time. For instance, the original 85/15 rules, which required that at least 15 percent of an institution’s financing come from outside the federal aid system was revised to 90/10. The distance-learning limits were repealed in 2005, fueling an explosion in exclusively online programs. “Gainful employment” regulations proposed by the Obama administration would address poor earnings outcomes of students in certain types of programs. But these would apply only to schools with extremely poor outcomes and are unlikely to be adopted under the Trump administration. And the cohort default rate rules have become less effective as institutions have become adept at helping students enroll in deferment, forbearance or IDR plans, delaying but not necessarily reducing the incidence of poor outcomes. Indeed, there are now consultants who specialize in helping schools “manage” their default rates, allowing institutions to escape sanctions without addressing the underlying problem of the inability of former students to repay their loans.

As a result, today’s accountability system provides little protection for taxpayer investment in higher education. Nor does it funnel federal loans toward higher-quality programs. Since a substantial fraction of loaned funds at low repayment rate institutions will never be recovered, the student loan program effectively subsidizes the least successful schools. Even without concern for low-income borrowers, responsible stewardship of federal dollars requires that subsidies be targeted to institutions that do well by their students.

Improving Accountability

Though college accreditation has traditionally emphasized inputs — everything from the degrees of the faculty to the size of classes — there is growing interest in student outcomes as a gauge of institutional success. This approach is more feasible with modern data and computing power. Indeed, the Department of Education now produces the College Scorecard, an online tool that provides school-by-school information about student outcomes, including college completion rates, average student loan burdens and repayment rates, employment rates and earnings.

Robust measures of student outcomes could also be used in accountability systems to promote good loan outcomes. Risk-sharing programs focus on exactly that — giving schools more skin in the game by requiring them to bear a portion of the taxpayers’ loss when students fail to repay their loans.

The idea hinges on the notion that loan outcomes are related to educational quality, and that quality can be improved. Ideally, institutions would respond by better matching students with programs, enhancing instructional quality, promoting degree completion, strengthening loan counseling and working to place students in better jobs after leaving school. Institutions that are already doing these things well would be able to keep a larger share of their federal loan dollars.

However, institutions could also respond in undesirable ways. They could reduce access to education for low-income students (who are more likely to struggle after college than their higher-income peers) or pass through the cost of risk sharing in the form of tuition increases or try to game their accountability metrics with potentially negative consequences for institutions serving low-income and minority populations. Hence, it is important to design any risk-sharing scheme to minimize unintended consequences and to reward institutions that serve low-income or underserved populations with effective programs.

We crafted a risk-sharing proposal with these concerns in mind. The plan improves on existing accountability measures by applying to a broader set of schools (where the scope is defined by loan outcomes rather than predetermined school characteristics), and by having a continuum of penalties ranging from small to significant depending on performance. In addition, we proposed to recycle budget savings from the risk-sharing program to fund a “mobility bonus” system to reward institutions that serve low-income students well.

We selected the cohort repayment rate — the fraction of a cohort’s initial principal that is repaid within five years after leaving school — as the core performance standard. Loan outcomes at the five-year mark are highly predictive of long-run outcomes. Specifically, the five-year repayment rate is highly correlated with student earnings outcomes, institutional quality and the return on federal loan dollars. The cohort repayment rate also has the advantage that is not based on defaults, which we believe are an increasingly unreliable measure of institutional success because of countermeasures taken by the schools and by the growing prevalence of IDR. It is also more difficult for institutions to manipulate than simpler yes/no binary measures that are used in the College Scorecard and elsewhere.

When institutions fail to achieve the performance standard, sanctions would apply. Our proposal would require institutions to reimburse taxpayers for a portion of the amount by which their students fail to achieve a minimum threshold for cohort loan repayment. Specifically, we propose five-year repayment thresholds of 20 percent — a rate of repayment consistent with borrowers repaying their loans over a 15-year period.

Based on the performance of the students entering repayment in 2009, about half of the institutions would incur a penalty under this framework. Penalties would be small for most of the institutions affected, serving mostly as a wake-up call to improve their students’ loan outcomes. But for the schools with exceptionally poor repayment rates, penalties could be significant.

The revenues raised from penalties could be used to finance bonuses to institutions with strong labor market outcomes among their low-income students, setting aside their repayment rates. This element is important because some schools may enroll particularly disadvantaged students who struggle to repay even when they have completed a valuable education.

For instance, our mobility bonus proposal would give institutions a fixed-dollar bonus payment for every low-income student in the undergraduate borrowing group who is making timely payments five years after entering the repayment phase. Bonuses would be given to the institution for each low-income borrower who earns $25,000 (roughly the median earnings of high school graduates) five years after entering repayment. The bonus payment structure would be an important way to mitigate the threat that risk sharing would cause institutions to reduce access to low-income students.

A Way Forward

Risk sharing and related ideas to increase accountability are gaining attention in high places.

For instance, the House Education and Workforce Committee’s Prosper Act would require that institutions have at least 45 percent of their students be in a positive repayment status to remain eligible for federal funding. A Senate Health, Education, Labor and Pensions majority staff white paper proposes that colleges with low repayment rates pay back the federal government, and that the government use the revenues collected to reward institutions with high rates of repayment among low-income students.

And though the Trump administration seems intent on rolling back accountability rules proposed by President Obama, the first Trump budget proposal does endorse the idea of evaluating colleges based on the rate at which their students repay their federal loans.

Indeed, in an era of fierce partisanship, few ideas seem as well placed in ideological terms to bridge the gap between progressive and conservative camps. No one supports a loan program that runs a deficit without delivering benefits to students inching their way up the mobility ladder. And no one wishes to burden millions of low-income families with the bill for worthless education.

The hard part will be to build a coalition behind effective legislation in a Washington besieged by lobbyists and distracted by tribalism.