My favorite course in law school wasn’t with a typical law professor. Food and Drug Law was taught by the former General Counsel of the FDA. He had a volcanic temper that we assumed was the byproduct of years of dealing with bureaucrats. In class, he abided neither tardiness nor squeamishness; the first topic was food sanitation and its flip side: adulteration. We learned about butter with chicken feathers and baby food chock full o’ cockroach fragments. We waltzed through classic cases like United States v. 184 Barrels Dried Whole Eggs, and United States vs. 915 Cartons of Frogs Legs. We covered official FDA filth tolerance levels for foods such as frozen broccoli (no more than 60 or more aphids and/or thrips and/or mites per 100 grams), pasta (fewer than 225 insect fragments per 225 grams), tomato juice (fewer than 10 fly eggs – or 5 fly eggs + 1 maggot – per 100 grams), canned mushrooms (fewer than 20 maggots of any size per 100 grams), and peanut butter (fewer than 30 insect fragments or 1 rodent hair per 100 grams). Tolerance levels are now available online for your perusal (avoid browsing while eating).

FDA sets tolerance levels based on Quantitative Risk Assessment (QRA) – a basic cost-benefit analysis. In the case of insect parts, QRA yields tolerance levels that are safe and “below which the defect is unavoidable under current technology.” And while insect parts in food may be disgusting, QRA gets trickier for known carcinogens. Carcinogen tolerance levels aren’t set according to average per capita consumption – food companies have made this argument in court and lost – but rather based on a level at which there is “no significant human risk.” In practice, what this means is a multi-stage risk assessment in which every assumption is conservative. As there may be as many as 20 distinct stages in a risk assessment, when FDA calculates a cancer risk rate of one in a million, that risk is only this high for, say, a child consuming the maximum possible amount of the food over an extended period. The actual risk for the average consumer will be much lower.

Which begs the question: if managing risk based on averages is unacceptable in food safety, why is it acceptable in higher education?

***

It seems like every week there’s a new article on the value of a college degree, always comparing the income of college graduates to students who never entered or completed college. I have three issues with this rash of reports. First, they never reference the obvious self-selection bias: that students with the talent, persistence and – probably most important – family background, wealth and support to complete a college degree are on track to earn much more in the workforce without regard to their college experience or credential. Second, these studies report on graduates who’ve been out of school for at least a decade, which means they graduated before the Great Recession and entered the workforce before an era of 50%+ unemployment + underemployment for college grads.

But third is the fact that the “college earnings premium” is only an average. Which means that there are plenty of college graduates whose experience is less than premium.

We know this because they’ve been making a lot of noise lately. They’re the degree holders unable to move out of their parents’ homes or start businesses. They’re making sub-optimal job decisions based on pressure to earn income in order to stay current on repayment. They’re the graduates seeking student loan forgiveness and raging about how student loan debt has become a cancer on the Millennial generation. They’re the force behind the “free college” movement that seems to have become Democratic Party orthodoxy (to the point that leading candidates for the 2020 nomination have felt compelled to announce programs). And only 38% of them strongly agree that “college was worth the price.”

Think about how student loan debt is working for the (average) student who graduates with $35,000 in student loan debt and owes about $400 every month. As of 2016, only 57% of the 22 million Americans with federal student loans were current on their payments; 43% were in default (3.6M), delinquent (3.0M) or in forbearance (3.0M). As if this weren’t bad enough, last month we learned that, due to a Department of Education coding error, actual repayment rates could be as much as 20% lower. Only 41% are paying as much as $1 toward their principal loan balance in the first three years after leaving school. Which means we have untold millions heading for wage garnishments with no hope of discharge through bankruptcy. Two years ago, the Federal Reserve Bank of New York recognized that the bottom 25% of bachelor’s degree holders are incurring debt, but earning no more high school graduates. Imagine if 25% of consumers were getting sick from our food supply. It’s now crystal clear that current higher education financing practices would fail an FDA Quantitative Risk Assessment.

It’s not only the FDA that has rejected risk management by averages. We’ve abandoned it in virtually every other major area of risk. Cars aren’t allowed on the road if they only protect the “average” driver or passenger in the event of a crash. Doctors don’t win malpractice suits by claiming that their treatment would have saved the “average” patient. Even auditors that allow companies to cook the books aren’t permitted to remain in business because their work would have been sufficient for the “average” client.

Rather than talking about college in terms of average outcomes, we need a new approach to managing student loan risk – one that makes conservative assumptions at each stage of a multi-stage risk assessment and protects not only the average student, but exactly those students who experience the worst possible outcomes.

In propagating income-driven repayment (IDR), the Obama Administration recognized this need, as well as the federal government’s unique ability to address this risk. (Unlike food or cars or accountants, the federal government foots the higher education bill.) Connecting student loan repayment to income level and permitting students to extend payments from 10 to 20 or 25 years has significantly reduced the risk for students who experience the worst outcomes.

But IDR has three major limitations. First, it doesn’t cover all student loans. Parent PLUS loans and private loans (representing at least 20% of total loan volume) aren’t eligible for IDR. Second, as with every government subsidy, there's a direct relationship between size of program and cost. IDR currently covers about one quarter of all borrowers and is estimated to cost at least $108B. With 25% of all borrowers enrolled and ~60% not repaying, one might conclude that, with IDR, we’re already 15% of the way down the road towards free college. The likelihood of expanding IDR – certainly during this Administration – is low.

Third and most important, we need to think about who is best positioned to reduce or eliminate the risks of student debt – or, returning to law school, who is the “Least Cost Avoider”? That would be colleges and universities themselves. And IDR does nothing to incentivize them to reduce student risk or change their behavior in any way. It’s as if the U.S. Department of Transportation has vowed to cover all costs associated with car accidents but allowed GM, Ford and Chrysler to continue to produce automobiles that are Unsafe at Any Speed.

***

If the FDA were put in charge of the Department of Education, every college and university would offer Income Share Agreements (ISAs). Rather than getting paid upfront regardless of student outcomes, ISAs are income-linked repayment contracts in which colleges and universities themselves front funds for student tuition and fees – either from their balance sheets, operating budgets, or external sources. Then schools (or their funders) receive repayment following graduation as a percentage of student income (always time- and dollar-limited).

ISAs, such as Purdue University’s Back a Boiler program, do three important things. First, students who would have experienced the worst outcomes with student loans now experience average outcomes because risk is shared across a pool of borrowers. Second, ISAs have great potential to replace the high-cost student loans – PLUS and private loans – that aren’t covered by IDR. And third, the availability and pricing of ISAs will send invaluable signals to students regarding likely return on investment at the level of program, credential and institution. In time, as ISAs become the standard “last mile” form of higher education financing (after students have maxed out federal IDR-eligible direct loans), these signals will change enrollment behavior – causing students to flock to high ROI programs, and shun low ROI programs – at both undergraduate and graduate levels. Low ROI programs will either need to be changed or shuttered.

So it’s heartening to see momentum building for ISAs. Two weeks ago, the American Enterprise Institute released findings from a survey of 400 college and high school students and 400 parents of future students to ascertain interest in ISAs. Although awareness of ISAs was limited at the outset, once respondents were educated about repayment scenarios, 53% preferred ISAs vs. only 29% who preferred student loans. Then last week, Senators Marco Rubio (R-FL) and Todd Young (R-IN) re-introduced a bill that would clarify the legal status of ISAs. It wouldn’t surprise me in the slightest if this Administration rolled out a new rule – call it a 10/90 rule – requiring colleges and universities to fund $0.10 in ISAs for every $0.90 they draw down in subsidized + unsubsidized Stafford loans.

While IDR has improved student loan risk management, higher education has a ways to go before it could be considered “FDA approved.” ISAs have the potential to bring higher education in line with how we manage risk in every other area of our lives. Over the next few years, the combination of IDR + ISAs will clear out the cockroach fragments from the nutritious baby food of higher education.

University Ventures is reimagining the future of higher education and creating new pathways from education to employment. Our portfolio companies are making higher education more affordable, pioneering entirely new approaches to learning, and helping employers think differently about how and where they discover talent. We believe entrepreneurs and institutions that produce meaningful educational outcomes will realize better economic returns. We believe that private enterprises can produce profound social impact, and that public-private partnerships can unlock the potential of mission-driven institutions. Our commitment extends beyond the world of traditional higher education, to the challenge of last-mile training and placement to help individuals showcase their skills and get great jobs in growing sectors of the economy. University Ventures is led by principals with decades of experience as entrepreneurs, investors, and leaders in higher education. Our approach draws upon the values and traditions of higher education to play a sustainable role in transforming the path from education to a stronger economic future for students, universities, and employers.