The U.S. is about to conduct a perilous economic experiment: In September, some 45 million borrowers will have to restart payments on student loans after three and a half years on hold. Whatever the consequences, one must hope they’ll inspire a deeper rethink of the way the country finances higher education.

The outlook isn’t good. Judging from data on their non-student debt, borrowers are already struggling. The median cost of servicing that other debt has increased 24% since the pandemic began, reflecting higher interest rates and larger balances on auto loans and credit cards. Delinquency rates, while still low by historical standards, are on the rise: As of March, nearly 8% of borrowers were at least 60 days late on payments, up from less than 5% in July 2021. Worse, the delinquencies are concentrated among middle-aged borrowers with high balances, who should be in their highest-earning years.

The best-case scenario is that tens of millions of people will have a few hundred dollars less to spend each month. The worst case is mass delinquency on student loans or other debt, perhaps with repercussions for the broader financial system. The reality will likely be somewhere in between.

What to do? So far, politicians—and the Supreme Court—have focused primarily on whether or not to forgive some of the debt, rather than on the much more important questions of how the debt arose, and whether so much of the risk of investing in education should fall on students in the first place.

Education is a valuable asset. It requires a big upfront investment and generates returns both for the individual and for the economy overall—in the form of, respectively, higher earnings and greater human capital. As long as individuals have enough incentive to invest on their own, everyone benefits. But if they perceive the costs or risks as unduly high, the whole economy might suffer.

To date, U.S. policy has focused mainly on subsidizing the cost—sometimes through grants and scholarships, but increasingly through federal loans. This leave a lot of the risk on the student: If you graduate into a recession and can’t get a job, or if a for-profit college sells you a worthless degree, the debt is still on you (and in most cases can’t even be discharged in bankruptcy).

The individual returns are still often attractive: Median wages for bachelor’s degree holders remain much higher than those of their high-school-educated counterparts. Yet recent research has found that income returns to college have declined for successive generations, and wealth returns have disappeared entirely for certain younger groups. High debt loads are partially to blame, and borrowers’ increasing struggles to pay demonstrate that for too many, the returns simply haven’t justified the risk.

Improvements in college quality, tighter accreditation standards and tuition regulation could all help boost the return-to-risk ratio. The best way to ensure a socially optimal level of investment, though, is for the government to take on part of the risk. To that end, it should borrow an idea from the economist Milton Freidman: End the reliance on traditional loans and switch to a model in which the state becomes a sort of equity partner in individuals’ education, footing a large part of the bill and receiving payments tied to the income that graduates manage to earn.

To some extent, the U.S. is already moving in the right direction. The government offers income-driven repayment plans, under which borrowers pay 10% to 15% of their discretionary income for 20 to 25 years, after which the remaining balance is forgiven.

Yet the existing programs have some serious flaws. The payments can still be unaffordable for many, particularly those with highly variable monthly income. Also, just signing up can be an administrative nightmare and individuals must re-enroll and recertify their income every year or face a steep penalty: capitalized interest. As a result, most federal student loan borrowers haven’t enrolled.

To remedy this, the government should make income-based repayment the automatic option in federal student financing and handle all payments through the tax withholding system, much like Social Security contributions.

When loan payments restart in September, they’ll offer further evidence that the risk of investing in education is misallocated. If all borrowers had been on well-designed income-based plans, there would have been no need for an emergency pause in the first place. Better to address the cause of unmanageable debt burdens, rather than contend time and again with the symptoms.

Kathryn Anne Edwards is a labor economist and independent policy consultant.

This column was provided by Bloomberg News.