The most consequential financial decision many Americans make before they are old enough to rent a car begins with a form. A student sits across from a financial aid officer, or more often now in front of a laptop, and agrees to borrow money for an education whose value cannot yet be known. The student does not know whether the degree will be completed, whether the labor market will reward it, whether family circumstances will interrupt school, or whether the institution will deliver what its glossy brochures imply. The college knows more about its completion rates, its programs, its students, and its outcomes than the student ever could, yet the student signs the note. The peculiar thing is how normal this arrangement has come to seem.
Roughly 43 million Americans now carry a combined $1.86 trillion in student debt, and more than one dollar in ten is at least 90 days delinquent (American Default, 2026; Education Data Initiative, 2026).
Americans have spent years arguing about loan forgiveness, repayment obligations, fairness to taxpayers, and the moral responsibility of borrowers. Those are legitimate questions, but they begin too late in the story. The more fundamental question is why we decided that the appropriate way to finance an uncertain investment in human potential was to use a financial product that behaves remarkably like ordinary consumer debt. That decision has shaped almost everything that came afterward.
Federal student loans have historically carried interest, delinquency penalties, credit consequences, complicated repayment plans, and unusually restrictive bankruptcy treatment. Borrowers were expected to navigate a fragmented collection system that once forced them to choose among a dozen overlapping repayment plans and relearn the rules every few years. A 2026 overhaul collapsed that maze into two paths, a Tiered Standard plan and a new income-based Repayment Assistance Plan, with the old plans set to disappear by 2028 (Study.com, 2026). The simplification is real. The central premise is not: the borrower accepts the risk, the institution receives the money, and the government administers the obligation.
The result is a system that asks young people to make a sophisticated investment decision before they possess the information needed to make it.
A mortgage finances a house. The house exists, can be inspected, appraised, insured, and sold. A car loan finances a machine whose approximate value can be estimated before the transaction occurs. A business loan often finances equipment, inventory, or an enterprise whose cash flow can at least be modeled. A student loan finances a prediction about a person who does not yet exist.
The prediction says that this particular student, after this particular program, at this particular institution, will enter some future labor market and earn enough additional income to justify the cost incurred years earlier. Completion matters. Program quality matters. Geography matters. Health matters. Family obligations matter. Labor-market conditions matter. Discrimination matters. Plain luck matters. The borrower cannot control all of these variables, and neither can the college. Yet American policy has traditionally treated the uncertainty as though it belonged primarily to the student.
Other countries made different choices.
Australia’s higher education financing system is perhaps the clearest example. Under its HECS-HELP scheme, nobody owes a cent until income clears roughly A$69,500 a year, and the balance grows only with inflation, not compound interest (WeMoney, 2026). Repayment is collected automatically through the tax system, not a monthly bill.
England uses a variation on the same principle. No repayment is required below roughly £25,000 to £28,000, and the loan is withheld through the same payroll system that collects income tax (GOV.UK, 2026). New Zealand runs a comparable model.
Germany solves the problem from the other direction. Public tuition is nominal, most support arrives as grants, and the loan portion of its BAföG program is interest free and capped at €10,010 total, no matter how many years a student spends in school (Handbook Germany, 2025).
None of these countries has discovered a magical way to make higher education free of cost. Someone must pay professors, maintain buildings, operate laboratories, provide technology, and support students. The important difference is how the financial risk is distributed. A system can place more risk on the individual student, more on the taxpayer, more on the institution, or distribute it among them.
America chose a model that placed a remarkable amount of the risk on the person with the least information and the least bargaining power.
That might have been easier to defend if universities themselves faced comparable financial consequences when educational investments failed. Historically, they did not. A college could recruit students, collect federal loan proceeds, provide an education of uneven economic value, and remain largely insulated from the borrower’s subsequent inability to repay. A new Earnings Accountability rule, finalized in July 2026, finally ties a program’s eligibility for federal loans and Pell grants to whether its graduates out-earn a comparison group (JD Supra, 2026). It is a start. It is also a pass-fail cliff rather than a continuous incentive, and meaningful institutional exposure to loan losses remains the exception rather than the rule.
The imbalance matters because institutions respond to incentives just as people do. Federal lending was intended to expand access to education, and it undeniably has helped millions of people attend college.
Yet research also suggests that increased loan availability pushes tuition higher. A widely cited Federal Reserve Bank of New York study found that roughly 60 cents of every additional dollar in loan limits showed up as higher tuition (New York Federal Reserve, n.d.). A broader review of the literature found a majority of studies supporting that same pattern, known as the Bennett hypothesis (Robinson, n.d.). That finding should complicate the usual conversation.
Political debates often assume that there are only two participants in the student loan story: the borrower and the taxpayer. One side worries about the borrower being crushed by debt. The other worries about taxpayers being asked to pay bills incurred by someone else. The institution that priced and sold the education somehow fades into the background. Imagine applying that logic elsewhere.
A bank that repeatedly made catastrophic mortgages would face questions about underwriting. An investment company that consistently sold products producing terrible returns would attract scrutiny. A manufacturer whose products routinely failed would eventually face liability, regulation, or market consequences. Higher education has often been treated differently because education is rightly understood as something more than a commercial product. That moral distinction should not become a financial exemption.
Education can be a public good and still be priced irresponsibly. A university can perform a noble social mission and still operate a program whose tuition bears little relationship to the economic opportunities available to its graduates. A degree can enrich a person intellectually while simultaneously leaving that person financially worse off. Society should be capable of holding all of these truths at once. The current debate about forgiveness tends to obscure this structural problem.
Debt cancellation can unquestionably help individual borrowers. Certain groups present particularly compelling cases: people who borrowed but never completed a credential, borrowers harmed by deceptive or failed institutions, people whose repayment outcomes have remained poor for years.
The racial disparities are especially striking. Twenty years after leaving school, the median white borrower has paid off 95 percent of the original balance. The median Black borrower still owes 95 percent of it. Black borrowers default at roughly three times the rate of white borrowers, 37.5 percent against 12.4 percent (Brookings Institution, 2024a, 2024b).
Yet forgiveness is fundamentally a treatment for accumulated damage. It does not answer the question of how the next generation should finance education.
A country could forgive hundreds of billions of dollars tomorrow and still recreate the same problem if tuition continued rising, borrowing remained loosely connected to program quality, and institutions faced little consequence when graduates could not repay. The floor would be clean, but the pipe would still be leaking.
A better system would begin with a different conception of what an educational loan is. The first principle should be simple: students should borrow less.
Public policy often responds to rising college costs by increasing the amount students are permitted to borrow. That solution can become self-defeating if additional credit merely allows prices to continue rising. Grants, direct public support, and tuition restraint should absorb a larger share of the initial cost, particularly at public institutions. The best student loan may ultimately be a small student loan.
The second principle should recognize that government does not need to make a conventional profit on education lending.
A redesigned system could eliminate real interest while indexing outstanding balances modestly to inflation or wage growth. Taxpayers would preserve the real value of the money lent without transforming the government into an institution collecting compound returns from young adults attempting to build productive lives. Australia provides a useful precedent for this type of indexation rather than conventional interest.
Critics will say this costs money the government does not have. They are not wrong, only incomplete.
Congressional Budget Office scoring of recent reconciliation packages shows how sensitive that number is. Different versions of the same bill moved federal outlays by tens of billions of dollars depending on how interest and forgiveness provisions were written (Congressional Budget Office, 2025).
A zero-interest, income-contingent design will likely cost more on paper than the current system. Weigh that against lower default and collection costs, higher completion rates, and the tax revenue that follows when more borrowers finish school and earn more. The evidence for that trade-off is real, if modest (Looney & Yannelis, 2020).
Reform is not free. The current system is not free either. It has simply hidden its price in delinquency, wrecked credit, and degrees never finished. The third principle should be automatic income contingency.
Borrowers should not have to become amateur experts in federal loan administration. Repayment should rise and fall automatically with income, ideally through payroll withholding or the tax system. Someone earning below a protected threshold should owe nothing. Someone whose education produces substantial earnings should repay more quickly. Someone whose earnings remain modest should pay modestly. The logic is almost embarrassingly straightforward.
Society finances an investment in future earnings. Repayment should therefore depend upon future earnings.
The fourth principle should establish an end point. A loan intended to finance education should not follow a borrower indefinitely through middle age and into retirement. Twenty years is a defensible horizon. Remaining balances after that period could be discharged automatically, provided the borrower has participated honestly in the income-contingent system.
The fifth principle should restore something closer to ordinary bankruptcy treatment.
American bankruptcy law recognizes that sometimes economic bets fail. Businesses fail. Mortgages fail. Medical bills overwhelm families. Credit cards become unmanageable. Bankruptcy is not intended to make these experiences painless, but it does acknowledge that permanent indebtedness is economically and socially destructive.
Student debt has long occupied a strange exception to that philosophy. Federal educational loans remain dischargeable only under a demanding undue-hardship standard, one that succeeds in an estimated 0.01 percent of bankruptcy filings (U.S. Congress, 2025).
Bills now pending in Congress, including the Student Loan Bankruptcy Improvement Act, would replace that standard with something more workable. Educational debt should not exist in a category approaching permanent obligation.
The sixth principle may be the most important: put the college on the loan.
A serious system of institutional risk sharing would require colleges and programs to absorb part of the losses when their students consistently fail to repay. The research model developed for this proposal suggests an institutional share of perhaps 20 to 50 percent of unrepaid principal, adjusted according to program-level outcomes rather than individual borrower performance.
That change would alter the incentives immediately.
A university considering a new graduate program would have to ask more than whether students were willing to enroll. Administrators would have to consider whether graduates could plausibly earn enough to justify the tuition. A program charging $80,000 for a credential associated with $40,000 salaries would suddenly become a financial concern for the university rather than solely for the student.
Universities might discover a new enthusiasm for controlling tuition.
Schools might also invest more heavily in completion, career placement, academic advising, and employer partnerships. Programs with chronically weak outcomes would become expensive to operate. The institution would finally participate financially in the prediction it asks the student to make. That idea carries a real danger.
Colleges facing financial responsibility for poor repayment could become more selective and avoid students they perceive as likely to struggle. Low-income students, first-generation students, older students, and students from historically disadvantaged communities could become undesirable financial risks. A badly designed risk-sharing system could therefore reinforce precisely the inequalities higher education is supposed to reduce.
The solution is not to abandon institutional responsibility. The solution is to measure it intelligently.
Risk sharing should operate at the program level and should account for the characteristics of the populations being served. A nursing program educating large numbers of low-income students should not be punished simply because those students entered college with fewer economic resources. A costly graduate program whose students consistently leave with weak employment prospects presents a different policy problem. Good policy distinguishes between serving difficult populations and selling weak products.
The distinction becomes particularly important when examining the borrowers who suffer most under the current system. Public attention gravitates naturally toward people with six-figure loan balances, but the most vulnerable borrower may instead owe $10,000 and have no degree.
That person received the liability without receiving the asset the liability was intended to finance.
Debt balance alone therefore tells us remarkably little about financial distress. A physician earning $250,000 with $150,000 in educational debt may be far more financially secure than a former student earning $32,000 with $12,000 in loans and no credential. Completion status, income, family wealth, and program value matter more than the headline balance. That reality also helps explain why the politics of student debt have become so confused.
Critics of forgiveness can correctly identify borrowers with expensive professional degrees and strong earnings who appear capable of repaying. Advocates for relief can correctly identify millions of borrowers whose educational investments produced little economic return and whose debts remain serious obstacles. Both groups exist simultaneously. A better system would not force public policy to pretend they are the same.
The moral argument for redesign is ultimately less radical than it first appears. Markets work best when risk is allocated to parties capable of understanding and influencing it. Students can influence their own effort, but universities influence admission, curriculum, tuition, program design, advising, completion support, and career preparation. Governments influence accreditation, subsidies, lending limits, bankruptcy rules, and accountability. Employers and labor markets determine much of the eventual economic return. Responsibility should follow control.
The student should bear some responsibility because education requires effort and choices. The taxpayer should bear some responsibility because an educated population produces public benefits. The institution should bear some responsibility because it designs, prices, and sells the educational experience. Shared responsibility is not the same thing as absolving everyone.
That is why the best student loan policy may have surprisingly little to do with forgiveness. A well-designed system would use grants to reduce initial borrowing, tie repayment automatically to income, eliminate real interest, establish a clear end point, allow meaningful bankruptcy relief, and require institutions to absorb part of the losses generated by consistently poor outcomes. The research underlying this proposal points toward precisely that combination: smaller loans, automatic income-contingent repayment, and genuine institutional risk sharing. The student at the financial aid desk would still be making a bet.
College always involves some uncertainty. Education cannot guarantee prosperity, and no financing system can remove all risk from adulthood. The purpose of policy is not to eliminate uncertainty. The purpose is to distribute its consequences in a way that reflects who had information, who had power, who benefited, and who made the promises.
America’s great mistake was not lending young people money to pursue education. The mistake was asking them to bear nearly all the consequences when everyone else turned out to be wrong.
A note on format, for anyone who made it this far. Most of what runs under this byline argues from evidence via narrative, earning trust through clarity, not footnotes.
This piece is different.
The numbers here are the argument: repayment thresholds in other countries, default rates split by race, how rarely a bankruptcy court actually discharges a student loan. Numbers like that only mean something if a reader can trace them back to where they came from, and a proposal that asks colleges to share financial risk and asks Congress to rewrite bankruptcy law should be able to survive someone checking the sources. So this one gets endnotes. Call it the exception that proves the rule.
For readers who want to run their own numbers, a companion calculator comparing repayment across all four systems, along with the full comparison tables and objections, is live at studentloans.harnessai.net.