To extend credit to people in the knowledge that they are unlikely to repay is decried by some as predatory lending. Those suspected of engaging in this practice will face scorn from regulators and opprobrium in the media.

Yet predatory is not an inaccurate sobriquet for government policy towards the funding of higher education. So doggedly have administrations around the world pursued the goal of giving everyone a degree, that they haven’t stopped to ponder whether this drive benefits taxpayers — or, indeed, students themselves.

As so often with government-induced disasters, the original motivation was laudable. Because human capital accumulation raises worker productivity and therefore salaries, higher education was viewed as a driver of increased prosperity and social mobility. Moreover, since higher productivity raises output and tax revenue, having more educated workers became not just a private but a social good in politicians’ minds.

However, the crisis and its aftermath have forced an unwanted reckoning. Even as enrolment rates have climbed strongly, rising from 49 to 69 per cent among US high school graduates — especially those from lower-income backgrounds — between the 1970s and 2016, and reaching 49 per cent among all 18-year-olds in Britain, the expected returns have failed to materialise for many.

A 2014 report from the Urban Institute, using a conservative methodology, found 25 per cent of US bachelor’s degree holders working in occupations for which they were overqualified. In Britain, as many as 16 per cent of those in employment between the ages of 16 and 64 were “overeducated” in 2015, up from 13 per cent in 2006.

Overqualification need not be a concern so long as it’s the product of choice. It is generally associated with lower median earnings than occupations which require a university degree. But plenty of less-than-well-paid jobs, such as journalism, political activism and creative writing, feature a preponderance of university graduates in their ranks.

The problem arises when overqualification stems from graduates’ inability to find suitable work.

It may be that their specific skills do not match the needs of firms – which might require more introverted computer scientists and fewer eloquent philosophers — or that the skills taught in universities do not add up to much valuable human capital. Economist Bryan Caplan has persuasively contended that college education is more a signalling device than a provider of hard skills. If that is the case, then encouraging university attendance will not benefit those at the bottom of the graduate pile, let alone those who fail to complete their studies.

The scale of the student debt burden in both America and Britain suggests that many are being failed by the system.

Outstanding higher education loans in the United States recently passed the $1.5 trillion mark, up from under $500 billion in 2006. In the UK, student debt stands at just over £100 billion, but it has more than doubled since 2011 and continues to increase rapidly. Aggregate amounts can only convey a crude picture, but the rapidity with which loans outstanding have mounted is at least suggestive that some borrowers are giving little thought to whether borrowing is a good idea.

On an individual level, the problem is not stratospheric indebtedness. Recently, the Wall Street Journal profiled a 37-year-old Utah orthodontist whose balance exceeds $1 million and will in a few years surpass $2 million, as his $1,600 monthly repayment doesn’t suffice to cover the monthly interest expense.

But that’s an exceptional case. Moreover, those who carry six-figure balances are often lawyers and medical doctors whose training raises their lifetime earnings by a multiple of the cost of their degree. Even making conservative assumptions about physician and attorney salaries ($150,000 and $100,000, respectively) and comparing them to the 75th percentile of U.S. income-earners ($65,000), the lifetime salary boost of a medical or legal education would justify an investment upwards of $500,000.

The amount borrowed is a poor predictor of whether a borrower will default. According to a recent analysis by the U.S. National Center for Education Statistics, 27 percent of borrowers entering higher education in 2003-04 had defaulted on their payments twelve years later. Such rates of non-repayment are extraordinary, on a par with payday and other high-cost short-term credit used in times of stress.

Perhaps surprisingly, default rates are highest among those who borrowed the least.

There’s an explanation for this puzzle. Those who never completed their degrees are over represented among loan defaulters. Their problem is not that they borrowed too much, but that they cannot make even modest repayments. The loans instead become a drag on their borrowing capacity until they are written off many years later. The stain on their credit score, on the other hand, can be permanent.

The key to solving the student debt mire thus seems to be not so much to reduce the borrowing needs of the typical student, but to ensure that those who will not benefit from higher education or are likely to drop out are not led to borrow for it.

Yet public policy currently achieves the precise opposite. The trend in both Britain and America has been to subsidize loan interest rates and link repayment to future income, writing off any outstanding balances after 25 or 30 years. Indeed, only 45 percent of post-2016 student borrowers in the UK are expected to repay their loans in full.

This system attracts more of the wrong sort of borrower, by lowering the cost of taking many years to repay and eventually cancelling whatever remains. If interest rates were instead adjusted for the borrower’s individual risk, as most other consumer loans are, those who expect to struggle to repay or to benefit little from education would refrain from borrowing.

Another problem is that the first beneficiaries from increased education spending, the universities themselves, have little skin in the game. Whether or not they furnish students with valuable and productive skills, their tuition fee revenue is unaffected. If universities instead had to bear part of the burden of repayment risk, they would be more reluctant to admit candidates whose likelihood of completion and professional success is low. Greater risk sharing might also lead universities to be mindful of cost inflation and focus additional expenditure on those items that enhanced the educational experience.

Importantly, however, governments that instituted greater risk-sharing should apply the standard symmetrically to all institutions. This means no special carve-outs for favoured universities or politically popular degrees, as these would again distort the incentives of university administrators.

It may be objected that these changes would result in the exclusion from higher education of some of the most vulnerable. But that implicitly assumes that those now pursuing degrees who would be excluded under a more market-based system do in fact benefit from their education. Default rates and the hardship of carrying a large loan balance for decades suggest otherwise.

Governments on both sides of the Atlantic are trapping many young people in debt at the start of their working life in pursuit of qualifications that, plainly, do not benefit them. If they weren’t lured into higher education, these people could begin their professional careers early and debt-free.

At a time of great transformation and opportunity, surely that presents a more attractive prospect than the costly disappointment of a failed university education.