I’m sometimes asked why I focus more on legal education than student debt generally because it’s a much bigger story. The easy answer is that I do write about student debt regularly, and I recognized early on that understanding how the U.S. student debt system works is crucial to informed writing on legal education.

The complex answer is that while student debt is the bigger story, it’s nevertheless fairly easy to summarize, yet once you grasp that summary, you discover that student debt doesn’t move much in the news. Moreover, most of what you hear or read doesn’t capture the magnitude of the problem.

Indeed, college-for-all boosters tend to trivialize student debt as a national issue, whether by ignoring delinquencies, pointing to the small percentage of debtors with large balances, claiming that only some credentials are overvalued but others are fine, refocusing on reducing dropouts, or limiting the crisis to graduate students. Most of these arguments treat student debt as the independent variable rather than as the dependent variable against debtors’ incomes. In truth, average balances of student debt have grown at every income level—as has the percentage of debtors—so it doesn’t matter how little debt some people owe if they can’t afford it. Consequently, it’s reasonable to refer to student debt as a bubble, but it’s more complicated than the other kinds you may have read about.

Here’s what I mean.

Student Loan Debt in the 21st Century

The Federal Reserve Bank of New York admirably tracks consumer credit growth longitudinally. You can see the rapid growth in student loan debt compared to other types of non-mortgage household debt since 2004.

Non-Housing Debt Balance

The NY Fed also tracks the distribution of student loan debt, the average balances (pdf), and its effects on other credit-based purchases. It’s all the grim news you suspected: Student loans sap consumers’ purchasing power and hampers household formation.

As for what will happen with student debt in the near future, the Office of Management and Budget (OMB) projects that federal student loans, which are the majority of all student loans by far, will continue to rise. In its fiscal year 2016 midsession review (pdf), the OMB estimated that between 2015 and 2025, the government will lend out an additional $1.4 trillion.

Projected Direct Loan Balances (OMB, Billions Current $)

On the other hand, my analysis of Department of Education data show that lending has fallen, likely due to declining college enrollments.

Amount of Federal Loans Disbursed

As a result, we can expect that federal lending will undershoot the OMB’s projections. That’s the good news: All the doomsday scenario of trillions upon trillions of dollars by 2030 or whatever aren’t going to come about.


Does the Government Make Money on Student Loans?

The answer is no, which I’ve found is controversial position to adopt.

Many student debtor advocates look to the interest revenue the government receives from student loans and conclude that the government is profiting at the expense of indentured workers. This belief ignores a few important facts.

(1)  Delinquencies have spiked.

Starting in 2013, data from the New York Fed showed that the delinquency rate for all student loans rose above 10 percent.

Percent Balance 90 Days+ Delinquent

In fact, by my estimate Department of Education data show that about 4.3 million debtors with older guaranteed loans and 3.2 million debtors with direct loans have defaulted on them. (There’s some overlap between those two figures, thankfully.) There are 40.8 million student loan debtors, so more than 10 percent have defaulted.

Many more debtors are in deferment, forbearance, or still in school. As of third quarter 2015, only 53 percent of all $1.182 trillion in federal student loans were in active repayment, and more than one-fourth of that debt was in some kind of income-sensitive repayment plan. When people forbear, defer, or default on their loans or choose a repayment plan that extends or forgives the total balance, the lender is not fully repaid, so it’s unlikely that the government is profiting from student loans.

Finally, the Federal Direct Loan Program lends as much money to graduate and professional students as their programs are willing to charge. When programs don’t need to compete for students, they absorb the unlimited federal loans and raise tuition costs accordingly. The academic literature refers to this phenomenon as the “Bennett hypothesis,” which posits that colleges and universities won’t expand their enrollments even if people are financially able to attend them thanks to subsidies. These debtors take out very high amounts, often six figures, but they can’t make enough money to repay their loans on conventional repayment plans. When they sign onto an income-sensitive repayment plan, it’s almost certain that their loans will be canceled in the future. This is what will happen to many of the law school debtors whom I write about, and it illustrates another reason I don’t think the government is profiting on student loans.

One of the biggest contributing factors to student loan non-payment is joblessness and low-paying jobs among debtors. Until that is remedied, the default crisis will continue.

(2)  Many graduates aren’t working in jobs that require college educations.

The best study on “malemployment” of college grads I’ve found is by Paul E. Harrington and Andrew M. Sum. In their two articles from 2010, they show that graduates who do not obtain jobs that the Labor Department believes require the skills and knowledge gained in college earn little more than high school graduates. Indeed, the New York Fed found that in any given year about one-fourth of college graduates earn about the same amount as the median high school graduate. College does not always pay off.

Regardless of whether one thinks that college builds irreplaceable human capital or just signals preexisting abilities, when graduates can’t find college jobs, their student loans are really just an income surtax, even if they’re on an income-sensitive repayment plan. The point of the loan programs is to increase national income and therefore revenue. It may appear that the government is making money under these conditions, but the long-term costs from sapping consumers’ spending and investing reduces the government’s income.

The foregoing argument also explains why I sometimes track household formation and housing vacancies on this blog.

(3)  Fair-value accounting works better.

Until 1990, the federal government used a “cash accounting” system to determine its budget. This meant that if the government lent out $10,000, it would’ve recorded a ten thousand dollar expenditure on its ledger even though it should’ve expected the money to be paid back in subsequent years. To remedy this, Congress passed the Federal Credit Reform Act, which switched the government’s accounting methodology to “accrual accounting” for its loans. This means that the Department of Education calculates the net present value of student loans using a statutorily mandated discount rate, which by law is the interest rate of U.S. Treasury securities. In other words, if lending a student money makes the government more than it would have by buying its own debt on bond markets, then it considers the loan worthwhile. While better than cash accounting, there are serious problems with this methodology.

For one, it confuses the government’s greater borrowing power with an assumption that the government never wastes money. Well, it does, e.g. the Iraq invasion. If the government can out-borrow and out-lend the private sector, then why not socialize all credit markets? I’m not talking credit unions; I mean no private banking whatsoever. That’s where arguments favoring accrual accounting leads us.

Two, some loan programs look ridiculously cheap by accrual accounting, like Grad PLUS loans to graduate and professional students. In 2014, the CBO estimated that the subsidy rate for that program was -30 percent, that is, $1.30 for every dollar lent out or better. If Grad PLUS loans are so profitable, why didn’t the private sector lend them out before? The reason is that the loans aren’t really profitable.

The alternative to accrual accounting is “fair-value accounting,” which the Congressional Budget Office (CBO) favors and treats the government’s lending programs as though it were a private lender. When the CBO estimates the costs of federal loan programs according to fair-value accounting, they lose money (pdf). The CBO’s methodology for fair-value accounting may be problematic, but it makes more sense than refusing to question whether all government loan programs are sound investments.

So What Happens to the Future?

Short answer: slow-motion default with large losses for the government.

As stated at the beginning, student loans are a different kind of debt bubble. Unlike a stock market or real estate bubble, almost all of the loans are either guaranteed or directly lent by the government. Because the government cannot go bankrupt, there will be no credit crisis, just a large write-down. As a result, it’s difficult to see how a recession or depression could result, despite the amount of debt.

The only rebuttal to calling student debt a bubble that has any legs is arguing that bubbles require secondary markets and a person’s credentials can’t be resold. I think this response just elevates form over substance. Like any other asset, a credential has a rental value, whether one considers it the skills and knowledge or the signaling capacity it grants to a student, but the full price isn’t in line with what graduates will earn over their lifetimes due to a deficiency in high-paying jobs. The federal loan program merely amplifies the credential bubble per the Bennett hypothesis.

All that having been said, there’s evidence that education borrowing is declining, and defaults will either persist or be absorbed into income-based repayment plans. I don’t think the percentage of defaults will increase much more. As analyses on student debt find that much of it will not be repaid, the discussion in Washington will become even more acrimonious than it is now.