Since the Affordable Care Act's passage (or, more specifically, with the passage of the Health Care and Education Reconciliation Act of 2010), the federal government has shifted to only directly providing student loans, rather than offering them through partner banks. The interest rates for those loans are set by law, and according to the government's accounting, they're high enough to make a profit. Here's the thing, though: they actually lose money.
That's true even though they're about to go up. Stafford loans, the most popular federal student loan, vary in rate depending on whether or not students are in need. Subsidized loans are available for low-income students, and can't be used on more than $5,500 a year in tuition and other costs, and can't accumulate to more than $23,000 in debt total. Those carry an interest rate of 3.4 percent, and the government pays the interest on the loan as long as the borrower's in school.
But that rate is set to double to 6.8 percent, the rate for unsubsidized loans (for richer students, or poor students with debt above the subsidized loan program's limits), on July 1. That means more debt buildup for student, and higher payments down the line.
That has Congress scurrying to find a way to avoid that sudden jump. Probably the most high-profile plan comes from Sen. Elizabeth Warren (D-Mass.), who proposes to set the subsidized rate at 0.75 percent for a year, the same as the Fed's discount rate at which it lends to banks. That's a bit of a bizarre comparison, since discount loans are generally overnight as opposed to student loans that last decades, and are actually a penalty rate that banks generally don't like to use (they loan to each other at a third of that rate). Nevertheless, liberal groups like MoveOn.org, Democracy for America, and the Progressive Change Campaign Committee have gotten 385,000 signatures in support of that measure.
President Obama has proposed setting the subsidized Stafford rate at the interest rate on 10-year Treasury bonds, plus 0.93 points (which would mean a rate of 2.88 based on last Friday's Treasury yields). House Republicans have proposed the same thing, but plus 2.5 points rather than 0.93. Inside Higher Ed's Libby Nelson has a great breakdown of the various plans here.
But just about all these plans have one feature in common: they set the interest rate on student loans below the market rate. And because they're below the market rate, that costs the federal government money. Contrary to popular belief, and many a breathless article, the government does not, in fact, book a profit on student loans. As New America's Jason Delisle has explained, that's because the Congressional Budget Office is required by law to use a bizarre and faulty method for determining the cost of government loans.
Just like any institution, the CBO determines the cost of loans by "discounting all of the expected future cash flows associated with the loan or loan guarantee—including the amounts disbursed, principal repaid, interest received, fees charged and net losses that accrue from defaults—to a present value at the date the loan is disbursed." To do that, it needs to settle on a "discount rate," which is usually the expected rate of return on the loan in question. Banks and other private institutions generally estimate that by finding loans with similar risks and maturities to the one being evaluated, and then using those similar loans' rates of returns.
The CBO does not do that. It discounts all government loans using the returns on Treasuries of similar maturity. So a 30-year student loan would be compared to a 30-year Treasury bond. But Treasuries are the safest bonds in the world. The U.S. government does not have a very high risk of defaulting, not least since it prints its own money. Student loans are much, much riskier. The default rate at four-year public colleges and universities is around 4 to 5 percent. To capture the true risk of these loans, you'd need to discount using the rate of return for another loan with similar risk. Comparing them to Treasuries make them seem safe no matter what the actual risk.
Delisle likes to explain this by comparing student loan risks to the risks of Greek government bonds (which hold a very high risk of default). The CBO methodology, says "would show buying up all the Greek bonds as profitable." That's because you're using Treasury bonds to pay for a much riskier bond, and counting the higher returns you get from the bond, but not taking the risk of default into account. Of course that's going to show up as profitable! But it's not how any business handles its books. "You're swapping safe assets for risky ones and the bigger the difference the bigger it appears the profit is," Delisle says.
Worse, it's hard to say what the actual market price of these loans would be, because unlike the federal government, private lenders actually ask basic questions of borrowers. The federal student loan program requires shockingly little in the way of information on the part of borrowers. "If you walked into a bank and said 'I'd like a loan' and then they asked for your credit score, and you say 'No questions asked, please,' they'd laugh you out the door," Delisle says. "That's how the government makes student loans."
Delisle guesses that the actual market rate would be between 12-18 percent, far above any rates the federal government currently charges. The government might be able to make an actual profit with rates below that, as it's better at making collections than private lenders. And if it started asking about credit histories it could go lower still. Sallie Mae offers a fixed-rate loan for undergrad that varies in rates between 5.75 percent and 12.875 percent, depending on your credit history. That's not too far from what the federal government is charging now.
Then again, maybe the federal government doesn't want to make a profit. Maybe it considers providing these loans a public good, and worth its while to provide even if it costs money. The question of whether subsidizing lower rates for programs like the PLUS loan and the "unsubsidized" Stafford loan, which are available to even the richest students, is good public policy, or better policy than boosting programs for the poor like Pell Grants, is an important one for policymakers to answer.
But this has nothing to do with making a profit or not making a profit. Delisle, for one, supports pegging the interest rates to the 10-year Treasury rate, just like Obama and House Republicans (who were probably influenced by his proposal). But he doesn't have any illusions that that reform is about making a profit.
"The fixed rate is well below market, and so you've got a program whose goal is to make loans to people at below market rates," he explains. "That's the only goal and it's accomplishing that goal. That is why Congress is really grasping at what the right rate should be, because they've already met their goal."
Update: As Malcolm Harris pointed out on Twitter, it's worth noting that the CBO's fair-value accounting analysis finds no subsidy for PLUS loans and unsubsidized Stafford loans, but a big one for subsidized Stafford loans, where rates are rising. Overall, there's a negative subsidy - profit. That's a way the government could be making a profit even under other accounting specifications, though obviously skeptics like Delisle dispute that. Thanks to Malcolm for the catch.
Update II: Delisle notes, contra Harris, that the degree of subsidy for federal loans depends a lot on the year you're looking at. 2010, for example, showed the government losing money, while 2013 showed a profit. The numbers also don't take administrative costs into account:
Some years fair value shows profit on some loans, but even CBO off the record doubts those profits due to "overly optimistic" doe collection rates. In 2010, fair value showed big subsidies, 2013 showed very small profit, but did not include administrative costs. Again, depends on the year.
I've gotten a few requests to change the headline of this post, which is fair enough - the latest CBO numbers do suggest a realer possibility of a profit than any previous estimates. But it's a disputed enough point, as Delisle's comments make clear, that the argument seems less about the plain facts than about their interpretation. I added a note at the end of the hed to encourage readers to continue to the bottom and get more of a sense of the debate around this.