Guest Post: Jason Delisle on the Federal Student Loan Program

by Reihan Salam

Editor’s note: Jason Delisle, director of the New America Foundation’s Federal Education Budget Project, has kindly agreed to share his thoughts on how we might reform the federal student loan program, with special reference to how we should account for credit risk in setting interest rates for student loans.

Why limit a government program that provides subsidies at no cost to taxpayers? That’s essentially the question Mike Konczal at the Roosevelt Institute asks regarding interest rates on federal student loans. After arguing that the loans are either profitable for the government or at least break-even he writes, “I’m not seeing the downside on [Congress] keeping [interest] rates lower. What am I missing?”

It sounds like a fair question, but it’s loaded.

Federal student loans are not profitable for the government, nor does the government break even on the loans. Using fair value estimates, student loans are the government’s most expensive loan program. In most years, taxpayers will spend about three times as much on the student loan program ($12 billion) as they do guaranteeing home mortgages through the Federal Housing Administration ($3.5 billion). Taxpayers will even spend less on average each year guaranteeing mortgage-backed securities through Fannie Mae and Freddie Mac ($5 billion) than they will making student loans.

If student loans do in fact cost something, then one “downside” to making them even more generous is the added cost. That shouldn’t be taken as a criticism of the loan program or a reason not to provide low(er) rates per se – the whole point of the program is to provide subsidies to what would otherwise be a non-functioning credit market. But cost is certainly one factor we ought to consider before expanding benefits.

Still, arguments like the one Mike Konczal makes abound. That’s because understanding the cost of government loan programs isn’t a straightforward task and even the most financially versed individuals get tripped up. The government’s official accounting rules make matters even worse.

The most common misperception centers on the low interest rates the government pays to borrow. For many (including Mike Konczal) this seems like a dead giveaway that the federal government can provide loans at lower costs than private lenders.

To see why the government’s cost of borrowing doesn’t capture the full cost of making a student loan, consider an example similar to one that Debbie Lucas at the MIT Sloan School of Management uses. Let’s say the government issues $100 million in 10-year U.S. Treasury notes to finance $100 million in student loans with 10-year repayment terms. Assume that after the 10 years is up, the student loan portfolio has suffered losses such that the U.S. Treasury bonds cannot be fully repaid with the loan repayments alone.

Does the government then default on its debts? Of course not. It taps taxpayers to make up the losses and repays bondholders in full. Note that this makes taxpayers equity investors in the student loan program – it is their money that will be used to absorb 100% of any losses on the loans to ensure U.S. Treasury bond holders are always repaid.

That highlights a key point: the interest rate on U.S. Treasury securities tells us what investors want to be paid to lend with zero risk of default. Federal student loans are not, however, free of default risk. The U.S. Department of Education expects that about 19 percent of loans made to students in 2013 will default at some point. Yes, cost estimates can build those default rates in, and Congress can charge an interest rate on student loans that more than fully offset such expected losses. But any unexpected losses, those that might occur if the economy weakens, wouldn’t be covered, placing the default costs squarely on taxpayers.

If you subscribe to the argument that we should measure (i.e. discount) the cost of government loans based solely on what the government pays to finance them with Treasury debt, you effectively treat the equity risk taxpayers bear in making the loans as costless. Or more aptly, you exclude it. You treat average expected losses as if they are certain to occur and assume the government won’t need to call on taxpayers to make up for unexpected losses. That is more or less the approach written into a 1990 accounting law that today still binds federal budget analysts at the Congressional Budget Office and every federal agency.

Of course, excluding a cost is not the same as the government having a lower cost, which brings us back to the original point.

Fair value estimates, which incorporate a cost for the market risk (the unexpected losses) inherent in any loan, show that the government has set the rates and terms on federal student loans “below cost.” That is, taxpayers are subsidizing borrowers. According to a 2010 Congressional Budget Office study, the fair value cost averages about $12 dollars for every $100 lent. An estimated $112 billion in new loans will be made this year.

In short, federal student loans aren’t profitable. They aren’t even free. Charging borrowers even lower interest rates, then, isn’t free either.