Latest / The Fiscal Policy Podcast with Fexingo: Government Budgets, Stimulus, and Public Spending / How Government Loan Guarantees Create Billion-Dollar Losses
Transcript
- Lucas: The federal government currently has about $3 trillion in loan guarantees outstanding. That's student loans, small business loans, export financing, housing — you name it. Luna: $3 trillion — but that's not the same as $3 trillion sitting in a bank account, right? Lucas: Right. It's contingent liability. The government promises to pay a lender back if the borrower defaults. The question is: how much of that $3 trillion actually goes bad? Luna: And the answer, historically, is a lot more than budget projections assume. The Congressional Budget Office has a running track record of underestimating losses on loan guarantees by about 30 to 50 percent. Lucas: Let's anchor on the biggest one: student loans. The Department of Education's portfolio is roughly $1.6 trillion. For every dollar of student loans issued, the government budgets a subsidy cost — the present value of expected defaults, interest subsidies, and administrative costs. Luna: But that subsidy cost is based on a model that assumes certain default rates, certain repayment behaviors, certain economic conditions — and those assumptions are often too optimistic. Lucas: Exactly. In 2020, the CBO projected that the Education Department's 2021 cohort of loans would have a subsidy rate of about negative 9 percent. Negative — meaning the government expected to make money on those loans, because interest payments would exceed losses. Luna: And then the pandemic happened, payment pauses, forgiveness proposals — and suddenly those 'profitable' loans are looking like they'll lose money. Lucas: This is a structural problem. The subsidy cost model is supposed to reflect risk, but it's calibrated on historical data that doesn't capture tail events. And because the budget records the subsidy cost upfront, any losses beyond that are essentially hidden — they show up as 'reestimates' years later, buried in the budget's fine print. Luna: If today's episode made you think differently about how the government accounts for risk, honestly, that's the kind of thing we try to do here. If it was worth the price of a coffee to you, you know the deal: buy me a coffee dot com slash fexingo. It helps us keep the show ad-free and independent. Lucas: Yeah, I'd second that. Listener support is what lets us spend a full episode on a topic like this. No ads, no sponsors — just the economics. Luna: Alright, back to it. So those reestimates — how big have they been? Lucas: In 2023, the Department of Education's loan portfolio required a $200 billion upward reestimate. That's $200 billion of losses that weren't in the original budget. And that's just one year. Luna: So the upfront subsidy cost was too low, and taxpayers absorbed the difference later. That's a classic budget gimmick: make the program look cheaper now, and let future Congresses deal with the bill. Lucas: Exactly. And it's not just student loans. The Small Business Administration's 7 loan guarantee program has a similar dynamic. The subsidy rate is set by Congress, and it's often below the actual loss rate. In 2022, the SBA's loan guarantee program had a net loss of about $2.2 billion — but the budgeted subsidy cost was only $1.1 billion. Luna: So the program loses about twice as much as the budget says it will. And because the losses are spread across many years, they don't trigger immediate alarm. Lucas: Then there's the Export-Import Bank. It guarantees loans to foreign buyers of U.S. goods. In theory, it's self-sustaining through fees and interest. But in practice, it's had years where losses exceed reserves, requiring a taxpayer bailout. The bank's default rate on large aircraft loans, for example, has been around 2 percent historically, but during the 2008 crisis it spiked to over 10 percent. Luna: And those spikes are exactly when the guarantees matter most — but also when the losses are hardest to absorb. Lucas: Part of the problem is how risk is priced. The government is supposed to charge a fee that reflects the borrower's default risk, but political pressure keeps fees low. Congress doesn't want to be seen as making it harder for students, small businesses, or exporters to get loans. Luna: So you get a system where the government underprices risk, issues too many guarantees, and then ends up with losses that could have been avoided if the private market were pricing them. Lucas: But there's a counterargument. The government's cost of capital is lower than the private market's, so it can offer guarantees that are cheaper for borrowers and still cover expected losses. The question is whether the political system can discipline itself to set fees at the right level. Luna: And so far, the evidence says no. The CBO has documented that the federal government's loan guarantee programs, on average, lose money even after accounting for interest and fees. Lucas: Let's talk about one specific program that illustrates the problem well: the Federal Housing Administration's mutual mortgage insurance fund. It insures mortgages for low-income borrowers. In 2013, the fund had a shortfall of $1.7 billion and had to draw from Treasury. The FHA had been setting premiums too low and underestimating defaults. Luna: And after that crisis, they raised premiums and tightened underwriting. But the tendency is to loosen again when the economy is good. Lucas: Exactly. It's pro-cyclical. When times are good, defaults are low, and the fund looks healthy, so Congress and the agency get pressure to lower premiums and expand access. Then when the downturn hits, the fund is undercapitalized. Luna: It sounds a lot like what happened with Fannie Mae and Freddie Mac before 2008. They were implicit government guarantees, not explicit, but the dynamics are similar. Lucas: That's the big parallel. Fannie and Freddie had a line of credit with Treasury and an implicit guarantee. They took on massive risk, underpriced it, and when housing crashed, taxpayers were on the hook for $187 billion. Luna: So the lesson is that explicit loan guarantee programs have the same moral hazard problem: the guarantor doesn't price risk correctly because the losses are socialized. Lucas: And the reform proposals are well-known. One is to require that the subsidy cost be calculated using a fair-value method — that is, discount expected losses using market interest rates, not Treasury rates. That would make the upfront cost look much larger. Luna: The CBO estimates that moving to fair-value accounting would increase the reported cost of student loan programs by about 50 percent. That would make it politically harder to justify expanding them. Lucas: Another reform is to require that the fees charged to borrowers be set by an independent risk model, not by Congress. That would take the politics out of pricing. Luna: But neither of those has gotten serious traction. Why? Lucas: Because loan guarantees are popular. They allow politicians to deliver benefits to constituents without showing up as direct spending. A $100 million loan guarantee costs the budget only a few million in subsidy cost, but it provides $100 million in apparent benefit. Luna: So it's a form of fiscal illusion. Voters see the loans, not the risk. Lucas: Exactly. And that illusion is hard to dispel. Even if you educate the public about the hidden costs, there's always a plausible argument that the program is necessary and the risks are manageable. Luna: Until they aren't. The question is whether the next crisis will be triggered by a wave of defaults on guaranteed loans — student loans, SBA loans, something else. Lucas: And whether, when it happens, the government will have the fiscal capacity to absorb the losses without cutting other programs or raising taxes. Luna: So the next time you hear a politician tout a new loan guarantee program, you know what to look for: what's the subsidy rate, how was it calculated, and who's going to pay if the assumptions are wrong. Lucas: That's the real story. And it's one that most budget documents don't tell you.