In the latest National Affairs, Jason Delisle and Jason Richwine make what they call ”The Case for Fair-Value Accounting.” This is the process of using the price of, say, student loans in the capital markets to budget and discount government student loans. (The issue also has articles walking back support for previously acceptable moderate-right ideas like Common Core and the EITC, showing the way conservative wonks are starting to line up for 2016.)
In the piece Delisle and Richwine make two basic mistakes in financial theory, mistakes that undermine their ultimate argument. Let’s dig into them, because it’s a wonderful opportunity to get some finance back into this blog (like it used to have back when it was cool).
Error 1: Their Definition of FVA Is Wrong
What is fair-value accounting (FVA)? According to the authors, FVA “factors in the cost of market risk,” meaning “the risk of a general downturn in the economy.” This market risk reflects the potential for defaults; it’s “the cost of the uncertainty surrounding future loan payments.”
These statements are false. There is a consensus that FVA incorporates significantly more than this definition of market risk.
Here’s the Financial Economists Roundtable, endorsing FVA: “Use of Treasury rates as discount factors, however, fails to account for the costs of the risks associated with government credit assistance — namely, market risk, prepayment risk, and liquidity risk.”
And the CBO specifically incorporates all these additional risks when it evaluates FVA: “Student loans also entail prepayment risk… investors… also assign a price to other types of risk, such as liquidity risk… CBO takes into account all of those risks in its fair-value estimates.”
This is a much broader set of concerns than what Delisle and Richwine bring up. For instance, FVA requires taxpayers to be subject to the same liquidity and prepayment risks as the capital markets. Remember when the federal government stepped in to provide liquidity to the capital markets when they failed in late 2008, because the markets couldn’t? That gives us a clue that there might be some differences between public and private risks.
Crucially, it’s not clear to me that taxpayers have the same prepayment risk as the capital markets. Private holders of student loans are terrified that their loans might be paid back too quickly, because they are likely to get paid back when interest rates are low and it will be tough to reinvest at the same rate. This is a particularly big risk with the negative convexity of student loan payments, which can be prepaid without penalty. Private actors need to be compensated generously for this risk.
Do taxpayers face the same risk? If student loans owed to the government were paid down faster than anyone expected, would taxpayers be furious? I wouldn’t. I certainly wouldn’t say “how are we going to continue to make the profit we were making?” as a citizen, though it would be an essential question as a private bondholder. Either way, it’s as much a political question as an economic one. (I make the full argument for this in a blog post here.)
Error 2: Their Definition of Market Risk Is Wrong
The authors like FVA because it accounts for market risk. But what is market risk? According to Delisle and Richwine, market risk is “associated with expecting future loan repayments,” as “[s]tudents might pay back the expected principal and interest” but they also may not. It is also “the risk of a general downturn in the economy… market risk cannot be diversified away.”
So the first part is wrong: market risk is not credit risk, or the risk of default or missing payments. The International Financial Reporting Standards (IFRS7), for instance, requires reporting market risk separate from credit risk, because they are obviously two different things. I’ve generally only heard market risk used in the context of bond portfolios to mean interest rate risk, which they also don’t mention. So if market risk isn’t credit risk or interest rate risk, what is it?
I’m not sure. What I think is going on is they are confusing the concept with the market risk of a stock, specifically its beta. A stock’s beta is its sensitivity to overall equity prices. (Pull up a random stock page and you’ll see the beta somewhere.) It’s very common phrasing to say this risk can’t be diversified away and is a proxy for the risk of general downturns in the economy, which is the same language used in this piece.
Market risk for stocks is the question of how much your portfolio will go down if the market as a whole goes down. But this has nothing to do with student loans, because students (aside from an enterprising few) don’t sell equity; they take out loans. If students paid for school with equity, in theory an economic downturn would lead to less revenue, since students would make less money overall. But even then it’s a shaky concept.
This isn’t just academic. There’s a reason people don’t speak of a one-to-one relationship between a market downturn and the value of a bond portfolio, as the authors’ “market risk” definition does. If the economy tanks, credit risk increases, so bonds are worth less, but interest rates fall, meaning the same bonds are worth more. How this all balances is complicated, and strongly driven by the distribution of bond maturities. This is why financial risk management distinguishes between credit, liquidity, and interest rate risks, and doesn’t conflate those concepts as the authors do.
(Though they are writing as experts, I think they are just copying and pasting from the CBO’s confusing and erroneous definition of “market risk.” If they are sourcing any kind of common financial industry practices or definitions, I don’t see it. I guess Jason Richwine didn’t get a chance to study finance while publishing his dissertation.)
Here again I’d want to understand more how the value of student loans to taxpayers moves with interest rates. Repayments are mentioned above. And for private lenders, higher interest rates mean that they can sell bonds for less and that they’re worth less as collateral. They need to be compensated for this risk. Do taxpayers have this problem to the same extent? If interest rates rise, do we worry we can’t sell the student loan portfolio for the same amount to another government, or that we can’t use it as collateral to fund another war? If not, why would we use this market rate?
Is This Just About Credit Risk?
Besides all the theoretical problems mentioned above, there’s also the practical problem that the CBO uses the already existing private market for student loans (“relied mainly on data about the interest rates charged to borrowers in the private student loan market”), even though there’s obviously a massive adverse selection problem there. Though not an error, it’s a third major problem for the argument. The authors don’t even touch this.
But for all the talk about FVA, the only real concern the authors bring up is credit risk. “What if taxpayers don’t get paid?” is the question raised over and over again in the piece. The authors don’t articulate any direct concerns about, say, a move in interest rates changing the value of a bond portfolio, aside from the possibility that it might mean more credit losses.
So dramatically scaling back consumer protections like bankruptcy and statute of limitations for student debtors wasn’t enough for the authors. Fair enough. But there’s an easy fix: the government could buy some credit protection for losses in excess of those expected on, say, $10 billion of its portfolio, and use that price as a supplemental discount. This would be quite low-cost and provide useful information. But it’s a far cry from FVA, even if FVA’s proponents don’t quite understand that.