07/11/2013 10:47 am ET Updated Sep 10, 2013

Sen. Warren Uses CBO Numbers -- and Still Gets Dinged by the WaPo Fact Checker

Close readers of my blog may recall my weighing in on the important but obscure issue of fair value accounting (FVA) and the federal budget. The issue came up again today in a Washington Post fact checker that (slightly) dings Sen. Elizabeth Warren for using CBO numbers (!!) that show an expected profit to the government over the next decade from its student loan program. I say Sen. Warren shouldn't have been dinged at all.

The budget currently scores lending programs, like those for student loans, by calculating the value today of both cash outflows and inflows over the life of the loan, accounting for default risk, prevailing and expected interest rates, and the timing of the flows. But because the government has unique characteristics as a lender relative to the private sector, the budget scores them differently than would, say, a private bank:

...private lenders face various disadvantages relative to the government and they reasonably want to be compensated for them, so they charge more for their loans. They can't borrow as cheaply as Uncle Sam, they can't tax, they can't print money, they won't make loans unless they expect to make a profit on them, and for all these reasons, they're more risk averse. So they add a risk-premium to their interest rate.

The FVAers want the budget to reflect that risk premium. Even though they agree that the government faces neither those risks nor their associated costs, it would face them if it were a private lender. To budget otherwise makes the government appear to be a less risky lender than the private sector. Which, as noted, they agree the government is... making these additional costs phantom ones.

In this regard, team FVA as cited by the WaPo get one important part of this wrong: they suggest that they current method fails to fully account "for the risk of default." Not so. As noted above, accounting for default risk is a critical part of the current method. Moreover, as the Center for American Progress has pointed out, these methods have provided an accurate accounting of the actual cash flows to and from the government:

Our review of federal government credit programs back to 1992 shows that on average the government is quite accurate in its risk pricing. In fact, the majority of government credit programs cost less than originally estimated, not more.

So, let's review: Sen. Warren, in an argument in support of affordable student loans, used official CBO numbers based on sound and accurate accounting, yet was scored by the WaPo factchecker as getting the numbers half-wrong (she got two out of four "Pinocchios").


OK, the fact-checker's not nuts. The reason he went there is because among those who raise questions about the use of the current method are the CBO themselves. That's a big deal in this town and I think it's fair to point out that the budget office doesn't necessarily agree that the added cost of market risk should be left out of the budget, even though the USG is not "the market" and doesn't face that same risk set.

On the other hand, I wouldn't say CBO's view is monolithic either, at least not in an historical sense. Former CBOers who are now my colleagues at CBPP as well as former CBO director Robert Reischauer strongly support the current method and oppose a change to FVA.

Here's the thing. Underlying this disagreement are two different accounting methods, and there are arguments for both. Honorable, knowledgeable scholars line up on both sides of the argument. The "correct" method for scoring government loans is thus a matter of opinion. Until that opinion is resolved, the correct thing to do -- the thing that is by far most commonly done by honest brokers in DC budget debates -- is to use CBO's published numbers.

That's what the Senator did and I encourage her to continue to do so.

This post originally appeared at Jared Bernstein's On The Economy blog.