Charles Lane had an editorial at the Washington Post while I was out about fair value accounting:
There is general agreement that federal budgets should include a dollar figure for the estimated lifetime cost of each year’s new lending. Congress adds up all the expected cash flow associated with a particular credit program — interest and principal payments, fees, expected defaults — and calculates its “present value,” based on a notional interest rate known as the discount rate.
Under current law, however, Congress bases these estimates on the government’s ultra-cheap cost of borrowing. That means the calculations are done as though everyone were as default-proof as Uncle Sam, which understates the costs and risks to taxpayers.
The Congressional Budget Office believes “fair value” accounting, which adjusts the discount rate to reflect the risk of widespread defaults during downswings in the business cycle, would be more accurate.
Yglesias responded here. This is entirely a fight over the discount rate to use for estimating government loans and whether we should be ”measur[ing] the costs of federal loans and loan guarantees at [private] market prices.” As those of you who follow the global warming economics debates closely know, slight changes in discount rates can have huge policy implications.
Jason Delisie of New America and Economics21 also argued for fair value accounting at The Agenda Blog in light of a post I wrote; he was also kind enough to invite me to a big Economics21 panel he organized on fair value accounting which is online here.
A few points.
1. I tried to make an example using swaps but it didn’t convey the point well, so let’s try it a different way. Imagine two identical firms A and B, who are making identical loans. They should use the same discount rate, correct?
Now imagine that A has a lower borrowing cost of capital than B for whatever exogenous reasons – ratings, savings, size, etc. It must be the case, provided that cost of capital is a monotonically increasing input into the discount rate (and I know of no model in which that isn’t the case), that A has a lower discount rate. A is the government here under these circumstances. That doesn’t mean that the discount rate should be the cost of borrowing per se, but financial logic dictates we don’t take the same market discount rate as B for A. I haven’t seen a sufficient answer to this argument.
2. It’s ironic that Lane pushes the idea of “downswings” so aggressively, because the whole reason the government now stands behind the mortgage market is because the entire private market collapsed during the downswing. Student loans weren’t going to go out and the securitization channel is a swamp of fraud, missing or forged documents and bad incentives on the servicing end. Meanwhile the government churns along without a crisis.
I’m not sure what to make of the argument that private lending channels couldn’t survive the downturn without their costs exploding and therefore the government, which is surviving fine, should use the discount rates of the private market – more prone to collapse! – because of a risk of a downturn.
3. Most of these arguments, especially in response to #1 above, are predicated that “shareholders” of the government, i.e. citizens, need to be compensated some amount X – equivalent to private market returns. It’s not clear to me why this is the case or how to determine it other than a normative argument about what government is. I brought up that this implies a normative arguments as to what the government is supposed to do in the panel and got some really bad looks from the wonks in the audience; Yglesias also came to the same conclusion in his response to Lane. But fair value accounting assumes that the government is just like any other firm, when in fact it has unique abilities (compulsion, scale, longevity, lack of a profit-motive, currency, etc.) that distinguish it from anyone private firm.
4. The two CBPP wonks at the panel, Richard Kogan and Paul N. Van de Water, went medieval on everyone else; it was like watching a kung-fu fight. That whole crew is awesome. Read their entire report, but note their points about “phantom costs” and that if we need a higher discount for risk-bearing is it weird that we don’t need a lower discount for risk-mitigating programs like Social Security and unemployment insurance.