I guess I was supposed to debate Simon Johnson in this NPR web chat. Simon blogs at Baseline Scenario and is a major star in the economics profession. I don’t feel particularly good about the chat.

1. The format seemed to me to produce a high noise/signal ratio. Some of this may have been inexperience on my part. But if I were to get the same impression from trying it a second time, I would abandon further efforts.

2. The title of the session was “Should U.S. Nationalize Banking System?” I took this literally, and the answer is obviously “no.” But the real question was, “should the U.S. nationalize failed banks?” There the answer is just as obviously “yes,” if by nationalize one means take over, carve up, and sell back to private investors. Simon and I were in violent agreement on that.

3. That leaves some tough issues, which we did not really address.

First, how do we decide what is a failed bank? To see why this is hard, think of an insurance company that has sold a lot of flood insurance policies on the Gulf Coast. If there are no serious hurricanes, they will be fine. But if they had to sell their insurance business to another company, no one would buy it: Given the probability of a hurricane at some point, other companies do not think that the premiums are sufficient to cover the expected losses and to provide adequate compensation for risk and return on capital.

The upshot is that our hypothetical insurance company is insolvent on a mark-to-market basis (it cannot sell its business at a price that would give it positive value), but it still could function as a going concern (as long as no big hurricane hits).

Think of the bank’s holdings of mortgage-backed securities as analogous to insurance that it has sold. The hurricane would be another drop in house prices of, say, 30 percent. If prices don’t fall that far, then the securities will have good value and the bank will be fine. But right now, the market is not sure that house prices won’t fall that much, and so the market value of those securities is such that the bank is insolvent.

What should the government do?

If you close the bank right away and sell its assets, then the taxpayers will lose money dumping the mortgage securities into a market that doesn’t want them. This is what I would do, because I believe in taking your losses early, rather than letting them build up. Again, I think Simon would agree. However, you have to realize that the investors who buy the mortgage securities may profit eventually, without the taxpayers getting any of the “upside.”

The alternatives, of having the government take some kind of equity stake in the banks, are even uglier. We’ve been doing that, and there are people who want to do it a lot more, and for the life of me I cannot see the attraction.

[UPDATE: This NYT story captures very well why valuing “toxic assets” is not possible. That makes it impossible to come up with a reasonable model for a public-private partnership via equity injection. It also means that if you put the bank out of its misery, there is a nonzero probability that its assets will not be as bad as the market thinks they are, so that after the fact you have shut down the bank needlessly. I would be willing to live with that, as long as clear rules are being followed.]