John Hempton writes,

If the bank runs for three years it will again be solvent. If it runs for less than six years it will be fully and adequately capitalised. This is in fact how the Japanese mega-banks recapitalised. I blogged about it here. At the spreads in America – which are several percent – the recapitalisation will happen much quicker than this and much quicker than in Japan. Indeed it is likely that with quasi government guarantees for bank funding and market rates for bank loans the spreads would be over five percent in America right now.

Thanks to Mark Thoma for the pointer. [UPDATE: Luis Zingales also is pondering the issue of rules and discretion.]

The performance of a financial institution depends on its cost of funds. As I pointed out in July, Freddie Mac and Fannie Mae look a lot better if their debt carries a low risk premium than if it carries a high risk premium.

What I think Hempton is saying is that if the government lends to banks for a few years using its low risk premium, then the banks will earn enough to rise out of insolvency.Hempton’s analysis raises the question: Suppose that the banks (and Freddie and Fannie) could come back if the government lent them money for a few years. Is there a public benefit in that?

Right now, the U.S. government enjoys borrowing privileges over everyone else. The spread between Treasury borrowing rates and private borrowing rates is high. Perhaps it should use those privileges to prop up banks. But perhaps there are higher rates of social return to be found elsewhere.

Ordinarily, I would want the market to decide where to invest. I think that entrepreneurial trial and error, with responsibility for failure borne primarily by the decision-makers, is the best approach.

In theory, we have a set of rules for when to close financial institutions. According to those rules, many banks should have been shut down by now.

Instead, regulators are using discretion to bail out banks. This use of discretion benefits some banks, but it makes it more difficult for other banks to compete. As a result, the risk premium paid by financial institutions rises. The higher the risk premium, the more banks get in trouble, and the more discretionary bailouts are undertaken.

I think we need to sort out the relationship between healthy banks and healthy markets. I think we could have healthy markets without healthy banks. That is, we could close a lot of banks and still have healthy markets. However, as long as the decision about what is healthy and what is not is made through weekend meetings among policymakers, private financial intermediation is bound to be driven out by government.

If we were using rules, then those rules might shut down banks that could survive if they were granted risk-free borrowing privileges by government. However, I would argue that such banks ought to be shut down. My hope is that under a system of rules, the risk differential between private and government borrowing rates eventually would narrow, so that there would not be the temptation to use the government’s borrowing privilege to bail out banks.