I propose we reinstate old-fashioned banking, where bankers know their borrowers, by tightly limiting what banks can do: nothing besides making loans to individuals and nonfinancial businesses--after old-fashioned due diligence--and simple hedging transactions. The standard would simply be whether the loan (or hedge) can be monitored by bankers and examiners who don't have PhD's in finance, and whether the risk is one that bankers would take if it was their own money
Others with similar proposals have called this a separation of "utility banking" from "casino banking." The utility banks get to have a government guarantee for their deposits, but they have limited asset choices. Casino banks can choose whatever assets they want, but they have no guarantee.
As you know, my view is that all attempts to regulate guarantees degrade over time. The casinos will try to figure out a way to gain access to guaranteed funds. And the utilities will try to figure out ways to invest in risky assets. And everybody will be hiring lobbyists to try to open up loopholes. Eventually, the regulatory fortress will be breached.
I am in the camp that wants to break up big banks, based on the hypothesis that is more synergy between politicians and big banks than between politicians and small banks. That is, I admit, a shaky hypothesis.
I also think that we ought to create a crime called "taking imprudent risks with government guaranteed funds." No amount of regulatory approval for an activity can serve as an excuse. If you are responsible for the failure of a government-guaranteed institution, and the records show that you could have prevented that failure with sensible precautions, you go to prison. By that standard, the CEO's of Freddie and Fannie in 2006 and 2007 would be in prison today. Top executives at AIG and a number of big banks, also.
Of course, the goal of the criminal statute is not punishment, but prevention. Right now, taking risks is a heads-you-win, tails-the-taxpayers-lose proposition. I want "tails" to mean "you could get sent to prison."
I agree with Bhide about the hazards of securitization, as opposed to "hands-on" lending decisions. However, I think he overstates the role of mathematical modeling in promoting securitization and understates (ignores, actually) the role of political lobbying.
Overall, I think A Call for Judgment can stimulate a lot of thought and discussion. But I found myself wishing that Bhide would have used a blog rather than a book as a vehicle of expression. I think it would be more satisfying to bat these sorts of ideas back and forth in the blogosphere.