1. Creating relatively mechanical (and therefore not politically gameable) rules that increase capital requirements with size.

2. Creating ways of credibly threatening large institutions with failure, for example with living wills and other mechanisms to forestall fears of chaotic failures.

3. Completely break up the GSEs. Sell their good assets to private banks, and put their bad assets in a “bad bank,” with the goal of eventually winding down its affairs completely.

4. Shift monetary policy. Part of this is getting a less political Fed that is willing to act more like the ECB. Part of this is technical. Targeting inflation using the CPI, for example, seems to create problems because the CPI doesn’t capture rises in asset prices, which can be driven by monetary expansion. One suggestion that I have heard is that rather than targeting interest rates or the CPI, the Fed should target nominal GDP or some other measure of price volitility that captures swings in asset prices.

The notion that the crisis resulted from shifty mortgage brokers, however, is ultimately implausible, and I think that potential for upsetting essentially harmless transactions by requiring risk retention is pretty high. Rather than telling fairy stories about agency costs, we need to look at the fundamental question of why there was such a huge appetite for risk. Hint: It wasn’t because fast talking mortgage brokers convinced Bear Sterns to buy their junk. They made the junk because Bear Sterns et al wanted it.