I ran across an interesting piece by George Selgin. He makes a point that a lot of people miss, including, in this case, Gary Gorton. History gives us useful information for organizing our thinking about financial crises (and other events of course), but many people attach too much weight to U.S. experience. This matters in particular for how we address the issue of whether financial fragility is inherent or induced.
If we confine our attention to U.S. financial history, we might think that banks are inherently unstable, as Diamond and Dybvig theorized. We also might think, if we look at the performance of behemoths like Bank of America and Citigroup, that too-big-to-fail could be an inherent problem in banking. But, as Selgin points out, the notion that banking arrangements are inherently prone to panics and too-big-to-fail problems flies in the face of experience in other countries - Canada in particular.
The kind of thinking Selgin objects to leads to inappopriate policy conclusions. Gorton in particular seems to subscribe to the idea that a fix which will make the banking system safe is some version of narrow banking - restricting what financial intermediaries can hold as assets to back any type of transactions account. But there are other - and seemingly better - ways to get safety (if you want it), as the Canadians have shown.