1. I am worried that breaking up the biggest banks would not make the system safer.
I do not believe that having smaller banks would have prevented the 2008 crisis or the next crisis. I do not believe that regulatory policy can prevent such crises. My only solution for systemic financial risk is to try to make the financial system easier to fix when it does break. I think it is easier to fix when there is less debt, so I would look for ways to reduce incentive to take on debt. Instead of subsidies for mortgage borrowing, how about subsidies for down-payment saving? Instead of favorable tax treatment for debt, why not favor equity–or at least be neutral between the two? etc.
For me, breaking up the banks is not a safety issue. It is instead an issue of restoring democracy and the rule of law. Really big banks are crony banks, whose interactions with government officials are at the highest levels. Instead, I would like to see the biggest banks dealt with by career civil servants who are following clear, predictable rules and guidelines.
2. I am worried that it would be difficult to define size.
Once you decide that banks above a certain size should be broken up, you need a definition of size. Among the problems with doing this, the first one that comes to my mind is accounting for derivatives. If you ignore derivatives, then in very short order banks will mutate their loan portfolios into derivative books. But if you try to include derivatives, then the whole notional-value vs. market-value argument is going to kick in. Also, gross exposure vs. net exposure.
Instead, I am attracted to using the amount of insured deposits as a measure of size. It is a clean measure that cannot be gamed using accounting transactions. It is a measure of the potential impact of the bank on the FDIC. Charging a risk premium that is graduated by size would be a reasonable rule-of-law approach to discouraging large banks. Banks could avoid the risk premium by spinning off branches. The giants that were assembled by mergers could be dis-assembled by spin-offs.
3. I am worried about large shadow banks.
You can do a lot of banking without a lot of deposits. You can finance with commercial paper. You can finance with repo. You can write a ton of derivatives. I do not think that this is bad per se. But, just as with large commercial banks, large shadow banks could acquire the political power of large commercial banks.
I welcome suggestions for dealing with shadow banking. I am not thinking about how to reduce the risk of shadow banking. My view is that systemic risk is systemic risk, and you cannot get rid of it by breaking up banks.
What I am concerned with is the political power that might be concentrated in a large financial institution. The problem is that, once you get away from deposits, measuring “large” becomes quite tricky.