Among the central topics of debate on the FinReg proposals is whether size matters.
Some (or many) think that downsizing TBTF is absolutely essential to the new legislation. The Volker rule which would limit activity is but a step in that direction, albeit an important step. Another, more decisive step would also limit the amounts of deposits that any banking instutution can keep and the amount of aggregate assets that it can have (on an off the balance sheet).
Others, like notably Paul Krugman in today's column, believe that size is not an issue. The basic argument here is that even if the financial institutions are small, in a pinch they'll have to be bailed-out anyway and a run on big banks is neither more nor less likely to happen than a run on small banks. In fact, both the Great Depression (and many bank panics that preceded it) as well as the more recent S&L crisis featured such runs on rather small financial institutions. And, as Krugman points out, it was very likely a mistake for the government to let the banks fail willy nilly during the depression.
So where do I come down on this? The former view - and here is why.
Krugman and others of the same view a missing the point and the problem of bailouts. It may not, indeed, necessarily be the case that we would let a bunch of small banks and financial firms fail in the next financial crisis without trying to prop up the system in some way or another - putting aside whether this is a good thing to do or not for a moment, the politics of the moment usually overwhelm sound thinking in these types of situations.
However, there are key advantages to having a world composed of small instututions from regulatory point of view at the time of a crisis:
1. When financial instututions are less inter-connected and have fewer tentacles, it is easier for regulators to operate them in receivership without much help from their management. In many cases, this should obviate the need for one of the most puke-inducing part of the AIG bailout where the chumps who caused all the trouble were nevertheless kept on board because they were the only ones who could help untangle the mess they created.
2. The most egregious players could be simply let go of, and made an example of. This did not happen in the last crisis - because the most egregious players were also some of the biggest. Lehman WAS made into a test-case, but that did not exacly ended well.
3. You could even experiment with different resolution regimes and test in real time which one is working better. It is highly unlikely that the whole system will be brought to its knees in a matter of days - what we'd probably have is an accelerating wave of failures that would be spaced out over the course of several months. This was the case in the current crisis - with Bear Stearns being the canary in the goldmine. But because Lehman and other systemically important players managed to put off the inevitable by a few more months through creative accounting, the regulators were basically caught flat footed both by the Bear Stearns failure and by the avalanche of near-failures that followed 6 months later. Imagine now, instead, that we'd have a slow at first but gradually building wave of small instututions going under: regulators would have a much better chance of developing a coherent responce by the time the crisis peaked in the Fall of '08.