14 June 2010

Something else about Kling's proposal

Forgot to mention this in the last post:

EconLog | Arnold Kling | Matching Narrative to Policy

2. My policy recommendations match my narrative. Assume that regulation will be clumsy, and aim for a system that is easy to fix as opposed to hard to break; get rid of all policies that encourage lenient subsidized mortgage credit. and break up the 10 largest banks into about 40 banks.
I like the easy-to-fix vs. hard-to-break distinction, but I'm less impressed with trying to break 10 banks down into 40.

My fear it that that this only works if you have diversity of behavior among the 40 smaller banks. If they're all pursuing similar policies, or if even 8 or 10 of them are pursuing similar policies, then you've still got big chunks of the finance sector all getting themselves into trouble at the same time. I don't know much about the S&L crisis, but I do know we had thousands of small banks rather than a few big ones and we got stuck with a ~$125B bailout and ~750 failed banks.

Having many small units is often more robust than a few larger ones (e.g. machine learning ensembles, and even mortgage-backed securities). But this is only true so long as there is statistical independence in the behaviors of the underlying entities. If Jim repaying his mortgage has nothing at all to do with Sally repaying hers then you're better off owning half of each than owning either outright. If their failure to pay is related you're screwed whether you own pieces of both or just one. (Approximately.) Same deal with a bank: if there's going to be a correlation between failures of banks then dividing them up doesn't help you much, and I think there's likely to be a high degree of correlation between bank strategies, if for no other reason than the "no one ever gets fired for buying IBM" effect.

I respect Arnold Kling a great deal though, and he undoubtedly knows more about this than I do, so maybe there's a very good explanation I'm missing. If so I'd really like to hear it.

2 comments:

  1. I don't know much about the S&L crisis, but I do know we had thousands of small banks rather than a few big ones and we got stuck with a ~$125B bailout and ~750 failed banks.

    I had a role at the Resolution Trust - the company chartered to clean up the mess. A very minor role - I worked for the contractor that provided Help Desk and LAN support. I got curious and read up a little bit.

    The organizations that precipitated the problem were not banks but Savings and Loans. There was a wave of deregulation in the S+L industry - suddenly the rules were relaxed and people could make money. The guys that started S+Ls were not bankers but guys out to make a quick buck. And they did, using the S+Ls as piggy banks. These guys just bought stuff and treated S+Ls the way the Mafia treats restaurants in Goodfellas.

    It wasn't - in my opinion - the number of thrifts that caused the problem but a lack of regulation and morals.

    ReplyDelete
  2. I don't think it's the number of institutions that's the problem (then or now), just that more firms without the right rules isn't going to cure anything.

    ReplyDelete