Archive for September, 2008|Monthly archive page

On the point of Regulation

In Money on September 27, 2008 at 17:37

A statement from SEC chairman Christopher Cox

The last six months have made it abundantly clear that voluntary regulation does not work.

Greenspan has argued that regulation is best done by counterparties and banks; not by the government:

I do believe bank risk managers and loan officers are more knowledgeable than government bank regulators. Bank loan officers, in my experience, know far more about the risks and workings of their counterparties than do the bank regulators that examine those counterparties.

He even punts at the end:

We have tried regulation ranging from heavy to central planning. None meaningfully worked. Do we wish to retest the evidence?

Greenspan implies these boom-bust cycles are inevitable in capitalism. However, forgoing regulation completely (on hindsight :)) seems to be like forgoing crime prevention efforts because we cannot have perfect crime prevention.


How Wall Street Lied to Its Computers – another nyt aritcle

In Money on September 19, 2008 at 05:29

Another nyt article. This time in the Bits Blog section of all places. Anyway, more reasons why the models failed:

  • The models looked at years of history instead of a few previous months. This makes no sense to me: “There is no data like more data” and all that.
  • The products were so complex that no one knew how to assign risk to them.

The article concludes that the blame also falls on the experienced non-quants in these investement banks for improper supervision.

Short Term Memory – Roger Lowenstein

In Money on September 7, 2008 at 22:27

Roger Lowenstein has written a nice simple take on the present housing mortgage crisis.

  • Models are not built on complete evidence; they do not know how to model unseen events.
  • Events become more correlated during stressful times.
    • LTCM assumed independent spreads which become correlated after the Russia default.
    • The present crisis resulted from a bad assumption that the housing markets among states (e.g., CA and FL) are independent. So the odds of subprime defaults in both states were assumed to be exponentially lower than in either of them.
    • (But these new investment vehicles (subprime, Alt-A, etc.) themselves were inherently bad products. I do not know what kind of default processes were assumed.)
  • However, modelers are not alone to blame. Regulators knew what can happen with unregulated growth after LTCM. However, follow-up was not done. Greenspan was strongly convinced (and still is :)) that investment banks are the best self-regulators. (He has also mentioned in some articles that he himself does not know what is the best method of regulation).
  • “The United States had never suffered a nationwide contraction in housing prices” – is this true??

My first thought on launching Chrome

In Tech on September 3, 2008 at 04:14