Monday, August 27, 2007

Supplemental Reading

In a prior post, I wrote:
Making the interviewer squrim and stutter and finally say, "I don't know," when asked how much their quantitative research group lost last week and the week before: priceless

It occurred to me from a reader comment that the significance of this question might not have been clear from the context. Last week, the Washington Post ran a good article talking about the massive losses encountered by funds that rely on computer algorithms for trading. Just for the record, I interviewed with one of the organizations mentioned in the article.

Another article that ran originally in the Wall Street Journal, includes a key issue that I've often wondered about:
The risks to quantitative investing may be rising. Even if they don't share the same statistical models, quant funds share similar approaches to the market. They are schooled in the same statistical methods, pore over the same academic papers and use the same historical data. As a result, they can easily come to similar conclusions about how best to invest.

In essence, these massive funds become something of an intellectual and algorithmic monoculture. Such a strategy might give the funds a short term edge, but it also means that the market can be shaken easily when the weaknesses of the strategies are exposed. There's no notion of resilience or immunity in that ecosystem.

I haven't thought deeply about this, but I've speculated that there exists one or more adaptive investment strategies that, when tuned to the weaknesses of the quant monoculture, could cause a massive transfer of wealth away from the inbreds. (assume faux evil genius facial expression) Maybe I should hunt down some texts on computational finance and start plotting?
blog comments powered by Disqus