Tuesday, May 20, 2008

Quant this!

Over the years, technical analysis has had its tools usurped by quantitative analysis and behavioral finance. But in an ever changing world, assumptions, especially those about tails and linearity, can kill the mechanical systems of the former. Hmmmm, sounds like portfolio insurance and Long germ Capital Management, doesn't it?

Here is an excerpt of an article found in the Financial News (US) and posted to a chat room by the world famous John Bollinger:

"Quantitative fund managers, who use computer models rather than human judgment to pick stocks, have continued to suffer since the credit crisis threw their calculations into confusion last summer.

They are fighting back with new models and new ideas, but are running into investor skepticism.

Three quarters of fund managers—quant and non-quant alike—agree the outlook for computer models is troubled.

It will be difficult for them to generate returns because they are all following similar market factors, according to a study last week from the CFA Institute (that's the fundamental analysis society to which all the Wall Street analysts belong), a trade body for fund managers.

Managers are responding in three ways, often in combination, the study found.

They are developing extra models, looking for fresh sources of information to feed into the models, and some are introducing a level of more traditional, human insight into their process."


End Quote


So, they are changing by adding more models? Garbage in, garbage out. At least they are adding a little human insight into the mix and guess what? that is technical analysis in an different dress.

Quants and all the propeller heads on Wall Street are a bigger herd than mom and pop retail investor.

2 comments:

Anonymous said...

I find the timing of these bear market rallies interesting as they tend to start at the beginning of a quarter (after as significant pull back) then run for a full quarter ignoring quantitative, economic, and sometimes technical analysis rules. Is it possible that financial manager's compensation schemes drive behaviors that should be incorporated into models? It appears there may be a significant amount of money in the market not always invested for the end-customer's best interest but instead for the fund manager's quarterly statistics.

Doug said...

great points "anonymous", definetely some food for thought...