Saturday, December 13, 2008

"Does the SEC do any quantitative research ?"

From The Big Picture:

"Given how easily identifiable the Madoff/Ponzi scheme was mathematically, I must ask a simple question:

Does the SEC do any quantitative research ?

There is little evidence that the SEC is using any of the quantitative methods — now so common on Wall Street — for searching out and indentifying fraud.

I would suggest to the incoming head of the SEC to put together a blue ribbon of math professors, quant scientists and algo specialists to develop a few basic programs that ferrets thru market, options, and perfromance data looking for aberrational data series, and leading to criminals and fraud artists.

>

See also:
Numbers follow a surprising law of digits, and scientists can’t explain why
Lisa Zyga
PHYSORG.com, 11:26 EST, May 10, 2007
http://www.physorg.com/news98015219.html

Benford’s Law Part 1 - How to Spot Tax FraudIntuitor.com
http://www.intuitor.com/statistics/Benford’s%20Law.html

Be Wary of Serial Correlation (December 2008)
http://www.ritholtz.com/blog/2008/12/be-wary-of-serial-correlation/

Multiple Red Flags in Madoff Case
GREGORY ZUCKERMAN
WSJ, DECEMBER 12, 2008,
http://online.wsj.com/article/SB122910977401502369.html

Bernard Madoff’s Prison Reading List
http://thereformedbroker.com/2008/12/13/bernard-madoffs-amazoncom-prison-reading-wishlist/

Math Models are tools. The must be used, and interpreted. Given Paul Wilmott's earlier post, I'm not so sure this is that easy except in hindsight. However, I have to admit, that this does explain how the amount of Fraud, Negligence, Fiduciary Mismanagement, and Collusion, which I believe happened here, could occur. Nobody was looking for it. I do like the idea of more ferreting.

This is a good point, but I still believe that my wide net idea will work the best.

No comments: