Robust Regression and Fama-French Factors, from Doug Martin

2019-10-16 · Doug Martin
Original source

Victor,

This morning I happened upon your article at DailySpec, was pleased to see the use of robust regression, and have the following comments.

1. The package WLE package is no longer available (CRAN archived in June 2018).

2. We have a new package RobStatTM on CRAN, which is a companion to the 2nd edition of our book Robust Statistics: Theory and Methods, published earlier this year by Wiley, see the Wiley site, where there are downloadable materials, especially scripts to reproduce all the examples in the book.

3. Robust statistics has been almost totally over-looked in quantitative finance. Last time I looked there was still only one paper of note in the Journal of Finance (Knez and Ready, 1997, who were the first ones to show that the Fama-French 1992 conclusion the returns are negatively related to size is driven be a small fraction of outlier in small size (log of market cap in $M) firms.

4. I have been doing research on robust statistics for a long time, especially applications to quantitative finance. And more so in recent years now that I am retired and have more time for research. If you are interested I could send you a published paper on robust betas, and a talk on Fama-French 1992 Redux with Robust Statistics (I gave the latter at Soros Fund Management this last June).

Best regards,

Doug

P.S. I used to play squash and followed your squash career many decades ago.