Investment factor growth has exploded to the point that are now close to 400 identified through research published leading finance journals. This factor explosion has been called the factor zoo and it is overcrowded. A careful review of the statistical testing procedures will tell us that it is highly unlikely that all of these identified factors will be true. Many are statistically significant only by chance given a multiple testing problems. Additionally, if we impose transaction costs on these factors, their profitability will be diminished and in many cases not useful.
Campbell Harvey and Yan Liu has written extensively on the statistical problems with the testing of factors and the incentive problem that is inherent with academic research. Negative papers are not published and papers that show "significance" may have been subject to the extensive testing of alternatives. There may be too many false positive. This requires increasing the threshold for significance. When these thresholds are applied, the factor zoo declines significantly.
The authors have now generated further work with the their paper "A Census of the Factor Zoo". It provides a google sheet list of all of the factors that have been studied over the last few decades. This information is now available to the public. See Factor Census https://tinyurl.com/y23ozzkc. They are suggesting a citizen science project that allows researchers to add to their database. This is google link is very informative and useful for any researcher who want to study factor investing. Understand the zoo and make better decisions.
No comments:
Post a Comment