• The Idea Farm
  • Posts
  • Elections, Managed Futures, and Four New Factors

Elections, Managed Futures, and Four New Factors

October 16, 2012

Below is our tenth issue of The Idea Farm Weekly!  If you are enjoying the emails please feel free to pass along to your friends and colleagues.

In this issue we include some research from Steve LeCompte at CXO Advisory, one of my all-time favorite resources.  Steve sifts through reams of academic studies and distills the findings into a coherent summary.  Considering how many studies come out on a daily basis (most of which are nearly unreadable), it is incredibly useful to have someone condense the material and even more importantly, offer a practicioner’s view on cautionary points.  We have included four recent summaries, and here is an example from the Four Factor study below:

—-

“The most commonly used models of stock returns are the Fama-French three-factor model (based on market, size and book-to-market factors) and the related Carhart four-factor model (adding momentum as a fourth factor). Is there a better model? In their September 2012 paper entitled “Digesting Anomalies: An Investment Approach”, Kewei Hou, Chen Xue and Lu Zhang propose a new factor model consisting of: (1) the market return in excess of the risk-free rate; (2) the difference in returns between stocks with small and large market capitalization size; (3) the difference in returns between stocks with low and high investment (measured as annual change in total assets); and, (4) the difference in returns between stocks with high and low return on equity (ROE). Using stock return and accounting data for a broad sample of U.S. stocks during January 1972 through December 2011, they find that:

Comparing new model factors to legacy model factors:The size factor generates a gross average monthly return of 0.31%, effectively the same as the Fama-French size factor.

  • The investment (change in total assets) factor generates a gross average monthly return of 0.44%, playing a role similar to that of the Fama-French book-to-market factor.

  • The ROE factor generates a gross average monthly return of 0.60%, playing a role similar to that of the Carhart momentum factor.

With respect to explaining commonly accepted stock return anomalies:

The investment factor largely accounts for the predictive powers of book-to-market ratio, net stock issuance, accruals, market leverage, long-term past return, earnings-to price ratio and composite issuance (change in market equity not attributable to stock returns).

The ROE factor largely accounts for the predictive powers of short-term past return, earnings surprise and financial distress.

Investment and ROE factors jointly contribute to the predictive power of idiosyncratic volatility.

  • Relative to the Carhart four-factor model, the new model:

  • Outperforms in explaining earnings surprise, idiosyncratic volatility, financial distress, net stock issuance and composite issuance (as well as, naturally, asset growth and return on equity) anomalies.

  • Performs similarly in explaining size/momentum, abnormal corporate investment and size/book-to-market anomalies.

  • Underperforms in explaining the accruals anomaly.

The stronger linkages to theory of the new model compared to the more empirical, datamined Carhart model support belief that its explanatory power will persist.

In summary, evidence indicates that investors and researchers should consider replacing legacy factor models of firm stock returns with one based on market, size, change-in-assets and return-on equity factors.

Cautions regarding findings include:

  • Progressive testing of many different factors and combinations of factors on the same sets of stock return data builds data snooping bias, such the the performance of the best models deteriorates when applied to new data.

  • Factor return calculations ignore trading frictions associated with periodic portfolio reformation and costs of shorting, thereby substantially overstating returns implied for strategies seeking to exploit these factors.

—

Here are five sample summaries sent to subscribers:

For more information and subscription details:

Contact:  [email protected]

AboutCXOadvisory.com presents financial markets models, research summaries, analyses and reviews designed for objective, unique and concise value to serious investors, financial advisors and money managers — a modicum of actionable conclusions filtered from a very noisy environment. The default approach is to challenge any and all conventional market wisdom with analytical skepticism. Part of CXOadvisory.com is subscriber-only, and part is free. The CXO Advisory Group LLC of Manassas, Virginia maintains the site.

CXO Advisory Group LLC founder Steve LeCompte is editor of CXOadvisory.com, responsible for all site content. His prior experience includes: project officer on the staff of Admiral H.G. Rickover at Naval Reactors; engineering and program management for complex systems with International Business Machines; senior consultant and executive manager for market research and new media development with units of International Data Group, including International Data Corporation; and, longstanding private investor in stocks, mutual funds, exchange-traded funds and options. He holds a B.S. in physics, summa cum laude, from Miami University and an M.S. in Physics from the University of Michigan. He is a graduate of the Program Management Course of the Defense Systems Management College.