Are commonly used stock market indicators more predictive for some subsets of stocks than for the stock market overall? In the November 2008 update of their paper entitled “How Predictable are Components of the Aggregate Market Portfolio?”, Aiguo Kong, David Rapach, Jack Strauss, Jun Tu and Guofu Zhou analyze return predictability for various subsets of the overall U.S. stock market, defined by portfolios sorted into 33 industry, 10 market capitalization and 10 book-to-market ratio segments. They consider 14 economic variables and lagged returns for 33 industries as predictors. Using economic indicator and industry/size/book-to-market return data from the end of 1945 through 2004, they conclude that:
- Out-of-sample (1966-2004) tests indicate pockets of substantial real-time return predictability, as follows:
- There is significant predictability for 23 (16) of 33 industry portfolios using the 14 economic indicators (lagged industry returns) as predictors. Predictability is strongest for construction, textiles, apparel, furniture, printing, automobiles and manufacturing.
- Economic indicators significantly predict returns for all size and book-to-market portfolios, with little difference across the ranges of size and book-to-market ratio.
- Lagged industry returns significantly forecast returns for the seven smallest market capitalization portfolios, with predictability clearly increasing as size decreases.
- Lagged industry returns significantly forecast returns for the two highest book-to-market ratio portfolios.
- Results are remarkably consistent across the out-of-sample evaluation period.
- Lagged industry returns are generally more powerful predictors of subset portfolio returns than are commonly used economic indicators.
- Degree of sensitivity to macroeconomic fundamentals/risk and information-flow frictions (indicated by industry concentration/capitalization) help explain variation in market subset predictability.
In summary, investors/traders may be able to enhance market timing results by focusing on the most predictable styles/industries (for example, via exchange-traded funds).
