Using the Industry-Based Fama-French Model to Evaluate Industry Portfolios

Abstract

In this paper I create the industry-based Fama-French 3-factor model by constructing the three Fama-French factors for each industry group separately. I then use this model together with the traditional Fama-French model to evaluate the returns of 41 monthly industry portfolios and I compare the results. Three sets of stock portfolios are used in my analysis: (1)Stocks in a single industry, (2)Stocks in industries in the same industry group, (3) Stocks in a industries not in the same industry group.

I show that the modified Fama-French model outperforms the traditional Fama-French model when assessing the performance of those industry portfolios. This is especially true for industry portfolios within small industry groups like mining where the regression R2 improved by 34.8%, 68.8%, 28% and 1052% for the underlying Mining, Coal, Oil and Gold industries. Actually in 16 out of the 18 non-manufacturing industry portfolios, R2 improved. These results imply that within each industry group asset space, the new industry-based model tends to be more efficient than the traditional Fama-French model.