Foster School of Business, University of Washington, 2011. – 29 p. – ISBN: N/A
This tutorial examines three factor models: the Fama-French factor model, a BARRA-type industry factor model and a PCA factor model. I discuss the mathematical aspects of each model and provide an R implementation, which I then use to construct minimum-variance weighted portfolios for each model. The resulting portfolios are then applied to new data.
Introduction and OverviewTheoretical BackgroundTime series factor models
Cross-sectional factor models
PCA statistical factor models
Algorithm Implementation and DevelopmentImplementation of the Fama-French three-factor model
Implementation of the BARRA-type factor model
Implementation of the PCA factor model
Computational ResultsResults for the Fama-French three-factor model
Results for the BARRA-type factor model
Results for the PCA factor model
Summary and Conclusions
R Script
Portfolio weights for BARRA-type factor model