The Inconsistency of Return–Based Style Analysis

Abstract

The authors demonstrate that the usual application of the return–based style analysis relies on commercially available indexes that exhibit extreme multicollinearity. The subsequent results are volatile and have little meaning. As a result, the authors argue that implementing return–based style analysis with commercially available indexes can result in accurate and inappropriate investment decision–making. Even with multicollinearity, however, they demonstrate that the results of the analysis can be meaningful as long as the explanatory variables properly capture the investment objective of the portfolio. The authors conclude that the only way to implement return–based analysis is to use portfolio–specific benchmarks that properly capture the investment objectives of the portfolio.