On the usefulness of linear factor models in predicting expected returns in mean-variance analysis

Jonathan Fletcher, Joe Hillier

Research output: Contribution to journalArticle

Abstract

This article examines whether using multifactor models to forecast expected returns leads to superior performance relative to a single-factor model in a domestic UK asset allocation framework. The evidence in the article suggests that when the investor faces no investment restrictions that using the conditional versions of the models tends to generate better performance compared to the unconditional versions of the models. Furthermore, the conditional version of the single-factor model tends to outperform the multifactor models. When the investor faces investment restrictions, the differences between the models become less substantive.

Original languageEnglish
JournalInternational Review of Financial Analysis
DOIs
Publication statusPublished - 1 May 2002

Keywords

  • factor models
  • mean-variance analysis

Fingerprint Dive into the research topics of 'On the usefulness of linear factor models in predicting expected returns in mean-variance analysis'. Together they form a unique fingerprint.

  • Cite this