In this paper, we evaluate linear stochastic discount factor models using a portfolio metric: the realized out-of-sample Sharpe ratio of mean-variance portfolios backed by alternative linear factor models. Using a sample of monthly US portfolio returns data spanning the period 1927 2011, we provide evidence that multifactor linear models have better empirical properties than the CAPM, not only when the cross-section of expected returns is evaluated in-sample, but also when this economic metric is used out-of-sample. When we compare a portfolio associated to a multifactor model and a portfolio associated to the CAPM, we document di¤erences of Sharpe ratio of up to 10 percent. EFM Codes: 310, 380. CAIR, Manchester Business School, and IGIER, Bocconi University. Address: MBS Crawford House, Manchester, M13 9PL, UK. Tel: +44(0)1612756406. Email: firstname.lastname@example.org. yManchester Business School and Universidad de Chile. Address: MBS Crawford House, Manchester, M13 9PL, UK. Tel: +44(0)1612754585. Email:email@example.com. zFinancial Mathematics and Computation Cluster (FMC2), School of Business, University College Dublin. Address: Bel
eld, Dublin 4, Ireland. Tel: +353(1)7164786. Email: firstname.lastname@example.org. Lozano acknowledges the support of Science Foundation Ireland under Grant Number 08/SRC/FM1389.
Download Full PDF Version (Non-Commercial Use)