41 Pages Posted: 23 Nov 2003
Date Written: January 21, 2003
Over the long-run from 1926 to 2001, the CAPM can account for the spread in the returns of portfolios sorted by book-to-market ratios. In contrast, using data covering the period after 1963, many studies find strong evidence of a book-to-market effect using conventional asymptotic standard errors. To conduct correct small sample inference, we estimate a conditional CAPM with time-varying betas and find that post-1963 book-to-market effect is statistically insignificant. We find some evidence of a book-to-market effect among medium-sized stocks, but not among the smallest stocks. We also find that while the momentum effect is robust to small sample biases, the reversal effect is not.
Keywords: book-to-market effect, value effect, conditional CAPM, momentum effect, reversal effect, time-varying beta
JEL Classification: C51, G12
Suggested Citation: Suggested Citation
Ang, Andrew and Chen, Joseph, CAPM Over the Long-Run: 1926-2001 (January 21, 2003). AFA 2004 San Diego Meetings. Available at SSRN: https://ssrn.com/abstract=346600 or http://dx.doi.org/10.2139/ssrn.346600