Measuring and modeling systematic risk in factor pricing models using high-frequency data

This paper demonstrates how high-frequency data may be used in more effectively measuring and modeling the systematic risk(s) in factor pricing models. Based on a 7-year sample of continuously recorded US equity transactions, we find that simple and easy-to-implement time series forecast for the high-frequency-based factor loadings in the three-factor Fama-French model gives rise to more accurate factor representations and improved asset pricing predictions when compared to the conventional monthly rolling regression-based estimates traditionally employed in the literature, in turn resulting in more efficient ex post mean-variance portfolios. As such, the methodology proposed in the paper holds the promise for important new insights concerning actual real-world investment decisions and practical situations involving risk management. © 2003 Elsevier B.V. All rights reserved.

Duke Scholars

Author Tim Bollerslev Economics