20 Pages Posted: 21 Nov 2009 Last revised: 29 Jan 2012
Date Written: May 31, 2009
The recent advent of high-frequency data and advances in financial econometrics allow investors to evaluate the accuracy of different beta (systematic risk) measurements. Benchmarking against the monthly realized beta formed by 30-minute data, we compare the popular Fama-MacBeth betas, the monthly realized betas formed by daily returns and our Hodrick-Prescott filtered betas, with the smoothing parameter, λ, set to 100. We find our filtered betas reduce the measurement error substantially relative to other beta measures. These results enable market participants to measure betas with greater precision and efficiency even with only daily returns in hand.
Keywords: beta measurement, high frequency, realized beta, Hodrick-Prescott filter
JEL Classification: C13, G10
Suggested Citation: Suggested Citation
By Andrew Ang