Vasicek and Blume Betas: Back to the Future Part I

A fundamental driver of the cost of capital is the systematic risk measured by the Beta coefficient. To achieve an accurate value of the cost of capital, the Beta estimates are required to exhibit a set of desired characteristics. Forecasted Betas should be unbiased, stable and, most importantly, good predictors for future returns. The standard Beta estimation technique is the ordinary least squares regression (OLS) of security returns against market returns over four or five years of monthly data. However, there is strong empirical evidence of the limited ability of OLS Beta to meet the desired characteristics. Previous research reveals considerable debate about the appropriate alternative techniques. This current analysis examines six Beta estimation techniques in relation to three essential criteria: bias, stability and predictive ability.