Excess returns calculated using nonstationary risk-free interest rates will
also be nonstationary and this may cause an unbalanced regression problem
in the estimation of Capital Asset Pricing Models (CAPM). Under such circum
stances, beta coefficients could be both biased and inconsistent. The impli
cations of these issues are investigated through a simulation study and an
empirical application using data on the FTA index and the 91-day UK Treasur
y Bill (T-Bill) rates. Although the simulation results are alarming, the em
pirical analysis suggests that the problem of unbalanced regression is not
likely to cause significant problems in estimating the CAPM.