Capital Asset Pricing Models on UK Securities using ARCH

This study tests conditional and unconditional versions of the CAPM using portfolios made up of security returns in the UK over the period January 1980-December 1999. The main objectives are to see if the GARCH betas differ from the unconditional betas, and to see if the market risk premium is positive. The CAPM tests are two-pass, where monthly returns are regressed on alternative beta estimates, and the time series mean of the coefficients is the average market premium. It is found that the GARCH and unconditional betas are correlated, either 0.475 or 0.575 depending on the method used. Using unconditional betas the average market premium is negative, but not statistically significant. Using conditional betas the average market premium is positive but not statistically significant. For some individual years a positive statistically significant risk premium is found. These individual years tend to correspond to periods when the stock market was particularly volatile which would tend to suggest that the model has value during periods of relatively high volatility.