Testing the Conditional CAPM and the Effect of Intervaling

Traditional tests of asset pricing undertaken within the CAPM framework have provided mixed results. One explanation for the supposed failure of the model is its inability to account for temporal dependence in unconditional residuals which can be induced by time-variation in volatility. This paper provides a test of the conditional CAPM utilising the GARCH-M framework thereby allowing for nonconstant variance. The paper utilises 16 years of daily data in the Australian equity market and extends existing research by investigating the effect of intervaling by conducting tests over different sampling frequencies. While some support is found for the GARCH-M specification, the inability of the "in-mean" parameter to achieve statistical significance is an empirical limitation. Despite this finding, the conditional CAPM is supported for weekly and monthly return data, however on balance, the greatest support for the model is found in the weekly return interval. Thus, we demonstrate that asset pricing tests in this context are sensitive to the return interval.