The existing literature contains conflicting evidence regarding the re
lative quality of stock market volatility forecasts. Evidence can be f
ound supporting the superiority of relatively complex models (includin
g ARCH class models), while there is also evidence supporting the supe
riority of more simple alternatives. These inconsistencies are of part
icular concern because of the use of, and reliance on, volatility fore
casts in key economic decision-making and analysis, and in asset/optio
n pricing. This paper employs daily Australian data to examine this is
sue. The results suggest that the ARCH class of models and a simple re
gression model provide superior forecasts of volatility. However, the
various model rankings are shown to be sensitive to the error statisti
c used to assess the accuracy of the forecasts. Nevertheless, a clear
message is that volatility forecasting is a notoriously difficult task
.