An investigation of the robustness of the day-of-the-week effect in Australia

The present study analyses the robustness of the day-of-the-week effect in Australia. The results show that sample size can distort the interpretation of classical test statistics unless the significance level is adjusted downward. After providing for sample size adjustment, the evidence of a day-of-the-week in Australia is considerably weakened. Further, error distribution specification tests also reveal widespread depar-tures-from ordinary least squares (OLS) assumptions. However, tests conducted using robust techniques that are insensitive to the failure of maintained assumptions provide evidence of a day-of-the-week effect which is very similar to the evidence from OLS regressions. This contrasts with Connolly (1989) who found that the evidence of the day-of-the-week effect in the US market was considerably diminished by the application of these techniques. Robust statistical techniques also show that the Australian day-of-the-week effect is independent of the US day-of-the-week effect.