A Nonparametric Distribution-Free Test for Serial Independence in Stock Returns: A Comment

Abstract

We have confirmed that Corrado and Schatzberg's (1990) criticism of our test for serial independence in stock returns (Ashley and Patterson (1986)) is correct. The corrected results still favor rejection of the null hypothesis that the daily returns for several stocks (notably Holly Sugar (HLY) and E Systems (ESY)) are serially independent, but only at the 12-percent and 14-percent significance levels, respectively. Evidently, this kind of test is long on simplicity and intuitive appeal, but short on power.