Testing for Unit Roots in Autoregressive-Moving Average Models of Unknown Order: Critical Comments

In this research note, we accomplish two objectives. First, we reexamine the reliability of unit root findings in the study by Said and Dickey (1984) and show that their results are internally consistent. Second, we provide new results from the reanalysis of the original data that were not included in their study and explain why their results cannot be generalized. In conclusion, we cast doubt on the continued usefulness of Augmented Dickey Fuller (ADF) test as a sound scientific method.