Some Pitfalls In Unit Root Testing

Since the seminal paper by Granger and Newbold (1974) on spurious regressions, applied econometricians have become aware of the consequences of unit roots in empirical analysis with time series data. Yet one can still find many published papers with unit root tests implemented in an inappropriate way. The objective of this Technical Note is to highlight the common pitfalls and best practices when testing for unit roots. In addition to the theoretical discussion, we provide examples using price data from Kenya, Mali, Togo, and South Africa to illustrate the procedures we think are worth following.