This paper investigates the finite-sample power of STAR-based unit root tests when the data generation process is a globally stationary three-regime TAR model. Unit root tests are proposed derived from STAR models that nest TAR models.

This paper investigates solving the spurious regression problem using an autocorrelation correction. It is shown that if the relevant data generation processes contain higher-order terms, this solution is not as effective as in the first-order case.