Bootstrap seasonal unit root test under periodic variation

Both seasonal unit roots and periodic variation can be prevalent in data. In the testing of under variation, validity existing methods, such as HEGY test, remains unknown. The behavior augmented test unaugmented is analyzed. It turns out that asymptotic null distributions statistics single at 1 or − when there are identical to no variation. On other hand, any coexistence 1, , i non-standard different from Therefore, exists, tests not directly applicable joint for concurrence roots. As a remedy, bootstrap proposed; particular, with independent identically distributed (iid) block implemented. consistency these procedures established. finite-sample illustrated via simulation prevails over their competitors’. Finally, applied detect various economic time series.