Bootstrap Seasonal Unit Root Test under Periodic Variation

TL;DR

This paper investigates the behavior of HEGY tests for seasonal unit roots under periodic variation, identifies limitations, and proposes bootstrap methods to improve testing accuracy in economic time series.

Contribution

It introduces bootstrap procedures for HEGY tests under periodic variation, establishing their consistency and demonstrating improved finite-sample performance.

Findings

Asymptotic null distributions are unaffected for roots at 1 or -1 under periodic variation.

Null distributions for joint roots are non-standard and differ from no variation case.

Bootstrap methods outperform existing tests in finite samples.

Abstract

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