Could regression of stationary series be spurious?

• Main Authors: Wong, Wing-Keung, Cheng, Yushan, Yue, Mu
• Other Authors: School of Physical and Mathematical Sciences
• Format: Article
• Language: English
• Published: 2024
• Subjects: Mathematical Sciences; Spurious regression; Stationarity
• Online Access: https://hdl.handle.net/10356/181224
• Institution: Nanyang Technological University

Spurious regression, commonly associated with independent and (nearly) non-stationary time series, has been extensively studied. However, the potential for spurious outcomes in regression involving stationary time series remains largely unexplored, representing a gap in the literature. To address this gap, we propose that regression of stationary time series may yield spurious outcomes and conduct a comprehensive investigation to verify this conjecture. Additionally, we present a remedy algorithm to mitigate spurious effects and improve model interpretability. Through extensive simulations, we validate our conjecture and demonstrate the efficacy of the proposed remedy. A numerical analysis further illustrates the practical utility of our approach. This study offers a fresh perspective on spurious regression and provides a practical solution to enhance the reliability of regression analyses involving stationary time series data.