Regulated seasonal unit root process
No Thumbnail Available
Date
2022
Journal Title
Journal ISSN
Volume Title
Publisher
Walter De Gruyter Gmbh
Access Rights
info:eu-repo/semantics/closedAccess
Abstract
Unfortunately, time series problems do not appear in data singly. We focus on the joint occurrence of nonstationarity, seasonality and bounded data. Seasonal unit root tests and bounded unit root tests already exist in the literature, yet when all these issues are combined their performance needs improvement. That is why we offer a testing procedure for bounded seasonal unit root processes. The combination of these tests is not straightforward as the nonlinearity coming from bounds causes the limiting distribution of the proposed test statistic to be multivariate Brownian motion while the others have univariate distributions. The simulation exercises reveal that the existing tests, which ignores the presence of bounds or seasonality, suffer significant size problems. Our statistic removes the size distortions and also maintain satisfactory power performance.
Description
Keywords
Regulated Brownian Motion, Regulated Time Series, Seasonal Unit Roots, Efficient Tests, Selection
Journal or Series
Studies in Nonlinear Dynamics and Econometrics
WoS Q Value
Q4
Scopus Q Value
Q2
Volume
26
Issue
3
