Size and Power of Tests for Stationarity in Highly Autocorrelated Time Series
U of St. Gallen, Economics Discussion Paper No. 2002-26
26 Pages Posted: 5 Feb 2003
Date Written: November 2002
Abstract
Tests for stationarity are routinely applied to highly persistent time series. Following Kwiatkowski, Phillips, Schmidt and Shin (1992), standard stationarity tests employ a rescaling by an estimator of the long-run variance of the (potentially) stationary series. This paper analytically investigates the size and power properties of such tests when the series are strongly autocorrelated in a local-to-unity asymptotic framework. It is shown that the behavior of the tests strongly depends on the long-run variance estimator employed, but is in general highly undesirable. Either the tests fail to control for size even for strongly mean reverting series, or they are inconsistent against an integrated process and discriminate only poorly between stationary and integrated processes compared to optimal statistics.
Keywords: Tests for stationarity, local-to-unity asymptotics, long-run variance estimation, mean reversion
JEL Classification: C12, C22
Suggested Citation: Suggested Citation
Do you have a job opening that you would like to promote on SSRN?
Recommended Papers
-
Tests for Unit Roots and the Initial Observation
By Graham Elliott and Ulrich K. Müller
-
Modified KPSS Tests for Near Integration
By David Harris, Stephen J. Leybourne, ...
-
Power of Tests for Unit Roots in the Presence of a Linear Trend
By Bent Nielsen
-
The Impact of the Initial Condition on Robust Tests for a Linear Trend
By David I. Harvey, Stephen J. Leybourne, ...
-
Reducing the Size Distortion of the KPSS Test
By Eiji Kurozumi and Shinya Tanaka
-
Trend and Initial Condition in Stationarity Tests: The Asymptotic Analysis
-
Break Date Estimation for Models with Deterministic Structural Change