Premium
Heteroskedasticity‐Robust Unit Root Testing for Trending Panels
Author(s) -
Herwartz Helmut,
Maxand Simone,
Walle Yabibal M.
Publication year - 2019
Publication title -
journal of time series analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.576
H-Index - 54
eISSN - 1467-9892
pISSN - 0143-9782
DOI - 10.1111/jtsa.12446
Subject(s) - unit root , heteroscedasticity , econometrics , mathematics , volatility (finance) , asymptotic distribution , test statistic , null hypothesis , statistic , unit root test , statistics , gaussian , series (stratigraphy) , consistency (knowledge bases) , statistical hypothesis testing , cointegration , quantum mechanics , estimator , biology , paleontology , physics , geometry
Time‐varying volatility and linear trends are common features of several macroeconomic time series. Recent articles have proposed panel unit root tests (PURTs) that are pivotal in the presence of volatility shifts, excluding linear trends, however. This article proposes a new PURT that works well for data that is both heteroskedastic and trending. Under the null hypothesis, the test statistic has a limiting Gaussian distribution. We derive the local asymptotic power to underpin the consistency of the test statistic. Simulation results reveal that the test performs well in small samples. As an empirical illustration, we examine the stationarity of energy use per capita in OECD economies. While the series are in general difference stationary, they could also be considered as trend stationary for specific time spans.