Review of International Economics · 2001 · 226 citations · 0 references
Social SciencesEconomicsEconometricsFinancial EconometricsTime Series AnalysisPanel Data AnalysisPanel Data MethodsPower AnalysisPurchasing Power ParityStationarity AnalysisTime Series EconometricsAugmented Dickey-fuller TestPanel Time Series AnalysisPanel Unit Root TestsPurchasing Power Parity DeviationsUnit Root TestingUnit Roots
Simulations demonstrate that when unit‐root behavior is rejected in a Levin and Lin panel test, it is incorrect to infer that all series are stationary. Recent tests proposed by Im, Pesaran and Shin, and by Sarno and Taylor, are also incapable of determining the mix of I (0) and I (1) series in a panel setting. This paper introduces a new unit‐root test that allows the researcher to discern which series are I (0) and which ones are I (1). The test has double to triple the power of single‐equation augmented Dickey–Fuller tests.