作者: Peter Pedroni
DOI: 10.1017/S0266466604203073
关键词:
摘要: We examine properties of residual-based tests for the null no cointegration dynamic panels in which both short-run dynamics and long-run slope coefficients are permitted to be heterogeneous across individual members panel. The also allow fixed effects trend terms, we consider pooled within dimension group mean between tests. derive limiting distributions these show that they normal free nuisance parameters. provide Monte Carlo evidence demonstrate their small sample size power performance, illustrate use testing purchasing parity post–Bretton Woods period.I thank Rich Clarida, Bob Cumby, Mahmoud El-Gamal, Heejoon Kang, Chiwha Kao, Andy Levin, Klaus Neusser, Masao Ogaki, David Papell, Pierre Perron, Abdel Senhadji, Jean-Pierre Urbain, Alan Taylor, three anonymous referees helpful comments on various earlier versions this paper. paper has benefited from presentations at 1994 North American Econometric Society Summer Meetings Quebec City, European Maastricht, workshop seminars Board Governors Federal Reserve, INSEE-CREST Paris, IUPUI, Ohio State, Purdue, Queens University Belfast, Rice University–University Houston, Southern Methodist University. Finally, I following students who provided assistance stages project: Younghan Kim, Rasmus Ruffer, Lining Wan.