Nelson and Plosser Revisited: A Re-Examination using OECD Panel Data∗

This paper applies various recent panel unit root tests to 14 macroeconomic variables. These variables, observed for the OECD countries since the 50’s, are the same as those considered in the seminal paper by Nelson and Plosser (1982) for the United States. Two distinct generations of tests are used. The first one is based on a cross-sectional independence assumption. If we consider only tests based on heterogeneous panel specifications as Im, Pesaran and Shin (2003), Maddala and Wu (1999) or Choi (2001), the conclusions are not obvious since the non stationarity is strongly rejected for 5 macroeconomic variables generally considered as non stationary. On the contrary, the results based on the second generation tests, which relax this cross-sectional independence assumption, are more in favor of the unit root hypothesis. However this conclusion must be moderated since the results strongly depend on the specification of the model and the method used to eliminate the cross-sectional correlations. •