A NOTE ON THE EXAMINATION OF THE FISHER HYPOTHESIS BY USING PANEL CO-INTEGRATION TESTS WITH BREAK

Abstract

One problem encountered when examining the Fisher hypothesis is that various policy changes and economic shocks may induce structural shifts in the long-run relation. We explore the argument that panel cointegration tests based on common correlated effect estimators have reasonably good power and size properties, even in the presence of structural breaks, if the timing of structural shifts roughly coincide to each other across individual group members. Using the data from Omay et al.. (2015), which pays special attention to cross-section dependence issue but ignores the possibility of structural break in the data, we provide support to the argument above.