Testing the Fisher Effect in OECD countries: An empirical investigation

Abstract

This paper tests the validity of the Fisher hypothesis for the OECD countries using a cointegration procedure developed by Gregory and Hansen (1996) that allows for the presence of a one-time endogenously determined structural break in the cointegrating vector. We are paying particular attention to the integration and cointegration properties of the variables, since meaningful Fisher effect tests critically depend on such properties. It is noteworthy that , contrary to the other empirical studies ,we test the hypothesis of stationarity of the variables using a relatively recent unit root test suggested by Ng-Perron (2001) due to its superiority to Augmented Dickey Fuller (ADF) and Phillips-Perron (P-P) tests. The results indicate that in the majority of the countries tested, their variables do not satisfy the integration properties and therefore we cannot proceed to cointegration techniques. For the rest of the countries where the integration properties are satisfied the cointegration procedure indicate that the full Fisher effect is present for only one country, Canada, while the partial Fisher effect holds for Belgium and Korea.