An empirical analysis of the Fisher effect in the G7.

Abstract

This study investigates whether the Fisher hypothesis holds in the G7 countries. The hypothesis involves a long-run one-to-one relation between the nominal interest rate and expected inflation. It has important policy implications since, if it holds, the monetary policy will have no influence on real interest rate as, in this case, any change in expected inflation will be offset by a change in the nominal interest rate, leaving the real interest rate unchanged. Firstly, we use unit root and stationarity tests to determine the order of integration and end up that all time series are I(1) besides from the inflation of Japan. So, then we apply Engle-Granger 2-step and Johansen cointegration test to examine whether there is cointegration among of the time-series for G7 beside from Japan. We found that for these six countries there is long-run, positive and more than one to one relationship between the nominal interest rate and inflation. Evidence for the full Fisher effect applies only for France and USA. The dynamic relationship between the two variables is also examined from the error correction model derived. We find that the two ECTs are statistical significant but inflation is more flexible in restoring long-run relationship equilibrium. Finally using the VECM applied Granger-causality test and Impulse Response Functions to investigate the interaction among the variables. Although the VECM there is a dynamic relationship among of two variables in all countries the Granger-causality test suggest feedback only for France and USA. And The IRF shows that there is positive relationship from inflation to nominal interest rate in Canada, France, Germany, Italy and USA which agrees with the theory of Fisher effect.