Working Paper, Federal Reserve Bank of Atlanta 2006-11
Empirical support for the long-run Fisher effect, a hypothesis that a permanent change in inflation leads to an equal change in the nominal interest rate, has been hard to come by. This paper provides a plausible explanation of why past studies have been unable to find support for the long-run Fisher effect. This paper argues that the necessary permanent change to the inflation rate following a monetary shock has not occurred in the industrialized countries of Australia, Austria, Belgium, Canada, Denmark, France, Germany, Greece, Ireland, Italy, Japan, the Netherlands, Norway, Sweden, Switzerland, the United Kingdom, and the United States. Instead, this paper shows that inflation in these countries follows a mean-reverting, fractionally integrated, long-memory process, not the nonstationary inflation process that is integrated of order one or larger found in previous studies of the Fisher effect. Applying a bivariate maximum likelihood estimator to a fractionally integrated model of inflation and the nominal interest rate, the inflation rate in all seventeen countries is found to be a highly persistent, fractionally integrated process with a positive differencing parameter significantly less than one. Hence, in the long run, inflation in these countries will be unaffected by a monetary shock, and a test of the long-run Fisher effect will be invalid and uninformative as to the truthfulness of the long-run Fisher effect hypothesis.