Long- and short-run Fisher effects: new tests and new results

Abstract

We find Fisher effects, both long- and short-run, in our full sample period and in the lengthy first sub-period, using Mishkin's monthly data for Treasury bill yields and the corresponding inflation rate. Recently developed cointegration techniques (Ahn and Reinsel, 1990; Reinsel and Ahn, 1992) are used to detect a long-run Fisher effect. The short-run effect is analysed within a cointegration framework, using a Granger causality test that we developed. In the two more recent, but comparatively short sub-periods since October 1979, our results indicate the Fisher effect is not present in either long- or short-run form. These methodologies and results differ from the conclusions of Mishkin's (1992) recent study.