On the Long-Run Fisher Effect: A Fractional Cointegration Approach.

Tests for fractional conintegration are employed to provide evidence on the validity of the long-run Fisher effect. We use post-war monthly data for the 3-, 6-, 12-month US Treadury bill rate. We conclude that the rejection of a "full" Fisher effect that results from the use of tests for integer conintegration is generally robust to the use of fractional alternatives. Similar conclusions emerge when tests for a "weak" Fisher effect are conducted. The results seem to be somewhat sensitive to the estimation method used for the fractional difference parameter.