Reassessing the fisher effect in South Africa: a dynamic bivariate and time-frequency approach to regional disparities in monetary policy transmission

Abstract

We re-examine the Fisher effect in South Africa using an integrated DCC-GARCH model and wavelet coherence analysis to assess dynamic inflation expectation–nominal interest rate relationships across regions and time horizons. A novel measure of expected inflation based on weighted exponential smoothing is introduced, integrating adaptive and forward-looking components. We use data from April 2008 to July 2024. DCC-GARCH results reveal significant regional disparities in monetary policy transmission, with rural areas exhibiting higher but less persistent correlations, while urban areas show lower but more stable and persistent correlations. Wavelet analysis confirms the effect operates at medium-term horizons (8-16 months). These findings underscore the heterogeneous spatial nature of monetary policy effectiveness and highlight the need for regionally nuanced approaches to maintaining price stability.