Revisiting Monetary Policy Reaction Function in Developing Economies: Multiscale-GMM Evidence of Policy Asymmetry from Ghana

Abstract

We estimate monetary policy reaction function for Ghana to comprehend the policy behaviour of Bank of Ghana (BOG) over the years using Generalized Methods of Moment technique within wavelet multiscale domain. The empirical results show an overriding evidence of asymmetry monetary policy rule in Ghana across the historical time domain and multiscale spectrum. The time-domain analysis reveals a generally weak policy response to overall output gap, while the multiscale result points to a statistically significant policy reaction to output gap in the short-run. The extent of policy easing during economic downturn appears to be stronger in the long run while policy tightening during economic booms is rather substantial in the short-to-medium run. On the other hand, policy response to inflation is stronger in the medium-to-long run and there is a compelling evidence of policy aversion to positive inflation gap. Yet, the magnitude of policy response to above-target (rising) inflation is well below unity across timescales, pointing to a clear non-adherence to the Taylor principle and by extension, a pursuit of flexible inflation targeting regime in Ghana. Notably, there is further robust evidence of high policy inertia, while the direct inclusion of exchange rate in the monetary policy rule appears to be superfluous in the context of Ghana. Moreover, monetary policy reaction to macroeconomic dynamics is however found to be homogenous across political regimes, construing lack of political regime-effect on monetary policy reaction function for Ghana. The empirical results thus transmit important and diverse policy implications for the central bank of Ghana to safeguard the price stability objective.