A dynamic regression panel approach to the determinants of monetary policy and economic growth: the SADC experience.

Abstract

Purpose – Empirically, the purpose of this paper is to investigate policy variables that determine monetary policy and economic growth of some selected countries within the economic bloc of Southern Africa Development Community (SADC). The selected countries are Botswana, Democratic Republic of Congo, Lesotho, Madagascar, Malawi, Mozambique, Namibia, South Africa, Swaziland, Zambia and Zimbabwe. Design/methodology/approach – Annual time series data for a panel of 11 Southern African countries spanning 1980–2015 were employed in the study. The major instrument of estimation is the dynamic regression panel model. In order to conform to econometric principles, robustness checks were carried out on the variables of interest so as to avoid spurious results. An estimation of impulse response and variance decomposition analyses were to complement the approach to the study. Findings – The result of the long-run dynamic panel regression reveals that GDP growth rate, inflation rate, exchange rate, money supply and oil and commodity prices do have profound impact on monetary policy within SADC. It was further revealed from the study that commodity price shock is the major exogenous determinant of monetary policy dynamics and the effect is transmitted via exchange rate channel to macroeconomics of the region; with inflation rate and money supply playing a major role in the transmission mechanism as it affects the economies of the countries in this region. Practical implications – The policy implication is that inflation is seen as a major challenge to the countries under review. Among other things, a hybrid of inflation and monetary targeting should be adopted to complement each other as policy combination within the region. Originality/value – The study accounts for the determinants of monetary policy vis-à-vis growth potentials of some selected countries in SADC, using a combination of dynamic regression panel approach and SVAR elements.