Monetary policy and uncertainty in South Africa

Abstract

Even though major advances in economic theory and modelling have in some cases furthered our understanding of how the economy works, the system as a whole has become more complex. If policymakers had perfect knowledge about the actual state of the economy, the various transmission mechanisms as well as the true underlying model, monetary intervention would be greatly simplified. In reality, however, the monetary authorities have to contend with considerable uncertainty in relation to the above-mentioned factors. This said, uncertainty has mostly been neglected in both the theoretical and empirical literature focusing on monetary policy analysis. Nonetheless, findings from a review of theoretical literature that does exist on this topic suggest that optimal central banks act more conservatively when faced with uncertainty. Similarly, empirical findings from the literature also favour conservatism. However, there is some evidence to suggest that this is not always the case. These results suggest that central banks do not always act optimally when faced with uncertainty. The limited number of industrial country cases examined prevents any generalised view from emerging. If anything, the literature findings suggest that central bank behaviour differs across countries. This thesis aims to contribute to the empirical literature by studying the effects of uncertainty on monetary policy in the developing country case of South Africa. In simplest terms, the thesis seeks to establish whether or not the South African Reserve Bank (SARB) responded optimally to uncertainty as suggested by theoretical models thereof. To this end, the thesis employs a theoretical model which resembles a structural rule-based approach. The optimal interest rate rule was derived given a set of structural equations relating to demand, the Phillips curve and the real exchange rate. To incorporate uncertainty, it is assumed that the coefficients are dependent on the variances of the exogenous variables, namely inflation, the output gap and the exchange rate. The uncertainty adjusted model allows us to investigate whether monetary policy is more aggressive or passive when uncertainty about the relevant exogenous variable increases. Inflation, output gap and exchange rate uncertainty estimates were derived through GARCH-model specifications related to the structural equations as defined in the theoretical model. The investigation considered both indirect and direct uncertainty effects with a sample period stretching from 1990 to 2011. The findings reported in this thesis provide strong evidence in support of the notion that uncertainty plays a significant role within the South African monetary policy landscape and contributes towards explaining the SARB’s actions. Furthermore, the results suggest that the SARB did in fact act optimally in responding more conservatively to target variable fluctuations on average. Also, the findings could potentially strengthen the case for inflation targeting as a monetary policy regime, as the results indicate a marked decline in the effects of uncertainty under inflation targeting than before.