An analysis of the influence of domestic macroeconomic variables on the performance of South African stock market sectoral indices

An econometric study was undertaken to determine the extent to which selected macroeconomic variables influenced stock market prices on the All-Share, Financial, Mining and Retail Indices of the Johannesburg Stock Exchange in South Africa from 1996:7 to 2008:12. The Johansen cointegration method was used to determine whether cointegrating relationships existed between the macroeconomic variables and the stock market indices. A Vector Error Correction model was estimated and found significant results that money supply, inflation, long and short- run interest rates, and the exchange rate all had an influence on stock market prices. Further, the VECM results for each sector indicated that certain macroeconomic variables had differing influences on each sector of the stock market. Impulse Response tests indicated that the selected macroeconomic variables caused shock to the sectoral indices in the short-run but that the effect was lagged. The Variance Decomposition tests conducted supported earlier evidence that the macroeconomic variables had a strong level of sectoral index determinacy in the long run. The results found that the All-Share and Financial Indices were negatively influenced by inflation and short term interest rates in the long run, while the Financial Index also had a negative relationship with long-run interest rates. Money supply was found to have a positive effect on all the indices. A weakening of the exchange rate was found to have a positive influence on both the Retail and Mining Indices, while it negatively affected the All-Share and Financial Indices. The long-run interest rate had a positive influence on both the Retail and Mining Indices. Overall the study finds that macroeconomic variables are important determinants of stock market prices in South Africa and that it is important to examine each sector of the stock market separately to capture what are different effects. The limitations of the study are that a different measure for exchange rates from the nominal rand/dollar exchange rate used here may yield more decisive results and provide insight into the link between exchange rate behaviour and performance of especially the retail sector.