This study attempts to identify the important variables that may affect yellow maize futures prices in the South African derivatives market. Data was obtained from the South African Futures Exchange, a division of the Johannesburg Securities Exchange. Weekly data on the rand-dollar exchange rates were obtained from the South African Reserve Bank (SARB). Monthly data regarding import volumes, export volumes, maize consumption and maize stocks in South Africa are available from South African Grain Information Service (SAGIS). Fifteen variables that may be used to forecast futures prices were identified from theory and similar studies. A correlation matrix of these variables with maize futures prices was determined at the 5% significance level. After applying various statistical analyses to test for autocorrelations, stationarity etc., only four variables were left with which to model the futures prices. The R2 of the remaining variables was only 12.21%, indicating a low goodness of fit. Applying the regression model to the ex-post prices clearly indicated that these variables that were identified do not adequately explain the movement in the futures prices. The primary reasons for the low accuracy of the model may be due to the use of the weather index for SA alone (a small contributor in a global market) and the linearity assumption underlying the selected dependant and independent variables may also be unrealistic. Further research is therefore needed to identify more appropriate variables with which to model yellow maize futures prices.
An explanatory model of South African yellow maize futures prices