Examining interest rate volatility and bond yields in South Africa
Abstract
This study examined interest rate volatility and bond yields in South Africa. Two independent variables were employed, the 91 days T-bill and short-term interest rate volatility. The short-term interest rate volatility was estimated by the use of the generalised autoregressive conditional heteroscedasticity GARCH (1,1) model. The average bond yields were used as dependent variables, with the following ranges: zero to three year bond yields (short-term) and 10 year and above bond yields (long-term). The sample period of the study spanned from January 2004 to December 2017 with a total of 168 observations. The choice for the study period was to incorporate periods before, during and after the 2007-2009 financial crisis. The relationship between interest rates, interest rate volatility and bond yields were examined by the use of the Cox-Ingersoll-Ross (CIR) one-factor (1985) model and the Longstaff and Schwartz (1992) two-factor model. Both these models have been used extensively in the term structure of interest rate studies in the literature. In order to evaluate these models the autoregressive distributive lag (ARDL) model was employed as the unit root results from the study indicated that the variables were integrated at different orders, namely I(0) and I(1). The results from the study demonstrated that there is a statistically significant positive relationship between short-term interest rates and bond yields in South Africa. These results are consistent with the expectation hypothesis theory that long-term interest rates are an average of the current- and future expected short-term interest rates. These results were obtained using the CIR (1985) one-factor model. The use of the Longstaff and Schwartz (1992) two-factor model revealed that there is a statistically insignificant positive relationship between interest rate volatility and bond yields in South Africa and the effect is greater for the (three to five year) medium-term bond yields. Investors make their decisions based on the prediction of the level of future interest rates. The results of this study verify that the expectation hypothesis theory holds in South Africa. This means it is possible for bond investors to predict the future changes in interest rates by a consideration made to the yield curve. Thus, these results may aid South African bond investors to hedge themselves against unfavourable interest rate volatility in the future, with reference made to the yield curve.