The Predictive Power of the Yield

Abstract

The purpose of this study is to determine if the information contained in the South African Yield Curve – being a profile of debt bearing instruments with different maturities – is a predictor of the South African Gross Domestic Product. The report is limited to the period following the 1994 democratic elections as this event signified a major shift in economic policy. The Term Spread – being the difference in yield between a short term instrument and a long term instrument – had been shown in a number of studies to be a significant predictor of inflation, economic activity and recessions. Most of these studies were carried out on first world economies. Literature showed that the addition of a stock market index and the level of the yield curve (the short term interest rate) to the model increased its performance. The numeric data – the GDP, Yields on T-Bills and RSA Government Bonds and the All-Share Index – used to complete this study is readily available from reputable institutions. This work shows that the Term Spread has weak predictive ability over the change in the South African GDP in the near term – up to two quarters ahead. The all-share index contains negligible additional information. When the 91 day T-Bill Rate (the short term rate) is added to the model, the predictive ability increases tremendously, both in terms of the explanatory power and the forecasting horizon – to the point where the contribution of the term spread can be considered non-significant. It is not clear why the T-Bill rate shows strong predictive power relative to the Term Spread in the South African context and represents an area for further research.