RDP 9204: The Term Structure of Interest Rates, Real Activity and Inflation 1. Introduction

In the last few years the slope of the yield curve has received considerable attention for its ability to forecast both real and nominal macroeconomic variables. Mishkin (1990a, 1990b, 1991) and Jorion and Mishkin (1991) have demonstrated that the slope of the curve beyond one year is a relatively good predictor of the change in the rate of inflation. On the real side, Stock and Watson (1989), Bernanke (1990) and Bernanke and Blinder (1990) have shown, using a vector autoregression approach, that the spread between the yields on long and short bonds helps predict future economic activity. Similarly, Estrella and Hardouvelis (1991) show that the slope of the yield curve helps predict the change in real economic activity over horizons out to twelve quarters. This paper examines the ability of the slope of the yield curve to predict changes in real activity and inflation in Australia.

If prices adjust instantaneously to monetary shocks, monetary policy should have no effect on real outcomes including real interest rates. Changes in the slope of the nominal yield curve should simply reflect changes in the expected future path of inflation. If prices are sticky, however, monetary policy can affect the slope of the nominal yield curve independently of the inflation channel. This results in changes in real interest rates and real activity. Section 2 of the paper presents a simple model which formalises this notion. It introduces a long term interest rate into a Keynesian model with long-run monetary neutrality. Expansionary monetary policy causes the slope of the yield curve to steepen as the liquidity effect of higher real money balances lowers short term interest rates. This steepening of the yield curve is accompanied by lower real rates and increases in activity in the future. In Section 3 of the paper the ability of the yield curve to forecast changes in real activity is examined empirically.

Consistent with the international evidence it is found that since 1982 the slope of the yield curve predicts a wide range of real variables measuring economic activity. In contrast to the index of leading indicators, which is useful in predicting economic activity over only the subsequent six months, the yield curve's predictive ability reaches its maximum at about the 18 months horizon. At horizons longer than 30 months it has little predictive power. The gradual increase in predictive power over time, followed by a decline, is consistent with the monetary explanation of the predictive power of the yield curve. The results also show that prior to 1982 the yield curve contained essentially no information about the future path of economic activity. The difference in the results between the two periods is attributed to financial liberalisation and changes in the method by which interest rates are determined.

Following the examination of the power of slope of the yield curve to predict changes in real output, Section 4 examines the yield curve's ability to predict changes in inflation. It is found that the yield curve provides no information about future changes in inflation in the short term (less than six months) but does provide significant information for forecasting changes in inflation over longer horizons. Regardless of the horizon, the spread between the 10 year bond rate and the 180 bank bill outperforms other yield spreads in forecasting future changes in inflation. Finally, Section 5 summarises the principal findings of the paper. It is argued that the results from both the activity and inflation regressions are consistent with the hypothesis that monetary policy temporarily affects the real yield spread, and thus activity, in the short run, but that in the long run its effect is on inflation.