Real Activity and Yield Spreads under the Consumption-based Asset Pricing Model

By using a flexible linear version of the consumption based asset pricing model, this paper shows the extraordinary capacity of the yield spreads to forecast consumption growth. More precisely, using both averages of yield spreads and average past stock returns, we are able to replicate the Economic Sentiment Indicator elaborated by the European Commission in order to predict turning points in the business cycles. In particular, available financial asset returns explain a striking 98.5% of the variability of the Sentiment Indicator for the European community. Interestingly, the contribution of the yield spreads by themselves is as high as 93.7%. 3 1.-Introduction Equilibrium asset pricing models analyse the interaction and relationships between aggregate economic variables and the return of financial assets. Although financial economics has mainly focus on the way these economic variables determine financial asset returns, the inverse line of research has also been conducted. In particular, the information content of security returns on the economic agent expectations regarding the future behaviour of aggregate variables such as GDP, consumption, inflation, and others, has been studied. Along these lines, this paper particularly focuses upon the predictive ability of the slope of the term structure of interest rates. This exercise seems to be justified given the previous empirical evidence in which the behaviour of the term structure of interest rates has been found to be an even better predictor of future economic activity than stock returns. In general, inverted zero-coupon curves tend to precede recessions, while upward-sloping curves tend to forecast expansions. The and Mishkin (1997) are key examples on this evidence. The importance of this field of research is in part motivated by the perception that structural econometric models fail when predicting real output evolution. These models are more complex that those based on interest rates and, as discussed by Harvey (1989), and Davis and Fagan (1997), some doubts exist about their stability and accuracy. As Hu (1993) points out, " the forecasting ability of the yield curve is quite impressive considering the cost and performance of many large-scale macroeconometric forecasting models that are also used to predict real output ". If the relationship between yield spreads and economic activity has been extensively documented from an empirical point of view, the theoretical explanations provided so far are mainly based on two reasons. First of all, the spread is an indicator of monetary policy and therefore it predicts future effects of central …