The Yield Curve and Macroeconomic Dynamics*

Abstract

We show that microfounded DSGE models with nominal rigidities can be successful in replicating features of bond yield data, including sizeable term premia and volatile long-term yields, which have previously been considered puzzling in general equilibrium frameworks. At the same time, sample moments of consumption growth and inflation can be fit relatively well. The improved model performance does not arise directly from the presence of nominal rigidities. However, this feature introduces (short-run) monetary non-neutrality, so that monetary policy affects consumption dynamics and bond prices. A high degree of 'interest rate smoothing' in the policy rule is essential for our results. At the core of dynamic macroeconomic models, we find equilibrium relationships which describe the allocation of quantities (e.g. consumption and investment) and, when imperfect competition prevails, the setting of prices. These equations are typically valid across time and across states of nature, and financial assets form the tool that is supposed to ensure their validity. Financial assets are therefore an integral part of macroeconomic models. Any well-specified model should be able to match financial data as well as macro data. It is therefore problematic that microfounded models have had a hard time explaining key features of asset prices. The most famous example of this difficulty is the equity premium puzzle but various features of bond yield data have also been char acterised as puzzling in the literature. A number of papers, including Backus et al. (1989), Donaldson et al. (1990), Den Haan (1995) and Chapman (1997), have con cluded that general equilibrium models cannot generate term premia of a magnitude comparable to what we observe in actual data - and may not even be capable of producing positive term premia. For equity, one could argue that fundamentals (the expected future profitability of individual firms) are unobservable and difficult to evaluate. Equity prices may therefore be thought to be subject to fluctuations disconnected from the real economy (information acquisition, fads, etc.) and one could hope that the inability of macroeconomic models to match equity prices does not represent a signal of misspecification. This argument is more difficult to construct for bonds. Bond yields should ultimately reflect expectations of future monetary policy decisions which, at least in recent years, are arguably more predictable than equity fundamentals. Thus, the ability of macroeconomic models to explain bond prices represents an important test of their empirical performance. In this article, we revisit the relationship between bond prices and macroeconomic fundamentals. Contrary to the papers cited above, that rely on flexible price models, we