Volatility-Induced Stationarity and Error-Correction in Macro-Finance Term Structure Modeling

It is well-known that interest rates are extremely persistent, yet they are best modeled and understood as stationary processes. These properties are contradictory in the workhorse Gaussian affine term structure model in which the persistent data often result in unit roots that imply non-stationarity. We resolve this puzzle by proposing a term structure model with volatility-induced stationarity. Our model employs a level-dependent conditional volatility that maintains stationarity despite the presence of unit roots in the characteristic polynomial corresponding to the conditional mean. An empirical macro-finance application is presented. We obtain term premia that are economically plausible and consistent with survey data. Compared to the Gaussian affine term structure model, we improve out-of-sample forecasting of the yield curve.