The Expectations Hypothesis of Term Structure of Interest Rates Revisited

This paper investigates the validity of the Expectations Hypothesis of the Term Structure (EHTS) employing standard forward-spot regressions which accommodate for the presence of time-varying term premia. The novelty of this paper is that the analysis is conducted by taking advan- tage of the following two properties of the Kalman Filter: first, it makes possible to model time-varying term premia as unobservables, and second it delivers recursive estimations of forward-spot regressions as more data become available. In fact, previous studies have modelled term premia by means of macroeconomic variables. To the extent that term premia are influenced by political and social climates which are difficult to ob- serve, it might be preferable to model them as unobservables, rather than by means of observed variables. Moreover, especially when tested over long periods of data, the EHTS might hold for certain periods while it might not for others. These periodic departures from and reversions to the EHTS cannot be detected by constant parameters models, which there- fore can provide only broad brush evidence. This paper shows that the recursive nature of the Kalman filter can be employed to construct a test for the EHTS which gives more refined evidence. The analysis is carried out focusing on the short-end of the US term structure spectrum.