A Flexible Non Linear Model to Test the Expectation Hypothesis of Interest Rates

Conventional approaches to examining the expectation hypothesis of interest rates assume a parametric linear specification among variables. In contrast, this paper tests the hypothesis using a flexible nonlinear inference approach proposed by Hamilton (2001). We examine the impact of the nonlinearity of interest rates to explain the variability of risk premia on market rates. It is assumed that the term structure of interest rates can be identified by two factors, the risk-free rate and its volatility. The results of the linearity test against nonlinear alternatives suggest that there is clear evidence of nonlinearity. Our empirical study shows that correctly accounting for the nonlinearity of the term structure of interest rates may explain the variability of risk premia and the specific characteristics of interest rate dynamics on the U.S. market.