A Defense of Traditional Hypotheses About the Term Structure of Interestrates

Abstract

Expectations theories of asset returns may be interpreted as stating either that risk premia are zero, or that they are constant through time. Under the former interpretation, different versions of the expectations theory of the term structure are inconsistent with one another, but I show that this does not necessarily carry over to the constant risk premium interpretation of the theory. Furthermore, I argue that differences among expectations theories are of 'second order" in a precise mathematical sense. I present an approximate linearized framework for analysis of the term structure in which these differences disappear, and I test its accuracy in practice using data from the CRSP government bond tapes.