TESTING THE EXPECTATIONS HYPOTHESIS WITH SURVEY DATA with an introduction of an analysis of surveyed interest rates

This paper is an empirical study of the properties of the term structure of interest rates. It tests statistically to what extent the forward interest rates that are implicit in the term structure can be used as a forecast of the future interest rates, i.e. it tests what is known as the expectations hypothesis of the term structure of interest rates (EH)1. It is tested in a great number of articles of Modigliani and Shiller (1973), Shiller (1979), Shiller, Campbell and Schoenholtz (1983), Friedman (1979), Fama (1984), Markiw (1986) and Campbell and Shiller (1987). Gerlach and Smets (1995) tested the EH for 17 countries at the short end of the maturity structure. In about half of cases (including Sweden) they could not reject the EH. USA and Austria are two countries where the EH does not hold. On UK data, MacDonald and Macmillan (1994) do not find support for the EH. In data from the USA it is often found that forward rates are worse predictors of future interest rates than the naive martingale method - that the future interest rate is the same as the interest today2. The null hypothesis in most tests of the expectations theory is a joint hypothesis - that the expectations are rational, and that the interest rate differentials between different maturities depend on expected interest rate changes.