Does the choice of interest rate data matter for the results of tests of the expectations hypothesis - some results for the UK

Using UK data for the period 1997:3 to 2005:5, this paper examines whether the expectations hypothesis is supported by recent UK data when the short-end of the term structure of interest rates is considered and whether the results of the tests of the expectations hypothesis are sensitive to the choice of data. The main results can be nicely summarized by considering five virtual researchers who test the expectations hypothesis using five different data sets for the 1997:3 to 2005:5 period for the 1 to 12-month maturity spectrum and who get quite different results. The main conclusion to be drawn from the analysis in this paper is thus that robustness check may be very important when testing the expectations hypothesis using the 1 to 12-month maturity spectrum of the term structure. Furthermore, the results suggest that the specific data set used in tests of the expectations hypothesis may be a candidate explanation of a rejection of the expectations hypothesis - along with the possibility that a time-varying term premium and/or a structural break are responsible for the rejection