Journal of Banking and Finance 15 (1991) 521-533. North-Holland An empirical examination of the pricing of European bond options Krister Rindell and Patrik Sandas* zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR Swedish School of Economics and Business Administration, Department of FinancelKrister Rindefl, SF-00100 Helsinki. Finland Received April 1990, final version received September 1990 This paper examines the pricing of European bond options, namely options on a live year Swedish T-bond. A continuous time version of the Ho and Lee model, derived by Heath, Jarrow and Morton, and the Black and Scholes model are tested using Hansen’s Generahxed Method of Moments. The results indicate that both models perform reasonably well. However, the Black and Scholes model exhibits a ‘time-to-maturity of the bond’ bias, not present in the Heath, Jarrow and Morton model. 1. Introduction Bond option pricing has been in vogue among financial economists for nearly a decade [see, e.g., Ball and Tourus (1983), Brennan and Schwartz (1983), Courtadon (1982) and Schaefer and Schwartz (1987), to name a few]. The interest in deriving a pricing model for such instruments was inspired by the Black and Scholes (BS) (1973) model for European-type options on risky assets. The scope of the research has been to derive a model with the same, very desirable, properties as in the BS model, i.e.: (i) No preference dependent parameters’ are present in the model. (ii) The model has a closed form solution. A solution to the first property was found by Ho and Lee (1986). Their intuition was to take the whole term structure, instead of individual bond prices or interest rates, as the state variable. They derived their model in an economy where the term structure follows a binomial process. It is well known that in such an economy options can be valued without the knowledge of investors’ preferences. However, binomial models do not have closed form solutions. Heath et al. (1987) derive a similar model in a continuous time economy. They show that the continuous time limit of the *Thanks are due to Lam Hansen, John Heaton and Masao Ogaki for providing computer programs for GMM estimation. written under a National Science Foundation Grant, 0378-4266/9@03.50 @ 1991-Elsevier Science Publishers B.V. (North-Holland)