An empirical examination of the Longstaff-Schwartz bond option valuation model

This article examines the suitability of the Longstaff-Schwartz two-factor model for practical use. Discussion of issues related to the implementation of the model addresses the problem of fitting the model to the initial term structure of interest rates. To asses empirical performance, the Longstaff-Schwartz model is used to value German interest rate warrants for the four-year period 1990-1993. The data set includes options with longer maturities, which are of particular interest for testing bond option pricing models. A three-step procedure is used. In the first step, the current term structure of interest rates is estimated. In the second step, the constant parameters of the model are determined, and the model is calibrated to the initial yield curve. In the final step, the theoretical values of the interest rate warrants are computed and compared with their market prices. The empirical results indicate that the Longstaff-Schwartz two-factor model has considerable predictive ability, although parameter estimation turns out to be time-consuming.