ON ESTIMATING THE YIELD AND VOLATILITY CURVES

Yield curve and yield volatilities are important inputs for pricing in terest rate derivatives, for generation of interest rate scenarios, et c. Nonanticipated errors in their estimates may essentially influence the resulting prices, yields and risks. In this paper we explore and c ompare several types of parametric and nonparametric regression models suitable for estimation of the two curves. In contrast to purely nume rical fitting procedures, these methods provide also an information ab out the precision of the fitted curves and a test of the goodness-of-f it of the postulated parametric model. The parametric models of yield curves are represented by the nonlinear and linearized Bradley-Crane m odel which is compared with Nadaraya-Watson and Priestley-Chao nonpara metric estimators and with cubic splines. The reported numerical exper ience is based on data from the Italian bond market.