In this paper we show how the Kalman filter (see Kalman (1960), Kalman and Bucy (1961)) could be used to estimate the partially time-dependent parameters of the extended Vasicek model introduced by Hull and White (1990). The results are compared with a method that uses market prices of a cap portfolio to derive implied estimators for the time independent mean reversion and volatility parameters of the stochastic process in two different ways: the resulting volatility structure and the effect of both methods on risk management. For the last comparison we apply the Approximate Full Valuation (AFV) method of Zagst (1997) to calculate the Value at Risk (VaR) of a cap portfolio for both estimation procedures.     «

In this paper we show how the Kalman filter (see Kalman (1960), Kalman and Bucy (1961)) could be used to estimate the partially time-dependent parameters of the extended Vasicek model introduced by Hull and White (1990). The results are compared with a method that uses market prices of a cap portfolio to derive implied estimators for the time independent mean reversion and volatility parameters of the stochastic process in two different ways: the resulting volatility structure and the effect of...     »