In this paper we consider the computation of state-prices in one-factor termstructure models. The contingent claim valuation problem is solved using the efficient method of finite difference, and the usual backward induction algorithm is reversed to produce an equivalent forward induction algorithm.
Given the set of state-prices a large class of contingent claims can be easily valued. Differentiating the state-prices with respect to the initial state, the volatility of the claims can be determined easily a well.
We consider applications to the calibration of termstructure models to yield and volatility input. We extend the state-price framework to include the prepayment feature of mortgage backed securities and give applications to MBS analysis.
By Nicki Søndergaard Rasmussen