A broad range of termstructure models is now available in RIO, including, but not limited to, the extended Vasicek and CIR, Black-Karasinski, and Black-Derman-Toy models.
Previously RIO calculated mortgage bonds values using a binomial Black-Derman-Toy model. Although simple and efficient the binomial model lacked the ability to match specific dates and rates. The new models are solved using finite difference methods and these techniques allows for a much more flexible and precise calculation setup [1].
In RIO the termstructure models can be calibrated to volatility input given in the form of quoted Black volatilities of caps and/or swaptions using the rw_Calibrate function.
The calibrated models may be verified or compared by computing option values and implied volatilities of caps or swaptions using rw_Cap and rw_Swaption respectively.
.
Use the rw_Swaption function and RIO's chart facilities for easy comparison of quoted Black volatilities for swaptions to the similar volatilities implied by the calibrated termstructure model. The calibration uses a simple four parameter description of the volatility curve, however, it is still able to fit observed quotes for almost all swap- and option maturities.
Finally the rw_VolCurve function computes a full set of statistics describing the calibrated model including yield curves, the spot- and forward volatility structure of zero coupon bond prices and rates, Black volatilities for at-the-money caps and swaptions etc.
[1] The models are calibrated to yield curves and volatility using a unique finite difference forward induction approach developed as part of a Ph.D. project sponsored by ScanRate. For more details see the paper "Finite difference computation of state-prices in termstructure models: with applications to calibration and MBS analysis" by Nicki Søndergaard Rasmussen.