>Using the bond price as
>the underlying (and thus using price volatility) is a very restricted
>approach. A more general approach is to use yield as the
>underlying and then use the bond yield formula to produce the price. This
>totally eliminates the problem of price converging to par as that is
>addressed through the bond yield formula.
Using the yield as the underlying eliminates the "convergence to par"
problem, but it presents new problems.
Others have already suggested that using HW, BDT, BK, BGM it's a better
solution.
As E. Derman pointed out you have 2 rates, the bond yield and the option
maturity discount: only the former varies stochastically.
While trying to correlate both rates may give some results, ultimately you
violate the call put parity. This sequence of problems has been exactly
what led to the BDT model.
You can read the article online at
www.goldmansachs.com/qs/doc/reflections_on_fisher.html
ciao -- Nando
------------------------------
Ferdinando Ametrano
Risk Map
Managing Partner
Via G. B. Vico 4
I-20123 Milano
ITALY
direct line: +39 02 43317503
fax: +39 02 43911424
mobile: +39 338 5911703
email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|