sk77
-
2009-07-23
The swaptions prices produced by analytical formula and by tree methods are significantly different for G2 model. The attached file is a slight modification of BermudanSwaption example from QuantLib distribution. I have run it with various number of time steps in the tree (I would recommend 72 or more). The analytical and numerical (tree) prices are close enough for Hull-White model, but they are significantly different for G2 and Cox-Ingersoll-Ross. Do you have an idea what could be the reason for it?