September 10th, 2009, 8:40 am
I am testing some perpetual call options on pages 109-110 of Collector's book. My code in C+ and I checked his results OK.Now, I have parameters S = K = 100, r = b = 0.1, sig = 0.2 which ensure that the parameter y1 can be either zero or one, in which case eq. (3.3) can blow up because we have terms like 1/y1 and 1/(1-y1).Is my example pathological or am I missing some assumption in McKean/Merton?//Another off issue is that when I price with larger values oF T in the time-dependent model the prices do not converge to the perpertual price... there's no reason (without proof) why they should.
Last edited by
Cuchulainn on September 9th, 2009, 10:00 pm, edited 1 time in total.