Has anyone had this problem with using the SABR model?? Problem persists even when I set Beta=1 (flat backbone)Beta=1 implies the ATM vol remains constant as the forward moves. I am calculating the delta by finite difference. So what is the problem?thank you

Could be convexity depending on what exactly you're looking at.

QuoteOriginally posted by: tjmuaATM Forward Delta NOT EQUAL to 0.5???Why do you expect the delta to be 1/2?Quote Has anyone had this problem with using the SABR model?? This is not a problem and this is not related to the SABR model. This is a fact related to (almost) all models. In the swaption world you can have a look at an article published in Wilmott last year and which give some examples of that fact:M. Henrard. Swaptions: 1 Price, 10 Deltas, and... 6 1/2 Gammas. Wilmott Magazine, pp. 48-57, November 2005As indicated in the article you have the best chance to have a delta close to 50% in a SABR model when beta=0 (and the worst chances with beta=1).

even for Black-Scholes this is not the case, see for example Know Your Weapon Part 1, free downloade on this website