Re: [Quantlib-users] swap-rate helper

[Luigi B. <lui...@gm...>]

On 02/20/2006 04:38:06 PM, mar...@li... wrote:
>     I have a question about the SwapRateHelper and the way it =20
> computes the final maturity date for a swap.
> In the function
>=20
>    void SwapRateHelper::setTermStructure(YieldTermStructure* t)
>=20
> initially the maturity date of the swap (stored in the private
> variable latestDate_) is computed as
>=20
>    latestDate_ =3D swap_->maturity();
>=20
> where the maturity of the swap is found in the usual way, i.e. by
> shifting the first fixing date by 2 working days and then shifting it
> again using the swap tenor. For me this is the correct way.
> But then, if QL_USE_INDEXED_COUPON is defined, the maturity date is
> found using the indexed coupon framework.

Marco,
	it might be a semantic problem :)
What is (re)computed is not the maturity of the swap---that is always =20
swap_->maturity() and is computed as you describe. What the code tries =20
to find out is rather the latest date up to which the term structure =20
needs to be extended in order to price the swap.

The two things are not the same. Take the example (not implemented, but =20
it gives the picture in an extreme case) of a 5-years CMS swap paying =20
the 10-years swap rate: the maturity of the swap is at 5 years, but the =20
latest date you would need on the curve is at about 15 years.

The same can happen with a swap paying Libor. In your example, the =20
maturity of the swap is indeed 22 Feb 11. But due to date adjustments, =20
the last coupon is a bit shorter than six months from its beginning; it =20
goes from 23 Aug 10 to 22 Feb 11.  On the other hand, the fixing of the =20
Libor rate for that coupon is performed over a period of six months =20
from 23 Aug 10; therefore, the term structure needs to be extended up =20
to 23 Feb 11.

Later,
	Luigi


----------------------------------------

I'd never join any club that would have the likes of me as a member.
-- Groucho Marx