New termstructure for bootstrap algorithm

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

New termstructure for bootstrap algorithm

jamal abid
Hi,
 
I am currently using Quantlib and would like to implement some new options with the "Term structure based on piecewise-constant flat forwards with libor-futures-swap bootstrapping algorithm". I would like to implement the same algorithm but with the so-called linear forward and quadratic forward algorithms. There is something that I still don't understand, the current implementation is called piecewise constant flat forward algorithm but where are these constant flat forward rates ? as far as I know in QL, a term strucure with loglinear interpolation is run along optimizing the reevaluation of each instrument till an accuracy is reached.
 
Thanks
 
the naif...

Send instant messages to your online friends http://uk.messenger.yahoo.com

Reply | Threaded
Open this post in threaded view
|

Re: New termstructure for bootstrap algorithm

Luigi Ballabio
On 05/30/2006 11:43:14 AM, jamal abid wrote:

>   I am currently using Quantlib and would like to implement some new
> options with the "Term structure based on piecewise-constant flat
> forwards with libor-futures-swap bootstrapping algorithm". I would
> like to implement the same algorithm but with the so-called linear
> forward and quadratic forward algorithms. There is something that I
> still don't understand, the current implementation is called  
> piecewise constant flat forward algorithm but where are these  
> constant flat forward rates ? as far as I know in QL, a term strucure  
> with loglinear interpolation is run along optimizing the reevaluation  
> of each instrument till an accuracy is reached.

Jamal,
        loglinear discounts are equivalent to flat (instantaneous)  
forward rates; you can verify it by manipulating the conversion  
formula, i.e., (LaTeX ahead)

D(t) = exp(-int_0^t f(\tau) d\tau)

Using the discounts was simply more convenient for the implementation.

Also, the piecewise-flat forward is not really "the current  
implementation"; it was the only one available up to a few releases  
ago, but now it's just a special case of a more generic piecewise yield  
curve. You can find more such curves used in the test suite, including  
curves which actually interpolate forwards.

Later,
        Luigi


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

For every problem there is one solution which is simple, neat, and  
wrong.
-- H. L. Mencken