Basket Valuation with QuantLib

[Steve Yalovitser]

Hello All,

 

I’m wondering which end of the framework to start at to add copula capabilities to Basket Valuation. It seems that the correlation matrix approach is baked in deeply into path generation.

 

As an aside, there seems to be support for cholesky decomposition but no clear use of it I see…

 

Cheers,

 

Steve

[John Kiff]

I don't know much about basket *valuation* but the cholesky decomposition is my workhorse for
evaluation the loss distribution that I feed into risk evaluations of basket-like products - e.g.,
CDOs. Basically, for my Credit VaR evaluations I use all of the following functions extensively:
qlCholeski, qlNormSinv and qlRand. It's still very much a work in progress, but you're welcome to
take a look at a note that I'm writing on such matters (www.magma.ca/~johnkiff/Loss distribution
math.pdf). John Kiff

--- Steve Yalovitser <[hidden email]> wrote:
> I'm wondering which end of the framework to start at to add copula
> capabilities to Basket Valuation. It seems that the correlation matrix
> approach is baked in deeply into path generation.
>
> As an aside, there seems to be support for cholesky decomposition but no
> clear use of it I see.


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

[Luigi Ballabio]

On 01/14/2005 03:38:21 AM, Steve Yalovitser wrote:
>
> I'm wondering which end of the framework to start at to add copula
> capabilities to Basket Valuation. It seems that the correlation  
> matrix approach is baked in deeply into path generation.

Steve,
        it's a long time since you wrote, I know. However, there are  
news---in the past couple of releases, I worked on generalizing the  
path generation so that correlation is no longer hard-coded. The  
current code should give you a chance to use copulas.

Later,
        Luigi

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

Lubarsky's Law of Cybernetic Entomology:
         There is _always_ one more bug.