Re: [Quantlib-users] CDS standard ISDA model

[Ben W. <ben...@ma...>]

The isda standard model is used to calculate the correct up front amount from a given spread. It is designed so that middle office and traders can agree on what the upfront amount is. The main assumption is that it has a flat hazard rate (rather than a term structure from a CDS curve). (see https://www.cdsmodel.com/)

I has other uses - one of my clients are given a CDS marks as a price as the mark-to-market each day- i.e. it is like  a bond price eg.... . 100+npv of the CDS on $100 face value. We need to turn that price into a spread, so we use the ISDA standard model to back solve from the NPV to the spread. We don’t need a CDS curve to do this which makes our lives much simpler. 

In order to get the exact same upfront amount, we do need to use the same IRS curves. It used to be LIBOR curves, but ISDA has switched to RFR rates some time last year I think.  Ihsmarkit publish these rfr curves (see  https://rfr.ihsmarkit.com/ - yes free market data) so that all the market participants are all on the same page.
 
Hope that helps

Ben 


-----Original Message-----
From: Christofer Bogaso <bog...@gm...> 
Sent: Thursday, April 13, 2023 7:12 PM
To: Ben Watson <ben...@ma...>
Cc: QuantLib Users <qua...@li...>; Luigi Ballabio <lui...@gm...>
Subject: Re: [Quantlib-users] CDS standard ISDA model

Hi,

Just curious on a question outside of QL.

What is exactly **ISDA standard model** for CDS, and how/if it is different from other textbook models (e.g. Hull)?

Any reference/book will be highly appreciated.

Thanks and regards,

On Thu, Apr 13, 2023 at 11:07 AM Ben Watson <ben...@ma...> wrote:
>
> Hi,
>
>
>
> I have adding the ISDA standard model to my CDS class. I have run into 
> one problem when using SOFR.USD to discount – I get this error
>
> return _QuantLib.CreditDefaultSwap_impliedHazardRate(self, *args)
>
> RuntimeError: yield term structure day counter (Actual/360) should be
> Act/365(Fixed)
>
>
>
> The C++ code has this comment
>
> “The yield curve should be LIBOR piecewise constant  in fwd rates, 
> with a discount factor of 1 on the calculation date, which coincides 
> with the trade date.”
>
>
>
> Looks like there is some error trapping on the daycount of the 
> discount curve. With the demise of LIBOR, ISDA has updated the 
> discounting defn to use  RFR’s, USD as an example is now OIS  Actual/360.
>
>
>
> See 3.2 in the document below.
>
> https://rfr.ihsmarkit.com/isda/document/RFR%20Interest%20Rate%20Curve%
> 20Spe cification%20-%206%20Currencies%20(August%209,%202022).pdf
>
>
>
> I could not find the line of code that throw the error, but maybe the 
> best way to fix this is to get rid of the day count checks.
>
>
>
>
>
> Regards
>
>
>
> Ben
>
> _______________________________________________
> QuantLib-users mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-users