#173 Schedule does not include termination date

[Cathie Jin]

The maturity date of a CDS contract is unadjusted and included in the last coupon period. The FixedRateCoupon doesn't differentiate the last coupon period from the rest.

For example, a quarterly paying 1-year CDS contract on Sept 30, 2014 maturing on Dec 20th, 2015 have the following cash flow from FixedRateCoupon. The last period is missing a day.

Payment Date	Accural Start	Accrual End	Accrual Days
December 22nd, 2014	September 30th, 2014	December 22nd, 2014	83
March 20th, 2015	December 22nd, 2014	March 20th, 2015	88
June 22nd, 2015	March 20th, 2015	June 22nd, 2015	94
September 21st, 2015	June 22nd, 2015	September 21st, 2015	91
December 21st, 2015	September 21st, 2015	December 20th, 2015	90

I used following to create the schedule for CreditDefaultSwap:

Schedule cdsSchedule =
        MakeSchedule().from(effectiveDate).to(maturityDate)
                      .withFrequency(Quarterly)
                      .withCalendar(calendar)
                      .withConvention(Following)
                      .withTerminationDateConvention(Unadjusted)
                      .withNextToLastDate(maturityDate);

Please advice if QuantLib has a special handling on CreditDefaultSwap cash flow generation that I'm missing.

[Luigi Ballabio]

What do you mean when you say that "the last period is missing a day"? Should the dates be different, or should an additional day be counted for the same dates?

[Cathie Jin]

the last period should be 91 days.

12/20/2015 - 9/21/2015 + 1 = 91

[Luigi Ballabio]

Ok. The day-count convention is act/360, right? So it's just a matter of adding one day and there won't be any other strange calculations going on?

[Cathie Jin]

That's all. Adding 1 day to last period with terminating date unadjusted. I don't think it matters what day count it is.

[Aparicio-Navarro Jose]

As I mentioned on Github, can you please check pcaspers branch 'CDS' to check you agree with what is done there? The aim of the branch is to replicate the ISDA standard engine and this point among others is dealt with there.
Thank you

[Luigi Ballabio]

This should now be fixed by the work on the ISDA engine.