Expose OvernightIndexedCouponPricer

[ralfkonrad]

No description provided.

[coveralls]

coverage: 72.639% (-0.003%) from 72.642%
when pulling 1c9494a on ralfkonrad:feature/expose_OvernightIndexedCouponPricer
into 2243398 on lballabio:master.

[lballabio]

Is this because you have another pricer to use instead of this one?

[ralfkonrad]

Not exactly. I'm currently working on harmonizing things between ORE and QuantLib. #2022 and #2020 are also part of this effort.

One (mid- to long-term) goal would be to replace the ORE implementation in here again by the QuantLib one and (if wanted) "donated" the CappedFlooredOvernightIndexedCoupon also to QuantLib.

Also, it is a bit inconsistent, that the OvernightIndexedCouponPricer is hidden "behind the scenes" but the (experimental) ArithmeticAveragedOvernightIndexedCouponPricer is exposed.

We might also move the pricer into its own header and source file.

Long story short: Exposing the pricer would make things easier.

[lballabio]

Ok, I'm sold. As we make it visible, I would rename it to something more specific though.

Also, may I ask if you can have a look at the ql/experimental/averageois/ and see what's still useful in there? From a quick look it seems that ArithmeticAverageOIS can probably be superseded by OvernightIndexedSwap now, and the same goes for the helper. If that's indeed the case, we might move the arithmetic pricer out of experimental and deprecate the rest.

[ralfkonrad]

Sounds good! I'll have a look.

[ralfkonrad]

Also, may I ask if you can have a look at the ql/experimental/averageois/ and see what's still useful in there? From a quick look it seems that ArithmeticAverageOIS can probably be superseded by OvernightIndexedSwap now, and the same goes for the helper. If that's indeed the case, we might move the arithmetic pricer out of experimental and deprecate the rest.

I think what is missing in OvernightIndexedSwap compared to ArithmeticAverageOIS is the

        MakeArithmeticAverageOIS& withArithmeticAverage(
                                       Real meanReversionSpeed = 0.03,
                                       Real volatility = 0.00, // NO convexity adjustment by default
                                       bool byApprox = false); // TRUE to use Katsumi Takada approximation

feature. Currently, we can't pass these parameters neither to the OIS or the OICoupon. So it can't be set in the ArithmeticAveragedOvernightIndexedCouponPricer and the defaults are used. However, I would like to keep this outside of this PR.

Two remaining questions:

1. Should we keep CompoundingOvernightIndexedCouponPricer inside ql/cashflows/overnightindexedcoupon.hpp or outside in its own header/source files?
2. Should we keep ArithmeticAveragedOvernightIndexedCouponPricer experimental or e.g. combine it with CompoundingOvernightIndexedCouponPricer into ql/cashflows/overnightindexedcouponpricer.hpp?

[lballabio]

I guess losing withArithmeticAverage is ok, those parameters belong to the pricer anyway, not to the swap. But yes, let's keep this out of this PR.

I would go for option 2.

[ralfkonrad]

Any idea why this check fails, I guess it is a false positive?!

ArithmeticAveragedOvernightIndexedCouponPricer::coupon_ is obviously used in the source file.

[lballabio]

False positive, not the first I see with that check on Codacy. I've shut it down.