Collateralized Cash Price — An introduction to the new settlement standard in Swaptions

Collateralized Cash Price will be effective on 26th November 2018. We discuss what has to be done to be well prepared.

Posted by Oliver Kahl on Wed, Nov 7, 2018
Tags: swaptions, python, cash vs. physical
Series: Cash vs. Physical Swaptions


For years the Euro Swaption Market was split into two major types of settlement methods: one being Physical and the other one being a cash settlement based on an Internal Rate of Return (IRR). The latter was causing that many serious valuation problems recently that a new market standard will be established, effective from 26th November 2018 onwards, where the IRR will be "swapped" for a discount rate associated with collateralized swap contracts: a Collateralized Cash Price (CCP). But let’s see what the problem is with the IRR and why the CCP perfectly suits into a world of mandatory collaterlization and clearing and therefore mimics a physical settlement quite closely.

Why was the original date for going-live in July delayed?

Basically there had been issues in the documentation with regards to the discount factors needed for the cash settlement amount. In the original documentation, those discount factors were linked to a static ISDA matrix, referencing EONIA rates for Euro. As those reference rates in the matrix are potentially replaced by new risk-free rates in the near future, the suggested documentation could run into trouble pretty soon. The industry therefore decided to link the discount rate to interest rates applied to variation margin by clearing houses via a so-called "mutually agreed clearing-house". That twist may turn out to be more suitable for whatever comes out of the decisions to be made around risk-free rates.

What is the difference between physical and cash IRR settlement and why is the latter causing trouble?

Let’s firstly turn our attention to the physical settlement of swaptions because probably the most natural way of thinking about swaptions is, when they put into existence one of the most traded derivatives contracts in global markets: a fixed vs. float interest rate swap. Typically the whole valuation machinery in financial institutions rests on swap curve which are physical by their very definition. Building an infrastructure for interest rate options on top of that layer is a quite straight-forward thing to do. Once up and running, our curves projects variable cash-flows out in the future by a forward estimation curve, while those variable cash-flows are discounted to today together with the already fixed cash-flows with another swap curve: the so called discount curve. In essence that means, in the physical world, we are perfectly consistent, all swaps and physical swaptions are valued with respect to the term structure in all relevant swap curves.

Turning to the IRR approach now, yields something significantly different: the swaption value at expiry is just determined by one single swap rate observation. That observation is input for the annuity or payoff as well as the discounting and therefore ignores any term structure in swap curves. Or more precisely, the IRR approach has just one term-structure: a — and just this single — flat swap curve.

Therefore, we must acknowledge that cash settled swaptions based on an IRR have an underlying that is different to the swaps we are generally referring to. Although correlation between the two is high, strictly speaking our IRR swap compared to the (physical) swap has to be modeled separately to capture its unique dynamics.

But why has one come up with such a "distorting" convention in the first place?

For this we fast forward things to the expiry of the IRR swaption and see that things get pretty easy with the IRR methodology: all we have to know is the swap rate fixing and a bit of plug & chug into the IRR formula (PV function in Excel for instance). Very convenient thing to do, so no discussion to arrive at the correct settlement amount. And to be fair, valuation differences, ignoring the speciality of the IRR swaption, tend to be limited. So probably the IRR method was initially agreed to make a settlement easy, while ignoring or tolerating a "small" valuation issue.

Why is the market now deciding to move to CCP?

There are quite a bunch of arguments around suggesting that move: Firstly, since 2008 the market focuses much more on "correctly" discounting cash-flows, by reflecting credit, funding and collateral involved in derivatives transactions. In that spirit, a CCP is more in line with current market standards. Another argument is a steep swap curve (where valuation differences are particular strong), and at the same time a low rates environment where dealers see many receiver swaptions deeply in-the-money. Especially those swaptions already substantially in-the-money cause severe valuation issues pushing bid/ offer spreads to untypically wide levels. Finally, more recent theory suggests that arbitrage is not only possible (as it was seen before) with trading in CMS style products, but also in vanilla products (i. e. swaption collars). A quite active market in zero-wide-collars — often traded with one collar settling IRR and the other physical — further support this argument. In that context switching standards is probably kind of "overdue" and finally was decided by ISDA and global dealers.

How does the CCP compare to physical settlement?

CCP mimics how valuation for the majority of physical swaps is done. Therefore the settlement amount due — according to CCP — is simply the value of the physical swap literally the second, when the options holder opts into exercising the swaption and thereby instantiating the swap. Basically, it is expected that there will be no pricing differences between CCP and physical settlement as long as those contracts are traded under a market standard collateral agreement. Documents are structured in a way that certain adjustments are possible with regards to some specialities in collateral agreements. But as soon as those documents are not in place or business remains uncollateralized things get complicated again: typically that is the moment where an XVA desk comes into play. But even in this more complex situation, it is an advantage to have the risk-free benchmark calculated accurately and then adjust for all the "va’s" afterwards.

What will happen to the old IRR Swaptions? Will there still be a market for them?

So far global dealers seem committed to further support the market for IRR style swaptions. It has to be seen how strong this commitment is and whether the market will dry up over a medium- or long term.

What has to be done to be ready for the switch?

That largely depends on where your swaption business is standing. We will therefore make case separations for three types of swaption businesses.

Case 1: Large book of physical, cash IRR, and CCP swaptions perspectively with already consistent valuation of IRR cash/ physical basis

That is the typical setup for a medium to big institution where valuation is already consistent for all products. Here the adjustment with the new market standard merely boils down to maintenance works. For example, calibration inputs are probably focused at IRR swaption at the moment because of the highest liquidity in that space. So probably that will be fed perspectively from CCP data sources as liquidity is expected to move to the new standard.

Case 2: Large book of physical, cash IRR, and CCP swaptions perspectively with IRR cash/ physical basis not implemented

That’s probably where most options approaching this lie. All of them have in common that they will incur some PnL effects sooner or later.

One option might be to trade out of cash IRR swaptions in advance or in the transition to spare implementation of a consistent framework. Afterwards input data can be simply switched to the new framework and consistency will be achieved for future trades. However, depending on the size of the legacy book and the moneyness of those trades, that can be difficult or costly in terms of what is quoted away from mid by dealers.

Another option on the menu might be holding on to the legacy book, while establishing a consistent pricing framework for legacy trades as well as new trades. That carries some implementation costs, but saves trading cost on closing positions on legacy trades. Potentially a more medium- or long-term strategy to get out of IRR swaptions can be attached to it as well.

One more opportunity might be utilizing providers for switching old IRR swaptions into new CCP swaptions. But at the moment those services are still in their infancy and it has to be seen how effective those will turn out to be.

Case 3: Book of physical swaptions only, intending to solely trade those in the future

The seemingly not affected use-case. Looking a bit under the hood, that assumption might turn out to be wrong. Typically an institution will still value their physical swaptions based on IRR swaption volatilities because those are associated with being the most liquid grid of input data. Once that is changed to CCP, valuation will move and therefore an effect can be seen. But probably the way to go in that scenario is to just take that one-off effect. Still, timing of the effect might be steerable, if one holds on feeding IRR volatility from legacy sources.

How does a consistent valuation in a book with Physical, IRR and CCP Swaptions look like?

That will be the main focus of the next articles in this series. Basically, we will do this in three steps: Firstly, we lie the foundation by obtaining and formatting the needed data. Step two is the heavy lifting, where we calibrate the model. Last step will be utilizing a SABR framework on the calibrated data to complete the volatility cube for all settlement styles.

This is a post in the Cash vs. Physical Swaptions series.
Other posts in this series: