LIBOR Fallback: is physical settlement an alternative? - Financial fiction

This could be classified as another episode of finance fiction!

In this period of consultation related to the LIBOR fallback, I would like to describe an option which, to my knowledge, has not been mentioned yet: physical settlement. It is not new by itself and not a panacea for the fallback, but maybe it could have a couple of niche applications.

To introduce that approach, I will take the case of caps/floors. If the fallback mechanism is based on compounding setting in arrears (1), as it is envisaged today, the cap/floor optionality changes dramatically. The current LIBOR caps are European options with expiry on the LIBOR fixing date; the post-fallback caps would become Asian options on the composition of overnight rate between the start accrual date and the end accrual date. This is a significant change in term of complexity and valuation mechanism. By luck, I wrote a formula that can price those instruments in a simple one-factor model more than 10 years ago in Henrard (2007). Recently Lyashenko and Mercurio (2019) have proposed an approach that simplify formulas when dealing with backward-looking rates, in particular for cap/floor.

A different fallback mechanism to the one currently planned would be to rely on a term rate or OIS benchmark. But those benchmark do not exist yet and there is reluctance on relying on them.

Can we do something somewhere in between the two? The answer could be a physical settlement of caps/floors. At expiry, the party long the option would have to decide if he wants to exercise the option. This is different from today as the cap/floor exercise is automatic and based in the LIBOR fixing. If the party long a cap exercises, he enters into an OIS for which it pays as a fix rate the cap strike and receive overnight compounded (plus the fallback spread). The dates of the OIS are the dates of the underlying theoretical LIBOR. The party long the cap can still enter into an ATM receiver fix OIS. He can do it with any counterparty. If he does so, he will receive a fix amount equal to ATM-Strike at OIS maturity, a pay-off similar to today's cap. In some sense, the LIBOR cap is transformed into an OIS swaption with physical settlement.

One of the big advantage of this method is that there is not need of OIS benchmark. It is better than compounding in arrears, even if based on the same mechanism, because you don't need to reinvent the wheel (the OIS market) inside your fallback. Just let the existing OIS market do its magic. This give the freedom to the parties to hedge (or not) the outcome of the exercise in the most liquid form.

Suppose that your exercise mechanism is in a cleared OIS. By trading the opposite OIS, you can eliminate the overnight compounding element. If the compounding feature stays into the cap and you trade a cleared OIS to offset the risk, the composition risk may be offset in global market risk terms, but will stay (twice) open in term of operational risk, it will generate double IM and may settle in slightly different ways (day convention).

In today's market, largely electronic, the number of trades is not a major factors. Adding a trade at each exercise date is not a problem. But reducing the notional and IM paid is a real challenge. Dividing the pay-off in small pieces that can be offset easily in the most liquid market is a real value added.

Such a mechanism may even be used for plain vanilla swaps. Suppose that you have a IRS with the same frequency on the fix side than on the floating side, e.g. GBP Fix 6M v GBP-LIBOR-6M. Instead of exchanging a fix coupon for a LIBOR coupon resetting on the coupon start date, the parties enter into an OIS. The party paying the fix rate in the IRS now pays fix on the OIS at the same rate (adjusted for day-count convention if needed) and receive overnight compounded (plus the fallback spread adjustment) for the LIBOR tenor. Like in the cap case, the overnight composition can be cancelled by any party by entering an ATM OIS. If any party do so, he removes completely the overnight composition problem/risk and is left only with the fix amounts. The two parties could even negotiate the ATM OIS rate between them. This would achieve the same result as the OIS benchmark but without the need of a benchmark and without the regulation that goes with it and would increase the liquidity of the pure OIS market on fixed tenors.

Such an approach would be better suited for institutions that have direct access to the OIS market and are familiar with its trading conventions and liquidity. This is not a mechanism to be used in retail, but could be used by large institutions.

Any taker to write a term-sheet for such an approach? Any institution really committed to the development of the overnight-linked market want to try trading those instruments?

Maybe I can sell the idea to Lego; they are used to small blocks that create amazing structures!

(1) As described in many blogs and notes, e.g. Quant perspective on LIBOR fallback, the compounding in arrears is not achievable/workable for most instrument types. In this blog, we suppose that the LIBOR and cap/floor dates align and we are in the lucky case where the approach is actually feasible.


Henrard, M. (2007). Skewed Libor Market Model and Gaussian HJM explicit approaches to rolled deposit options. The Journal of Risk, 9(4). Preprint available at SSRN:

Lyashenko, A. and Mercurio, F. (2019). Looking forward to backward-looking rates: A modeling framework for term rates replacing LIBOR.


Popular posts from this blog

Alae Iacta Est (2)

Making money on LIBOR fallback (5)

Fallback compounding in arrears won't work