Quoted daily for each of the main currencies across multiple tenors (1m, 3m, 6m, 12m), Libor represents the rate for which one bank would lend money to another for a set term. It is calculated as the arithmetic average of the submitted rates for each tenor and currency from a panel of banks, with the top and bottom quartile submissions discarded from the calculation.
But – the rates submitted by the banks never linked to actual transactions; they were only ever provided as indicative levels. This made it possible for traders across banks to conspire to change the final average rates calculated by agreeing to submit rates higher or lower than they would otherwise have done.
Indeed, before and during the financial crisis several banks were accused and subsequently fined for manipulating Libor.
It was this practice along with the lack of transparency in linking rates to transactions and the subsequent significant reduction in volumes of interbank unsecured lending that ultimately led to a loss of confidence in Libor and plans to replace it.
The question is, with what?
To enable the market to move away from Libor, three main areas of concern must be addressed:
1. Identify alternative risk-free references (RFRs) to replace Libor
2. In the case of these replacement rates, write fallback alternatives into contracts and specify how they will be used in any existing contracts that reference Libor
3. Adjust the RFRs (essentially overnight rates) to deliver a term rate representation
1. Identify alternative RFRs to replace Libor
The good news is that the five most important markets have already decided on the RFRs that they will use, as shown in the table below.
Country Libor currency Replacement RFR Working group association
Secured Overnight Financing Rate (Sofr)
Federal Reserve Bank of New York
Reformed Sterling Overnight Index Average (Sonia)
Bank of England
Tokyo Overnight Average Rate (Tona)
Bank of Japan
Euro Short-Term Rate (Ester)*
European Central Bank
Swiss Average Rate Overnight (Saron)
Swiss National Bank
*not yet being published
You can find details of these and how they are calculated online, but they all share the following characteristics:
• They all reference actual transactions in the market
• The rates are representative of highly liquid traded markets
• They are overnight rates
Currently, four of the five rates are being published daily. Ester will go live from October 2019, with an indicative rate being published daily up to that date.
Traded instruments referencing the new rates
With the publication of the new RFRs, and the aim of transitioning Libor-based transactions over, several instruments referencing them have been introduced.
In the US, with Sofra live, the CME is now actively trading a full list of futures contracts that reference the new rate and regulators are encouraging institutions to use the RFRs. Hence Fannie Mae made a large issuance which then fuelled the interest rate swaps (IRS) market for basis and outright Sofra swaps.
In GBP there is an active market in Sonia futures, but EUR, JPY and CHF futures are still in the offing.
As for IRS and basis swaps, Isda publishes its weekly numbers and year-to-date stats but numbers are still very low (in billions in terms of notional, but with outright trades this year just counted in tens). Given the uncertainty around how the new RFRs will relate to existing Libor, it is understandable that volumes remain low: participants are still unsure how to hedge Libor-based derivatives with RFR-based equivalents. This means that current volumes are being driven by outright hedging of newly issued RFR-referenced transactions.
2. Going forward: the need for benchmark fallbacks
At present if an Ibor rate is not available the calculation agent would have to obtain quotes from major dealers in the relevant interdealer market. But if an Ibor has been permanently discontinued (as will be the case post 2021), major dealers will be reluctant to quote for something that doesn’t exist. Plus, any quotes that are obtained will probably vary significantly across markets and are unlikely to continue to be quoted for long-dated contracts.
In December, Isda published the results of its Benchmark Fallback Consultation1 for derivatives referencing GBP Libor, CHF Libor, JPY Libor, Tibor, Euroyen Tibor and BBSW. Others were referenced but not directly covered.
Isda sought feedback on two topics:
• How to link RFRs as overnight rates to Libors quoted as term rates;
• Given that Libors are term loans between banks, and so have a credit spread incorporated into the quote, how best to calibrate this?
If these two issues could be resolved, Isda would be able to adjust the fallback options to existing and future Ibor contracts to allow them to continue after Ibor’s demise.
Multiple possible solutions were considered, but the two methodologies outlined below were agreed as routes forward.
3. Adjusting RFRs to term rates compounded
Setting in arrears
The most popular view is to use a compounded setting in-arrears rate for the adjustment RFR. There are four clear benefits of using this methodology:
• Reflects the actual daily rate moves for the relative period;
• Less volatile than a single spot rate as it is calculated as an average;
• An understandable concept for most market participants;
• Relatively common market practice mirroring other traded instruments such as the RFR OIS swaps in EUR/USD/GBP.
But it is not without disadvantages:
• Setting in-arrears means the information needed to determine the rate is not available at the start of the period;
• Actual rate moves may not reflect prior expectations for any given period.
The first disadvantage was viewed as having limited impact since several instruments already trade and fix in a similar fashion in the market. But having a fixing in advance does introduce unnecessary convexity adjustments and mismatches as to how the rate trades in the market.
Adjusting for spread
There was a strong consensus view for using a historical mean/median approach to adjust for the spread. The Isda paper maps out the advantages of this methodology:
• Reflects the current market condition at the time a fallback takes effect and avoids a cliff edge, but transitions to a longer-term average market condition;
• Captures the tendency for interbank rates to ‘mean revert’ to long-term means;
• Reduces the effect of market distortions or manipulation at the time of triggering a fallback;
• Transparent, given it is based on readily available historical information.
But again, there are a couple of disadvantages:
• Unlikely to be present value (PV) neutral on the calibration date;
• Requires a long history of Ibor and RFR fixings.
The simplicity of the methodology, along with its transparency and the fact that it would be difficult to manipulate, makes it a very appealing approach going forward.
The fact that it may not be PV neutral could be mitigated through a longer transition period allowing spreads to migrate towards the average. Similarly, tax implications could be taken into effect along with an understanding of the impacts on counterparty credit risk and CVA modelling.
Further consultations will take place over the coming months to gather feedback on the remaining reference rates: USD Libor, EUR Libor and Euribor, following which further fallbacks will be developed for inclusion in standard contracts.
Subsequent to that, the exact parameters and formulas for the historical mean/median spread approach will be determined, along with decisions as to whether to use mean or median and on the exact historical length to be used. This will most likely take the form of iterative consultations with market participants before the details are finalised.
A word of warning, however.
Isda, although one of the main authorities on the derivative markets, is not the only authority publishing fallback alternative suggestions. The US working group (the Alternative Reference Rates Committee) is currently consulting on fallback language for floating rate notes, loans and securitisation products. In the UK the Bank of England-convened working group on sterling risk-free rates is conducting its own consultation for cash products, while the euro risk-free rates working group is looking at a single solution for Euribor-linked contracts, to cover swaps and bonds.
These alternatives appear to favour a forward-looking approach – the complete opposite of the ISDA fallback approach. In any event, until all outcomes are known, it’s hard for market participants to know what, if any, fallbacks to use.
That doesn’t mean we can’t start planning now.
Calypso is well-positioned to help
As noted, a fair few unknowns still need to be clarified. But from what we know of the current position – and of the direction in which we are heading, it is worth looking at areas where Calypso can help.
As highlighted in the table above, the new RFRs have all been identified and, except for Ester, they are also trading. These indices will follow a uniform methodology and can be captured in Calypso based on client demand. They should be covered off as a replication of the daily OIS indices which preceded them.
Sonia futures are already trading in the market and are fully captured by Calypso. They will not change in terms of definition, so will remain fully supported going forward.
Sofra futures started trading on CME in 2018 with average daily volumes increasing six-fold from May to November. There are two types: the 3-month Sofra future, daily compounding referenced quarterly (mimicking Euro dollar futures); and the 1-month Sofra future, average daily in-arrears over the reference month (mimicking Fed Fund futures). With some minor changes to the naming conventions in Calypso, these are now fully supported in the latest version of Calypso and are currently in production at several clients.
Other futures related to the relevant RFR will come to market in the short to medium term. As they do, Calypso will explore with our clients how best to support these.
Several RFR-referenced bonds came to market in 2018 – notably agency bonds referencing Sofra in the US, and GBP-denominated bonds referencing Sonia. We continue to review our compounding calculators to ensure we can work with our clients to support any new index-referenced bond types and we are also working with primary issuers to cover off their proposed issuances for 2019.
RFR as Ibor replacement
With the exact calculations and details around the Ibor fallbacks not yet known, we are working with what we do know to see how best to support this switch-over in Calypso. Many of the standard calculations will probably resemble those we already use, although the exact process still needs to be clarified.
Going forward we hope to consult with our clients on their anticipated transition planning and work out the best way to replicate that in Calypso.
With the increase in volumes for RFR Futures we can constantly check that our potential curve set-up and boot strapping will support any new future types that our clients might need. With reference to IRS and basis swaps, given that the new indices follow pre-existing market conventions, any new index referenced in an IRS or basis swap will be verified so we can look to best support our clients’ needs.
At present the only curve type we see being used is a CME futures-based curve with Sofra 1 month or 3-month futures as the underlying. But as the fallback spread calculations are finalised, we expect a more liquid IRS and basis market to take off, using such instruments as underlying in a curve.
Unknowns and food for thought
In summary, a few areas remain in which we need further clarity:
• Final methodology to generate the spread between IBOR tenors and RFR
• Adjustment of existing derivatives to conform to the fallback option
• Repricing of adjusted contracts
Basis trading and levels will be closely monitored through 2019 as the fallback option details are finalised. ISDA aims to publish the results of sensitivity analysis to provide a better understanding of the mean/median approach – and short-end basis trading will likely become quite volatile.
The effect of setting in arrears will also be watched closely. Two immediate thoughts spring to mind:
• The replacement rate setting in arrears could have a larger impact than originally anticipated. We are thinking here of range accruals, FRAs and any other IR instrument that needs the IBOR rate close to its fixing time;
• Even vanilla instruments on Ibor rates now become payoffs on compounded rates, which could be an issue – especially during the current coupon period.
And one final thought.
The results of a recent survey published on Risk.Net2 revealed that 10% of users don’t even know if they have Libor-affected contracts. This appears to stem from the hidden implications of Libor in fixed income cash products
– which the recent consultation reports on fallback have not even begun to explore yet, derivatives being the focus to date.
In short, there is still a lot of work to be done – and sadly, no magic wand to get us there. Over the next two years we could benefit from working together to ensure that we transition smoothly into the post-Libor world.
If any of you are interested in swapping thoughts and sharing information about the new RFRs, please contact us on [email protected]
We can make sure we are all heading in the same direction.