Consistent with our reply to ISDA’s July 2018 Consultation on Certain Aspects of Fallbacks for Derivatives Referencing GBP LIBOR, CHF LIBOR, JPY LIBOR, TIBOR, Euroyen, TIBOR, and BBSW (the July 2018 Consultation) and ISDA’s May 2019 Supplemental Consultation on Spread and Term Adjustments for Fallbacks in Derivatives Referencing USD LIBOR, CDOR and HIBOR and Certain Aspects of Fallbacks for Derivatives Referencing SOR (the May 2019 Supplemental Consultation), Chatham continues to believe the key criteria for determining appropriate fallbacks are minimizing value transfer, ensuring ease of operational implementation, continuing hedging utility, and robustness.
Chatham continues to believe the key criteria for determining appropriate fallbacks are minimizing value transfer, ensuring ease of operational implementation, continuing hedging utility, and robustness.
Chatham has evaluated the fallback alternatives offered by ISDA in this final consultation against the key criteria and prefers a compound setting in arrears rate and a forward spread approach that keeps valuations present value neutral on the trigger date. In our previous responses, Chatham consistently preferred the forward spread approach because it constitutes a market-based mechanism for price discovery. In our response to the May 2019 Supplemental Consultation, we no longer supported the historic mean/median approach for calculating the spread adjustment because we observed changes in U.S. market conditions since the release of the July 2018 Consultation. Specifically, Chatham observed that forward spreads between LIBOR and risk-free rate (RFR) proxies have meaningfully converged to their static historic averages.
Based on the results of ISDA’s fallback consultations, the majority of respondents appear to favor the historic mean/median approach and, therefore, we have provided responses to the questions regarding the specific parameters as if the historic approach is chosen. Within the historic approach, Chatham recommends using the historic median with a five-year lookback period.
Given the heterogenous needs of derivatives end users, a one-size-fits-all approach could lead to substantial market disruption. It is critical to remember that the primary function of derivatives markets is to serve the needs of end users for price discovery and risk mitigation. Chatham recommends that ISDA thoroughly consider the implications if even a small minority of market participants are contractually unable or voluntarily choose to forego incorporating fallbacks into legacy derivative instruments. In addition, given the voluntary nature of this contemplated Protocol, ISDA should consider providing additional information to the market as to how existing ISDA Definitions would operate in the event LIBOR is permanently discontinued and a more robust fallback is not agreed to in the contract. Lastly, Chatham believes that given the uncertainties as to how certain IBOR alternatives will work in practice for end users, a flexible Protocol that accommodates different approaches may allow for more market participants to adhere as they can maintain flexibility to align their cash and derivatives instruments.
Responses to Questions
Primary questions relating to the historical mean/median approach to the spread adjustment:
Question 1: Which option do you support? Please differentiate between different IBORs if views differ. We strongly encourage you to limit your response to the options listed in I and II above. However, if you strongly prefer a different option, please explain that and explain why you prefer it over the options above.
In Chatham’s response to the May 2019 Supplemental Consultation, Chatham continued to favor the forward approach but no longer supported the historic mean/median spread adjustment approach, because we observed that U.S. market conditions had changed since the release of the July 2018 fallback consultation. Specifically, Chatham observed that forward spreads between LIBOR and risk-free rate (RFR) proxies had meaningfully converged to their static historic averages, which is detailed in the response to the May 2019 Supplemental Consultation. This observation remains true. Chatham, however, recognizes that the majority of respondents have supported the historic mean/median approach and, therefore, ISDA has presented this approach as the primary method for calculating the spread adjustment. Given this approach as the primary method being considered, we tailored our response to focus on the specific considerations of historic mean/median approach.
The historic mean/median approach aims to capture the cyclical nature of markets and, over time, revert to the mean. Therefore, it is important to understand the relevant time scales for the market to complete a cycle. After the recent financial crisis, it took more than five years for markets to stabilize. Even today, many markets remain in an unusual position of low interest rates. While a 10-year lookback would be attractive due to its inclusivity of different market regimes, Chatham believes that it is difficult to properly test a 10-year lookback due to the presence of the Financial Crisis and lack of longer-term data. Testing with the data currently available would give too much weight to the crisis relative to the rest of the 10-year lookback period. If more historical data were available, it is likely the lookback would result in a more stable and accurate credit spread. Given the limitations of the historic data, however, Chatham recommends the use of the five-year lookback period which better captures the weight of events and excludes the 2008 Financial Crisis in the historic lookback period.
Chatham recommends using the historic median. In our historic scenario analysis, which is detailed in Section 2.2.2 of our response to the July 2018 Consultation, the median historic credit spread resulted in fallback rates that were more similar to the replaced IBORs across currencies and different historic averaging lengths.
Question 2: Would you oppose and/or be harmed by using an option other than the option you supported in response to question 1? If so, which option and why?
Chatham favors the use of (I) a median over a five-year lookback period. Chatham does not anticipate significant harm to market participants if option (II), the trimmed mean over a 10-year lookback period, were used. Derivatives end users, however, may experience value transfer when using a fallback rate that is markedly different from the IBOR it is replacing.
Question 3: Is consistency across IBORs important? Is it critical, very important, somewhat important or not important at all? Please explain.
There are potential problems with using different fallback methods across different currencies. Differing fallback language on a floating-floating cross-currency swap that is not harmonized could cause issues around calculation periods, cash flow dates, etc. For example, in an XY cross-currency swap where X uses a fallback provision with a Compounded Setting in Arrears Rate and Y uses a fallback provision with a Compounded Setting in Advance Rate, there will be a mismatch between when the payment on each side of the cross-currency swap is known. All things considered, minimizing value transfer in order to preserve the economics of the original agreement is a higher priority than the operational challenge of managing inconsistencies across fallback methods.
Question 4: Which is more important to you — your top preference or consistency across IBORs (assuming you could not have both)? Please explain.
As stated in the response to question 3, minimizing value transfer should be a higher priority than concerns about managing inconsistencies across fallback methods. In our analysis our top preference and consistency across IBORS were not in conflict. In a situation where our top preference and consistency were in conflict, then it would be preferable to have our top preference over consistency.
Other technical questions relating to the historical mean/median approach to the spread adjustment:
Question 5: Should the transitional period described above be included in the spread adjustment?
A transitional period should not be included. A transitional period is a mechanism that allows the IBOR fallback rate to transition from the IBOR rate on the date of the discontinuation to a historic average after one year. Chatham believes this transitional period is too long, does not reflect actual LIBOR movements and that there are other potential transition mechanisms.
Looking at the history of LIBOR spreads, the spread returns to its average over a period of a few months. As an example, see the historic spread between 3M USD LIBOR and term-adjusted SOFR provided in the Excel sheet by the Brattle group.
Movements in the spread typically take a few months to return to a more long-term value. The spike that occurred during the 2008 Financial Crisis would not be accounted for by a transitional period because it took more than two years to return to its long-term value.
Chatham also believes that the transitional period allows for market speculation around the proposed discontinuation date. Given the discontinuation date is approximately known, speculators may try to manipulate the spot spread around the discontinuation date. In this case, a speculative spread would be locked in and effect payments for the following year.
Not including a transitional period means that, on the discontinuation date, there may be a jump from the IBOR rate to the IBOR fallback rate. Because LIBOR is already a model-driven rate, it is also possible that the LIBOR submissions will drift to the LIBOR fallback rates in the period before the discontinuation. By not including a transitional period, the LIBOR submission process may naturally provide a smooth transition on the transition date.
Question 6: Should outliers be excluded? Please explain the rationale for your answer.
Outliers should not be excluded. Exclusion of outliers is useful when there are large erratic deviations. The data is “noisy,” so the observed values are not representative of the true values. Exclusion of outliers is also useful if the data is “skewed,” because in this case a simple average is not a good representation of the center of the data.
Outliers can be excluded by removing or trimming data. A median can be considered a fully trimmed mean. All data except the central value has been removed.
The LIBOR and RFR calculation methodologies currently include data trimming. Values near the top and bottom of each day’s LIBOR submission pool are dropped, and the remaining values are averaged. The data underlying RFRs are also trimmed. For example, bilateral repo transactions, which make up SOFR, are trimmed; the bottom lowest values are dropped. The other repo transactions that make up SOFR are not trimmed. The volume weighted median of the remaining repo transactions is then taken to be SOFR.
The spread of LIBOR and term adjusted RFR is therefore a measure of the difference between a trimmed mean and a trimmed median. Excluding spread outliers will have an effect that is like further trimming the LIBOR submissions and repo transactions. Trimming the spread will not be representative of the difference between the current definitions of LIBOR and the RFR.
Question 7: If outliers were to be excluded, to what extent should the data be trimmed? For example:
- the top one percent and bottom one percent of the observations could be removed for the calculation.
- a winsorized mean could be used, pursuant to which observations that are in the top one percent could be set to the 99th percentile value and observations that are in the bottom one percent could be set to the 1st percentile value for the calculation.
- the maximum and minimum could be removed for the calculation.
- only observations within +/- three standard deviations could be used for the calculation.
Should trimming be symmetric (i.e., trimming the same proportion of observations from both tails of the distribution) or asymmetric (i.e., one tail of the distribution should be trimmed more than the other)?
Please explain the rationale for your answers.
As explained in the answer to question 6, data should not be trimmed.
Question 8: If negative spreads have been historically observed for an IBOR/RFR pair, are there compelling reasons to exclude such observations from the calculation of the spread adjustment? If so, what are they?
Chatham concludes there is no reason to exclude a specific observation of a negative spread in the historic data.
Spreads are usually positive and reflect credit risk over time. Negative spreads before the use of negative interest rates were rare. However, negative spreads can reflect unrealized expectations of lower interest rates. If the market expects a cut in interest rate which does not happen, then the spread could be negative. There does not appear to be any reason to exclude negative spread values arising from unrealized expectations of lower rates.
Negative spreads can also indicate more demand for the RFR transaction than LIBOR transaction. For example, the September SOFR spike would make the 1d USD LIBOR – SOFR spread negative. While this was an unusual event, it did happen and may happen again. This negative spread is capturing an actual widespread market dynamic.
Question 9: Negative spreads can be prevalent for some IBORs. If negative spreads have occurred frequently enough that the spread adjustment is itself negative, are there compelling reasons to not implement a negative spread adjustment?
Chatham concludes there is no reason to not implement a negative spread adjustment. Chatham considers the lifetime of the LIBOR fallback and adoption of the Protocol to be the main issues to address regarding negative spreads.
First, how long should one consider the fallback valuation to be trustworthy? If the answer is that the fallback valuation should be valid for a few years past the LIBOR discontinuation date, then negative spreads, which have been in the market for a few years, should be implemented in the fallback. Without any more information, the best prediction of the near future is that it will look like the near past. If the answer is that the fallback should be valid for 50 or more years past the LIBOR discontinuation date, then negative spreads, which have only been in the market for a few years, should not be implemented in the fallback. Over a longer time horizon, rates and spreads are more likely to return to a positive long-term average, as negative rates and negative spreads have only been more prevalent in the past few years.
Second, negative spreads may affect the adoption of the Protocol. To increase the likelihood of Protocol adoption, and without further information about central bank monetary policy, Chatham believes historic spreads should be used regardless of whether it is positive or negative. An unadjusted spread could potentially give rise to a LIBOR fallback rate that is closer to the LIBOR rate on the discontinuation date. This will increase the likelihood of Protocol adoption as most contracts will end within a few years of the anticipated LIBOR discontinuation date.
Primary questions relating to the compounded setting in arrears rate approach:
Question 10: Is it necessary to apply a backward-shift, lockout or similar adjustment to avoid making payments on the same date as the date on which the fallback rate is known? Please note in particular if you would not be able to transact without an adjustment.
In our response to the July 2018 Consultation, Chatham stated the main disadvantage of the Compounded Setting in Arrears Rate from an operational perspective is that the relevant rate would not be known to derivatives end users until the end of each calculation period. This challenge alone represents a dramatic change for end users whose internal finance programs, accounting processes, and treasury departments are structured around knowing the relevant rate at the beginning of each calculation period. Often, treasury departments require approvals prior to making payments. This operational requirement can take several days and relies on the relevant rate being known before the calculation period end dates to make timely payment. While this challenge could be mitigated by applying a backward-shift, lockout, or similar adjustments into contracts, the shift to setting the rate at the end of the calculation period requires significant changes to these internal processes and decision-making frameworks. Due to these significant changes, Chatham supports the need of an adjustment.
Question 11: If an adjustment is necessary, do you support using a two-Banking Day backward-shift, a two-Banking Day lockout or a different adjustment? Please explain your answer.
Chatham and its end-user clients use derivatives to hedge financial risks associated with interest rates, foreign currencies, and commodities. Differences in fallback language for cash instruments and derivatives create a divergence that introduces basis risk, operational difficulties, and accounting challenges. With that said, Chatham is concerned about the potential divergence between fallbacks for cash instruments and derivatives. When entering into derivatives transactions to hedge cash instruments, best practices are predicated on having close alignment between LIBOR cessation triggers, replacement indexes, and spread adjustments.
In the U.S., for example, there have been different conventions used in SOFR issuances, but no single market structure has emerged. Until certain market conventions emerge as standard, end users may be hesitant to adhere to a Protocol which prescribes this type of an adjustment without certainty that it will align with their underlying hedged item. In other words, end users will be more likely to adopt the Protocol if it aligns with the dominant related provisions in the cash markets or allows for flexibility to align with their existing agreements. With the ability to select different adjustments, counterparties can adopt the Protocol and have the freedom to align the economics of their derivative with the underlying hedged item.
With the ability to select different adjustments, counterparties can adopt the Protocol and have the freedom to align the economics of their derivative with the underlying hedged item.
A two-Banking Day backward-shift is the most robust adjustment for the sake of valuations as it will cover the entire period over the life of the instrument and capture all the rates. With a two-Banking Day lockout, fixings would be consistently left out each period.
The difference between the proposed adjustments is usually small due to low daily volatility. While Fed Funds swaps use a lockout, there would not be much difference in the swap value if it used a backward shift unless the lockout period overlaps with a Fed rate adjustment. In this case, the Fed Funds rate can change by tens of basis points overnight.
Repo rates, however, are different from many other interest rates in that they have occasionally changed by hundreds of basis points overnight. For repo rates, the impact of the different adjustments is more obvious, and the backward shift captures the full economics while better matching OIS swap conventions. If flexibility cannot be accommodated in the Protocol, we favor the two-Banking Day backward-shift for this reason.
Question 12: Which cities should apply for the purposes of the two-Banking Day backward-shift or lockout?
Chatham supports applying Banking Days of the city for which the overnight RFR is published (e.g., London for fallbacks for GBP LIBOR, Tokyo for fallbacks for JPY LIBOR and New York for fallbacks for USD LIBOR).
It is also important to consider, however, that when entering into derivatives transactions to hedge cash instruments, a primary goal is to align the economics of both instruments. Many LIBOR loan agreements stipulate London as the city for LIBOR lookbacks, with some stipulating New York and London. Differences in fallback language for cash instruments and derivatives create a divergence that introduces basis risk, operational difficulties, and accounting challenges. As mentioned in previous answers, end users will be more likely to adopt the Protocol if it aligns with the dominant related provisions in the cash markets or allows for flexibility to align with their existing agreements.
Question 13: Would either option be problematic or would you be able to transact if either option were implemented for derivatives fallbacks? Please explain.
Given that Chatham and its end-user clients use derivatives to hedge financial risks associated with their cash instruments, it will be problematic if a company cannot structure a derivative to match the economics and mechanics of a loan they are hedging.
In addition, Chatham believes there are certain derivatives products for which it may be challenging to transact using definitions incorporating fallbacks. Chatham has specific concerns regarding how the fallback methodologies will impact interest rate caps and option products more broadly. Our concerns relate to difficulties in valuing an interest rate cap using the proposed fallbacks. Interest rate caps are a common hedging strategy for end users to set a ceiling on the interest rate payable when borrowing using floating-rate debt. Significant challenges in valuing the product could inhibit the use of caps as a hedging strategy among some market participants.
Technical questions relating to the compounded setting in arrears rate approach:
Question 14: For what products would a two-Banking Day backward shift or lockout not work? Is there any way to address the problems using the “compounded in arrears rate”?
In-arrears products, such as in-arrears swaps and caps, pay the rate from the next interest period instead of the period that just ended. They are not widely traded, but they would not fit in well with a fallback that uses the compounded in arrears rate. Without using a forward-looking rate to calculate the in-arrears products, the only other option would seem to be to change the payment dates to better match up with the original in-arrears product economics.
Another scenario where using the compounded in arrears rate would be difficult without further adjustment is in certain cap agreements where the payment is due in the middle of the interest period. This is a common practice for interest rate caps that hedge floating rate CMBS loans. For example, the Calculation Period would be from the 15th to the 15th of each month, with payments for the current period are paid on the 9th of the month during the Calculation Period. The purpose of this structure is to allow sufficient time for payments to be processed by the servicer and for administrative steps to occur before further payment is made to note holders. This is possible with a forward-looking term rate that is known at the beginning of the period, but the compounded in arrears rate will not be known until the end of the period. Given this, the derivative payment dates may require a change to align with any changes required to the loan and bond payments. Such a change cannot be facilitated solely by the fallback to compounded in arrears rate.
Question 15: Is it problematic to use the Calculation Period instead of the IBOR period?
Using the calculation period instead of the IBOR period can cause problems and possible hedging mismatches. These problems are due to the different nature of IBORs and overnight rates. For regular swaps and caps, these mismatches will be relatively small.
Question 16: Is two Banking Days the correct length of time for a backward shift or lockout? If not, what is the correct length of time?
As stated in the response to question 11, there have been different conventions used in SOFR issuances (as well as other RFRs), but no single market structure has emerged. Until certain market conventions emerge as standard, companies using derivatives to hedge an underlying cash instrument may be hesitant to adhere to a Protocol which prescribes this type of an adjustment without certainty that it will align with their underlying hedged item. In other words, end users will be more likely to adopt the Protocol if it aligns with the dominant related provisions in the cash markets or allows for flexibility to align with their existing agreements. Given the current uncertainties and unknowns, a flexible Protocol that can accommodate different approaches would allow for end users to maintain flexibility to align their cash and derivatives instruments.