IBOR Transition and FRTB cross dependencies highlighted

FRTB brings significant change. With the aim of harmonising capital standards for market risks across jurisdictions and preventing significantly undercapitalised trading book exposures, changes include a revised Trading/Banking book boundary and a revised Internal Models Approach (IMA) approval process (P&L attribution and back testing) with new market risk measures (Expected Shortfall, Default Risk Charge, Non-Modellable Risk Factors). Further changes include a revised Standardised Approach (SA) with a sensitivity-based methodology and a new role in determining the IMA floor, as well as more stringent requirements in terms of policies, procedures and reporting.

After multiple delays, the international Basel Committee on Banking Supervision (BCBS) finally set the deadline for FRTB implementation in 2022 – Q1 for SA and Q4 for IMA.

Overview on dependencies and agenda

There exists a cross dependency between FRTB regulation and IBOR transition requirements defined by central banks. As per FRTB requirements, the new definition of the trading desk requires institutions to have a well-defined and documented business strategy. Institutions need to revise their product strategy to restructure the existing IBOR-based derivatives portfolio and ensure a switch to the RFR-based strategy. Meanwhile, IBOR reform raises important challenges for FRTB implementation, in particular for the risk factor definition and for IMA capital modelling and computation. With the uncertainty around LIBOR publication beyond 2021 and the insufficient volumes of RFR-linked transactions, issues arise in terms of Interest Rate (IR) data and liquidity which may result in material increases in market risk capital requirement.

With their overlapping timelines, both IBOR and FRTB projects are of major importance and require close cooperation between all stakeholders, including front office, market risk managers, IT teams, as well as quants and risk modellers. Fundamental to ensuring a smooth transition to RFRs will be communication and coordination so that everyone is aware of the cross dependencies between these projects. Resources are key. For example, IT resources need to ensure reliable test environments for both projects. Additional resources are required by quants and risk modellers to work on the construction of robust IR models that follow modelling requirements and adhere to the principles of modellable risk factors as required by FRTB.

Impact on market risk modelling: what are the challenges and how will the industry and regulators respond?

From a modelling perspective, the capital requirement computation under SA is less challenging than IMA since all the price determinants, as well as capital aggregation formulas and parameters, are defined by the Regulator. Therefore, cross dependencies with IBOR reform should be limited to an operational challenge: adapting the pricing systems to the new rate curves.

However, the cross dependencies between IBOR and FRTB become more significant where a bank gets regulatory approval to determine its regulatory capital requirements for market risk using IMA.

IR risk factors: definitions and modelling

As defined by the revised market risk framework, risk factors are variables that affect the value of an instrument, for example an equity or IR spot rate. Hence RFRs impact the IR risk factor definition.

By 2022, LIBOR will be less liquid, if not unobservable, and RFR liquidity will increase as they become the reference IR. Therefore, all the IR models currently based on LIBOR market data calibration will need to be amended. At this point in time, the industry is not expecting material model changes since robust modelling assumptions shall remain the same. However, RFRs are overnight fixings and RFRs by term structure are computed by averaging the daily compounded spot rates over the previous months. This means that the rate is backward-looking and will not be suitable for structured products such as caps and floors based on forward-looking LIBOR. Until the regulators provide a final definition of term structure RFRs, this change might require banks to amend their pricing models. Additionally, IR model parameters are subject to re-calibration in order to take account of the new instruments quoted with respect to RFRs, such as Credit Default Swaps (CDS).

Given the tight timeline, as well as the nature and volume of changes implied by the IBOR transition, market participants – particularly from the banking industry – are pleading for regulators to be flexible. It is of the utmost importance that institutions can properly review their models and prioritize the changes consistently with their materiality. For example, changes that would be qualified with different levels of priority include:

• The switch to euro short-term rate (ESTER) discounting by central counterparties (CCPs) on July 27th, 2020: this is considered an important change which requires pricing systems to be adapted by this date.
• RFR data quality and backfilling: this should be considered as a top priority as the computation of risk measures will depend on the reliability of RFR quotes.
• Model re-calibration requiring RFR market liquidity should come later in the agenda. Indeed, credit models based on the CDS spread currently quoted with respect to LIBOR, will be subject to re-calibration when the CDS spread is quoted with respect to RFRs. As published by the FED mid-July 2019[1]: the publication of pre-production estimates of the Secured Overnight Financing Rate (SOFR) going back to 1998 allows banks to study SOFR data characteristics and make  comparisons with other rates such as the US repo and the effective federal funds rate (EFFR). For further examples of data proxies, please refer to our article[2].

Expected Loss: Data quality challenges and risk factor proxies

The Expected Shortfall (ES) computation involves three calculations: ES on the reduced set of risk factors during the current period, ES on the reduced set of risk factors during the stressed period, and ES on the full set of risk factors during the current period. The reduced set of risk factors needs to explain 75% of the fully specified ES model.

On the one hand, while RFRs are important risk factors, banks may want to include them in the reduced set of risk factors. However, the lack of data on RFRs during the 12-month stressed period raises the challenge to find the appropriate proxy for RFRs during this stressed period. On the other hand, with LIBOR becoming increasingly illiquid, IBOR benchmarks will also suffer poor data quality problems and will therefore need to be proxied during the current period.

Expected impacts on regulatory capital requirements

The latest FAQ published by the BCBS[3] in June 2020 provided some relief regarding the forbearance on regulatory capital requirements:

• Both IBOR and RFR price observations can be used in the Risk Factor Eligibility Test (RFET) for a period of time that should not exceed 12 months after LIBOR discontinuation date.
• Regarding the computation of the ES: the RFRs deemed modellable risk factors, can be used to compute the ES on the current period for full and reduced set of risk factors. IBOR benchmarks can be used to compute the ES on the reduced set of risk factors during the stressed period.

Therefore, the RFRs seem to be fully modellable. However, they remain subject to RFET and the reduced set must be approved by the regulator and meet the data quality requirements. For example, the identified reduced set of risk factors must be able to explain a minimum of 75% of the variation of the full ES model. In case a RFR is to be considered as non-modellable or in case one of these conditions[4] is not met, banks would have to compute Non-Modellable Risk Factors (NMRF), which may have a material impact of the capital increase.

External cross dependencies impacting the future of both projects

The COVID19 crisis might have an impact on the IBOR transition agenda and on FRTB implementation: the market turmoil in March 2020 has led to increased doubts on the market liquidity of RFRs. The current situation begs the question whether the end of 2021 is an appropriate date to switch to a new rate world and cease IBOR benchmark publications? Particularly as no one can guarantee transition in the current situation. Plus, RFR trading volumes are still low and the market players are waiting for a catalyst from the Regulators to highly improve to liquidity if the RFRs.

Officially, the agendas are not subject to any changes. But if we want to keep to these deadlines, more coordination is required to ensure RFR market liquidity and to align supervisory strategy and banking industry competence.

[1] https://www.federalreserve.gov/econres/notes/feds-notes/historical-proxies-for-the-secured-overnight-financing-rate-20190715.htm

[2] https://financialservices.mazars.com/ibor-transition-modelling-of-sofr-risk-factors/

[3] https://www.bis.org/bcbs/publ/d503.pdf

[4] For further details, refer to articles 31.12 to 31.25 of the FRTB final rules: https://www.bis.org/bcbs/publ/d457.pdf

Article written by Mariem Bouchaala.

6 August 2020