The Challenges and Opportunities of FRTB

 
Major South African banks have rather timidly started their journey towards implementing the Fundamental Review of the Trading Book (FRTB), the most significant transformation of the trading market risk framework for the last 20 years. This new capital requirement standard will have significant impacts on the cost of derivative transactions for corporate treasuries, as the increased capital charge will have to be transferred and shared with corporate clients.  Are corporate treasurers fully aware of the main challenges ahead, and how can they use this complex regulation to redefine their business model and move ahead of their competitors? 

FRTB represents a significant and revolutionary change to the existing framework for calculating market risk capital. Following the 2007-08 financial market crisis, which exposed the weaknesses of the Basel II and VaR-based framework, the Basel Committee introduced a set of incremental revisions to the Basel II market risk framework to address the most pressing deficiencies, which were issued under the Basel II.5 denomination.

At the same time, a fundamental review of the trading book was also initiated to tackle a number of structural flaws that were not addressed by those incremental revisions, with the main purpose being to ensure that the standardised and internal model approaches to market risk deliver credible capital outcomes and promote consistent implementation of the standards across jurisdictions. 

The initial FRTB paper was issued in 2013, followed by various iterations until issuance of the latest version in December 2017.  The new rules were initially set to come into force globally by December 2019, but recent delays and complexities to obtain alignment on the rules have pushed the implementation date back to January 2022.

Banks are allowed to implement FRTB under the Standardised Approach or the Internal Model Approach.

FRTB Internal Model Approach (IMA) 

The results of various quantitative  impact studies conducted by banks in the European Union (EU)  showed that FRTB IMA will represent an increase of 1.5 to the current Market Risk Capital charge, which is still less punitive than the whopping 2.4 factor for banks operating under the standardised approach.

The most visible changes brought by FRTB are the replacement of Value-at-Risk with expected shortfall as the basic risk measure for internal model approach, the redefinition of the boundaries between trading books and banking books, and the creation of a new desk level and profit & loss attribution testing regime for internal model approval. 

Furthermore, a bank under IMA must apply a separate capital add-on for risk factors for which it has insufficient data to model, and that separate capital is represented under the label NMRF -  Non Modellable Risk Factors. NMRF identification and their treatment is creating significant operational challenges to banks currently going through the design and implementation of the IMA. Finally, there is a consensus across industry groups that NMRF will account for 30% of total market risk capital under IMA.

FRTB Standardised Approach (SA) 

The standardised approach or sensitivity-based approach rules are based on  using sensitivity of the instruments to underlying risk factors such as delta, vega or curvature to calculate the market risk capital. Those sensitivities are further bucketed based on metrics such as tenor or credit quality. 

The bucketing prescribed by FRTB is not similar to those used currently by most banks in their current risk framework, hence banks will have to re-implement large areas of their trading and risk platforms to meet FRTB SA requirements.

Although FRTB SA appears far less expensive and time consuming to implement compared to the IMA, there are still major challenges such as the treatment of sensitivities on indexes. The rule requires a bank to break down the index into individual components and calculate the sensitivity on those components. 

Overall, the SA rules will save banks considerable time and effort, but will come at a heavy cost on capital charges. The challenge for most banks at this early stage is to decide whether or not they should even consider the IMA rules, and under which business strategy and for which desks.

FRTB implementation challenges

Data

It is not surprising that data is emerging as the source of many worries that banks are facing in their FRTB programme. Under the SA rules, the mapping and bucketing of data to the specified requirements, or the transformation of sensitivities calculated under the current regime to match FRTB rules is a daunting task, even for smaller banks. The sourcing of data for less liquid products and to prevent them falling into the residual-risk add-on highly punitive capital charge is also driving significant effort from banks. 

For those implementing IMA, the data requirements to classify risks as NMRF as well the sourcing of the relevant amount of historical data for the multiples liquidity horizons are the main challenges. Given the serious risk faced by desks under IMA to fail back-testing or profit & loss attribution testing, banks have an extra incentive to make certain that the data required to ensure success on the testing are readily available and accurate. 

Analytics

For banks considering the SA model and looking to leverage its existing sensitivity-based VaR model, there is a complexity to consider, given the difference between most banks’ sensitivity calculation and the prescribed FRTB formula. Some banks might have to duplicate their analytics at a significant cost, with a set of calculations for FRTB and another set of sensitivity calculations for internal risk management, unless the resulting discrepancies between the two set of formulae are minor. 

Some banks are considering the option to build complex transformation rules to convert their current sensitivities into FRTB compliant ones, with major model validation questions already being raised by the regulator for those following that approach.

Computational 

The current market risk framework under Basel II.5/III requires calculation of VaR and Stressed VaR using a single methodology and liquidity horizon. The new framework under IMA, requires multiple liquidity horizon per risk categories, which will basically increase by more than a factor of 10 the computational requirement to calculate internal model market risk capital.

The challenges are forcing banks to re-assess their trading and risk architecture, with techniques to accelerate processing time such as adjoint algorithmic differentiation (AAD), In-Memory aggregation and  grid technology with graphic processing unit (GPUs) considered in isolation or in tandem to tackle the massive computational challenge of FRTB.

FRTB business impact

CRO level and data architecture

Basel II.5/III and BCBS239 regulations have increased the role of the Chief Risk Officer (CRO) in areas related to data sourcing, governance, and aggregation for the purpose of risk management. FRTB increases that trend, with the CRO taking into his/her hands additional responsibilities to ensure alignment between risk and finance.  In order to ensure this alignment under FRTB, data sourcing, management and validation must be controlled by the front office, with policies for data ownership/custodian amended to fit that purpose.

A key interrogation and concern for many banks is the role and ownership of producing risk metrics and capital calculation. In the current framework for most banks, the CRO is in charge of defining the risk framework, operationalising and running the production of risk and capital calculation, hence they are ultimately the owner of data used for risk and capital calculation. 

Given the changing role of the desk heads in the FRTB universe, there is a clear trend to transfer the responsibility of data for risk and capital under the ownership of the front office, with the CRO in charge of risk framework definition and implementation, but with the daily production transferred to front office.

The BCBS239 principles will require banks to opt for data to be owned by the front office in this instance, keeping data where it originated and not disseminating it across the organisation. This change will trigger a move to a decentralised risk model and data architecture, which will be at the opposite of the direction taken by banks designing and implementing centralised data architecture to comply with BCBS239. 

There is a clear indication in the industry that FRTB will help banks simplify their risk architecture and move to a decentralised data architecture model for IMA or a basic centralised approach for banks opting for FRTB SA.

Finance/Product control 

In the current framework, the finance function of most banks is responsible for capital and profit and loss reporting, while the risk function looks after the risk, capital models definition and operationalising. Given the FRTB requirements, the finance function is unlikely to have the skill base and the tools to continue carrying out the final capital calculation and reporting. As a result, some banks might move their capital reporting function to the risk team. 

Front office/Desk level 

Under the FRTB regime, desk heads will be required to be more autonomous in the process of P/L calculation and attribution, and not rely as usual on finance and product control. The P/L attribution and testing will likely move to the risk team, with some shared responsibilities with front office and desk heads.

The desk head will need to have total control of the data used in the P/L attribution, capital calculation and back testing, which will drive a re-alignment of responsibilities between risk, finance and front office. 

Basis risk trading 

FRTB is likely to increase significantly the cost of hedging for banks and corporate treasuries when, for instance, a single stock is hedged with indexes, or when a four-and-half year swap is hedged by a five-year swap. Under the FRTB regime, the current accepted flexibility to hedge Overnight Index Swap (OIS) rate with Libor will come with an extra cost, as it punishes with extra capital anything that does not offset perfectly.

A direct consequence will be a crowding of the market, with all dealers focusing their liquidity position around a common benchmark to the detriment of less traded products of benchmarks. That will naturally increase the cost paid by clients to obtain perfect hedges or support an increase in basis risk charges.

Overall implications for corporate treasurers

FRTB will have a short-term impact on the cost charged to banking clients as a result of the market risk capital charge increase. Banks have historically been the natural providers of liquidity to markets, even for the most illiquid products. The additional capital charge required to support trading on illiquid instruments will reduce liquidity in emerging markets, and transactions on those markets will come with an added premium. All these factors will drive a standardisation of trades used for hedging purposes, and in some cases, treasurers will have to arbitrage between perfect hedge effectiveness at a high cost and ineffective hedge at a far lower cost. 

As banks are reviewing their operating models and discontinuing some business lines as result of the new market risk capital charges, corporate treasurers will have to rethink their own operating models and how to finance and hedge their activities. The cost of derivative transactions is expected to become uneconomic in some cases, and treasurers will have to contain those unhedged risks on their balance sheet and become more creative in structuring natural hedges.

Although FRTB is still a couple of years away from becoming the standard for market risk capital charges, all banks are currently busy reviewing and re-designing their trading activity operating models. In that respect, corporate treasurers need to be fully aware of those upcoming changes, in order to de-risk any ‘cliff’ effect resulting from the sharp increase of derivative transaction costs and the impact on their ability to finance and hedge their operation from January 2022.