Response to consultation on Regulatory Technical Standards on the exemption from the residual risk add-on own funds requirements for certain type of hedges
Q1.Do you agree with the distinction between instruments with residual risks that are characterised by a non-SbM risk factor, and the rest of the instruments? Please elaborate.
We do not agree that the application of this rule should be restricted to solely non-path dependent options, to the extent that these options contain a non-SbM risk factor, which has been sufficiently hedged.
Therefore, the Industry suggests that hedges of the non-SbM risk factor should be exempt from RRAO charges, regardless of whether the product that is being hedged contains some form of path dependency (e.g. Bermudan Options).
Q2.Do you agree with the requirements set out in Article 2 for instruments with residual risks that are characterised by a non-SbM risk factor? What is your preferred option between option A and option B in point (d) of that Article? Please elaborate, highlighting operational challenges that you may face under the two options.
We overall agree with the conditions to benefit from the derogatory treatment (that remains subject to supervisory approval) set out in Article 2. However, this treatment is subject to a very cumbersome internal policy, which incorrectly refers to counterparties (and to trading desks as this notion of trading desks is only valid under FRTB A-IMA and does not exist under FRTB A-SA), in particular in the case of market making activities (see our answer in Question 4, where we propose to remove the entire paragraph 2 of Article 2); In addition, we wish to highlight the existing and permanent exclusion for hedges in the alternative correlation trading portfolio (ACTP) (CRR2 Article 325u paragraph 2(b)(ii)) does not require any conditions to be met;
We agree that the hedge must reduce the sensitivity of non-SbM risk factor(s). However, both options A and B are not acceptable as they contradict the level 1 text on the possibility of partial hedging (as per CRR3 Article 325u paragraph 6: [… the sensitivities of the combined positions and the risks remaining unhedged in the combined positions, taking into account in particular the possibility that the original position can be hedged by a partial amount]:
- Option A requires a significant reduction that could be interpreted as a reduction of more than 50%;
- Option B proposes a hard threshold of at least a 50% reduction.
While a bank may choose to hedge significantly or fully the sensitivity to non-SbM risk factor(s), it may also voluntarily choose to (considering risk appetite) or have to only partially hedge the original position and the remaining exposure towards the non-SbM risk factor(s) will be managed dynamically within position limits as per Article 2 paragraph 1 (e);
We believe that there is no need to justify that the hedge meets quantitative effectiveness criteria (as it would have been the case in the context of hedge accounting treatment under IFRS for positions in the Banking Book). Indeed, an eligible hedge will always reduce the economic risk of the bank because it must: 1) reduce the risk of the non-SbM risk factor(s) and 2) not introduce any other residual risk other than those stemming from the mapped non-SbM risk factor(s) (which it will reduce) as per Article 2 paragraph 1(c) in order to comply with the objective of Article 325u(4a) to “solely hedge the market risks of positions in the trading book that generate an own funds requirement for residual risks and are subject to the same type of residual risks as the positions they hedge”;
Therefore, we propose a modified version of Option A (Option A’) in point (d) of that Article, where the words “significantly” and “significant” are removed to allow for partial hedging as allowed in the level 1 text:
- Option A’: (d) the institution’s sensitivity towards the non-SbM risk factors is "significantly" reduced as a result of the hedge. The assessment of the achievement of such "significant" reduction shall be done at the level at which the hedge is performed in accordance with point (f)(iv); <See attachment for exact edit>
Q3.Do you agree with the requirements set out in Article 1 for instruments with residual risks that are not characterised by a non-SbM risk factor? In which cases, other than back-to-back positions, do you think hedging instruments would meet the conditions referred to in Article 1? Do you think there are alternative objective ways of assessing whether instruments currently falling under the treatment set out in Article 1 act as a hedge? Please elaborate.
We believe that when a back-to-back transaction has taken place (i.e. the terms of the client trade are offset by a hedging transaction), then the residual risk add-on should be computed on the net notional of the two positions, given that the net risk to the portfolio has been eliminated. Therefore, a zero RRAO charge across the client and hedge instrument is appropriate (and not only the hedge instrument), as per the case of perfect hedges exemption in the Level 1 text of CRR Article 325u(4)(c): “ the instrument perfectly offsets the market risk of another position in the trading book, in which case the two perfectly matching trading book positions shall be exempted from the own funds requirement for residual risks”.
Furthermore, we consider the requirement set out in Article 1 paragraph 1(c)(iii) as overly burdensome, given that trading desks constantly monitor the profit and loss (PnL) and hedge effectiveness of their portfolio as a whole which provides all the necessary information. Thus, additional parametrization of their book to a specific product or subset of a population will not add any value. In addition, the notion of trading desks does not exist in the FRTB SA framework and is only valid in the context of FRTB IMA as per Article CRR Articles 102(4) and 104b(1). Therefore, we suggest removing this requirement in the context of the RRAO framework under FRTB SA.
We note that Article 1 paragraph 1(c)(i) and Article 1 paragraph 1(c)(ii) are highly impractical and not in line with the real practices of a market making desk, as explained in more detail in Question 5.
Q4.What are your views in relation to the requirement to consider whether an instrument has been taken in the interbank market, as a way to distinguish the hedge from the hedged instrument? Which are the cases where the hedge is not performed with the interbank market? Please elaborate.
We acknowledge that the interbank market is a standard practice for hedging however, it is not the sole route that institutions choose to hedge their risks. During the EBA Public Hearing, it was confirmed that there is no requirement to hedge only with interbank counterparties (but this is rather seen to help supervisors to identify hedges). For example, institutions may seek more structured hedge solutions versus non-banks / real money investors (e.g. funds) that are natural providers of risk-reducing solutions for banks (for example in the case of securitisations), who may wish to have exposure to the specific payoff, at arm’s length.
In addition to the topic of the hedge counterparty, it is important to describe the two main trading activities (where banks use CMS spread options for simplification purposes):
- The most common strategy based on client servicing (strategy 1) is where banks hedge an exotic portfolio of multi-look instruments traded with real money investors with vanilla single-look CMS spread options used as hedges,
- Another strategy is more typical and dynamic market-making activity using CMS spread options (strategy 2)
Based on the current draft consultation paper, the requirements for the hedge to be recognized as eligible in the client servicing (Strategy 1) seem to be possible to be met. The hedging documentation framework is very cumbersome to benefit from the hedging exemption and is not suitable for the market-making Strategy 2, putting at risk the possibility to benefit from the hedging RRAO exemption from a capital perspective.
Indeed, for Strategy 2, it is important to note that banks act as market makers and they hedge dynamically throughout the day. Namely, they quote instrument prices to balance bid and offers, such that trades hedge each other, ensuring that the bank is not left with a large directional position. For example, if traders notice a surge in client demand and are delivered a net long inventory, they increase the price to incite clients to sell and decrease the size of their inventory. As a result, the client flow becomes a natural provider of hedges in a market-making activity. Moreover, the client flow may include banks as well as non-bank entities; when a bank aims at hedging a residual position from another bank, then it becomes a client from this other bank’s point of view. Thus, we wish to highlight the case where a bank offers a complex instrument for a client and hedges it only with an interbank trade, which is a very simplified, stylized representation of a much more complex process that should not be a criterion to define hedges. This is a simplified way to represent a typical client servicing strategy that applies to Strategy 1. However, in case of Strategy 2, the representation of a market-making activity is not as simple as is in Strategy 1.
In table 2 below we outline a representative example of a book for a market making desk. Trades may flow in different directions and are influenced by the market maker’s price so that the book remains broadly flat risk.
In the example below,
- Some of the trades were made with ‘clients’, namely the counterparties came to the market making desk in order to have exposure
- Some of the trades have been made with the intent of ‘hedging’, although this is not visible in the book and not recorded anywhere after the trader booked the trade (i.e. the trader’s intentions cannot be recorded and banks should not be expected to do so)
As can been seen in table 1, trades can be made either between banks and/or non-bank counterparties.
<See attachment for details on table 1>
As per the current consultation, the categorization of the portfolio into types of counterparties can be used as additional information that compliments the description of the hedging strategies in Strategy 1 but the type of counterparty used to reduce the exposure to the non-SbM risk factor for Strategy 2 should not be part of the criteria to identify eligible hedges.
In terms of Strategy 2, requesting banks to document and track the counterparties of trades is highly impractical. Namely, a bank may transact with hundreds of counterparties and each of these trades can be a potential hedge to another instrument. As a result, it is not feasible to track the counterparties of all these trades.
In the case of Strategy 1, such information on the counterparty is not relevant to the determination of a hedge in the market risk framework. From a market risk point of view there is no requirement on the credit quality of the hedge counterparty, as opposed to the credit risk management (CRM) framework (a hedge is effective regardless of the counterparty). The type of counterparty will have an impact on the counterparty credit risk, so hedges may have an additional impact on the CCR or CVA RWA. However, the counterparty should have no bearing on the RRAO charge or corresponding hedges. We also note that the CRR3 does not refer to counterparties. The justification for requesting counterparty information in this RTS relies on the ability to better identify hedges but as previously mentioned, this is not in line with the existing practices of a market making desk’s trading activity of Strategy 2.
Therefore, we believe that hedges should be exempt from RRAO charges to the extent that they at least partially reduce the non-SbM risk factor, regardless of the exact type of counterparty used to mitigate the risk in a case of market-making activity.
Consequently, we believe that Article 1 paragraph 2 point (b) should be at least changed from:
“(b) whether the trade was made in the interbank market;”
To:
“(b) the industry of the counterparty (e.g. interbank, funds, other)”
Finally, Article 1 paragraph 2 point (d) should be omitted as the rationale for executing a hedge transaction against a counterparty should not be relevant to inform the determination of the RRAO charges. The same comments apply to Article 2 paragraph 2 point (b) and (d).
Ideally, as the CRR3 amendment does not refer to the nature of the counterparty, he whole paragraph 2 of Article 1 and of Article 2 should be removed.
Q5.What are the material cases where institutions hedge an instrument with residual risks using other instruments with residual risks? Does the proposed regulation address those cases? If not, how can the assessment of the hedge be performed in those cases? Please elaborate.
Case 1: Generic case where regulation does not consider the dynamics of market making desks
In general, the proposed regulation does not consider the dynamics of market making desks, which makes the proposal impractical for such desks. As previously mentioned, trades are hedged dynamically throughout the day to prevent open directional positions. The proposed regulation suggests identifying hedged instrument and hedging instrument, but in reality, such a one-to-one relationship usually does not exist. We instead suggest considering gross position and net position for a certain type of instrument, and then identify the set of trades which are risk reducing as eligible to be called a ‘hedge’.
The Industry proposes the selection of a metric (e.g. sensitivity or stress) representative of the risk to differentiate between long and short positions. The smallest of the two legs would then be excluded from the residual risk add-on as it is providing the hedge. This proposal provides the benefit that there are no requirements to mark individual trades as “hedges” based on trading intent or counterparty and the solution works dynamically, even if trading flow reverses over time as should the previous “hedges” become risk additive if net risk changes direction.
In table 2 <see attachment for details on table 2> we illustrate an example of a sample calculation where the overall short position is risk-reducing and should be considered as the hedge. RRAO RW is applied only on the ‘hedged’ long positions.
Case 2: IR Bermudan Options
Institutions use interest rate (IR) Bermudan options and are typically hedged with other IR Bermudan options in order to limit their exposure to the switch value / Bermudan Tax non-SbM risk factor. This risk factor is measurable, and the hedge efficiency can also be demonstrated. As such, even though there is a finite set of predefined exercise dates, the Bank’s ability to hedge the non-SbM risk is not limited. Therefore, exemption from RRAO charges for the hedges would incentivize prudent hedging strategies that benefit safety and soundness.
IR Bermudan Options Overview
A Bermudan Option gives the holder the right to exercise at pre-specified dates; European options give the holder the right to exercise only at expiry. Consider the premium of European options with maturity aligned to each call date of a Bermudan option; The premium of the Bermudan Option ought to be higher than the maximum premium between the equivalent series of European options.
The delta in premium between the Bermudan Option and the maximum European Option is called the “Switch Value” and is governed by a model parameter typically called “Mean Reversion” used in the Bank’s model. An additional, non-SbM, risk factor called “Berm Tax” ensures that the “Switch Value” the model produces matches prices seen in the market. Therefore, a counterparty that is long the option, i.e. long the switch value, is short the Berm Tax risk, as this is an adjustment that discounts the modelled switch value to hit prices in the market.
Banks hedge this non-SbM risk factor by trading Bermudans against Bermudans. For example, a long client Bermudan position could be hedged by taking a short Bermudan position in the Interbank market or with Hedge Funds to minimize banks’ exposure to movements in the “Switch Value”.
Banks always ensure that their choice of hedging closely matches the client trade characteristics including first and final expiry and exercise interval (distance between dates the option holder can exercise their option). Strike is also taken into account. This ensures a balanced hedging strategy that allows for an active management of the Berm Tax non-SbM risk factor in conjunction with the early exercise feature Bermudan Options contain.
We also wish to highlight that a Bermudan Option’s SbM risks (IR Delta & Vega) can be hedged with vanilla instruments. Hedging with other Bermudan Options ensures that the non-SbM risk factor is also hedged, which is a more prudent risk management practice. However, given that RRAO risk weighted assets (RWA) is currently applicable to the hedge of Bermudan Option trades, which is introducing the wrong incentive as it results in penalizing this prudent risk management practice.
Hedging with Bermudans vs basket of vanilla instruments
Delving further into the hedging practices, it is important to distinguish the effectiveness of hedging a Bermudan Option with another Bermudan with similar but not exactly matching terms (e.g. notional, exact first and last day of exercise, strike etc.) vs hedging it a portfolio of vanilla instruments from an SbM and non-SbM risk factor perspective.
Firstly, a portfolio of vanilla instruments does not provide a hedge for the non-SbM risk factor. Furthermore, changes in the market conditions (e.g. moves in the yield curve or the volatility surface) impact the Greeks (SbM factors) of a Bermudan differently as compared to a portfolio of vanilla instruments used to statically hedge the Bermudan’s risks upon inception. Changes in the market conditions could change the max European component of the Bermudan’s premium, which will lead to changes in the Bermudan’s Delta and Vega risk profile. These changes will be requiring a rebalancing of the portfolio of vanilla instruments to re-optimise the hedging strategy. A Bermudan option hedge would exhibit similar dynamics to the client trade and would reduce the need for dynamic rebalancing of the portfolio of European options making it a more exact and efficient hedge.
A stylized example of a Bermudan trade hedged statically with a portfolio of vanilla instruments (Hypothetical portfolio 2) versus a Bermudan trade partially hedged with a Bermudan with similar terms and a smaller portfolio of vanillas (Hypothetical portfolio 1) can be found in the attachment. The example shows the effectiveness of the hedging strategy under current market conditions and under a scenario whereby the IR Curve & ATM vol surface have been bumped up moderately.
It is evident that the Bermudan hedge allows for better risk management reducing the need for portfolio rebalancing of vanilla instruments.
<see attachment for illlustration on hypothetical portfolio 1 and 2>
Under the current rules, hedging with vanilla instruments (only SbM risk factors hedged) attracts less RWA as compared to a Bermudan Option hedged more effectively (albeit partially) with another Bermudan (SbM and non-SbM risk factors hedged). This is considered a non-intended consequence of the regulation and the current RTS could remediate this issue by providing a RRAO relief for the relevant hedges.
Proposal
We recommend to allow for positions that reduce the Bank’s exposure to the IR Bermudan non-SbM risk factor to become exempt from RRAO charges that will incentivize prudent hedging strategies. Thus, we propose the following change in the Annex II of the RTS:
“ANNEX II
1. options where the pay-offs depend on the path followed by the price of the underlying asset and not just its final price on the exercise date;
2. options that start at a predefined date in the future and whose strike price is not yet determined at the time at which the option is in the trading book of the institution;
3. options whose underlying is another option;
4. options with discontinuous pay-offs;
5. options allowing the holder to modify the strike price or other terms of the contract before the maturity of the options;
"6. options that can be exercised on a finite set of predetermined dates; " <edit can be found in attachment with suggestion to remove>
6. options subject to behavioural risk.”
Another example we wish to highlight are the FX barriers. When a bank has a large parity risk (i.e. a barrier with a big discontinuity) it may well try to hedge it with exactly the same or a similar barrier in the interbank market. Similarly, when offering a more complex structure to a client, deconstructing it into simpler barrier and vanilla types and at least partially hedge in the market is a strategy that may be applied.
Case 3: Structured Note Issues
Structured notes with RRAO payoffs are usually fully or partially hedged by a portfolio of bilateral over the counter (OTC) derivatives with similar RRAO payoffs (e.g. Linear Combination Notes). This is very similar to the ‘market risk offsetting’ condition of the back-to-back criteria to cover originating trade and a collection of hedging transactions with similar market risk profile up to notional size.
For example, consider a bank issuing structured notes to the financial industry (pension funds, insurance companies, other banks, etc.). The bank hedges the note issuance using bilateral derivatives transacted with other bank counterparties.
In the examples below, the underlying is simply denoted as 'X' (this could be any asset class, e.g. Interest Rate, FX Rate, Equity Index etc.) The derivative payoffs are kept generic.
In the first example, the payoff is denoted as f(X), whilst in the 2nd example, the payoff is a simple sum of payoffs g(X), h(X). This could also be extended to a linear combination of pay-offs. In all cases, the payoffs are assumed to be subject to RRAO of 0.1%.
<see attachment for details on Examples 1 & 2>
Case 4: Products hedged with other products containing an exotic underlying
The derogatory prudential treatment defined in Article 2 only covers the scope of instruments bearing other residual risks (an exotic underlying is not treated as a non-SbM risk factor under such a consultation) and the proposed draft regulation suggests identifying hedged instrument and hedging instrument for the same type of residual risk with similar products. Such a one-to-one relationship excludes cases where RRAO charge is related to other reasons, particularly in cases where the instruments are hedged with other instruments that are referencing an exotic underlying.
Thus, the proposed regulation does not address hybrid situations where the hedging instrument references an exotic underlying whereas the hedged instrument is subject to RRAO charge because bearing other residual risks.
As a material example, Autocallable products are path-dependent products that contain other non-SbM risk factors which are also best hedged using instrument that might be considered as exotic underlying. For example, variances swaps are used respectively to hedge stocks volatility embedded in Autocallable products.
Variance swaps overview
For holders of Autocallable products is well known that any significant jumps in the price of the underlying asset can skew the volatility. Variance swaps are well suited for hedging such exposure on volatility. Unlike vanilla options, variance swaps do not require additional hedging whereas options may require recurrent delta-hedging. We believe that when variance swaps are decomposed with the underlying’s sensitivities, should not be in scope of exotic underlying for the RRAO or at a minimum, variance swaps should be excluded from the RRAO charge when they are used as hedges[1].
Therefore, reducing the risk arising from a client deal by hedging it with an instrument that might be considered exotic underlying is penalised even though it does not introduce any additional risk. The sole purpose of these hedging instruments is to cover the risk arising from client transactions. However, the current RTS state that these instruments are applicable to RRAO charge. Therefore, we suggest that institutions should identify such products, incorporate their hedging strategies in their internal policies, and demonstrate the reduction in the non-SbM risk sensitivity as set out in the proposed RTS, Article 2.
Proposal
We propose to allow for instrument positions with exotic underlying that reduce the Bank’s exposure to be exempted from RRAO charges which will incentivize these prudent hedging strategies. This proposal requires that the perimeter of instruments under the derogatory treatment of Article 2 is not limited to instruments bearing other residual risk but also encompasses hedging with instruments referencing an exotic underlying “. In this context, we also propose that Article 2 is amended as such:
[“By way of derogation from Article 1, hedging instruments meeting the conditions set out in Article 325u(2), point (b) of Regulation (EU) No 575/2013 because of their exposure to risk factors that are not included in the sensitivities[1]based method (‘SbM’) laid down in Part Three, Title IV, Chapter 1a, Section 2 of that Regulation (‘non-SbM risk factors’), shall be subject to the exemption laid down in Article 325u(4a) of Regulation (EU) No 575/2013 where all of the following conditions are met:
(…)
(c) the hedging instrument is not exposed to any other residual risk other than those stemming from the non-SbM risk factors mapped in accordance with points (a) and (b); < see attachment for edit and suggestion to remove>
c) the "non-SbM risk factor" terminology refers to both non-SbM risk factors contained in instruments bearing other residual risk (such as correlation risk in spread options) or to exotic underlyings contained in instrument with exotic underlying
(..)”]
[1] Industry response to the EBA draft RTS on the specification of what an exotic underlying is and which instruments are instruments bearing residual risks (see attachment for link)