The EBA defines a ‘commercially reasonable replacement trade’ as “the replacement trade entered into on a netted risk exposure basis, on terms consistent with common market practice and making best efforts in order to obtain best value for money”.
Article 7.1 (a) of the BRRD refers to “the applicable terms of the relevant netting agreement”. According to the close-out process in standard master agreements, the non-defaulting party is under no obligation to enter into actual replacement trades. Therefore, the EBA’s approach does not seem in line with Article 7.1 (a). To respect the terms of standard master agreements, the counterparty should have the right to come up with quotes rather than actual trades.
Counterparties may have an economic interest in coming up with prices for replacement trades which are not ‘commercially reasonable’ from the perspective of the bank in resolution; as such, the resolution authority needs to take the risk of predatory pricing into account. Indeed, the authority will want to avoid having to enter into long commercial negotiations with the bank’s counterparties. If the counterparty has several replacement prices from different market participants (four or five, for instance), this would provide the resolution authority with some reassurance that the trade would be commercially reasonable.
Furthermore, we recommend that the EBA clarifies that replacement prices can be given for a group of closed-out transactions that can be netted, rather than necessarily for each transaction.
Finally, we welcome the definition in Article 1.5 of these technical standards, which refers to the independence of the valuer. This is important to ensure a fair process.
The RTS should state explicitly that the notification of closing out of the derivatives contracts would be undertaken at the same time as the bail-in announcement. Otherwise, this may encourage counterparties to “run” to avoid an imminent resolution and lead to a liquidity squeeze, thereby impeding the bail-in process. Also, if not announced at the same time as the bail-in, the notification could be seen as equivalent to market signalling and may be a breach of market abuse rules.
Moreover, the wording of Article 2.1 may need to specify more clearly that the master agreement will be terminated. The current wording could lead to uncertainties around the termination date. If that date is too far out, this uncertainty may have a negative impact on the legal opinions for netting under the relevant master agreements and impair the nettability for regulatory capital purposes (cf. 295 et seq. of the Capital Requirements Regulation). We recommend the following addition: “of its decision to close-out the derivative contracts and to terminate the master agreement”.
The sequence of events outlined in the draft RTS is not entirely clear. Therefore, we would like to emphasise that in the event of an ‘open bank bail-in’, valuations would need to be finalised and the haircut applied over the ‘resolution weekend’. Operationally, it would be very difficult for counterparties to come up with replacement trades and for these to be analysed by the authority (to ensure they are commercially reasonable) over a weekend. For structured transactions and derivatives traded in illiquid markets it will be difficult to obtain adequate pricing of replacement trades within the required time frame.
In practice, we believe the resolution authority should:
1) Conduct a provisional valuation of derivative contracts during the weekend. The bail-in of derivatives should be carried out on the basis of this provisional valuation.
2) Subsequently, the replacement trades approach could be used to adjust the valuation, as defined in Article 5. The resolution authority will need to set a deadline that leaves enough time for replacement trades to be found but not too long to avoid market volatility and market movements against the bank in resolution.
We understand why the EBA chose the replacement costs approach which takes into account the losses incurred by counterparties to replace the terminated contract. With this approach, which is standard in the close-out rules contained in master agreements, the EBA seeks to ensure that the creditor is not worse off in resolution than in insolvency. Nevertheless, this approach in the context of bail-in would have negative implications.
As mentioned above, inviting counterparties to come up with replacement trades would put tremendous pressure on markets and could lead to predatory pricing, which would as a result increase the losses of the bank in resolution – losses to be borne by shareholders and creditors. Moreover, the replacement cost approach means that the bank under resolution’s portfolio would be broken apart. As a result, the re-hedging would be potentially more costly and the bank could be exposed to open market risk in a volatile market, which would further increase the bank’s losses. The destruction in value would be such that the resolution authority is likely to conclude that it is not worth bailing-in derivatives.
It is very complex to determine an adequate methodology for a swift valuation of derivatives in the event of bail-in. As explained in our general comments, in addition to the destruction in value, the bail-in of derivatives presents contagion risk.
Therefore, when it comes to deciding on exclusions, we expect the resolution authority to consider various options, depending on the causes of failure and the resolution strategy. It might decide to preserve the bank’s derivatives portfolio; or choose to do a portfolio novation to another bank. Novation of derivatives would preserve the value of portfolios for counterparties, respecting the NCWO principle. Also, novation of whole portfolios of derivatives would be less destructive of value for shareholders and creditors of the firm in resolution than closing out certain portfolios and finding replacement hedges.
In light of the replacement costs approach chosen by the EBA, we understand the preferential treatment given to commercially reasonable replacement trades. However, this approach has shortcomings.
As explained in our answer to Question 1, according to the close-out process in standard master agreements, the non-defaulting party is under no obligation to enter into actual replacement trades. Even if the counterparty had to come up with several quotes rather than actual trades, it would be difficult to obtain these in a very short timeframe and under a stress scenario, especially for structured and illiquid transactions. Moreover, the authority would need time to verify that these are commercially reasonable.
In practice, the authority should conduct a provisional valuation to determine whether derivatives should be bailed-in and to determine the bail-in amount (see our answer to Question 5). The replacement trades approach does not seem operationally realistic to inform the bail-in decision, but it could be used in a second stage.
We consider that internal models should provide a reasonable valuation of derivatives, given that valuation models are approved by regulators under the Capital Requirement Regulation (CRR) and respect prudent valuation rules developed by the EBA. The valuations arising from these models reflect all available and appropriate market information, and are subject to IFRS fair value and regulatory audit.
To determine swiftly the value of derivatives, the valuer could look at IFRS Fair Market Value (FMV), with the following adjustments:
• Exclude DVA, or adjust it, such that it reflects the actual losses to be absorbed by the derivative liabilities.
For the purpose of valuing the net derivative liability to be bailed-in, DVA should be excluded from/adjusted in the FMV calculation because DVA represents the fair market value of the bank’s default risk on its own liabilities which may not accurately reflect the loss that a counterparty might suffer under a resolution.
For example, a derivative liability with a theoretical default-risk free value of €100mn, may be recorded on the balance sheet as a liability of €95mn after DVA of €5mn has been factored in as part of the fair value to reflect the market’s perceived default risk of the institution. However, in insolvency the counterparty’s claim would normally be based on the full €100mn FMV of the liability and under a resolution, the derivative counterparty might not be required to absorb any losses (depending on where the derivative liability ranks in the creditor hierarchy and the amount of capital that is left when the bank has been taken into resolution).
The fair value of DVA is set by the market and should reflect the losses that uncollateralised derivatives would suffer under resolution. The DVA will change depending on how the market perceives the bank’s loss absorbency capacity (i.e. once the Minimum Requirement for own funds and Eligible Liabilities and the Total Loss Absorbency Capacity are publicly known). The resolution authority would need to make adjustment to the valuation process accordingly.
• Use Prudent Valuation Additional Valuation Adjustments (AVA) under existing EBA requirements to add an extra layer of conservatism and so that financial valuation and capital accounting are aligned.
Under IFRS, derivative liabilities are measured at fair value using quoted prices in active markets where data is available (i.e. Level 1 in the IFRS hierarchy). However, active markets do not always exist, particularly for tailored derivative contracts that are traded over-the-counter (OTC). Therefore, IFRS allows banks to calculate fair values for these instruments using valuation techniques (i.e. Level 2 and 3 in the IFRS hierarchy).
Valuations for Level 2 transactions are largely based on active markets for constituent parts, whereas valuations for Level 3 transactions can contain a larger number of unobservable inputs. However, banks are expected to use the maximum amount of observable market data available to calculate a derivative’s fair value and to test the validity of their models.
AVA is an additional regulatory adjustment to accounting fair value that is subtracted from regulatory capital. AVA is calculated as the difference between the valuation recorded for accounting purposes and a ‘prudent valuation’ of fair value positions. Under the ‘core approach’ (applicable to large banks) a series of adjustments are applied to the fair value of positions based on a conservative market value with a confidence level of 90%. For example, for a derivative liability that is trading in active markets, the 90th percentile ask price would be used rather than mid-market prices or prices assessed to be reasonable for IFRS accounting purposes adding an additional layer of conservatism to the valuation.
In addition to the adjustments specified under Article 5(4)(c), we would recommend that the resolution authority verifies the ability of market participants to trade at the indicated bids by looking at whether these stakeholders have enough liquidity or whether they are in stressed position. Authorities should also seek to understand the economic motivation behind the trade, to see whether these would be carried out with the sole aim of moving the market.
We agree with the EBA that liabilities of a bank under resolution to a CCP are likely to be exempted from bail-in. Indeed, positions against CCPs are, in most cases, overcollateralised. Also, closing-out liabilities to a CCP would limit hedging opportunities for the bank under resolution at a critical time, when the firm would have to re-hedge its book.
If the resolution authority does decide to close-out derivatives with a CCP, it should not deviate from the default procedures as these have been developed to protect the CCP and prevent systemic contagion. While we understand that Article 5.2 of the RTS would only apply in very exceptional circumstances, this procedure might have risky implications for the CCP.
Regarding the timing, we have the same concerns as regards the early termination date as described in our response to question 2.
Considering the volume of outstanding positions held by a Global Systemically Important Bank (G-SIB), we do believe that there would be increased market volatility following the notification to close-out and that this could adversely affect the termination value.
In practice, we consider that the provisional valuation provided for under Article 7(2) would be necessary to ensure a swift resolution process.
As explained in our answers to questions 1-4, an ‘open bank bail-in’ involving derivatives would not be feasible if the resolution authority was required to wait for final valuations based on commercially reasonable replacement trades before bailing-in net uncollateralised derivatives.
The provisional valuation would need to be based on IFRS fair value, with certain adjustments, as explained in our answer to question 5.
No, we support the idea of a provisional valuation.
The default management procedures already establish a clear process to auction and close-out portfolios of a clearing member. We do not support the idea of an early determination in relation to CCP claims, as it would introduce a lack of transparency in the market.
Transactions with a CCP will be overcollateralised in principle. In the event that a bank has an unsecured liability at the point of resolution, before applying the close-out of those liabilities based on a provisional valuation, the resolution authority should have to account for the cost of closing access to a CCP at a point when the institution might need to continue hedging its risks following the bail-in. In that sense, the bank should be allowed to meet its collateralisation requirements before a decision is made on the bail-in of these derivatives.
In the event that a portfolio of derivatives with a CCP is undercollateralised at the point of the provisional valuation, the resolution authority should consider the cost of closing-out these derivatives. The bank under resolution would lose access to a CCP at a time where it still needs to hedge its risks. The entity under resolution should be allowed to re-establish its collateralisation before a decision is made on the bail-in of these derivatives.
If the CCP determination process holds back the finalisation of the bail-in, then we agree with the proposal of using the NCWO principle as a backstop.
Yes, we consider the guidance sufficiently clear.
We agree that these elements would be the main drivers of the destruction in value. In addition, the RTS should also require the resolution authority to consider the impact of bailing-in derivatives on the ability of the bank in resolution to re-hedge, the timing for re-hedging and the net impact on capital of having one-sided un-hedged risk in the days subsequent to a bail-in, in a potentially volatile and illiquid market.
While we understand the objective of a precautionary buffer, we do not see the need for having a specific buffer for the valuation of derivative liabilities. The final draft RTS on valuation (EBA/CP/2014/38) mandate that any provisional valuation should include a buffer for additional losses. Any uncertainty regarding derivative valuation could be built into this buffer, rather than defining a separate precautionary buffer specifically for the valuation of derivatives.
The resolution authorities should be required to consider the impact on the firm’s franchise value from the termination and close-out of derivative contracts. This is particularly important for banks with an ‘open bank bail-in’ strategy, where the franchise value of the ‘new’ recapitalised bank will have an impact on the market value of shares issued to bailed-in creditors and will thus directly impact losses imposed on those liability holders.
