Dealers have welcomed new European proposals to provide flexibility in the calculation of derivatives counterparty risk capital requirements for hard-to-measure clients, but they say the rules are too restrictive to be of significant benefit.
In June, the European Banking Authority (EBA) proposed amendments to the method for calculating the Basel III credit valuation adjustment (CVA) risk charge for counterparties with illiquid or no credit default swaps (CDSs) of their own, and where no proxy CDS of comparable peers exists. Banks, however, say the fact that the new method will apply only to a small set of counterparties means any advantages it offers will be small.
“The first thing which is good with the regulatory technical standard is it allows banks to use more fundamental credit analysis. What is a bit frustrating is that this option is limited to the case where none of the counterparty’s peers have a credit spread,” says a CVA desk head at one European bank. “We have 75% of our counterparties for which we don’t have CDSs, and we believe that probably 10% of them would benefit from the change.”
While the proposals provide some relief for the calculation of loss-given-default (LGD) numbers – a key input in the capital charge along with probability of default derived from market-implied credit spreads – the CVA head says that the changes may not provide much, if any, capital relief overall.
“In my opinion, [the capital impact] is completely neutral. There is a 50-50 likelihood that the exposure will either increase or reduce,” he says.
Credit spreads are a key input for the calculation of Basel III’s CVA risk capital charge, which covers future volatility of CVA risk.
As soon as even one of the counterparty’s peers has a spread, then just because of that you lose all flexibility and have to come back to the mechanical three-factor approachCVA desk head at a European bank
Under Europe’s current CVA rules, banks can use either a standardised or advanced method for calculating CVA risk. The EU implemented the rules in 2013’s Capital Requirements Regulation (CRR), in which Article 383(1) says that for banks on the advanced method that are trading with counterparties without a CDS spread, “an institution shall use a proxy spread that is appropriate having regard to the rating, industry and region of the counterparty”. A regulatory technical standard (RTS) released in 2013 clarified that while the three criteria had to be used primarily to calculate the proxy spread, other inputs could supplement the calculation.
In February 2015, however, the EBA issued a report on the subject of CVA that included the results of a benchmarking exercise. This found that banks used proxy spreads for 75% of counterparties but that as of October 2014, 20% of possible combinations of rating, industry and region lacked a liquid CDS spread.
If a client was a B-rated financial counterparty in Asia, for instance, there was no proxy CDS spread to use. In this example, a firm might have to drop one of the categories – possibly obtaining a spread by looking at B-rated financial counterparties globally. But critics say that if a B-rated Asian financial CDS contract suddenly becomes liquid, it must be used to calculate the proxy spread, causing the overall credit spread suddenly to move.
As such, the EBA recommended that firms be allowed to use alternative approaches “based on a more fundamental analysis of credit risk to proxy the spread of those counterparties for which no time series of credit spreads are available, or any of their peers, due to their very nature”. On June 21 this year, the agency published a draft amending RTS to implement the recommendation made in the 2015 report.
But while banks say the option to use fundamental analysis is a welcome change, they argue they are heavily restricted in where they can apply it.
“There is one word that is really annoying us. The RTS is telling you that when no time series of credit spreads is observed in the market for any of the counterparties, a bank is allowed to use a more fundamental analysis of credit risk. All that is very good, except for the words ‘any of’,” says the CVA desk head at the European bank. “Take an energy producer using coal in Indonesia. Maybe you will find another one using coal in Indonesia, and if by some chance if you have some information on the credit spread in the market … you are not allowed a more fundamental analysis of the credit risk.”
If there are sufficient observations then banks would have to suddenly revert to the original method, which is known as a three-factor proxy approach.
“It is a bit binary because as soon as at least one of the counterparty’s peers has a credit spread liquid on the market, you come back to the three-factor proxy, which is extremely mechanical. Either none of the counterparty’s peers has a credit spread and therefore the bank is allowed a lot of flexibility – which we believe is the right thing to do – or, as soon as even one of the counterparty’s peers has a spread, then just because of that you lose all flexibility and you have to come back to the mechanical three-factor approach,” says the CVA head.
In some cases, there may not be sufficient spread data for all relevant tenors, thus allowing for the application of the alternative approachEuropean Banking Authority spokesperson
Others agree that the changes are restrictive but are hopeful they will help promote a more liquid CDS market.
“This is quite restrictive, but it is restrictive for a reason and I see it as a consistent regulatory approach. The right long-term solution is a wider and more liquid CDS market, as is acknowledged by the EBA paper,” says a derivatives valuation adjustment quant at a second European bank.
A spokesperson for the EBA points out that there has to be sufficient data to generate proxy spreads for all relevant tenors, which could provide flexibility because fewer counterparties are likely to meet that threshold. But this depends on the granularity with which counterparties are categorised.
“In some cases, there may not be sufficient spread data for all relevant tenors, thus allowing for the application of the alternative approach. In addition, the assessment is performed on the basis of an increased granularity. The definition of these more granular categories is left to the institution. Examples in recital 2 include reference to “funds, such as pension funds, collective investment funds or alternative investment funds, but also infrastructure project entities”. In our view, this represents an appropriate trade-off allowing for increased flexibility for firms where this is needed,” says the EBA spokesperson.
One element in the draft RTS that dealers are positive about is the recommendation to allow banks to use the CDS spread of a parent as a proxy where it is more appropriate than its subsidiary.
“This helps for cases where the subsidiary does not have a CDS spread and is in a different region to the parent. Previously, a mapping to a more generic sector, region and rating was required. Now the parent CDS can be used,” says a regulatory expert at a third European bank.
The draft RTS also proposes to allow institutions to reflect the seniority of the netting set in their expected LGD calculations. This is to allow firms to amend their standard 60% LGD input, which is based on senior unsecured debt, if the derivatives netting set is likely to get a different recovery rate in a default scenario.
Some countries, including Germany, require senior debt to be legally subordinated to operational liabilities such as derivatives to allow that debt to count toward the total loss-absorbing capacity buffer. This would mean derivatives recovery rates would be higher than those for senior debt, so the LGD should be lower than 60%. This also had implications for how banks hedged their CVA toward German banks such as Deutsche Bank.
But dealers point out that firms which only use full revaluation to calculate CVA risk – where the entire trade or portfolio is revalued for different inputs – are allowed to take this approach. The draft RTS recommends adjusting the LGD formula for paragraph 1 of CRR Article 383, which covers firms who use full repricing, but it does not amend paragraphs 2b and 2c, which are for banks that do not use that method.
“They haven’t explicitly made any amendment to those two formulas in 2b and 2c, so how is a bank that is not on full revaluation going to be able to use this?” says the regulatory expert.
The EBA spokesperson says that while the amendment does focus on the regulatory formula in paragraph 1, it may be possible to reflect this in paragraph 2 as well.
“This is mainly due to the fact that our mandate to develop RTS on proxy spreads (Article 383(7) of the CRR) refers to paragraph 1 and there is a limit to what we can do from a legal point of view. This being said, considering that all formulas in paragraph 2 are obtained by derivation of the regulatory formula in paragraph 1, there is a sense that the proposed amendment also applies to the formulas in paragraph 2. We stand, of course, ready to further discuss with the European Commission better ways of reflecting this directly in the legal text for improved legal certainty,” says the spokesperson.