Regulators face a tricky task in setting appropriate capital charges for counterparty risk. While it is widely recognised that this type of risk needs to be better capitalised, many are worried that some of the provisions in Basel III may cause as many problems as they solve.

The depth and scale of the financial crisis prompted a strong response from regulators. Basel III raises minimum Tier 1 capital levels to 7% and global regulators look set to add a surcharge of 2.5% of risk weighted assets on top of that for systemically important financial institutions (SIFIs). 

Regulators have also recalibrated the capital required on the frontier between market and counterparty risk – which banks had seriously underestimated – with a new charge aimed at capturing the volatility in counterparty risk. Banks are already looking at ways to reduce this risk and minimise the capital charges, with some banks revisiting the idea of selling off risk via securitisations. Regulators are, unsurprisingly, nervous.

Bankers would be disingenuous to deny that this risk must be addressed. They argue, however, that the capitalisation and hedging framework put in place will cause as many problems as it solves. They say it will lead to higher prices for credit, add to volatility in the credit markets and increase the phenomenon of 'death spirals' when markets are in crisis. Further, they fear that the push towards central counterparty (CCP) clearing could turn a capital problem into a liquidity crisis.

Thankless task

Regulators face a thankless task. With every step they take, howls of protest can be heard from the banking industry. It is also a complex task, for the crisis revealed the extent to which market and credit risk had collided, exposing banks to much greater risk than they thought. Trading books, primarily monitored for changes in market risk, were found to be more vulnerable to deteriorations in credit quality than previously estimated.

The crisis also demonstrated that asset correlations among financial institutions were higher than originally estimated, with most institutions simultaneously hit by market price volatility. Similarly, 'wrong-way' risk emerged, particularly among financial guarantors whose creditworthiness was eroding at the same time as the value of the assets they had guaranteed.

As counterparty credit risk (CCR) moved to the top of the agenda, an arcane (and largely accounting) practice called counterparty or credit valuation adjustment (CVA) was thrust into the spotlight. It attributes a market value to counterparty risk present within an over-the-counter (OTC) derivatives contract. If the price of an interest rate swap is mostly based on the expectations of interest rate movements during the lifetime of the transaction, the CVA attempts to capture – and pass onto clients in the up-front cost of the swap – what it would mean to the bank if the credit quality of its swap counterparty deteriorated.

According to the Basel Committee on Banking Supervision, roughly two-thirds of losses attributed to counterparty credit risk during the financial crisis were due to CVA losses and only one-third were due to actual defaults. What this represented in cash losses has not been defined. What really concerned regulators, however, was that under Basel II the counterparty default and credit migration risks were addressed but mark-to-market losses due to CVA were not.

The response to improving capital allocation against CCR in the OTC markets is several-fold. Basel III increases the current correlation assumptions (and therefore the risk weightings) of SIFIs by adding a 1.25 multiplier. It increases the risk weightings of CCR for banks using an internal model (expected positive exposures, or EPE) through a set of specific measures such as the concept of stressed EPE, and will require banks to put aside capital against the worst-case exposures. Finally, it adds a new Pillar 1 capital charge aimed at capturing the value-at-risk (VaR) in CVAs.

For banks with big derivatives portfolios – including trillions of dollars-worth of uncollateralised OTC hedging instruments such as interest rate and currency swaps – the CVA charge will have a significant effect. According to ratings agency Standard & Poor's, a large corporate and investment bank can have balance sheet trading assets of more than $500bn and notional levels of OTC derivatives in the hundreds of billions of dollars.

Industry-wide, it involves mind-boggling numbers. Figures from the International Swaps & Derivatives Association (ISDA), put the level of OTC derivatives notional outstanding, excluding foreign exchange derivatives, at $419,000bn at year-end 2010.

Double capital charge

Basel estimates that the CCR capital requirements will double under the new regime. The Bank for International Settlements (BIS) says it has not released any figures from its quantitative impact study, however, banks' own figures show how the new CVA charge will increase risk-weighted assets (RWA). For example, according to presentations for third-quarter 2010 results, Deutsche Bank expects an additional €60bn in RWAs – on top of €51bn under Basel II, Barclays an additional €50bn (on top of €45bn) and BNP Paribas an additional €20bn (on top of €27bn). S&P's says the bulk of these additional RWA come from CVA.

The rules are at least partly intended to fulfil global policy-makers' requirement that more OTC derivatives business is collateralised or is cleared by a CCP as set out in G-20 commitments. While banks will also be required to hold capital against their exposure to CCP default funds, cleared trades will attract a low risk weight, proposed at 2%, and will not be subject to the CVA capital charge.

“The problem for many banks will lie in large legacy books of long-dated uncollateralised derivatives, and around the business that cannot be cleared,” says David Murphy, head of risk and research at ISDA. “The industry has moved quickly on clearing, but there are still going to be a lot of transactions that will not be able to go into central clearing.”

According to figures released by ISDA in May, the level of cleared interest rate swaps, for example, exceeded 50% of interest rate swap notional outstanding at the end of 2010, up from 21% at year-end 2007. Still, the volume of uncleared interest rate swaps outstanding at the end of 2010 stood at $116,000bn.

The regulators' aim of reducing systemic CCR is further complicated by the fact that, for the business that remains outside of clearing, end-users of hedging instruments – including corporates, sovereigns and financials – have lobbied hard to avoid any requirement to post collateral, saying this makes genuine hedging uneconomic.

CVA and risk managers at banks – most unwilling to speak on the record because of banks' sensitivity about regulatory topics – say they understand the need to reduce counterparty risks, but think Basel has got its numbers wrong. Many are also worried that regulators and policy-makers have failed to think about the unintended consequences of the CVA capital charge.

Costs to clients

“We think Basel has underestimated the numbers,” says one European head of CVA. “Our calculations show that, on average, the new charge will most likely triple our CCR capital requirements, and for some individual counterparties the capital charge could increase by up to seven times. This is much more than was required to cover our CCR losses during the crisis.”

He adds that client business will be adversely affected. “We have a profit-and-loss hurdle – which has effectively doubled – so we will have to make considerably more on a trade with a client. This means it will be much more expensive for clients to do any hedging business that we cannot push into clearing, or will make some business uneconomic for the bank.”

Dr Jon Gregory, a partner at consultancy firm Solum Financial Partners, who was global head of credit analytics at Barclays Capital until 2008, says a fundamental issue with CVA is the conceptual framework of the charge, which forces banks to mark CVA to a spread and hedge using the only fully eligible instruments – single name credit default swaps (CDS).

The problem is that only large counterparties have a liquid CDS curve and therefore a market-price spread. So, to hedge against other counterparties' business, banks will have to buy proxy protection using a CDS index, against which they will only achieve 50% of the capital relief available from a single name CDS.

“XYZ Pizza Company is never going to have a spread, so banks have to 'give' it a spread – one not based on market prices – and then mark this proxy spread to market,” says Mr Gregory. “This means hedging is imprecise.”

Small corporates penalised 

Bankers say that, as a result of the charge, smaller corporates will be penalised. “Fewer than 400 out of our 3000-plus clients have a liquid CDS,” says the head of CVA at another bank. “For more than 2600 clients we will be able to get 50% capital relief, at best, using a CDS index. This will mean that we have to charge more to those clients that we cannot easily hedge or buy protection on.”

Some banks say they are already putting in place frameworks that encourage derivatives traders to do shorter-term, more hedgeable business and to cherry-pick which clients they do business with. “We are already incentivising our traders to prioritise some types of business over others; we are finding it difficult to justify long-dated or uncollateralised business,” says the European head of CVA.

The practice of actively managing CVA has been common among most major US banks for several years, where it was driven by US accounting laws. JPMorgan, for example, first began looking at CVA in the 1990s. It now has a central CVA desk which manages all the firm's CVA exposures as a single credit portfolio.

But until recently, CVA management has not been widespread throughout the banking industry. According to a survey carried out in May by risk management software company Quantifi, only about 27% of firms actively manage and hedge CVA.

The pivotal role that can be played by a CVA desk was amply demonstrated by documents released in July 2010 by senator Chuck Grassley of Iowa, the ranking Republican on the Senate Finance Committee. The previously classified documents revealed from which banks Goldman Sachs had bought protection from insurance giant AIG (although how valuable Lehman’s protection, or even Citi’s, would have been is an interesting question). Crucially, while Goldman had bought $1.7bn of protection on AIG versus exposure exceeding $2.3bn, it was the CVA desk which really neutralised this risk; its hedging activities swung the bank to a net positive position of $2.9bn.

There are concerns that the additional cost may even act as a perverse incentive for some clients not to hedge.

David Murphy

One large European bank says that it only began looking at CVA recently because it saw that its derivatives desk was suffering from “adverse selection” and getting a disproportionate amount of higher risk business such as cross-currency swaps. The head of CVA says that the proportion of such instruments to, say, vanilla interest rate derivatives, was so high that it asked corporate clients why they were putting so much cross-currency business the bank's way. “They told us that some other big banks were pricing CVA into transactions, making it too expensive to do business with them, so clients were coming to us because we were cheaper,” he says.

“There are concerns that the additional cost may even act as a perverse incentive for some clients not to hedge. This is not the intended outcome, to say the least,” says ISDA's Mr Murphy.

While the CVA charge will hurt the biggest derivatives dealers the most, Basel III means they are not the only ones that will have to get to grips with CVA. Any bank with a sizeable corporate business could well have a significant derivatives portfolio, and the charge will hurt many second-tier investment banks that have broadened their derivatives business in the past few years. The portfolio may well be full of AAA names such as the World Bank, but on an uncollateralised basis that business will be less attractive under Basel III.

Securitising CVA

Banks are already investigating ways to reduce CVA risk in order to reduce the associated capital charge and there have been reports in the media about the potential to securitise CVA. Whether or not securitisation is an economic choice depends on a bank's risk management strategy and its cost of capital: investors will expect a chunky coupon for taking such risk.

It is not a new idea. Several years ago there was a handful of contingent CDS – which hedged the CCR of a single name – and in the early 2000s, UBS, Deutsche Bank and ABN Amro all sold portions of their derivatives CCR in transactions called Alpine, Eirles Four and Amstel, respectively. Then the driver was to free-up credit lines and increase derivatives trading capacity. This time around, the most sophisticated derivatives houses are said to be revisiting this idea to reduce capital requirements.

Reports suggest that Royal Bank of Scotland (RBS) has executed such a transaction this year and that last year UBS did another in its series of Alpine transactions. Both banks declined to comment.

Olivier Renault, who works at boutique investment bank StormHarbour on structured credit and balance sheet solutions for clients, says it is unsurprising that large dealers with complex derivatives books are looking to revisit this technology – and not just because of the additional capital charge CVA will generate.

“Even on big credits, very often there is a mismatch that leaves the bank still holding risk,” says Mr Renault, who previously worked in Citi's credit derivatives structuring team on regulatory capital, risk transfer and balance sheet optimisation transactions. “For example, if the subsidiary with which you've done a derivative has gone bust, there is a chance that the CDS of the parent company will not get triggered.”

Seeking hedging efficiency

A person with knowledge of the UBS transaction says it was this difficulty of accurately hedging – not capital relief – that was the driver for the bank's transaction. He says the bank bought protection from a single investor on 0% to 8% of the first losses on a $2bn portfolio containing about 400 counterparties – a lot of which were financials – based on the bank's derivatives receivables. The reference portfolio was created based on UBS's potential claims under the derivatives ISDA agreements and as deliverability it will use the bank's claims under ISDA.

“UBS runs a market risk-neutral trading book. For every bit of credit risk it takes the bank hedges externally, so no credit risk is borne by the bank on its potential claims under ISDA. The problem this causes the bank is that some of the credits underlying its portfolio are not hedgeable in the market place and do not trade in CDS. This trade allows UBS to rebalance its hedges such that the names that are in the portfolio now have a direct hedge.”

The regulators are aware of the transaction, says the banker, but formal approval was not needed because UBS can apply the Supervisory Formula Method under Basel II. “It was an extremely clean transaction which achieved full risk transfer for the bank. The transaction took a year to execute with a very sophisticated investor who carried out his own due diligence on the pool and the methodology. The attraction for them was to achieve an above market return for taking that risk.”

Many think regulators will try to stop securitisation of CVA. Naturally, with the ghost of structured products past still haunting regulators and policy-makers, the idea of a collateralised debt obligation (CDO) of a CVA makes many people nervous.

Damiano Brigo, Gilbart professor of financial mathematics at King’s College, London, and formerly managing director and global head of the quantitative analytics team at Fitch Solutions, says this kind of transaction is extremely difficult to model and therefore hard to know the securitisation has been priced correctly and the appropriate amount of capital relief gained.

“To model a CDO of a CVA, you'd need to understand very clearly the volatility of the portfolio and the default risk of the counterparties – you need to be able to see their default probabilities and their volatilities over time. You'd also need to model accurately the correlation between the default of each entity and what drives the portfolio as well as the correlation between the default of one counterparty and the default of another.”

Mr Brigo does not subscribe to the view that it was CDO modelling per se which caused the financial crisis, but he acknowledges that the model is flawed – and that this makes a CVA securitisation a huge challenge. “We still don't have a satisfactory technique to properly model a standard CDO. So to do a CDO of CVA – which is much more complicated because of all the portfolio drivers and the correlations between the portfolios and the default – is that much harder still. It is possible, but it requires an exceptional technical and methodological effort.”

CVA pressure cooker

Solum’s Mr Gregory believes that as long as supervisors ensure that transactions are done to genuinely transfer risk and not to increase capacity, there could be an argument for doing them.

If banks don't find ways to really get some of this risk off their books, then the banking system is going to become a pressure cooker of CVA, with nowhere for it to go.

Jon Gregory

“Regulators see securitisations of CVA as more complex and more dangerous than the securitisations that blew up the world,” says Mr Gregory. “Yes, they are difficult to model and price, but the crisis did not happen because of CDO models; it happened because of the originate-to-distribute business model. If transactions take risk off bank books and give it to investors – and are not a basis for ramping up derivatives trading – then securitisation could be a good thing.”

Mr Gregory is concerned that if regulators shut the door on securitisation because of the financial crisis, risk will build up to dangerous levels. “If banks don't find ways to really get some of this risk off their books, then the banking system is going to become a pressure cooker of CVA, with nowhere for it to go.”

In June, the UK's Financial Services Authority (FSA) issued a proposed guidance on the Supervisory Formula Method and Significant Risk Transfer. It adds third-party rating requirements for the senior notes retained by banks on a transaction seeking capital relief by selling the riskiest portions of their loan books.

The guidance – aimed at banks using the Internal Rating Based approach (which allows banks to model their own risk exposures and come up with a figure for their own RWAs) – seeks to ensure that capital relief is appropriate, says the FSA. Bankers fear that the additional cost this will add to the process will wipe out up to £8bn ($12.9bn) of UK banks' capital credit and deprive lenders of a crucial channel for offloading loan risk.

While the guidance is aimed at the banking book, many worry that as a signal of the UK regulator's approach to securitisation it bodes ill for attempts to offload risk from the trading book using a similar mechanism.

Paul Hawkins, who works on banking policy at the FSA and who was involved in putting together the proposal, maintains that the FSA is not against risk being transferred, “but all risks should be adequately capitalised. It remains important to the FSA that firms do not get disproportionate capital relief from securitisation transactions.”

Basel adds volatility

Mr Gregory also fears that the Basel framework will increase credit market volatility. He believes that increased hedging by CVA desks will push out spreads and contribute to so-called 'death spirals' when markets are stressed. The Bank of England identified this dynamic at work last year as banks sought CDS protection on peripheral eurozone sovereigns.

“Last year, when CVA desks were all trying to hedge their sovereign exposure, the European iTraxx blew out and became heavily correlated to the swap curve, all because of CVA desks hedging mark-to-market positions. It shows that CVA hedging can move the markets,” says Mr Gregory. “And that was with only a proportion of banks hedging CVA. What will be the effect of all banks doing it?”

Who is going to be selling protection at times of market distress? It is hard to see how there will be adequate liquidity in a one-sided market.

Olivier Renault

StormHarbour's Mr Renault also wonders who will be the seller of protection in such times. “When the new rules are in force there will be a double whammy as both loan books and derivatives books rush to hedge,” he says. “But who is going to be selling protection at times of market distress? It is hard to see how there will be adequate liquidity in a one-sided market.”

To VaR or not to VaR?

Some market participants are totally opposed to the VaR approach. “The notion of trying to capture CVA VaR is crazy,” says Mr Gregory. “Banks have all this CVA that cannot be marked-to-market because it has no spread, but they will be forced to give it a proxy spread and mark that to market. And then they will have to charge VaR on that pretend mark-to-market. It is nonsensical.”

Mr Gregory thinks it would be better to treat CVA like a loan book. “You look at the ratings and at the probability of default, put a reserve against it and manage that,” he says.

Some bankers agree. “Our CVA reserve is not counted under Basel III. This is counter-intuitive, as it calculates expected loss over the full life of the trade. If we have a big loss in one year, we can avoid going bust in that year because we already have the reserve to cover us over the remaining life of a trade. A few people think the CVA capital charge is either not required at all, or should take account of the CVA reserve that already exists on the balance sheet,” says one head of CVA.

Enough liquidity to go around?

The logical conclusion of additional CCR capital charges and other measures is to ensure that more OTC business is collateralised and to push as much as possible into central clearing. This will have a seismic impact on the amount of liquid securities required to be posted as initial and variation margin.

David Silver, director of global banking and markets treasury at RBS, says this is also a balancing act. “A credit exposure can be mitigated by collateral. This does not eliminate risk, rather it transforms the risk from capital to liquidity risk. The question for regulators is how to strike that balance.”

According to an impact study by US regulator the Office of the Comptroller of the Currency, as much as $2000bn – the equivalent of three quantitative easing programmes – could be “locked” up in third-party custodial accounts if banks are required to post initial margin when trading with each other under US proposals on margining for uncleared OTC derivatives trades between swaps entities.

It is not only in the OTC markets where liquidity may be at a premium. A BIS study released in June suggested that G-14 dealers could face a cash shortfall in very volatile markets when daily margins are increased by CCPs, triggering demands for several billions of dollars to be paid within a day. “These margin calls could represent as much as 13% of a G-14 dealer's current holdings of cash and cash equivalents in the case of interest rate swaps,” said the BIS.

Many bankers fear that the succession of rules (including liquidity buffers under Basel III) using liquid assets as a credit or liquidity risk mitigant could simply overwhelm the supply of those types of assets.

There is no doubt that the effects of capital and liquidity regulations will be felt in the global economy. Credit will be more expensive and banks will have less capital to allocate to clients. Policy-makers must decide how much of a trade-off they will accept.

“The regulatory calibration will be critical,” says Jeroen Krens, managing director in front-office risk management at RBS. “It's a matter of getting risk management and capital measures in place that make the financial system safer, but limiting any unforeseen consequences to the broader economy.”

PLEASE ENTER YOUR DETAILS TO WATCH THIS VIDEO

All fields are mandatory

The Banker is a service from the Financial Times. The Financial Times Ltd takes your privacy seriously.

Choose how you want us to contact you.

Invites and Offers from The Banker

Receive exclusive personalised event invitations, carefully curated offers and promotions from The Banker



For more information about how we use your data, please refer to our privacy and cookie policies.

Terms and conditions

Join our community

The Banker on Twitter