By Satyajit Das, derivatives expert and the author of Extreme Money: The Masters of the Universe and the Cult of Risk Traders, Guns & Money: Knowns and Unknowns in the Dazzling World of Derivatives – Revised Edition (2006 and 2010)
This four part paper deals with a key element of derivative market reform – the CCP (Central Counter Party). The first part looked at the idea behind the CCP. This second part looked at the design of the CCP. The third part looks at the risk of the CCP itself and how that is managed.
The key element of derivative market reform is a central clearinghouse, the central counter party (“CCP”). The CCP is designed to reduce and help manage credit risk in derivative transactions – the risk that each participant takes on the other side to perform their obligations (known as “counterparty risk”). The CCP also simplifies and reduces the complex chains of risk that link market participants in derivative markets. However, the proposal relies on the ability of the CCP itself to manage risk.
Risque Matters …
The CCP holds the credit risk of cleared derivatives. All participants in the clearing system have exposure to the CCP, specifically its risk management systems.
The basic methodology is that used in exchange traded derivatives. The CCP receives an initial margin or deposit from all parties to a transaction that acts as surety or a security bond against performance. The contract is marked to market daily or more frequently, if market conditions dictate, to establish gains and losses. Parties must post margins to cover the losses on open positions. If a party fails to meet a margin call then the CCP closes out the position, replacing it in the market. The CCP will use the margin it is holding to cover the replacement cost.
The CCP is reliant on risk models and the ability to value contracts. There are significant issues in pricing and valuing contracts and, for some products, reliance on complex models.
The CCP risk management process assumes availability of market prices. In the OTC market, not all instruments trade with liquidity and reliable market prices may not be available. In 2009, Robert Pickel, then Chief Executive of the derivative industry body ISDA (International Swaps & Derivatives Association), told members of the U.S. House Agriculture Committee that some derivative contracts trade infrequently even if they have standardised economic terms.
Under the CCP, only a few instruments will be capable of being marked to market against actual prices. For some instruments, it will be mark-to-model based on inputs that may be validated from market prices. In other cases especially more complex products, it will be a case of mark-to-make-believe or mark-to-myself.
There are significant problems even with standard models for conventional derivatives. One interesting issue is the risk of counterparty credit risk and the extent to which this should be factored into the valuation.
Subsequent to and in response to the global financial crisis, there have been significant changes in the way cash flows are discounted back over time in derivative transactions. Prior to the crisis, it was commonplace to discount back cash flows at the swap rate, which provided a reasonable proxy for the cost of funding for major banks around the world active as derivative traders. The sharp and significant differences in the funding costs of individual banks during and following the crisis have resulted in changes in models. A complicating factor is the use of collateralisation arrangements to enhance the credit of counterparty.
The current trend is for dealers to discount future cash flows in uncollateralised trades at the rate at which each bank can borrow. For collateralised trades, cash flows are discounted at the relevant overnight index swap (OIS) rate. Unfortunately, there is no agreement between dealers on specific aspects of the models.
For example, a US dollar transaction that stipulates the posting of dollar cash collateral should be discounted using the federal funds rate. But this becomes complex where the trade is backed by a variety of different collateral, involving a variety of credit risks, currencies and trading liquidity. While dealers agree the discount rate should in theory be based on the cheapest-to-deliver collateral, it is near impossible to determine what specific security this might be over the entire life of the transaction. It is not clear how the CCP will resolve these issues in the absence of agreement amongst dealers.
For exotic products, the risk of inaccurate market prices is significant. There may no agreement on pricing models and inputs further complicating valuation. David Goldman, a former credit strategist, described quotes for credit default swap (“CDS”) prices in the following terms: “The business looks like the window of a Brezhnev-era Soviet butcher shop. Mouldy scraps hanging in the window. Old women lining up at 4am to try and buy credit protection on General Motors. What are reported as trades are really ways to establish prices to satisfy the auditors.”
CCP risk management relies on models that are variants of the Value at Risk (“VAR”) to establish the level of initial margin consistent with risk. The models are based on historical data and also assume price behaviour of assets inconsistent with actual performance under conditions of stress. These are the same class of models that proved problematic in the GFC.
Some products present special modelling challenges. Small changes in market prices may have large valuation effects; for example, in knock-in and knock-out options or digital options. Similarly, CDS contracts are triggered by defaults. Unexpected and rapid deterioration in the credit condition of an entity can trigger large changes in value – known as “jump to default” risk. Such rapid changes in value are difficult to model and capture in risk management systems.
These problems mean that initial margins may be too low, increasing the risk that the CCP is inadequately protected against counterparty default. Alternatively, the initial margin may be set too high creating disincentives for legitimate risk management activity.
Where a margin is not paid, the mechanics of close-out assume the ability to replace the defaulted contract with a new counterparty at current market prices. This assumes an active market with liquid trading. In the aftermath of the Lehman Brothers’ bankruptcy filing, market liquidity diminished sharply and price volatility increased. It was practically difficult to replace contracts. Market prices and valuations were significantly different from model valuations. It is not clear how these risks will be managed by the CCP.
The CCP will, it is assumed, aggregate all positions across instruments and asset classes for each clearing party. Margins will be based on netting and cross margining across the portfolio of trades. The CCP risk models will need to incorporate correlation between different asset classes and products.
There are important differences between different products and asset classes. For example, a CDS is different from an equity option. The CDS, a form of credit insurance, provides a binary outcome conditional upon default of the reference entity. In contrast, equity prices and the behaviour of equity options more closely approximate a continuous distribution of outcomes.
These differences create modelling problems for formal relationships between asset classes, products and price distributions. Relationships are also likely to be highly unstable. Tractable correlations developed under benign and stable conditions may prove misleading under conditions of stress. These risks may undermine, perhaps severely, the ability of the CCP to manage its risks. Lack of liquid markets in many OTC products may distort prices and compound the problem.
The CCP also requires high quality operational systems to manage its trading, payments, collateral management and risk oversight. All market participants subject to clearing will also need commensurate operational capabilities to manage liquidity demands and the collateral management processes.
Gross and Net of It…
There are two possible clearing models, with different risks. In the first, all participants deal with the CCP directly lodging, margins with the designated clearing entity (“gross clearing”). The second entails non-clearing participants dealing via a CCP clearing member (also known as a clearing broker or in the US a futures clearing merchant) (“net clearing”). In net clearing, non-clearing members have no direct relationship with the CCP when trading. They lodge margins with the clearing member who deals with and is accountable to the CCP for payments and contract performance.
The CCP sets standards for and regulates clearing members. In a net margin arrangement, the relationship between clearing members and clients is entirely negotiated. Key elements agreed include the level of margins, the form of collateral permitted, netting of positions, the timing for meeting margin calls and the clearing fees. Clearing members may also provide credit facilities, funding margin calls on behalf of clients, enabling trading without credit enhancement.
Commercial negotiations focus on the margin levels and type of permitted collateral, including haircuts on securities. Clearing members may cover some or all the margin requirements on a client’s behalf, based on its own internal offsets with the CCP. It may also rely on offsets with the client, cross margining other transactions such as futures, bilateral trades and prime brokerage business. It may also rely on revenues from other business with the client, pricing the clearing function on a holistic basis. Competition between clearing members may reduce risk management standards, reducing the effectiveness of the CCP.
A net margin arrangement creates complex inter-relationships between cleared and uncleared trades as well as different margining and netting models. Assume a transaction involving a cleared OTC derivative and a related uncleared non-standard derivative over it. The cleared derivative requires a CCP margin. Where the two transactions are transacted through a dealer who also acts as a clearing member, the dealer may not require collateral on the uncleared derivative using it own risk model to offset the two positions. This does not result in a lower margin requirement on a client’s cleared transaction, but cuts the total margin paid across cleared and uncleared trades.
Most existing futures exchanges use net clearing. This reflects the administrative and operational complexity of gross clearing. Dealers also favour net clearing, as it creates a profitable business for them clearing non-member trades. In existing exchange traded markets, most of the profits from futures broking comes from not from execution but clearing, including crucial access to client funds that can be reinvested at a profit.
Dealers will push aggressively for net clearing, enabling them to develop a significant business clearing OTC derivatives trades for non-clearing parties. They will argue that this is essential to offset the losses from moving OTC derivatives trading to the CCP.
Net clearing means that the CCP structure will resemble that set out in the Diagram 3 below. In practice, this means that there will be two separate layers of risk – at the level of the CCP and one at the level of the clearing members.
Given that most inter-dealer OTC derivative trading is already collateralised to a substantial degree, the CPP arrangement only formalises these arrangements. For other OTC derivative participants that trade through clearing members, the risk remains with these entities. Given the dominant position of a few firms in OTC derivatives trading and eventually in clearing, this may not reduce risk concentrations significantly as sought.
Diagram 3 – CCP Structure with Net Clearing
Risk Taming …
ISDA’s case against the CCP is based on the fact that OTC products are difficult if not impossible to clear. ISDA argues that the CCP’s ability to clear contracts is conditional upon liquidity and availability of market prices. Pickel on behalf of ISDA testified that this made “it difficult for [the CCP] to calculate collateral requirements consistent with prudent risk management.”
The U.K. Financial Services Authority (“FSA”) also argues that some OTC derivatives may not be capable of clearing. In its December 2009 report Reforming OTC Derivative Markets: A UK Perspective, the FSA did not support mandatory clearing because “the clearing of all standardised derivatives could lead to a situation where a …CCP… is required to clear a product it is not able to risk manage adequately, with the potential for serious difficulties in the event of a default.”
The CCP’s ability to manage risk effectively is questionable, at least for all products. This reflects the lack of availability of prices, limitations of market liquidity and inherent product attributes that may be difficult to model and mitigate. Rejecting the trading of CDS on the futures exchanges, Howard Simons, a Chicago exchange trader, identified the problems of risk management of certain OTC derivatives: “The clearing members of the CME [Chicago Mercantile Exchange] think trading this stuff is the stupidest idea in the world. I didn’t work my whole life so some investment bank can take all our capital. Do I look like Hank Paulson?”
Where products can be cleared, commercial, CCPs may undercut each other on margins and initial deposit requirements to gain market share, in the process undermining the stability of the system itself. Riccardo Rebonato, an experienced risk manager at Royal Bank of Scotland, noted: “In a world where CCPs are competing for an undifferentiated product – clearing – the main differentiating factor for an outsider is going to be the margin and some CCPs may be tempted to compete on margin. But margin must be compatible both with the systemic resilience of the new hub-and-spoke system and with considerations of commercial viability. LCH.Clearnet chief executive Roger Liddell recently criticised newer US rival International Derivatives Clearinghouse for “reckless” behaviour in setting low margin to win business.
On 12 May 2010, the Basel-headquartered Committee on Payment and Settlement Systems (“CPSS”) and the Madrid-based International Organization of Securities Commissions (“IOSCO”) published 15 recommendations for CCPs. The guidelines were vague on risk management issues, only stating the need “more complex models and methodologies” to calculate risk exposure and margin requirements and requiring methodologies to “be reviewed periodically by a qualified, independent internal group or third party”.
CCP risk management may be based on the attributes identified by poet e. e. cummings: “all ignorance toboggans into know and trudges up to ignorance again.”
Earlier versions of this piece have been published as “Tranquillizer Solutions Part I: A CCP Idea” and “Tranquilizer Solutions: Part 2 – CCP Risk Taming” in Wilmott Magazine (May and July 2010)
So true about the liquidity issues and problems with modeling jump-to-default risk for CDS.
I would love to see yourself or someone else address the inherent problems in CDS valuations. For instance, the pricing models tend to assume a set recovery rate that has historically proven to be often incorrect. Do you think that credit event binary options might be a way around the problem, or do these compound the speculatory nature of credit-based derivatives?
Binary options solve one issue by eliminating uncertainty on the recovery rate but create additional problems.
In theory, CDS can be hedged by a position in a corporate bond. The hedge handles the recovery rate automatically.
There is no way to hedge a binary CDS with a corporate bond.
Thanks for a long winded discourse on a very simple problem: you have two kinds of derivatives: interest rate swaps which present no problems, and credit default swaps which present insuperable problems so long as the bankers can pretend they are anything other than insurance. Since they are insurance there is no justification for permitting payoffs to those lacking an insurable interest, i.e. bond holders. You need to regulate CDS as insurance and you need a transaction tax to limit unregenerate gambling. Doing anything else is rearranging Titanic deck chairs.
To quote Carlyle (The French Revolution), a lie cannot be continued indefinitely. Nevertheless, those profiting from the lie will continue lying. It is time to laugh them out of office.
In meltdown conditions or even a significant downturn, the CCP won’t have the capital to replace all the failed counterparties and will itself fail. If the price of these instruments reflected their true costs, they would be seldom or never used. This is the central fact. Everything else is superfluous.