While the dominant topic of conversation when discussing margin hikes (or reductions) usually reverts to silver, ES (stocks) and TEN (bonds), what everyone so far is ignoring is the far more critical topic of real margin risk, in the form of roughly $600 trillion in OTC derivatives. The issue is that while the silver market (for example) is tiny by comparison, it is easy to be pushed around, and thus exchanges can easily represent the illusion that they are in control of counterparty risk (after all, that was the whole point of the recent CME essay on why they hiked silver margins 5 times in a row). Nothing could be further from the truth: where exchanges are truly at risk is when it comes to mitigating the threat of counterparty default for participants in a market that is millions of times bigger than the silver market: the interest rate and credit default swap markets. As part of Dodd-Frank, by the end of 2012, all standardised over-the-counter derivatives will have to be cleared through central counterparties. Yet currently, central clearing covers about half of $400 trillion in interest rate swaps, 20-30 percent of the $2.5 trillion in commodities derivatives, and about 10 percent of $30 trillion in credit default swaps. In other words, over the next year and a half exchanges need to onboard over $200 trillion notional in various products, and in doing so, counterparites, better known as the G14 (or Group of 14 dealers that dominate derivatives trading including Bank of America-Merrill Lynch, Barclays Capital, BNP Paribas, Citi, Credit Suisse, Deutsche Bank, Goldman Sachs, HSBC, JP Morgan, Morgan Stanley, RBS, Societe Generale, UBS and Wells Fargo Bank) will soon need to post billions in initial margin, and as a brand new BIS report indicates, will likely need significant extra cash to be in compliance with regulatory requirements. Not only that, but once trading on an exchange, the G14 "could face a cash shortfall in very volatile markets when daily margins are increased, triggering demands for several billions of dollars to be paid within a day." Per the BIS "These margin calls could represent as much as 13 percent of a G14 dealer's current holdings of cash and cash equivalents in the case of interest rate swaps." Below we summarize the key findings of a just released discussion by the BIS on the "Expansion of central clearing" and also present a parallel report just released by BNY ConvergEx' Nicholas Colas who independetly has been having "bad dreams" about the possibility of what the transfer to an exchange would mean in terms of collateral posting (read bank cash payouts) and overall market stability, and why a multi-trillion margin call could result in the biggest buying spree in US Treasurys... Ever.
First, for those who are unfamiliar, here is what happens when an exchange (or a Central Counterparties) proceeds to trade OTC derivatives with any given counterparty (from the BIS):
CCPs typically rely on four different controls to manage their counterparty risk: participation constraints, initial margins, variation margins and non-margin collateral.
A first set of measures are participation constraints, which aim to prevent CCPs from dealing with counterparties that have unacceptably high probabilities of default.
The second line of defense is initial margins in the form of cash or highly liquid securities collected from counterparties. These are designed to cover most possible losses in case of default of a counterparty. More specifically, initial margins are meant to cover possible losses between the time of default of a counterparty,8 at which point the CCP would inherit its positions, and the closeout of these positions through selling or hedging. On this basis, our hypothetical CCP sets initial margins to cover 99.5% of expected possible losses that could arise over a five-day period. CCPs usually accept cash or high-quality liquid securities, such as government bonds, as initial margin collateral.
As the market values of counterparties’ portfolios fluctuate, CCPs collect variation margins, the third set of controls. Counterparties whose portfolios have lost market value must pay variation margins equal to the size of the loss since the previous valuation. The CCP typically passes on the variation margins it collects to the participants whose portfolios gained in value. Thus, the exchange of variation margins compensates participants for realised profits/losses associated with past price movements while initial margins protect the CCP against potential future exposures. Variation margins, typically paid in cash, are usually collected on a daily basis, although more than one intraday payment may be requested if prices are unusually volatile.
Finally, if a counterparty defaults and price movements generate losses in excess of the defaulter’s initial margin before its portfolio can be closed out, then the CCP would have to rely on a number of additional (“non-margin”) resources to absorb the residual loss. The first of these is a default fund. All members of the CCP post collateral to this fund. The defaulting dealer’s contribution is used first, but after this other members would incur losses. The default fund contribution of the defaulting dealer would be mutualised among the non-defaulting dealers according to a predetermined formula. Some additional buffers may then be available, such as a third-party guarantee or additional calls on the capital of CCP members.
Otherwise, the final buffer against default losses is the equity of the CCP. In order to calculate initial and variation margins, CCPs rely on timely price data that give an accurate indication of liquidation values. Clearing OTC derivatives that could become unpredictably illiquid in a closeout scenario could impose an unacceptable risk on the CCP.
Table 1 summarises the risk management practices of SwapClear, ICE Trust US and ICE Clear Europe, which are currently the main central clearers of IRS and CDS.
The gist of the BIS paper focuses not so much on the inboarding costs and concerns of migrating hundreds of trillions of products to CCP - a topic evaluated much more in depth by Nick Colas - the BIS does instead look at a hypothetical example of what may happen in the case of a "risk flaring" episode, and how much variation margin G14's may need to post:
As shown in the left-hand panels of Graph 2, estimated initial margins can vary significantly with prevailing levels of market volatility, especially for CDS. The upper left-hand panel shows, for example, that Dealer 7 would need to post $2.1 billion of collateral to clear its hypothetical IRS portfolio in an environment of low market volatility, similar to that prevailing before the recent financial crisis. This would grow by around 50%, to $3.2 billion, if volatility increased to the “medium” level seen early in the crisis, just before the rescue of Bear Stearns. And it would grow by around 150%, to $5.3 billion, if volatility increased to the “high” level seen at the peak of the crisis, amidst the negative market reaction to the US Troubled Asset Relief Program (TARP) and before government recapitalisation of banks began in the United Kingdom. In comparison, the bottom left-hand panel shows that initial margin requirements for the hypothetical CDS portfolio of Dealer 7 would increase by around 160% or 325% from $0.6 billion if the prevailing level of market volatility increased from low to medium or high. The total initial margins that the CCP requires clearing members to post are $33 billion (low), $70 billion (medium) and $105 billion (high) for IRS and $6 billion (low), $20 billion (medium) and $35 billion (high) for CDS.
Nevertheless, it seems unlikely that G14 dealers would have much difficulty finding sufficient collateral to post as initial margin. The diamonds in the left-hand panels show collateral requirements relative to dealers’ unencumbered assets, with different colours again representing different levels of market volatility. Even the requirements based on high levels of volatility do not exceed 3% of the unencumbered assets of any dealer for which it was possible to estimate this figure. Although many unencumbered assets held by dealers do not presently qualify as acceptable collateral for initial margins, some of these could be swapped for assets that do qualify.
By contrast, dealers may need to increase the liquidity of their assets as central clearing is extended. The centre panels of Graph 2 show similar patterns in potential variation margin calls as prevailing levels of market volatility change. In the worst case, variation margins could be several billions of dollars, which would have to be paid in cash within a day. These margin calls could represent as much as 13% of a G14 dealer’s current holdings of cash and cash equivalents in the case of IRS. A five-day sequence of large variation margin calls that could be expected with a probability of one in 200 would equate to around 28% of current cash and cash equivalents in the worst case.
These results also have direct implications for the liquidity provisions of CCPs, as they would have to pay variation margins in the case of default of a clearing member. Access to central bank funds in distressed circumstances would help to ensure that CCPs could make substantial variation margin payments in a timely manner.
With a probability of one in 10,000, non-margin resources at risk from the failure of one particular dealer, two particular dealers or any dealer with sufficiently adversely affected portfolios would respectively be 20%, 37% and 42% of total initial margins for IRS, and 36%, 46% and 65% of total initial margins for CDS. If prevailing levels of volatility were high, these figures would equate to $21 billion, $39 billion and $44 billion for IRS, and $13 billion, $16 billion and $23 billion for CDS. By comparison, the G14 dealers contributing to default funds had equity of around $1.5 trillion as of 30 June 2010.
Alas, the problem is that the bulk of this "equity" is, for lack of a better word, worthless, as it is based on such assets as intangibles, and MTM-locked up assets, whose true value is far, far lower than where banks carry these. And of course, the need to sell them would come precisely at a time when everyone else would be selling. Which means that in the event of a market lockup, there would be no one on the other side of the trade, meaning the entire CCP experiment would likely collapse spectacularly, as nowhere near enough cash is available.
Next, we look at one of the "percentile probability" charts to determine just where the system is weakest, because these uber-6 sigma events tend to become the norm when TSHTF. According to the BIS, the absolute worst case scenario from a risk management standpoint is a 0.002% probability event, at which dealers could see $160 billion in total margin shortfalls across the IRS and CDS book. And there are those who wonder why banks are stockpiling cash for a rainy day...
The BIS issues a rather ominous warning at this point:
Even after incorporating expected shortfalls into initial margin requirements, however, a sizeable gap remains between the total margin shortfalls (relative to total initial margins) that could be expected with very low probabilities for CDS and equivalent shortfalls for IRS. CCPs clearing CDS may wish to make an adjustment to default fund contributions to ensure that this is taken into account.
Will dealers do this? Of course not.
While there is much more in the full BIS paper extract (found here), we were less than impressed with the methodology used to construct hypothetical CDS and IRS portfolios. In a nutshell, the BIS assumed a hedged book and matched-maturity positions: something, which every OTC trader knows, absolutely never happens, as the whole purpose of derivatives is to take low margin risk positions that coincide with the herd, and thus, not hedge (otherwise what is the point?). As such, we believe, that the full potential shortfall on the up to $600 trillion in gross notional is the full net exposure in the market at any time. Which we are convinced is well over the $160 billion 0.002% case (according to some estimates, between CDS (this one is easy - just look at weekly DTCC data) and IRS (this one is far more complicated), the net notional at risk at any given moment is anywhere between $2 and 8 trillion. And this is capital that the G-14 supposedly have handy for a rainy day?
And next, moving away from dry academia, we shift to one of our favorite authors, BNY's Nicholas Colas, who coinicdentally, discussed precisely this issue in his Friday edition of his Mornina Markets Briefing:
Bad Nightmares and Good Collateral
Summary: The rulemaking around Dodd-Frank is far from over, but one area of new regulation drawing a lot of attention is what to do about over-the-counter derivates trading. It is a huge market – some $600 trillion at the end of last year – and dominated by interest rate contracts, where the notional value is $465 trillion. Just a little perspective – the entire value of the S&P 500 is $12 trillion. If even a portion of this trading moves to a quasi-exchange structure, it will require significantly more collateral than is currently used to support this market. That is strongly bullish for sovereign debt, should these changes come to pass. This dynamic got us wondering what “good collateral” really means anymore. U.S. dollars and Treasury securities are the bedrock of trading collateral, but those assets might not work as well for this function in the next financial crisis as they have in the past. The reason is that sovereign debt is increasingly losing its “risk-free asset” status as developed countries – not just the U.S., mind you – issue more debt to stimulate their economies and avoid taking the pain for previous mistakes.
My longest lasting repeat nightmare, which this year celebrates its third decade festering in my psyche, is that I have failed a class in business school and therefore don’t actually have my degree. For the first 10 years after graduating I kept my diploma under my bed, so vivid was that particular dream. The actual class that causes this lingering worry was ‘The Pricing of Illiquid Securities,” focusing primarily on exotic mortgage backed bonds. It was part fixed income analysis, part options math, and wholly difficult to understand. I almost failed it. Almost. Thankfully it was pass/fail, and the transcript clearly has a “P.”
But pricing illiquid assets has become a popular form of reality TV, from PBS’ Antiques Roadshow to Pawn Stars to Auction Kings . The formula is largely the same – walk in with something obscure, and an expert will tell you what it’s worth and/or give you cold, hard cash for the item. The analysis is a combination of authentication, historical sleuthing and market analysis of likely buyers and the price they will pay. The head of a rare doll might fetch $10,000. An entire motorcycle, even if owned by a minor celebrity, might only be $5,000. Every item is different and has to be appraised on its own history and merits.
There has been a recent flurry of activity in one capital markets dedicated to oddball illiquid securities – the pricing and trading of over-the-counter derivatives such as interest rate contracts. The reason for the attention is the Dodd-Frank Wall Street Reform and Consumer Protection Act, which gave the Securities & Exchange Commission and the Commodities Futures Trading Commission the power to restructure the ways in which OTC derivatives trade and settle. The new rules aren’t out yet, and likely won’t be available until July, according to various press accounts. That said, there are a few “Hard points” to consider:
- The global market for OTC derivatives is huge. According to the Bank for International Settlements, the notional amount of total contracts was $601 trillion at the end of 2010.
- It is primarily an interest rate market. Of the $601 trillion, just over 75% ($465 trillion) is tied up in interest rate contracts, most of which are swaps.
- This interest rate swaps market is still growing. The poster child of troublesome OTC derivatives, Credit Default Swaps, is down to just $30 trillion in notional value from $42 trillion in December 2008. At the same time, the market for interest rate contracts continues to grow, up to the previously mentioned $465 trillion from $433 trillion in December 2008.
The reason all is this is significant is simple: a change in how these instruments trade, from strictly OTC to something that more closely resembles an “exchange” could involve market participants posting consistent and predetermined collateral in order to clear trades. That’s not the way it is done now. The major banks that drive this market have freedom to determine what collateral is needed both to initiate a trading relationship and to keep it going. See here for more details: http://www.risk.net/risk-magazine/news/2074073/margin-proposals-lock-usd2-trillion-assets.
All this brings up a whole host of interesting issues, such as how much collateral the major players in OTC derivatives will have to pony up in order to keep trading. It is impossible to come up with a number at this point, given that the rules are still being written and the market structures to support a new trading regime are not yet assembled. July 2011 was the initial target date for proposed rules, but it appears to be slipping. See http://www.businessweek.com/news/2011-06-01/wallstreet-asks-swaps-regulators-to-re-propose-new-rules.html.
Underneath this bubbling surface of regulatory confusion, however, there is an equally interesting existential topic: what is “good” collateral, anyway? In some ways, it is probably easier for a pawn shop to determine the appropriate price for a used electric guitar than it is for an exchange to decide what assets a market participant needs to post in order to transact buy/sell orders. In the spirit of a thumbnail case study, we went to the CME Group website and pulled what kinds of assets they consider “Good Collateral” and the haircuts they give certain types of assets. Before we review that data, however, it makes sense to consider what makes some collateral better than others. A quick list:
- Good collateral should be something that everyone agrees is valuable. Basically, it is anything that you would rush to pick up off the street before someone else got to it. Gold, developed country currency, and fixed income instruments are all good collateral.
- It shouldn’t vary too much in price, regardless of market conditions. Ideally it would appreciate slightly in value when financial troubles strike, since that is most likely when counterparties fail and you need the collateral to ensure a trade can clear.
- It should be very liquid. Again, markets only really worry about the value of collateral when things are going south. That’s not the time to find out that South Florida condo real estate isn’t anyone’s idea of solid collateral.
The attached chart shows some assets that the CME considers “Good Collateral” and the haircuts it gives to those assets. The bigger the haircut, the more of the asset you have to post to support your positions. You can find them here: http://www.cmegroup.com/clearing/financial-and-collateral-management/. A few observations:
- U.S. Government and agency paper is “King of the Hill.” Only two assets get no haircut: U.S. dollar cash and U.S. Treasury bills. Agency debt is a 3% haircut, and longer dated Treasuries are 3.5-5.0%.
- Then comes developed country currencies and debt, at 5-9% haircuts.
- Gold and the Mexican peso aren’t often put in the same risk category, but they are here. Both receive a 15% haircut, which means that in the eyes of the CME they have equivalent appeal as collateral.
- At the far end of the equation are the Turkish lira (20%) and equities (30%).
Two things pop out to me from this quick analysis:
- If there ever is an exchange-like trading mechanism set up for OTC derivatives, there is going to be a real run on U.S. government paper. The CME’s list of assets and haircuts tells the story – Treasuries are the most efficient way to fund collateral. The notional amount of interest rate swaps alone – some $465 trillion – is enough to swamp the $14.3 trillion of total government debt outstanding, let alone the $9.7 trillion that is actually available for purchase.
- The whole notion of good collateral is very much anchored in the thought that U.S. sovereign debt is “risk free.” Whether or not that is true in the absolute sense is irrelevant. Remember that collateral needs to be at least crisis-resistant and preferably negatively correlated to asset prices during financial stress. With the U.S. government currently at loggerheads over how to deal with the Federal Debt Ceiling and the most likely path is to simply issue a lot more government paper, the time could be coming where Treasuries no longer fulfill the purpose of “Good Collateral” during crisis. They are just as likely to be the cause of a financial storm rather than provide shelter from the rain.
Bottom line: instead of wondering how to nudge the silver market (lower) or the ES and TEN contracts (higher), perhaps it is time for the key exchanges, which are obviously captured by the very same G14s on whose tithes their existence depends, to actually proactively engage in some risk-mitigation when it comes to the one biggest threat: not that of the measly several billion dollar silver market, but of the $600 trillion IRS and CDS market, which is and continues to be the biggest ticking timebomb in capital markets. And on the other hand, if anyone is wondering what will cause the biggest run on US government bonds... ever... then as soon as every dealer is forced to be on a CPP, all one needs to do is a massive "risk-flaring" collapse which sends everyone scrambling to provide collateral. And since there is a 40-to-1 ratio of notional outstanding in OTC derivatives to total US debt, well, readers can do the math.