I am sticking my neck out a bit on this post, since the credit default swaps market doesn’t garner much coverage, so any readers who are involved in this busines are encouraged to comment.
Yes, there are frequent references to what changes in CDS prices mean about the credit-worthiness of particularly names, but there is parlous little attention paid to the health and activity of the market as a whole.
Despite the use of the term “swap,” CDS are really insurance contracts. A protection seller (effectively, the insurer) agrees to make a payment to the protection buyer if specified bad things happen (a “credit event” usually defined as bankruptcy or failure to pay) to a “reference entity” which can be a company (“single name”) or an index. See here for more detail.
Now while it may look like the risk being traded here is default risk, there is a second risk: counterparty risk. CDS are the largest credit derivative product, and they are traded solely over the counter. That means that the CDS agreement is only as good as the protection seller that wrote it.
When Warren Buffet described derivatives as “financial weapons of mass destruction” he wasn’t worried about speculators blowing themselves up, but about counterparty risk:
Unless derivatives contracts are collateralized or guaranteed, their ultimate value also depends on the creditworthiness of the counterparties to them. In the meantime, though, before a contract is settled, the counterparties record profits and losses -often huge in amount- in their current earnings statements without so much as a penny changing hands. The range of derivatives contracts is limited only by the imagination of man (or sometimes, so it seems, madmen).
So why should we be concerned about CDS? Even though they have been around since the mid 1990s, they have grown explosively since 2002, in a (until a few months ago) benign credit environment:
Notice that the notional amount outstanding is $45 trillion. While the economic exposures of derivatives are a fraction of the notional amount, this is still large enough to focus the mind.
Who is standing behind these contracts? As Ted Seides tells us:
Banks are the primary sellers of CDS, totaling 40% of all written CDS and representing notional exposure of $18.2 trillion.[xv] Banks claim to run hedged books, effectively serving as a market-maker in the CDS market. As should be evident from the events in subprime, even the most sophisticated systems are often unable to fully hedge risks of this size and degree of complexity. If printed materials are any indication, banks may be asleep at the switch. The “Counterparty Considerations” section in the Credit Derivatives Primer of market share leader JP Morgan is a single paragraph on the last page of the volume, which proclaims “the likelihood of suffering (counterparty default) is remote.”[xvi] (italics added)
Hedge funds appear to be in over their heads as well. According to printed statistics and consistent with anecdotal evidence, hedge funds are sellers of 32% of all CDS, insuring exposure of $14.5 trillion.[xvii] Recent estimates indicate that the entire hedge fund market is approximately $2.5 trillion in net assets under management. Thus, hedge funds are bearing risk in excess of their ability to pay the piper if anything goes wrong.
Although, as noted above, there is comparatively little reporting on the CDS market, we are seeing some signs of stress.
The last two years have seen the growth of CDS based on asset backed securities and ABS indices, such as the ABX and the CMBX, based on the CMB, an index representing commercial real estate securities. CDS were also one means to achieve credit enhancement in collateralized debt obligations. Given the sharp decline in the ABX indices and downgrades of CDOs, it’s a no-brainer that anyone who wrote protection on them has taken large losses. Case in point: Swiss Re recently announced that it had taken over $1 billion in losses on two CDS referencing mortgage-backed investments.
Consider another troubling development: banks are already worried about counterparty risk in the money markets, as witnessed by the uncharacteristically large spread between T-bills and Libor. It seems inconceivable that banks wouldn’t have similar worries about CDS exposures. We have indirect confirmation via this story in the Financial Times, “Trading in derivatives slows to a trickle“:
Liquidity in some of the world’s biggest derivatives markets has dried up this week amid increasing fears over the health of the international financial system.
Over-the-counter trading in derivatives of equities, credit and interest rates have all seen much lower volumes as problems in financial markets have prompted investors to sit on the sidelines…
Analysts said flows had slowed to a trickle this week – even lower than in the summer when the credit squeeze was at its peak – as investor appetite for risk had diminished amid talk of potential bank defaults…
David Brickman, head of European credit strategy at Lehman Brothers, said: “Generically, trading volumes [in credit derivatives] are a lot lower than they were in the summer.
“The theory is that if people can’t trade bonds, they’re going to go to CDS [credit default swaps]. But in an environment like this you can’t get liquidity on single-name CDS either. That just leaves the indices.”
And a more colorful confirmation comes from a reader:
To remind you we have been short the ABX CMBX and CDX since last January. To keep this short I will summarize my points.
1. Our counterparties are Bank of America and Barclays. We only have ~20 million in equity and were allowed to go short the market 150M notional by posting 3% margin. Presumably the numbers are similar for someone who wanted to go long (should be a little higher).
2. After July hit we were no longer allowed to take ‘new’ short positions. We were originally told that we are allowed to close out our shorts at any time but not allowed to open a new one. So we couldn’t roll into the on-the-run series or move our shorts up the capital structure as the environment continued to deteriorate.
3. My original thought was that they wanted to artificially prop up the market because the market makers may have too much of a net long exposure – and they are extrememely concerned – or they have someone on the proprietary desk that got themselves into a pickle. So I thought this would pass in a few weeks and we could move our position around.
4. Today I still cannot open a new position. In the past few months a few friends have been let go from some of the larger banks and have shed light on the topic. The reality is that they are concerned about collecting from hedge funds that were long. Although this is just hearsay from friends – it completely matches my own experience. To further the point – working with the collateral team at BOA in pay as you go contracts – they are a mess. They are overdrafting accounts – forgetting to pull money when its due – they can’t reconcile a mistake in less than 8 weeks. It’s a debacle.
Conclusion:
The credit teams for these banks may have been allowing hedge funds to trade at excessive leverage without enough controls. They most likely were basing margin requirements on the volatility of the security (or contract). I am a firm believer that this type of analysis blinds analyst to real risk and gives them unjustified confidence in how ‘secure’ they are – but that’s another story. I am not going to pretend I can estimate these losses or give accurate numbers to even shed light – but this is definitely and ‘unknown unknown’ with a high potential impact. As Nassim Taleb would say – this has the makings of a Black Swan.
What seems odd, given all the foregoing, is that Swiss Re, admittedly a new player but one with a small book, is the only concern to have reported CDS losses. Someone has to be on the other side of the eyepopping trades entered into by Paulson & Co. and for that matter, Goldman, which has been short mortgage-related credits. That raises the question of how much latitude financial firms have in marking their derivative books (and auditors have no hope of getting to the bottom of their economics).
The other open question is what happens to the CDS market as banks continue to shrink their balance sheets? CDS have become integral to the way a lot of players measure and manage credit risk, but if the market becomes illiquid, there will be a lack of reliable price information and a dearth of protection sellers to write new contracts.
If you’d like to worry even more about CDS, this post from “CDS: Phantom Menace” from Sudden Debt is insightful.
26 comments
Comments are closed.
I have always assumed that the big banks marked-to-market their positions and traded collateral based on who owed whom what. Perhaps this is not the case with CDSs? But if it’s the case, the risk is only the end-of-the-financial world risk: one day there is such a big move in the mark-to-market that one bank in the chain can’t pay and goes belly up. Needless to say, bankers don’t care about end-of-the-financial-world risk, and a lot of the money they make is based on the fact that they can make trades which effectively treats this risk as 0.
Yes, the CDS markets might be in problems – and if it does get there, we would remember subprime as the good old bad days.
The counterparty problems is even more severe, becasue (IIRC, from data about a year back), the exposure is mostly concentrated with I think three big investment banks, who account for most of the volume (who knows where they part it though)
There’s a few points:
– for all the talk to hedging, someone, ultimately, has to have a naked position. These may be “hedged” by a correlation trade – which might or might not work, as correlation is not really market observable and default is a jump event.
– should one of the major cptys go down, a lot of parties who thought they were hedged will all of sudden have naked positions.
– chances are, that if one of the big boys goes down, it’s due to massive credit events – and possibly vice versa. The old “in crisis, correlation of everything goes to 1”. So, you’d find yourself not only with naked position, but quite possibly with a naked position that just dropped a lot (or even someone making a contingent claim). I can imagine a very nice domino effect there.
a – collatelar agreement may or may not be in place.
Also, the problem is that because default is a jump event (that is, it’s not a continuous slide of value towards a recovery value), if the cpty and the underlying are correlated (and again, crisis => rho =1 )the cpty just won’t be able to post the collateral, just at the time you’d need it most.
So far, two CDS-related blow ups are known: Merril increased its CDO exposure by $600M last month after a counterparty default (the counterparty wasn’t named), and the Carina CDO that went into liquidation was a synthetic that wrote about $1.03B in protection on subprime MBS to Deutsche Bank. The greatest risk is not a big house blowing up but rather a hedge fund or synthetic CDO defaulting, and those events while disturbing can probably be absorbed by the major houses though not by smaller players.
Counterparty default reverts risk back to the protection buyer.
Credit default swaps don’t create credit risk in aggregate. They don’t increase the amount of net credit risk in the system – they only redistribute the outcome. They either transfer risk as intended or fail to transfer risk as intended. There is no net systemic risk change as a result.
In this sense, the systemic problem is typically overstated.
Very good post Yves.
The hedge fund participation is the fatal defect in the CDS system.
The part that was worthy of particular attention was this:
“According to printed statistics and consistent with anecdotal evidence, hedge funds are sellers of 32% of all CDS, insuring exposure of $14.5 trillion.[xvii] Recent estimates indicate that the entire hedge fund market is approximately $2.5 trillion in net assets under management.”
So hedge funds have written $14.5 trillion in credit insurance.
The question is:
Assuming the economy went into a recession as severe as let’s say 1990-1991 (and that is being terribly kind), what kind of loss rate (%) would be experienced on the credit insured by CDS?
What about a loss rate from the 1982 or 1974 recession?
Therefore, how much of that $14.5 trillion would the hedge funds be called upon to pay in claims?
“They don’t increase the amount of net credit risk in the system – they only redistribute the outcome. They either transfer risk as intended or fail to transfer risk as intended. There is no net systemic risk change as a result.”
How is this argument different from the original explanation that the subprime crisis was overstated as the CDOs distributed but did not increase risk that has proven false? Isn’t the issue that risk transfer ability increases risk taking and that indeed increases the overall systemic risk when the “insurance” fails? Or is CDS simply a wealth creation method without any economic substance in which case the argument makes sense as all the illusory positions would net out as there never was any actual risk transfer (albeit no doubt the positions were leveraged and probably have some interesting LTCM-like tax consequences)?
I would say, dd, that the argument/explanation is not so much false as it is irrelevant. While the distribution of risk within the CDS structure is a zero sum, it is subject to risks loaded onto the structure from events outside its universe.
As to risk transfer increasing risk-taking, that is commonly reflected in the ratings of various offerings as provided by Fitch and the other ratings agencies. The perception that an offering is ‘insured’ does in fact affect investors’ willingness to assume risk, and its absence leaves them exposed to more risk than they bargained for. But your remarks suggest you know this already.
check out this guys article about CDS:
http://suddendebt.blogspot.com/
thx
Significant counterparty risk is there for unsecured exposure. If a bank receives collateral (that is marked frequently, or overcollateralized), then the bank doesn’t have to look to the counterparty to make it whole if it defaults. The collateral covers losses if the counterparty ever goes under. This is the only way banks can get comfortable trading with lowly rated or unrated entities, like hedge funds.
The counterparty risk groups are relatively sohpisticated, so they do their own leg-work on credit (and not just rely on ratings). Even if they are wrong both on credit risk and market risk (MTM), they close out positions when counterparties fail to post, so the exposure period is only a day or two for the grace period and a few more days after that to liquidate collateral.
dd- (re my earlier comment). There are two separate issues. First, I agree that structured finance increased the amount of credit created at the origin (e.g. subprime mortgages), due to the expanded financing mechanism. But second, given the amount of credit that was created in that process (e.g. subprime mortgages), the leverage inherent in structured finance doesn’t additionally increase the risk inherent in the underlying assets (e.g. subprime mortgages). In this example, structured finance increased the amount of subprime mortgages created, but it doesn’t amplify the risk inherent in those subprime mortgages – it only distributes the resulting losses. This is relevant because the media tends to spin and exaggerate the amplification of the aggregate risk result due to underlying asset deterioration, due to the transmission of actual losses through leveraged financing structures, which is not a correct analysis.
>>Credit default swaps don’t create credit risk in aggregate. They don’t increase the amount of net credit risk in the system – they only redistribute the outcome. They either transfer risk as intended or fail to transfer risk as intended. There is no net systemic risk change as a result.
The above would be true if CDS were only used to offset the risk by a grantor of credit on the specific credit granted. Since most CDS written nowadays are on indices, the CDS risk can be many times the amount of the underlying credit. The easiest way to think of it is to look at the futures market. The reference for Nymex traded oil is the price in Cushing, OK, but the actual oil traded in that location is a small fraction of the value of the Nymex contracts written that use it as a reference.
Structured finance amplified the inherent risk by re-engineering a liability (ie a loan with no or little chance of repayment) into an “asset.”
If there is a run on CDS policies, could this lead to the implosion of the Hedge Funds?
If so, that worse-financial-crisis-since-1929 prediction takes on a compelling urgency.
” The above would be true if CDS were only used to offset the risk by a grantor of credit on the specific credit granted. Since most CDS written nowadays are on indices, the CDS risk can be many times the amount of the underlying credit.”
This proves my point further. The net result of the payoff between two counterparties of an index CDS is 0 – one ‘loser’; one ‘winner’. There is no net credit risk created in the system.
The fact that the notional amount of such contracts exceeds the aggregate underlying credit in the system is irrelevant. The net transfer of gains and losses is 0.
Also irrelevant to the same point is whether its an index CDS or a single name CDS.
The more relevant analogy is that the existence of stock options in aggregate does nothing to affect the underlying stock market capitalization in aggregate (other than through peripheral interaction such as delta hedging, which is besides the point). Credit default swaps are essentially put options. They don’t affect the credit risk that is generated by underlying reference assets.
Anon 12:52 PM
I really can’t make sense of what you’re saying. If a brand new hedge fund is created and sells protection through CDS and it turns out that bad realizations on those CDS contracts can bankrupt the hedge fund (i.e. the margin imposed by prime brokers was insufficient), we now have a new entity, the hedge fund that is subject to bankruptcy and therefore credit risk. This is credit risk that did not exist before, right?
You have a student loan for $100,000, and you’re flat broke. I make $200,000 worth of bets with third parties that you will pay off your loan. Then I give you $100,000. Profit.
The bonus is, I’m a white knight rescuer and the counterparties look like greedy speculators seeking to profit from the misfortunes of the downtrodden, etc. etc.
If the notional amounts of CDSs are starting to exceed the total amount of debt, as the linked article states, then what’s to stop, say, a sovereign wealth fund from playing this game?
If we’re not at that point yet, we surely will be soon. The graph in the linked article shows CDSs amounts growing exponentially, and surely total credit market debt is not increasing as fast.
Anon 1:10 p.m.
A simplified example: The counterparty has bought protection. Suppose the reference asset is a subprime mortgage (or some structured finance variation of it). The underlying credit risk is the failure of the subprime mortgage in some sense –default or otherwise. The rest is a transfer of an ‘insurance’ payoff from the protection seller to the protection buyer. The protection buyer suffers losses equal to the subprime exposure apart from the swap, less any “insurance” realization from the swap. The protection seller’s bankruptcy is the result of all such payouts. The payouts are a cost to the protection seller and a gain to the protection buyer – a wash in aggregate. But the subprime mortgage loss is a net economic loss in terms of aggregate system credit risk.
Anon 2:19 PM
But what if the seller which is bankrupted by the CDS losses has other liabilities the resolution of which will be tied up in bankruptcy court for years (i.e. in the absence of CDS contracts the seller could have made good on its other liabilities). Then the CDS contracts by allowing a new entity to risk insolvency in the basis of a single credit risk has increased the number of assets that are subject to be tied up in bankruptcy proceedings.
anon 2:47
Fair point. Still the ultimate proceeds collected by the protection buyer offset the payout of the protection seller. The rest of the seller’s problem and costs are an additional risk of being in the business.
Can someone explain to me how the notional amount of CDS outstanding can be orders of magnitude larger than the underlying debt? What is the source of a leverage in the system?
Nov 30 (Reuters) – Yield spreads on indexes of commercial mortgage bonds narrowed on Friday on speculation that a federally supported plan to curb home foreclosures will improve conditions in all U.S. credit markets.
The yield spread on the CMBX-NA 2 index of “A” rated bonds narrowed by 35 basis points from Thursday to a bid side of 255 basis points and an offered side of 270 basis points, one investor said.
The U.S. Treasury is expected to next week announce a plan with top mortgage lenders and servicers that will freeze payments for many subprime borrowers facing higher rates that would otherwise send them into foreclosure. Soaring defaults on subprime bonds sparked a credit crunch that since mid-year has spread to commercial and corporate borrowing.
Anon of 6:45 PM,
Have a look at the Sudden Debt link at the very end of the post. The author provides some detail on the margining practices, such that they are. Also look to the back-and-forth between the author and a CDS trader in the comments section, which provides further information.
“The more relevant analogy is that the existence of stock options in aggregate does nothing to affect the underlying stock market capitalization in aggregate (other than through peripheral interaction such as delta hedging, which is besides the point). Credit default swaps are essentially put options. They don’t affect the credit risk that is generated by underlying reference assets.”
Anon is the stock options market really analogous? My understanding (happy to be corrected) is that indeed CDS do impact subprime underlying assets via their packaging into CDOs (that contain CDS as an insurance mechanism that boosted the rating). So movements in the CDS and ABX indexes can impact overall CDO (and SIV) valuations. Hence the money market and mutual fund withdrawal from the markets (and that too then impacts the CDO valuations) but it all begins with the structured product’s valuation being tied to the quality of “insurance” backing up the CDO valuations. In that context the CDS market impacts the CDOs which in essence are indexed bonds (with the CDO proxy for index as it is composed of slices of many debt instruments but oddly the subprime ends up superweighted at present). So like bonds, CDOs are greatly impacted by CDS as these stand as proxy to the derivatives used to hedge publicly traded bonds and the result was the Delphi & GM freeze up. CDS and ABX indexes may play in their own way play a greater role now that liquidity is gone from the CDO and ABCP markets as now that is all that is trading.
So in a nutshell the CDS market does impact the pricing of the CDO market and and the suprime component. The discovery of the subprime (and perhaps worthless insurance) triggered a self-feeding loop where the prices in one impact the asset valuations in the other.
Then this analysis could be totally incorrect as this those markets are so opaque and this is what I have learned from public filings and news reports.
And thank you both anons for the interesting discussion.
Then too, I would be remiss in not thanking our host for such a wonderful blog.
Thank you, Mr. Smith
Protection Seller are the one taking the payments and protection buyer are the one paying