I hate it when I get contradictory information from supposedly reliable sources, which happens upon occasion when the topic is CDS.
The latest Institutional Risk Analytics newsletter is again after one of its favorite objects of ire, credit default swaps. One of its beefs is that the collateral posting rules are a joke. Key excerpts:
The under-collateralized wagers that are CDS contracts threaten the solvency of financial institutions around the globe, this despite the efforts to reform the system via enhanced clearing mechanisms, increased collateral and margin requirements……
But alas, one of Wall Street’s dirty secrets is that most of the CDS dealer banks don’t post margin with one another at all! If CDS dealer banks were actually compelled to post real, effective collateral with other dealers to back performance, then the entire CDS market would collapse. Indeed, that is what is happening right now, in slow motion. As leverage in the global banking system is being forced down, the CDS market is being squeezed out of existence by a market that can no longer ignore the inherent contradictions in these OTC options…
the op-risk issues with CDS are as nothing compared to the pricing and funding problems with these contracts. In that sense, the focus on back office problems facing CDS and indeed all OTC derivatives has been a canard, a distraction from the real issue, namely the bankrupt intellectual basis for the CDS contracts themselves….
What Geithner and Fed Chairman Ben Bernanke failed to tell the Congress, President-elect Barack Obama and the American people is that CDS dealers don’t post any effective collateral at all with other dealers. So much for the legal requirements for safety and soundness in 12 CFR. If Barack Obama and the Congress ever needed a final reason to strip the Fed of all regulatory responsibility for financial institutions, the coming nuclear winter of CDS unwind is it…..
You see, in the make believe world of interdealer CDS, when a “margin call” occurs, no cash or securities actually change hands. Instead the CDS dealers merely shuffle some paper around and effectively rely on the overall credit standing of the other banks, much as they do in foreign exchange or money market transactions. And virtually none of these dealers even attempt to model the actual default risk in CDS!.
Mind you, this is from a guy who sells a very high end research product to top tier institutional investors and knows the regulatory scene well.
However, this left us a little perplexed, since we had understood that the reason the Lehman CDS settlement went well was that the counterparties had to post margin when Lehman BK’d, so the stress event was before-hand (hence all the ugly market action in October due to the need to sell something in order to post margin). And we recall seeing a great chart that showed CDS collateral calls to hedge funds, and they shot up in October.
So this undermines IRA, no? Not exactly. It appears that hedge funds are kept on a short leash, at least when on the protection writing side, and are subject to pretty tough collateral posting rules:
…..margining for hedge funds tends to be somewhat more stringent. They typically post collateral at 100% of their current exposure, and furthermore might also be asked to post collateral to cover close-out risk on their contracts for a certain number of days going forward. The estimation of forward exposure is done through forecasting future scenarios.
But this factoid does not prove or disprove the IRA contention, it merely says that hedge funds who write CDS do indeed pony up quickly. Moreover, the ISDA does everything it can to burnish the image of the CDS market, so like the National Association of Realtors’ comments, you need to read their comments about the market with a fistful of salt.
And to give you an idea of the general utility of information about CDS, I found this old post: “Barclays: Counterparty Risk in Credit Default Swaps Only $36 to $47 Billion.” AIG disproved that estimate.
it seems natural to me that broker/dealers do not post collateral with each other: all of them have government backstops and bankruptcy is no more an option, hence there is no credit risk besdes the sovereign backing the bank.
hedgies on the other side cannot borrow from the fed and do not have implicit guarantee on their obligations, so they should be subjected to stricter collateral rules.
this double standard system is just the reflection of a double glass mirror.
IRS, like many commentators, has a prejudice against the CDS market I fear, and it isn’t going to let the facts get in the way of a good panic. The vast majority of inter-dealer CDS are done under ISDA masters that cover all derivatives dealing between the counterparties, so net margin is posted on the entire portfolio of equity, credit, interest rate and commodity derivatives between the counterparties. Typically this is done daily with low thresholds (so that if the net exposure is greater than, say, $20M, collateral is posted). Dealers also control their potential future credit exposure (how much money they might be owed if there is a move in the market) using limits based on simulations.
The only issue that does arise here is that CDS can (do not always but _can_) jump in value on default. The margin process works well if there is a slow slide to default with the credit spread widening. It works less well, obviously, if you get a point default from a tightish spread, as in Parmalat. Given where spreads are right now, there aren’t many credits this applies to at the moment…
Well therein lies the problem …
No one knows for sure. 60 trillion? 70 trillion? worth of CDS and the rules are not clear, let alone the exposure of institutions.
If there is one thing the market hates it is uncertainty and we have that in spades which is exactly why we need to nationalize the banks and either net these derivatives out or declare them void via a National Security Finding.
An analysis will have to be made as to the viability of a CDS market whereby they are traded as to ascertain market value. That analysis would certainly include standardization of contracts and qualification of parties eligible and under what circumstances.
they don’t have to post collateral or on margin, they are net zero anyway?
i find it surpising they “don’t” have to post collateral, and would be more willing to accept that they are “not asked” to post collateral, witness “NYmag’s” article on Lehmans. At some point, Jamie Dimon asks them to “post collateral” to open the doors another day, presumably some assessment of CDS risk is covered under that blanket margin call?
what david murphy says, plus some anecdotal evidence:
i worked at one of the big investment banks the past six years. about three years ago all our traders (yes, across the entire bank) were told for a period of some weeks that we could not do any more trades with jp morgan because we had hit our risk limits on derivatives. note that this limit covers the “gap” exposure (one-two day movement in prices that is not covered by collateral). anyway, the point being that the bank took this risk sufficiently seriously (at the height of the boom) to feel it necessary to cause serious disruption to its trading activities.
i don’t think there has ever been any kind of “gentleman’s agreement” between brokers when it comes to counterparty exposure. senior management have long enough memories, and know a lot more than the public does (e.g. about lehman having been insolvent for a few months after the august 98 russia default, but it lied on its quarterly financials).
what i can believe though is there could be serious losses because of operational screwup. for example, trades get closed over the phone on a recorded line or by email. legally speaking, this verbal contract is sufficient. but then they are supposed to get confirmed in writing (including all the details of the trade that didn’t get discussed on the phone). most banks have an enormous backlog of unexecuted swap confirmations. when it comes to more esoteric hedging (e.g. total return swaps sold to hedge funds), the risk of a counterparty denying a trade’s existence is very real.
as for margining under isda, all i can say is that each trade gets recorded on the bank’s i.t. systems, and then gets picked up by a specialised margin management group. at that point it ceases to be a concern for front office traders. whether the margin actually gets managed correctly i have no clue, although i suspect any bank that was lax about this 12 months ago, is much more sharp about it now!
David Murphy has described it well. Pretty much everyone has collateral agreements (or CSAs, Credit Support Annex to the ISDA agreement), and they definitely tend to cover CDSes (although they can be not included, as it’s an enumerative document, and you need to state explicitly what is covered).
The few cpties that don’t tend to post collateral at all, are AAA rated parties, but there’s generally a clause where the party has to start posting (in one huge jump) should they be downgraded.
The key thing is the difference between gross and net exposure.
Net seems to be 1/10th of gross.
It also seems like the netting is done at the broker/dealer level. I imagine the brokers themselves can be brutal regarding collateral with their clients.
Most of this seems to net out to be less of a big deal.
BUT — I fail to see how it helps the real economy in a material way.
Plus, the numbers are big enough that very minor accounting differences could net to enough billions to be a problem. Note that there should be mirror accounting and it isn’t trivial to get the same price on each leg in the accounting statements.
Whatever the 90% of gross exposure that gets netted out does — someone needs to explain how the economic benefits outweigh additional systemic risk. Can’t create a synthetic bond? Buy something real.
Slightly different but importnat point on this: CSA’s have no set way of valuing a contract (and I can’t remember whether there’s an arbitrage clause in the ISDA), so it’s not uncommon that you get an argument about the value of the collateral posted. More so these days, with the market being dislocated.
Even vanilla IR swaps can have significant different valuations if using different yield curves. So, it’s not like a plain margin call where the amount is indisputable.
vlade: the downgrade trigger is a very good point, and another source of potential catastrophe. the trigger also arises on rated synthetic cdos (and no doubt many analogous transactions) where the rating agencies would demand a step up in cash margining from the sponsor bank to the spv in the case that the sponsor bank gets downgraded (i think the key rating trigger is usually d/g below single-a). it is one of the many ways in which the technical impact of a credit rating downgrade can make it become a self-fulfilling prophesies.
re valuations – the vast majority of interdealer transactions are on relatively “vanilla” transactions with “standard” (market consensus) terms where there is wide market consensus on valuation and an easily available underlying market reference price. the more esoteric / non-standard transactions are typically done between dealers and end-buyers of risk (hedge funds / insurance cos / dumbass banks / etc), in which case the dealer bank will make sure it has total control over valuation (hedgies / ins cos) or the transaction is fully funded (clns sold to dumbass banks / etc).
@Bena re valuations:
That worked sort of ok when the times were good. Now I’m seeing quite a few arguments over valuations on even vanilla IR swaps (that’s why I mentioned it). To simplify, the value of IR swap depends on how you do your curves. Last year has blown some people’s assumption from the water (say assuming minuscule basis on different tenors), and so can come with valuation hugely different from other people – even though both of them are using their tried-and-tested methodology. And I ommit trvialities like how you interpolate on the curves, what you include in the short end of the curve, how and whether you overlap etc. etc. (these tend to drive much smaller differences than the basis blowout).
the concerns i have for the CDS debt construct are –
~lack of transparency, as in finding out the counterparies on CDS for US debt has not been possible (in my research).
~ebbing to no mark to market, as aptly stated above, when credit ratings downgrade, for example.
~the vast quantity of these debt items.
it doesn’t seem to difficult to imagine a scenario for implosion. perhaps the sovereign bond revolt in Ecuador spreads, or…
I think the confusion here is due to the failure to distinguish between initial margin and mark-to-market margin.
What happens when a hedge fund sells a CDS is that they have to post initial margin with the dealer, perhaps 15%, to account for the hedge fund’s theoretical lack of relative creditworthiness. When a dealer sells a CDS to a hedge fund, they do not post the equivalent initial margin.
(When you buy a CDS you do not have to post initial margin, generally, since you only owe premium and if you don’t pay then you don’t get the benefit of the payout if a credit event happens).
Once the trade is on, both parties post margin to each other to account for the change in value of the CDS on a daily basis; this is the mark-to-market margin.
So it is perhaps true to say that dealers don’t post margin on the initial risk transaction, but do on the variation movements.
In the Lehman case, there was a very big change in the value of the CDS after the bankruptcy. This amount was margined because it was a change in the mark-to-market. Normally you don’t see such big fluctuations in the value of a CDS so in a plain vanilla case where there is less risk of bankruptcy there is less margin but there is also less exposure.
The thing that is stunning to us outsiders, and that the insiders seem to take for granted, is that the collateral isn’t posted until after something bad happens to change the valuation formula. Which, of course, is the same time the counterparty is least likely to be able to post collateral due to the bad things that are happening.
On the positive side, you have given me an idea for a similar arrangement next time I refinance my mortgage – I will simply promise to execute a deed of trust after I default. That way I can use the house as collateral for other things in the mean time. This should be completely legal and acceptable as long as my models show I am not likely to default.
If banks don’t require margining of CDS contracts that’s news to me. I feel like an idiot for posting collateral every morning when they call for it. For every contract there is an initial haircut, used to be as low as 1% of the notional but is much higher now. Plus 100% of the current exposure of course. These indeed are updated every day by every counterparty. People used to be lackadaisical about margin but I guarantee every credit officer who wants to keep his job is watching their margin requirements like a hawk. Call it early and often is the new manta.
I have a question that may not necessarily fit here, but how do people account for the recent increase in the US Treasury 10-year CDS price? It’s at 68.4 basis points as of yesterday, up from 2 in July of 2007.
Now the easy answer is that the likelihood of default has increased substantially through the bailouts and anticipated future stimulus. But a friend is arguing that the supply is way down as foreign banks are going under. But if the risk is limited, what would prevent other banks from picking up that slack? My position is the price increase is based almost entirely upon an increase in demand and increase perceived risk. His position is that the supply is way down. One right, one wrong? Both right? What do people think?
But a friend is arguing that the supply is way down as foreign banks are going under.
A lot of shops selling vol and insurance have been under, shall we say, mild stress recently. The governments have decided to backstop to every financial firm in the world while running 10% of GDP deficits. Chicken, egg, who knows, but I tend to blame the chicken.
The LEH CDS settle is a false flag operation by the Fed to send a signal to the market that “nothing to see here.” It has worked, it would appear, but for the blogs. It is virtually the equivilant to Wells Fargo raising its dividend into the short rule as it modified the way it records charge offs / delinquencies.
We could clear it up if we new what collateral the Fed held? Waiting for the bloomberg FIA request to come through….and waiting…
David Murphy: The vast majority of inter-dealer CDS are done under ISDA masters that cover all derivatives dealing between the counterparties, so net margin is posted on the entire portfolio of equity, credit, interest rate and commodity derivatives between the counterparties.
If you read the whole of the IRA’s post, you will find that this is precisely the situation that the IRA is worried about. There are two problems:
(i) posting collateral on the net portfolio obligation in no way guarantees that an institution has the liquidity to cover obligations on a specific default. We need to remember that CDS values “jump” by design — i.e. the actual event of default is all but guaranteed to precipitate a large change in value. While the market handled Lehman well, the market could not possibly have had more warning signs of that default — Lehman’s problems had been leading the business section of the news for months.
(ii) Given the valuation problems for even plain vanilla derivatives right now (see vlade), how is it within anyone’s imagination that the net margin posted as per ISDA protocols is meaningful? The ISDA is getting a good bite in the butt from it’s implicit reliance on the efficient markets theory.
Every single major collapse was precipitated by interdealer margin calls: Bear, Lehman, AIG. And we probably just haven’t heard about Citi yet. While the press hasn’t connected these margin calls with CDS (except in the case of AIG) and the ISDA agreement probably prevents the margin calls from being attributed to any specific type of derivative, I think the IRA is just reading the writing on the wall.
Very interesting take on CDS and bank reserves…banks used cds to expand balance sheets and thus no way to unwind the market without bankrupting the banks. The increase in reserves is exploding becasue the patient is dead..the capital holes are so enormous
The plot with collateral thickens as each bank (as previously stated) has “netting” agreements by legal entity across all products with their counterparties. Hence 5 long and 5 short = 0 w/ counterparty a. Key issues that have historically muddied the water beyond confirmation delays have been the ability to “novate” positions away from conterparty a without counterparty a’s knowledge hence 5 – 5 does not equal zero. Further muddying the water is the fact that most IB’s look at their portfolios using a very high level aggregate risk analysis that also nets longs/shorts …. the old saying ” clever by half” is really the more appropriate way of looking at many of these transactions. For Wall Street Banks these products have been a tremendous fee generator without any of the balance sheet funding implications (i.e, capital funding or regulatory capital). Forget expecting collateral to compensate for the scale and madness of these products. The real pain of these products lies in their “Heisenberg-like” qualities … on equity prices and overall firm stability. Naughty, naughty, naughty.
Mike S … NJ Earth
Thanks for the reference to the itulip piece. If correct, it greatly facilitates my understanding of the centrality of CDS to the current mess. That CDS is of central importance, I have no doubt. I agree with IRA on that. But heretofore, I had thought it was the speculative use of CDS unrelated to an underlying credit that was the real issue; that and the lack of initial margin posted by the major players. The itulip piece helps me understand why interbank lending has ground to a halt: they all know that at their core they are all rotten because of the immense amount of loans they made only because they had the (now seen as illusory) protection of CDS.
Would be interested in what Yves thinks of the itulip piece.
CDS are not the problem…nor is portfolio-margining…the problem is that the desks taking on risks have had no effective controls from their rah-rah credit officers and risk managers for the past 5+ years.