IFR: What lessons were learned from the correlation market disruption of April and May and what is different in terms of the effect from an actual default such as Delphi compared to the previous auto downgrades on the markets?
Andrew Feldstein (BlueMountain Capital): I think that they are both really interesting and important topics. I think they are two different subjects. The current issue around Delphi is very much about the mechanics of settlement and less about market pricing. The events of April and May were very much about a big disjointed price movement that caught many people unaware. April and May was more about a surprising depth to the structured credit bid, to a degree that people had not anticipated, and an imbalance between supply and demand of risk and some triggering events that led to the dislocation in the market.
What we are talking about now is that, while there are pricing dislocations that are still going on as a consequence of May, as they work themselves through, the Delphi issue is about how the industry comes together to settle probably around a hundred thousand contracts. And the thing about over the counter credit derivatives is that for every long position there is a short position. By definition for every buyer there is a seller, so if you are able to collapse everything down so it would be a fair and equitable settlement process, everybody would get the benefit of their bargain. But it could be very operationally burdensome. What the industry is struggling with now is how to ensure that everyone gets the benefit of their bargain, while avoiding extreme operational burdens.
Lisa Watkinson (Lehman): One of the lessons learned from the April/May dislocation was the ability to bring a lot of new investors into what was traditionally more of a pure correlation player market. I know personally of at least 20 to 30 brand new counterparties that had never even been involved in a CDO before, had never been in the correlation space, and had never in many cases been in investment grade. So it did bring in a new buyer base, thinking of this more as a relative value play in many cases, as opposed to a pure model-based player. I think that has been one of the more interesting things we have found are the distressed and high-yield funds and some real money insurance companies coming in.
Derek Smith (Deutsche Bank): Certainly a number of new players came in. A lot of people talk about position balances and I think that was definitely present in the market. But we should not ignore that there was a fundamental change in what was happening in the market. The autos started trading with much more volatility and much wider than the rest of the market, and there really was a de-coupling. There was a fundamental re-pricing in the market.
The reason it happened so quickly and probably the reason it initially over-reacted was because of position imbalances. But there was something fundamental going on, where people had on a trade, both dealers and to a large extent hedge funds, and it was just the wrong trade at the time. So the market re-priced itself, probably mostly appropriately and if you look at where correlation is now, it is not that fundamentally different. I think that is interesting to note.
IFR: Did any major players pull out or back off for any significant period of time?
Feldstein: We saw liquidity from all of our major counterparties straight through May. It was actually a very pleasant surprise, given the turmoil, that we saw no major counterparties retracting their liquidity. I would agree with Derek in terms of there being credit issues that either exacerbated or were exacerbated by, depending on which one you want to start with, the technical imbalance.
But I think the other lesson learned is that while the models that we all use to have a common language to talk about these instruments enable us to trade quickly, at the end of the day it is still credit. At the end of the day these portfolios are between a 100 and 200 actual companies. They are not thousands upon thousands of consumer receivables. I think the lesson learned by a lot of people is one that I think we had always understood, which is that you don’t trade credit based upon a model. You use a model as a common language and you trade credit based upon credit views. I think that ethic has been instilled a little more soundly in the market since May.
Richard Hrvatin (Fitch): As a rating agency we are not looking to price these deals – basically what we do is aggregate many credit default swaps into a synthetic CDO structure. When we developed our model, a similar type of Monte Carlo model for sizing up the credit enhancement levels, we studied what was in the market. But everyone is using the same set of assumptions and there is a danger in that, everyone giving too much faith in the model itself and not looking at the fundamental credits. We look at a ratings-based default probability, at what is the long-term horizon of this credit and what is the default probability. So the way our models work differs from the way the market models are working.
IFR: Is there any confusion about that?
Hrvatin: I don’t think so. I just think that there are two different motivations of how our model is being used versus how the market is using the model.
Watkinson: I think that really all that these models hope to achieve is some sort of defined hedge ratio between tranches and a hedge ratio between idiosyncratic credits.
IFR: Have hedging policies changed significantly since April/May? Is more active trading increasing volume in the market?
Smith: Any time a parameter that people assume is relatively constant turns out to be not so constant, people are a little more hesitant to take risks on the parameter. To the extent that correlation is moving a lot, dealers might be hesitant to take as much correlation risk.
Feldstein: The important assumption that goes into the model to make trading easy is that there is a constant correlation across all of the credits of the portfolio. And one of the things that we mean when we say we do the credit analysis is we actually think about how the credits are correlated to each other. And it would obviously be very difficult to have a liquid market if everybody, each time they traded, had to agree on a big correlation matrix before they agreed a trading price.
So we trade in the liquid market and we use the index as our delta and we use commonly accepted deltas simply to facilitate trading and to take out macro spread positions. But after the fact, we almost always will modify our index hedge by substituting varying amounts of single names based upon our view of the single-name credit risk and of the relationship of the different names to each other. So in effect, we are superimposing our own correlation matrix based upon our own fundamental credit use. A necessary simplicity of the way the market trades is a constant correlation. But the world doesn’t actually work that way.
IFR: When you superimpose your single-name hedges and your own correlations afterwards is it much of an issue how much the markets move or how quickly you do that? Does everybody else do that in as vigorous a way?
Feldstein: There are varying approaches to how you hedge these instruments starting from “not at all”. The vast majority of the mezzanine risk buyers are just long and some portion of equity risk buyers are long as well and just really do not hedge at all. Or if they do, they pick a name or two every now and then that troubles them and they just hedge that credit.
There are also very model-based hedgers who basically run their model each day and then just do what it says. And there are market participants like ourselves who are fairly active hedgers, running a book which is flat spread-risk in our opinion, but where our hedges are based on our fundamental credit views, which change as frequently as daily or hourly. So if we trade something in the correlation market and that alters our fundamental credit position the instant we do it, we generally very shortly thereafter reverse out longs or shorts that we would rather not have. But you do not do that just at the point of executing the correlation trade. It is an ongoing process as your views change and as the actual relationships among the credits in a very large portfolio change.
IFR: How is Delphi settlement likely to work?
Smith: I think this is an issue not about market dislocation but operational risk and a realisation that the market has grown much faster than one might have thought. To a large degree the market has already come to this realisation in terms of its move even with relatively small names defaulting in the index space to cash settlement, just because the number of contracts are so large.
Now we’re facing a situation with a bankruptcy in a company that has a large amount of single-name outstanding risk. The important thing to preserve in going to a cash-settle mechanism is to preserve the nature of the CDS contract as it was intended, so you can buy protection knowing you are protecting either loans or any of the bonds that you are hedging, but to combine that with something that is practical and allows for easy operational settlement. So the challenge is to construct a robust auction mechanism that is fair, transparent, and ideally allows for a large number of bonds to be transacted around that price.
Feldstein: I think it is going to be a binary result. If the auction process is good, the more people join the auction process, the more people will follow. If the auction process is perceived to be flawed and certain large market participants start opting out, that will cause others to opt out for fear of being left in the minority and being disadvantaged relative to the people that opted out. So I do not think there is a middle ground. It is either a whole lot or not very many.
IFR: Is everyone indicating that they will join in?
Feldstein: It is a fairly fluid situation where I think it is apparent to us that the dealer community is working extremely hard to come up with a solution that is fair to all market participants and does not disadvantage any particular type of market participant. And basically it ensures that everybody gets the benefit of their original bargain, what they intended to have, while at the same time avoiding what would be a very operationally burdensome several months if we actually went to physical settlement on all of the contracts.
But the big sticking point is how comprehensive is the settlement process with respect to the instruments that are included. And the dividing line is whether single-name contracts are included in the settlement process or whether they are not. Where you have market participants who very actively trade single-name products as well as index products, whether it’s the index or tranches on the index, those participants will be extremely reluctant to submit to a process that the settles some of their contracts but not others, because they would run the risk that there would be a different settlement price, and that market participant may have gone into this situation thinking that it was perfectly hedged with respect to Delphi because it had single-name positions which offset its index product positions.
So the most important thing to that type of market participant is less what the determined settlement price is and more that it is the same one for all of their instruments. And that type of market participant could ensure that effectively via a physical settlement if it just went to physical settlement of all of its contracts, and in fact, most market participants who have paid attention to their documentational basis risk and managed it well, could say the same thing. So at the end of the day everybody protecting their own interests, unless we can come up with a solution that works for everybody, will simply opt for physical settlement to protect against that basis risk.
Smith: As an important side note it seems to be quite likely, almost no matter what the outcome is, in terms of whether we actually get this auction developed in time, that this experience in Delphi will lead to the majority of new CDS contracts being written in cash-settled form.
Feldstein: Although we will still have the issue of hundreds of thousands of legacy contracts.
Smith: And potentially there will be a mechanism to exchange them developed.
IFR: How do you expect exchanges to work?
Smith: The market has gone through a series of these types of issues where it grows to a certain point where the old documentation and old frameworks did not work. And time and time again it has found a way to come to an amicable solution for everyone and move on.
Feldstein: In the long term these things are very good for the market because they refine the roles of engagement and the contractual relationship in a way that make the market better.
IFR: Could there be a fall-off in single-name liquidity in the near term? Have people been going away from single-name trading at all, or is it just as busy as ever?
David Carlson (Bear Stearns): I think it is just as active. Fundamentally, the contracts work extremely well. It is still the best tool for managing credit risk, so while we refine some of the settlement mechanisms, the fundamental value of the product is still largely intact and I think you will see volumes continue to grow.
Watkinson: I know the industry’s goal is for the next roll-date in March to have a cash-settlement mechanism for the indices. It would then make sense probably to have that be the case for single-names as well.
Feldstein: I think that if the indices went cash-settled without a cash-settlement option for single-name CDS, you would see a trailing off of liquidity in single-name CDS – as very active trading participants no longer viewed those as interchangeable. So a lot of trading strategies would take on basis risk that they did not previously have, and a number of the trading strategies that rely on trading the indices and single-names are those that are creating a lot of liquidity in the market. So I think that the community of market-participants recognises that and I think it would be unlikely that the index would come out with a cash-settlement mechanism without some plan for either a standard contract to trade in single-name cash-settlement or at least some kind of bifurcation where you can check a box.
IFR: You talked about new participants coming to the market. Are there any who are deterred by issues like these? Do you all feel you have to do a lot more education with potential different types of users?
Carlson: In the big picture, these are minor issues that are being worked out and they have been worked out pretty well in the past. Compared to the benefit of the product and what you can do with your portfolio and manage credit risk, there are things you need to pay attention to and learn and understand the nuance. But they are not things that once you get your hands around them are fundamentally going to prohibit you from getting into the market. In the short run, some controversies may back people away. But I think we get to these things pretty quickly. This will be resolved in a couple of weeks, maybe a month at the longest, and then we will have a new framework that will apply going forward.
Smith: I think we are all happier if we can bring in more participants, but the market is so large now that the best way we can continue bringing in more participants is to make sure we are certain of all the constituencies in our current market. Our market is big enough, has enough critical mass, that even if we did not have any other users it is extremely important to make the market continue to function as well as we can make it function.
IFR: How is the novation protocol going in terms of relations between hedge funds and dealers ?
Feldstein: From our perspective the novation protocol says the right things, we do not really have any problems with it. We think that we have actually abided by the rules of the protocol all along, before there was a protocol. Whenever we assign a trade, both sides are notified that day. We have done mass assignments where all of the notifications were handled the day the assignment took place.
The protocol effectively is putting into a statement what was already in the contracts between the market participants. In some ways it gives the dealer community extra leverage over market participants who are a bit recalcitrant and it gives them the ability to fall back on something other than the bilateral contracts that they have with each other.
But, most of the master agreements between dealers and those on the buy-side already gave the dealers the right to do everything that the protocol says. So all this does is makes it an even more public and obvious requirement.
It helps the enforcement, so we think the protocol is a good thing. I think we would like to sign a protocol in connection with a lot of other agreements and protocols including, for example, cash settlement mechanisms and other things going forward. And we think that should probably be done once. We probably should encompass as many of the things that we are changing as possible, but in terms of substance, I think the protocol is spot-on.
Watkinson: There is nobody that lacks motivation here. I do not think there is a controversy as it relates to getting it done and making it right and adhering to the new novation protocol. Having DTCC be up and fully running was the other critical factor and I think that is going to ease the whole straight-through-processing element of it. It just forces a quick turnaround to be able to fix that infrastructure problem.
IFR: Are you all happy with the way the discussions with the regulars went and are they happy with the timetable and the new matrix for measuring the change in the confirmation backlog?
Carlson: Bilaterally everyone was at various of stages of doing the same process and the good thing about this is that it has got everyone to the exact same spot in a short period of time. So I think from that perspective it has been a very healthy thing. And shining light on the issue is very easy when people who are in day to day business push these issues aside.
Smith: In terms of the timetable, my chief concern is that we address strategic issues as well as tactical issues. To the extent that we are self-imposing overly strict deadlines, I think we risk focusing on crisis fixing and therefore disrupting the market potentially for a period of time and under-addressing strategic fixes.
Feldstein: I think that is a very important point. We have been very vocal about the need to fix the confirm backlog problem for well over a year now. So we think it is great that the Fed got involved and rallied people and applied moral suasion to get the right amount of attention on this. But you can separate it into two issues: there is a confirm backlog that needs to be fixed now, and people’s attention should be on that. And then there is a need to make sure that we do not get right back in the same soup three months hence.
But I wonder whether the deadlines for some things ought to be framed in terms of having made the investments in the right technology, having made the investments in the right business processes, having made the investments in training people, having committed to compensation packages for people who are operating in the middle and back office at these dealers that keeps them there and retains them, Because one of the biggest issues that has led the market to where it is now is an over-investment in the front office relative to the investment in technology in mid office and back office. And that has to do with people and retaining and training the good people and providing them the technology that they need, and the ultimate solution has to focus on the business process and the workflow. To me that that is the way the targets and the deliverables should be framed.
IFR: Would that be too difficult to monitor for the regulators?
Smith: One way you could do it would be instead of unsigned confirms being the primary target, it could be something like the number of trades that are electronically matched or straight-through processed, or on a percentage basis.
IFR: Moving on to new products, what do you see as being issues with developing products like structured finance CDS, and how do you see those developing compared to the way that simple CDS and index trading developed? Are you all happy with the documentation, and the way that CDS of ABS trade?
Hrvatin: In the CDO context a lot of times it is really on a case-by-case basis how we analyse the individual swaps that go into these fields. For example, we have been seeing for the first time this year, synthetic CDOs purely of ABS assets, either all RMBS or all CMBS assets. A lot of that has to do with new documentation – that pay-as-you-go template that came out in summer. So we tried to shoe-horn those assets into a template that it really wasn’t meant to go into.
There are operational issues there that we need to understand more than if they are cash funded. It is a lot easier if you just have one big credit default swap on the entire portfolio. This way you do not make payments until you breach into a level of credit enhancement. And now what we are also seeing is what they call hybrid CDOs – half-cash and half-synthetic – which further complicates the matter.
Carlson: Away from the hybrids, where there are some tricky elements, I think the fact you now have the standard single-name contract to pay-as-you-go, is creating liquidity around the single-name market which is allowing you to do synthetic CDOs. We have done a number of fully sold deals in which all the underlying collateral was pay-as-you-go swaps and we have started to look at single tranches. I think market is still at its early stage and there is further evolution to come. But now you have the pieces of a puzzle that work and you can start to do some of the structures and innovations you have been able to do in the corporate world.
Smith: I think thus far there is a distinct difference with how ABS and preferred CDS contracts are being defined relative to corporates, and it might lead to a testing of that contract in the future. Specifically, they are generally triggered when there is a deferment of payment, which is a lot more of a soft-credit event than we are used to having in corporate swaps. In a preferred CDS contract as it is currently being written, the contract is triggered if there is a coupon deferral. And the coupon deferral does not accelerate anything in the capital structure of a corporation and it is something that is allowed under the documents of the existing security. Yet that is triggering a CDS contract. Presupposing that you are going to get physically delivered the preferred deal, that is OK potentially. But it is very different from how the CDS market has evolved in the corporate market and even in the emerging market space, which is that only if you see hard credit events and only near the end of the lifespan of that entity if possible, do you want to see a triggering of those contracts. If every time there is a deferment – an impairment – and that triggers a contract, to me it acts a lot more like a total return swap on that instrument rather than a CDS contract.
Watkinson: It is the difficult nature of the underlying product itself. It is a very different contract than a corporate hard credit default where you have bankruptcy or failure to pay. Alluding to modified restructuring then as far as the PCDS, the preferred credit default swap, it does get more complicated as it relates to what are you trying to hedge and whether or not it actually meets the characteristics of the hedge. Or what risk are you trying to take on, and does that meet the characteristics It is a bit different with PCDS and asset-backed CDS.
With PCDS it is different in that it does not look like corporate CDS, so if you are going to try to trade a corporate CDS against PCDS or asset-backed CDS, there is a whole different risk it introduces in an additional basis.?
At the end of the day though, who are the natural users and what are they trying to achieve? I agree that in PCDS right now there is not a completely standardised document and I think we need to get there and that we will get there quickly. The indices have presented such a wonderful forum for the industry to get together and cooperate and understand what their problems are and how to fix them and implement a fix that could hopefully reach the longevity and liquidity we are trying to achieve.
I think the market place is hoping that with the launch of the ABS index, the single-name credit defaults swaps will grow out of it. It is the opposite of how the corporate market, where we started with single-name CDS and then created the index. This will be build it and hope that they will come. With the PCDS there has been an index launched and I think that the industry is very close to getting together and actually coming up with a standardised contract – that is in the works as we speak. One other thing. In the very near future there will be a group aggregated to get together to talk about loan-only CDS. There have been some complications with a standardised contract there as well.
ABS CDS and PCDS have growth opportunities that are magnificent. I think loan only CDS reaches a whole new audience; you now have the entire capital structure, everything except equities that you can trade. I think that is a worthy goal to get to, but I think for loan only CDS, the players are very different yet again. So what are you trying to achieve? What are you trying to hedge? What is the exposure you are trying to gain? And if you put loan-only CDS into a synthetic CDO, how does that work?
Hrvatin: To add to your point about going down the entire capital structure with the exception of equity, if I can throw out one more acronym: EDS. Whatever happened on that? We spent a great deal of time on equity default swaps as a proxy for a credit event. We came up with a model of how we can do a CDO of equity default swaps. We built it and no one came; it just seemed to die.
Carlson: I think EDS was created without a natural end user. There were some equity buyers, but it was created by derivatives guys. It was an interesting concept to do with CDOs and it was marketed quite heavily. I think for the growth of these markets, for CDS of ABS, PCDS, loan-only CDS, the thing that will determine whether it is successful or not is if you go back to the underlying market and see whether you transferred risk – not as a derivative person might think but as the core cash risk players think.
So I think on synthetic ABS with the pay-as-you-go swap it is not what a corporate credit derivatives guy would create, but it is something where ABS players feel like they have transferred that risk. And as to the quirky nature of ABS – I do not think you can get away from that in the contract, I think you have to address it. So I think where you create a contract where the underlying market sees the value of the risk transfer in it, then I think you create the liquidity around it to do synthetic and portfolio trades. So I think that is the key success factor. Maybe you have deferral on preferred swaps, because deferral in the preferred market is a major event – maybe having that feature in the contract is quite important.
Feldstein: I actually think that you have to look at whether the contract itself has a natural universe of both buyers and sellers. And I think in the case of EDS, it did not have both. It was created because people thought they could create synthetic CDOs and the creators thought they could take a nice arbitrage out of it and it turned out there was nobody on the other side who wanted to accommodate that. So there were buyers but not sellers, or rather potential buyers of protection if the EDS CDO market could have developed. I worry about ABS CDS for the same reason.
I think there may be some hedgers who would like to shed ABS risk. The pay-as-you-go contract, if I understand it correctly, applies to specific bonds. And Derek was talking about the fungibility or the generic nature of corporate CDS, and my understanding is a lot of the CDS that go into the synthetic ABS CDOs, where there is a natural risk taker base are if you are going to do pay-as-you-go, link to very specific bonds. The C-class US$14m note issued in 1999 of the MBNA Trust V say. Well, where are you going to find somebody on the other side that wants to trade in that very specific bond that is buried in somebody’s portfolio somewhere?
The guy who owns the bond may just want to hedge it, but he could also just sell the bond. So I think the natural two-way flow in those pay-as-you-go contracts might be very different from the natural two-way flow in a contract that represents very generic risk. I worry about the long-term viability of ABS CDS for that reason. And for a second reason which is you are not really sure what kind of risk you are getting. You take a home equity loan -- does it stop paying because of interest rate factors? Does it stop paying because of credit factors? Does it stop paying because people are doing irrational things with their home equity loans? Who knows why it stops paying?
I am a little more interested in preferred CDS. And I think that the interesting parts or the interesting trades in preferred CDS would not trigger on a deferral of coupon, but would solely trigger on the same thing that would trigger a corporate credit default swap. So then you truly, in bankruptcy, you could actually say: “Look, I think the preferred is going to recover something,” so I think it isn’t a bad trade to be long the preferred versus short the senior. Or I think the preferred is going to get zero and there are people on the other side of that view. So if you can get away from the deferral as a triggering event, the preferred actually has some interesting applicability because it is just one more part of the capital structure. That is exactly the same reason I think loans are interesting.
Smith: We have a template for a close proxy in subordinated CDS. Some subordinated debt allows for deferral of coupons. Agencies, for example, without any cross-acceleration of debt and without any credit event, and that is a very actively traded market – the senior subordinated market.
Watkinson: The experience so far at Lehman for PCDS has been that we have traded well over US$3bn and have 52 swap counterparties. So it seems fairly robust in that there are buyers and there are sellers – and it has been good two way. I will echo again what you guys are saying that you have to have both sides to thrive and to grow and to have that applicability elsewhere, so I think that the end gain from my firm’s perspective is to find where ultimately all the supply – the buyers and the sellers do meet. Right now it seems it is with this deferral and with the ability to deliver the entire capital structure. But if that is not the case, we are all open to see what the right answer is.
IFR: What sort of a cross-section have you had among the 52 users?
Watkinson: The fascinating thing with the PCDS contract has been the number of real money guys that entered first versus the traditional derivatives player, which would be the hedge fund. It has been an incredibly broad mix of the two kinds of counterparties, but with real money being a fairly large chunk of that.
Smith: There is a very similar issue with loan CDS – I think it could be a huge market. I think it is a classic trade-off of direct transfer of the physical risk versus fungibility of the contract. I think the two chief issues are how callable it is, whether the callability is dependent upon what happens to the underlying loan and second how many underlying loans are brought as a deliverable set. And those two things are a direct trade-off between direct risk-transfer of the exact loan that the hedger is doing versus the fungibility of the contract on the other side, and maybe the ability to broaden the market to have active participants have it in CDOs.
IFR: Do you need indices there as well, or do you just need to be able to offer people whatever they want to use to hedge?
Smith: I am sure if the contract is right we could have indices. I do not know why not.
Watkinson: I think there could be a lot of neat applications for having an index, but yes we have got to get the contract right first. And I think just as in several of these other new markets, I do not think anybody has a strong axe. In the previous new product developments, many times people would walk in the door with a strong axe. I think the lessons to be learned from the last three years of product innovation and development are let's just get it right, and so let’s get a preconceived notion of what is right off the table and let's get a consensus.
IFR: How are options developing?
Watkinson: We have traded US$300m PCDS options. But for options on single-name corporate CDS overall, and options on the indices the growth trajectory since May has been phenomenal in our experience.
Carlson: I do not think the volumes are to the point where you would think they would get and I think the market suffers a little bit from the same phenomenon we have been talking about where there are not a lot of natural buyers and sellers. A lot of the attractions for options in the cash market are that they provide leverage.
The CDS market already gives you leverage, so you don’t need a CDS option to do that, so you really do need a client base that is using options to hedge things away from just taking a view on volatility. You are hedging some exposures and you need an asymmetrical risk-profile to do that and you don’t have a big client base doing that, so then you go to the hedge-fund community and it is difficult to trade volatility when you do not have liquidity and you are just trading volatility in an abstract fashion. So I think that until you see that market develop you are not going to have a huge growth in the volumes in line with what you’ve seen on the core indices and single-names.
Feldstein: We trade index options very actively; we see the index option market as incredibly robust and quite liquid. We usually get every size we would like to trade from multiple counterparties within a few minutes notice. We find that it is very different in the single-name market for exactly the reason that you point out. Some single-names, the expected ones, have relatively liquid option markets. Not surprisingly GMAC can trade fairly liquidly in the option market.
But I think when we trade options in the single-name market we are not comfortable really doing it as a volatility trade per se, that is we do not think we can gamma trade it and actually cover our transaction costs. So we are usually using it when we trade, which is not the case in index. The index market is so liquid and the bid-offer is so razor thin now that you can actually trade your gamma in the index option market. In single-names, we are generally using it to express some kind of directional view, either monetising an entry point, or for a difference in view about an entry point or an exit point to a single-name view that we have. And therein the problem is simply that it would only be by chance that the dealers we communicate with happen to find somebody on the other side of that view.
I think for a period of time, in late 2004, when people were really gearing up their options businesses, dealers were so keen to get the business going that they were providing a lot of liquidity in the single-name options market, which was very nice for participants like us. But I think that stagnated a bit when the dealers realised they were making liquidity for us and they really never developed a robust market – there were not enough other participants in the market to have an exchange of views about delta levels and volatility levels for different options. And the fortunate thing about the way the credit options market has worked so far in terms of being able to ramp up and ramp down is they are all generally three to six-month contracts. So the option market could pick up or wind down very quickly depending upon the entrance or exit rates of participants.
IFR: Are there applications for more fixings?
Watkinson: I will say that we are thinking about some products that a fixing would help with the constant maturity – a variance swap market maybe? From an innovation standpoint that probably would get us there, the ultimate would be a futures contract. But that obviously has met with a lot of its own problems. But in some of these third-generation products it might be useful.
IFR: Are people just not interested in credit futures contracts?
Watkinson: From my experience, there is a fundamental reason that it is difficult to launch a futures contract which we can get around – do a constant maturity and make it different than the index as far as having the term structure or time decay, so it has constant maturity. But it always goes back to “How do you settle a credit event?” There has been a lot of talk about it. I think that this contract is working so well, this meaning CDX in almost a futures contract’s form, that it is difficult to rally behind and put a lot of thought into that one fix.
Feldstein: I think the credit event is certainly one issue, although you could probably develop some kind of auction mechanism. I think a futures market for the indices probably is not too far off. But if you look at all successful future markets, they all fundamentally rely upon an extremely robust and transparent underlying market, whether you are talking grain or treasuries or what have you. And I think the single-name credit market has become very liquid and very transparent, but it does not, for most underlying credits, even approach the liquidity in the Treasury or equity market.
So, I think it would very hard for participants to trade futures on anything other than the indices, which are very liquid and very transparent, and a handful of names where people would be comfortable with fixings. But I do not know that there are too many market participants that would be comfortable with fixings on more than 10 to 20 names, if even that many.
Smith: I think that as we move through the issues that we have currently in the markets over cash settlement there will be a natural robust process for fixings. I think there will be a next generation of products that will evolve, assuming we find buyers and sellers, like constant maturity swaps which I think would be pretty useful. But given that backdrop I do not actually see the generational gain that we would have from having a futures market. The market is so liquid that I do not see an incremental need.
IFR: Are the differences between Europe and the US over issues such as the restructuring clause much of an issue any more?
Smith: If we do go to cash settlement it is really hard to see how we can retain a restructuring clause at all, because the issue you have is that every contract on a restructuring trigger could have a different set of deliverables. So if you are trying to come up with a single mechanism for settling contracts, it would seem to be very difficult to do that and still have a restructuring clause.
Hrvatin: As far as CDOs go, we have got a matching group in Europe in our London office and there are totally opposite types of CDOs that they work on. Something like 70 % or 80% of their business is doing synthetic CDOs. In the US it is just the opposite – 70% of our CDO business is doing cash CDOs, which is more traditional – kind of buy and hold investors that buy cash CDOs. It is a complete reversal of the way the European CDO market works. You now see a little bit of a convergence. Actually, with the whole hybrid structure, I asked the question of someone: “I really do not see the economics in the deal, why are they doing that?” And he almost made it sound like it was synthetic training for cash-flow people, so you have a deal that is half-cash, half-synthetic and you are going to get them more comfortable with the whole synthetic structure.
Carlson: I think there is a theme of convergence, certainly on the synthetic CDO front, because I think for the account base in Europe and Asia, such as the banks and the insurance companies, a lot of them have the same issues that they have in the US, where loan growth may not be as robust, and there is the asset- liability gap for insurance companies. Initially the Europeans were more comfortable with synthetic CDOs. The Asian investor base, not only may they be more comfortable, but the fact that you could get the investment in their local currency much easier, or have shorter tenors than a cash deal; there was a natural early demand for a synthetic product to do rather than a cash CDO.
I think you are seeing the US investor base becoming more educated on synthetics and more accepting, I think the accounting in the FAS 133 treatment, putting synthetics on par with cash will be a big motivator and accelerator for US players to get involved in a product. So while the trade types may be different to reflect the local market needs, I think the theme of cash and synthetics coming together and simply looking at overall structured credit as a tool for continued growth will continue.
IFR: Away from the mechanics of the market, what would constitute a serious credit event that would affect the sector? For example, what if GM goes bankrupt? Would there be new issues for the market or would it be more like the issues we have talked about already, but with a broader scope?
Feldstein: The obvious answer is that General Motors is the one that is on everybody’s mind. The big lurking risk is that there are a handful of credits that have been very popular in the synthetic CDOs – the bespoke CDOs that have been distributed predominantly in Europe and Asia. And the likely candidates which were investment grade or just barely investment grade-rated names that had a higher spread. So GM or GMAC, Ford, Delphi and – but probably not to the same extent – DaimlerChrysler. Those kinds of companies are very prevalent. And if one of those defaults or a couple of those default, you could see the subordination, and therefore the ratings in a number of these synthetic CDOs cut very quickly.
Hrvatin: The credit enhancement levels for an investment grade portfolio are markedly different than from a non-investment grader. They are very non-linear, so one or two names can blow right through those credit enhancement levels.
Feldstein: And I think it is exacerbated by some of the kinds of deals that were done more recently. The CDO squared deals in particular, where effectively GMAC or GM does not just show up once – it might show up 10 times, so that a default at GM could on its own pierce the subordination of the CDO squareds.
Hrvatin: Actually the whole CDO squared market is kind of leverage on leverage. And talking about this non-linear effect, the correlation assumption just becomes hyper-sensitive for the CDO squared.
Like you said, GM could show up 10 different times. If you have a low correlation assumption and that same name shows up over and over again, it is not going to have that non-linear effect on credit enhancement levels. But just slight differences in your correlation assumption are going to have huge skews in that.
Andrew Feldstein: I think that the thing that has really been keeping a lid on global credit spreads for the last two years has been the bid for credit risk by CDO buyers. And I think if that lid blew off the kettle, because GM defaulted, you would see a massive widening of spreads globally, as the appetite for credit risk by this new set of buyers would probably dry up very quickly. And they may actually become the next sellers of credit risk pretty quickly as they try to get out of their CDO squared deals.
There was a lot of talk in May about the idiosyncratic risk that was faced by hedge funds and dealers as a result of the autos. I think that that idiosyncratic risk very quickly becomes macro spread risk if one of these very popular names, like GM, were to default, because I think you would find a lot of structured credit holders who had Double As or Single As suddenly holding Double Bs or Single Bs. And there may be panicked selling on their part, which would probably blow the lid off global credit spreads.
Carlson: I think there is a lot of fear about synthetic CDOs and the amount of volume that has been done. But I think the way those deals are structured, at least away from the CDO squared market, is that the impact of one single name on the portfolio, because of diversification, is not as great, especially at the mezzanine layer. And a lot of the older deals have rolled down the curve and so their ratings cushion is larger than people would think.
We just looked at Delphi and we looked some old deals and there is a cushion where there’ll be no impact. You are not going to have the downgrade at the mezzanine layer, so I do not think you will have the big meltdown on credit spreads due to people panicking in synthetic CDOs. It is a mysterious faceless market that people can reflect on, so there might be some bad deals here or there or CDO squared deals with a ton of overlap. Although most of the deals I have seen have not had that much overlap, it is not to say that someone has not structured something that is going to do extremely poorly. But I think if you look back at 2000 when you had Enron and WorldCom – I mean that was painful for a couple of deals, but for the majority of the deals, people weathered it and got through. And some are probably happy if spreads double. I still think that idiosyncratic risk will be the more likely problem, or the knock-on effect of having exposure to GM for a credit you did not think had it. I do not think synthetic CDOs are going to be the culprit one way or the other.
Smith: When you think about the implied default rates that the market is pricing for the next five or 10 years, there has certainly been a significant move out in maturity, we have seen a large increase in the number of seven to 10 year synthetic CDOs being done. If you look at those implied default rates, I think it is reasonable to think that if actual default rates are significantly higher in the short term than those priced in, then there very well could be stress on that market. And that could easily lead to stress in the single name markets and broad credit levels in general.