The lessons from Delphi's default

The bankruptcy of US car parts maker Delphi last October proved a real test of the market's ability to settle outstanding contracts. With the spectre of GM looming large this year, Sarfraz Thind reviews the lessons learned from the landmark Delphi settlement process


It has taken a while for the issue of settlement of credit derivative contracts to come to the fore. While participants have considered proposals for improving the settlement process, it is only in the last year that the market has really been tested on this issue. The increased number of defaults in the latter half of 2005, following the relatively benign credit cycle in the two preceding years, and the continued growth of credit derivative volumes has now made this one of the top priorities for the market.

In its most recent Quarterly Review, published in December 2005, the Bank for International Settlements highlighted the risks posed as a result of the enormous growth rate of credit derivatives which, in volume terms, has far outstripped the growth in underlying cash bonds in recent times. The requirement for physical delivery, stipulated in the majority of credit derivative contracts has, said the report, increased the risk of squeezes "in which the demand for the debt of a firm exceeds the supply of such debt, resulting in the breakdown of the normal pricing relationship between credit derivatives and the underlying debt contracts. This in turn may cause traders to withdraw from the market, thereby draining liquidity."

The settlement issue was brought into sharp focus last October following the default of US car parts manufacturer Delphi. Already under pressure from the New York Federal Reserve Bank to clear the huge backlog of outstanding unconfirmed trades, the credit derivatives market faced the problem of settling the enormous volume of contracts referencing Delphi, estimated by Merrill Lynch at $28 billion.

Huge buying pressure saw the firm's bond price spike to 71% in the three weeks following the bankruptcy announcement (see chart 1), as market participants scrambled to source deliverable Delphi bonds from the $2.2 billion outstanding.

With fears of a potential market dislocation, the 14 largest credit derivatives dealers met and agreed on a cash settlement process for the outstanding index and index tranche trades referencing Delphi. The objective was to relieve the technical pressure on the bonds and prevent a massive short squeeze which would have left protection buyers - who receive the difference between the settlement value of the bond and par - holding on to essentially worthless credit derivative contracts.

Dealers settled on the credit event auction process originally developed with data provider Markit and credit derivatives e-trading platform Creditex. This process was used for the post-bankruptcy settlements of car parts supplier Collins & Aikman as well as Delta Air Lines and Northwest Airlines in mid-2005.

In all, 574 participants signed up to the International Swaps and Derivatives Association (Isda)-drafted Delphi protocol supporting the auction. They represented nearly all the relevant parties with exposure to Delphi.

"The most important thing on Delphi was that you got all 14 major dealers involved," says one head of credit trading. "The process of physically settling millions of dollars of outstanding contracts without full participation could have radically skewed the price."

With the volumes of outstanding Delphi contracts dwarfing that of Collins & Aikman, Delta Air Lines and Northwest Airlines, changes were made to improve the procedure used in the earlier credit fixings.

Basis risk

Participants were concerned that, because many of the index contracts outstanding on Delphi were hedged with single-name credit default swaps, which had not been cash settled, contract holders would be exposed to basis risk. Hence, unlike in the previous auctions, dealers also had the opportunity to trade the underlying bonds at the final cash settlement price, submitting two-way executable prices with a 2% bid-offer spread on $10 million notional bond values.

This particular change has been praised as an advance on the previous auction methodology. However, some dealers have complained that the $10 million notional amount on the cross-trades was too high, given that the average Delphi index contract would have traded in the order of $2 million in the market.

"The whole process is designed to ensure smooth settlement, not to take on bonds where you don't need to and in sizes which could double your position the other way," says a head trader at a European bank. "If you got lifted on bonds through a bad directional play on $10 million of notional, it would hurt."

Participants have suggested that $5 million would have been more attractive given the positional risk that individual dealers were committed to taking in the auction.

In addition to entering a tradable bid and offer price, participants were also required to represent the bond position they would have held if the contracts had been physically settled. Dealers were therefore restricted to being able to put an order reflecting the net position of all their index trades and those of their participating clients.

"This expedited the closure of residual single-name Delphi positions, avoiding some of the basis risk associated with settling index and single-name trades at different levels," says Mazy Dar, head of electronic platforms at Creditex.

In total there were $186 million of market order bids submitted and $285 million market order offers, with 16 matching trades, leaving an imbalance of $99 million open interest - the difference between the total market bids and market offers - on the offer side.

The second stage of the auction involved matching limit order bids to the open interest - in this case the open interest was filled with limit orders to buy. The final cash settlement price of 63.375%, taken from the final limit order used to fill the open interest, was around five basis points out from where Delphi bonds were trading immediately prior to the auction.

The net sell market order surprised some who thought that the auction would have been bid up by those anticipating a short squeeze. "I thought there would be better buyers of the bonds to settle the contracts," says a head trader at one of the large credit derivatives houses. "But if the buy-and-sell order had been more lopsided this could have resulted in a massive imbalance, which would have been extremely bad for the market."

In the event, while there were at least two large hedge funds which abstained from signing up to the protocol, the high turnout of both dealers and their clients meant that the open interest was not as large as it could potentially have been. However, say participants, the open interest was partly curtailed by the fact that the auction was only used to settle the outstanding index contracts. Had single-name credit default swaps also been included, the open interest could have been much larger.

But while the auction served to take a lot of the sting out of the market, the influence of technicals on the value of the underlying bonds prior to the auction, and the threat of a short squeeze, remain points of concern for future credit event settlements. The rise in the Delphi bond price following its bankruptcy was contrary to the usual pattern following most bankruptcies (see chart 2).

And the huge turnover in Delphi's bonds prior to the auction was not only caused by nervous protection buyers looking to source deliverable bonds, but by speculators anticipating a short squeeze in the market.

"Everyone knew that the 70 price was bogus," says a credit derivatives trader. "But no one wanted to sell short with the threat that they couldn't cover position. A short squeeze is not good, but it would have been more of a problem if the price had ended up the other way around: in other words if the final settlement value was much higher than pre-auction levels."

Whether the auction took enough of the sting out of the market is open to debate. By the end of December, Delphi bonds had dropped to the low 50s, though they rallied back to around 56% by the middle of January.

At the end-of-year price, unhedged protection buyers holding Delphi index contracts would have realised around 30% more than they achieved in November's auction price. One indicator of fair value of the bond can be seen in the recovery value assigned to Delphi's debt. Fitch Ratings, for one, put the company's senior unsecured debt in its R6 recovery rating category, signifying a 0-10% recovery band.

"Post-auction and with the subsidence of speculation regarding a possible short squeeze, bonds traded down 15 to 20 points from levels that appeared to be unjustifiable on a fundamental basis," says James Batterman, senior director at Fitch Ratings in New York. "In fact, current levels are still significantly higher than many published estimates, including our own."

But whether anything could have been done to prevent the price squeeze, short of imposing a trading curfew on Delphi bonds, is doubtful. "Traders will always try to engineer squeezes to make money," says an unnamed source at a major credit derivatives technology firm. "But you have to be brave to engineer a squeeze that's five points away from where the market ends up."

With single-name contracts still being cash settled, the broader issue of basis risk in the settlement process also remains a concern. As the market has seen greater amounts of basis trading around index and single-name default swaps, so the issue has become magnified, and many in the market have called for cash settlement for all credit derivative contracts.

"We need a process that will ease the settlement of gross positions and still maintain a common settlement mechanism across single-name credit default swaps, index swaps and tranche contracts," says Bryan Mix, managing director in the credit group at Goldman Sachs in New York.

Synthetic CDOs

And it is not only single-name and index contracts where there is a basis risk. The popularity of Delphi in many synthetic collateralised debt obligations (CDOs) - which remain cash-settled - during last year (Delphi was the nineteenth most referenced entity in Fitch's synthetic CDO index) highlighted the importance of agreeing a uniform settlement mechanism across all credit derivative contracts.

At present the market is debating the possibility of cash-settling all credit derivative contracts. For dealers this has the major benefit of easing the operational burden on the banks' back-office departments, as well as helping to avoid potential market dislocation.

Yet cash settlement may not be universally accepted. Firms that already hold an underlying deliverable, which may be hedged against a credit derivative position, would not benefit particularly from entering into a cash settlement mechanism and being subject to any potential vagaries of the cash settlement procedure.

The credit derivative settlement issue is currently a top priority for the dealer community and for the legislators, says Bob Pickel, chief executive of Isda. The derivatives trade organisation called a meeting at the end of January to discuss the issue of cash settlement of all credit derivative contracts, and participants remain hopeful that a more comprehensive solution can be found.

"We want to move away from the ad hoc protocols to a systemic mechanism," says Mix at Goldman Sachs. "There have been conversations regarding this and we may see something in the first half of this year." Hopes are that there should be something concrete in place for index contracts at least by the time of the next iTraxx index roll in March.

With the default of US power company Calpine in December the credit derivatives market's attention was once more focused on the settlement problem. At January's credit event auction for Calpine, the final price settled was 19.125 with total open interest at $45 million on the offer side.

The terms of this latest auction have been developed out of the Delphi credit event fixing. Again each participant will submit limit and market orders and indicate the size and direction of their orders. Participants will also be able to trade underlying Calpine bonds at the cash settlement price, on $10 million notionals. However, with Calpine referencing only a fraction of the volumes of credit derivative contracts of Delphi, the likelihood of serious market disruption remains much smaller.

A bigger shadow looming over the market is the growing likelihood of default for General Motors. GM, with over $100 billion of bonds in circulation, and its subsidiary General Motors Acceptance Corp (GMAC) are amongst the most referenced names in the credit derivatives world. A GM and GMAC default would undoubtedly cause major disruption to the market.

That said, participants believe the potential problems of settlement over GM are not quite as great as for Delphi. While a default may hit the market hard, GM deliverables should in theory be easier to come by than Delphi's, where the ratio of outstanding credit derivatives to cash bonds was substantially higher.

In preparation for this eventuality the market has been working to achieve a consensus on how to best advance the settlement process. With the growth of the derivatives market showing no signs of abating, the sooner that participants agree on a permanent solution to this issue, the better.

Credit derivatives: the sprawling monster

The growth of the credit derivatives market in the past few years has been phenomenal. Credit derivative volumes ballooned by 60% to $10.2 trillion in the first half of 2005 compared with the first six months of 2004, far outstripping the growth rate of the underlying cash business, according to the Bank for International Settlements in its December Quarterly Review.

The market has also widened, with the more recent subset of credit derivatives such as index-related products showing huge growth potential: index trading has grown exponentially over the past two years. According to a recent report by John Tierney, head of US credit derivatives research at Deutsche Bank, outstanding contracts in the bank's index products in 2005 grew to 20% of the overall business, from 11% the year before. In pure trade volume terms, around 60% of Deutsche's credit derivative trades were index linked in 2005, up from 35-40% in 2004.

And the regulatory framework has thus far not been able to keep pace with the changing dynamics of the market. "In 2004, during discussions about credit fixings, several institutions raised the possibility of using a fixing to resolve a credit event," says Mazy Dar, head of electronic platforms at Creditex. "However, it was not a primary consideration at the time because the default rate was low by historical standards.

Participants have been working aggressively to find solutions to the growing settlement issues in the market. One particular development, to which many have been lending support in recent times, is the trade unwind service offered by financial technology company TriOptima. In 2005 the service was responsible for terminating over 71,000 index and single-name credit default swaps with a notional value of $1.226 trillion.

TriOptima has also helped to reduce the settlement burden for credit derivative trades post-bankruptcy. TriOptima was used to tear up $23 billion of single-name trades referencing Delphi in October following the company's default, representing around 70% of the outstanding volume of contracts on Delphi, says Susan Hinko, managing director for North America at the company. Participants have been exploring the possibility of improving the tear-up service to include CDOs and tranched products in future.

As the credit derivatives market continues to grow, so does the work for those involved in the business.

But it is not all bad news. "No one wants to stop the market growing," says the head of credit derivatives at a major European bank. "The credit derivatives market has become one of the major means of dispersing risk. Four years ago if a GM defaulted it would probably have taken a bank with it."

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact or view our subscription options here:

You are currently unable to copy this content. Please contact to find out more.

You need to sign in to use this feature. If you don’t have a account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here