Contents

CDR Liquid Index data as at 4 June 2007

Source: Credit Derivatives Research

CDR Liquid Indices
The CDR Liquid indices represent the CDS levels of the most-liquid names in their respective markets and ratings classes. Liquidity is measured by the number of bid-offers a credit receives. Index values are simple averages of on-the-run five year CDS levels.

CDR Global Market Depth™
The CDR Global Market Depth Index is a daily measure of how many names are actively traded. Liquidity is measured by the number of bid-offers a credit receives. Index values are counts of the number of names that exceed CDR's Liquidity Floor.

CDR Global Market Activity™
The CDR Global Market Activity Index is a daily measure of activity within the global CDS market. Liquidity is measured by the number of bid-offers a credit receives. Index values are simple averages of total bid-offers of all names that exceed CDR's Liquidity Floor multiplied by CDR's Global Base Liquidity Constant.

Funds raise market manipulation concerns

Hedge fund calls for amendment of the CDS on ABS template

New York-based hedge fund Paulson & Co. and other funds involved in credit protection buying in the CDS of ABS market are calling for the revision of contract documentation. Specifically, they are seeking to preclude non-economic transactions by credit protection sellers with the underlying ABS structures that have the effect of manipulating the CDS market.

Michael Waldorf, senior vp at Paulson, explains: "We are concerned about an issue surrounding CDS on MBS that would appear to leave the market open to manipulation." He is, however, keen to stress that the issue has nothing to do with loan modification.

"For the record, Paulson does not oppose the modification of sub-prime loans. To the contrary, we support the modification of sub-prime loans because it is the fastest way of addressing sub-prime problems and helps keep borrowers in their home," Waldorf says.

Instead, the hedge fund is concerned that sellers of CDS on ABS credit protection could gain access to the servicing rights of the underlying MBS structures and could engage in what can be termed 'non-economic transactions' that have the result of pushing around CDS prices and manipulating the market. Paulson has consequently written to ISDA to propose amending the relevant documentation to prohibit such access.

The theory is that a CDS protection seller could either buy or hold the residual equity or servicing rights in the underlying structures, and this would enable them to buy defaulted mortgages out of the trusts at par rather than at their reduced value, thereby injecting cash into the trust. This would mean that the trust's securities would be more likely to pay off and consequently the protection seller would be less likely to have to make payments under the CDS contracts – which, thanks to leverage, could be far more costly than the cash injection payments.

Waldorf adds: "We saw this as a possible issue for some time, but until recently did not regard it as a practical one because we believed it was market manipulation, which is illegal. However, our view as to the likelihood of these transactions changed when Bear Stearns introduced a proposal to ISDA to modify the CDS on ABS template to allow free rein to sellers of credit protection to enter into transactions with the underlying mortgage-backed securities trusts."

Bear Stearns' proposal has since been withdrawn. "Anything we do to keep borrowers in their homes like the EMC Mod Squad has the potential to negatively impact someone who is short the sub-prime market," says Tom Marano, global head of mortgages and asset-backed securities at Bear Stearns. "The ISDA contract states that all rights and rules of the underlying reference obligation apply. In response to a suggested change that would have modified those rights and obligations, we proposed a clarification of what the rights were in order to ensure participants understood the terms of the underlying documents. When market participants said they believed the documentation was already adequate, we withdrew our proposal."

Paulson has now enlisted the help of several other funds that are active protection buyers to lobby for a clarification of the issue in the CDS of ABS documentation wording. The fund has hired Harvey Pitt, former chairman of the SEC, to advise on the issue and, according to Waldorf, Pitt's opinion is also that this is illegal market manipulation – irrespective of whether the contracts permit such transactions. "Nevertheless, other market participants don't have to agree with our conclusion on legality to realise that the current position is undesirable for the market as a whole, as it will ultimately dissuade credit protection buyers from trading," says Waldorf.

For its part, ISDA says that the trade association and its membership are in the process of discussing the points made in Paulson's letter. Robert Pickel, ISDA's ceo, comments: "ISDA's documentation and its extensive work in respect of privileged information relating to the credit derivatives business make expressly clear that parties must not only comply with the law, but must also observe sound practices and principles in the conduct of such business."

MP

US futures' details announced

First-to-launch credit index contract will differ from European model

Both the Chicago Board of Trade and the US Futures Exchange announced details of their credit derivatives contracts last week. CBOT aims to begin trading its index contract on 25 June, while USFE's agency-based credit futures are planned for launch in Q4.

The CBOT CDR Liquid 50 North American Investment Grade Index futures contract will be based on the CDR Liquid 50 NAIG Index. The Index is developed and maintained by Credit Derivatives Research (CDR) and also published weekly in SCI. The CDR Liquid 50 NAIG, which includes 50 North American investment grade reference entities, aims to offer a transparent, unbiased selection process based upon a defined set of rules, including minimum levels of liquidity. For this reason, the index is reconstituted every three months to ensure it includes the most liquid entities from the OTC market.

The fact that the index rolls on a quarterly basis fits well with standard futures market practice. Indeed, CDR has moved its roll date up by a month so that it coincides with the standard settlement dates on the futures contract.

The CBOT contract will differ in other ways to the futures being traded on Eurex, according to Dave Klein, CDR manager of credit indices. "It is a self-contained contract that utilises a simple mechanism, whereby everybody knows that the recovery rate is 40% and consequently what the maximum running spread on each name is at outset. There is therefore no split off into another futures contract when a credit event occurs," he says.

CBOT says it has done extensive pre-launch research into the potential demand for the new contract. "There is a very clean and defined structure that we employ when looking at new products and part of that is talking to market participants and potential end users. We have been encouraged to go ahead with our product and its design by the overwhelmingly positive response we have had from both the buy-side and the sell-side," notes Gene Mueller, md financial instrument research and development at CBOT.

The exchange says it intends to create a market maker programme to help ensure that a two-sided market will be readily available for electronic trading. Credit Market Analysis (CMA) will provide pricing information for all the underlying component issues within the CDR Liquid 50 NAIG index and its pricing data will be used to calculate the index value on the futures' quarterly expiry dates.

Meanwhile, USFE (formerly Eurex US) says that it will list the first credit derivative futures on Federal Agency debentures, beginning with a government tranche of the CDX index including CDS on Fannie Mae and Freddie Mac. The contracts were designed in conjunction with Bear Stearns and aim to allow for the creation of synthetic agency notes, as well as spread trades against corporate and sovereign debt.

"Default swaps from both Fannie Mae and Freddie Mac are included in the tens of trillions of dollars referenced to the CDX family of indexes," says Satish Nandapurkar, ceo of USFE. "Yet there has never been a distinct government tranche to represent the highest quality credit risk. USFE is pleased to provide fixed income investors with a new, on-exchange opportunity to hedge risk in this area."

The other two major Chicago exchanges had been waiting on regulatory approval before announcing launch dates for their credit futures products. CME received approval from the CFTC yesterday, 5 June, regarding its initial contract, which is to be based on an index of bonds of 32 North American investment grade, high volatility names. The exchange plans to list single name contracts at a later date.

The CBOE says it expects to hear from the SEC shortly and hopes to launch credit products by the end of the month. It will first offer single name credit default options and then credit default basket options.

MP

RelVal opportunities revealed in European LCDS

Recovery risk proportionate to the number of senior secured loans on an issuer's balance sheet

Research published last week by BNP Paribas suggests that CDS on European leveraged loans offer strong relative value opportunities versus high yield products.

"Despite weakening fundamentals, and despite the unfavourable asymmetry of the risk-reward profile at current tight market levels, we believe that there are still pockets of value in speculative-grade corporates," explains Andrea Cicione, credit strategist at the bank.

In order to assess the relative attractiveness of leveraged loans and crossover bonds, the research explored the default probabilities and recovery rates implied by the iTraxx Crossover S7 and LevX indices. Assuming respective recovery rates of 40%, 85% and 40% for the iTraxx Crossover, LevX Senior and LevX Subordinated indices, the implied default rates for the indices are all roughly the same – about 3.5%.

According to BNP Paribas, this number indicates a risk-neutral (or pricing) probability, and cannot be directly compared to the historical default rates. However, its marked difference from the latest actual default rates suggests two things. First, the market is expecting actual default rates to pick up; or, second, the sub-investment grade market commands a significant risk premium.

In addition, Cicione says: "There is a third possibility: that there is still value in sub-investment grade credit, even at the current historically tight spread levels. Rather than making such a directional call, we believe it is more interesting to try to understand where relative value is between the different asset classes."

Most of the companies in the Crossover index not only have unsecured senior debt outstanding (referenced by the CDS in the Crossover index), but leveraged loans as well. "In our view, it is unlikely that the recovery rates for senior unsecured bonds would be the same, in case of default, as that of second and third lien loans, which are secured and senior to the bonds, and have tight covenants," Cicione adds.

From this perspective, it appears that the Crossover index spreads are too tight compared to the LevX Subordinated index. Assuming a slightly lower recovery rate of 30% for the Crossover names, in order to keep the implied default probability unchanged at about 3.4% spreads would have to be about 220bp. A 20% recovery would require spreads to be about 255bp to be fair relative to LevX Sub levels.

In the context of single names, the report argues that the greater recovery risk resides with the issuers that have a higher loans-to-'loans+bonds' ratio: the higher the proportion of senior secured loans on a company's balance sheet, the lower the likelihood for the (relatively junior) unsecured bond holders to recover a significant fraction of their investment in the case of default.

Looking ahead, Cicione concludes: "Risk-reward is deteriorating, driven by a combination of higher leverage and a potential profit slowdown that looms ahead as the global economy slows. For now the party may continue for a bit longer given still strong liquidity. However, should winds change, both will encounter choppy waters, but the leveraged loan market may provide better refuge compared to the high yield market."

MP

Lehman adds Mustique

Innovative CDO-squared launched with mitigated high sub-prime exposure

Arranger Lehman Brothers is currently marketing Securitized Product of Restructured Collateral Limited SPC's Mustique Series 2007-1. The CDO-squared's headline exposure to sub-prime collateral is substantially reduced, thanks to the make up of its reference portfolio.

The deal, which is rated by Fitch Ratings, offers six series of notes: €19.6m triple-A rated Class A2As; €19.6m double-A plus Class A2Bs; €19.6m double-A Class Bs; €13.1m double-A minus Class Cs; €19.6m single-A Class Ds; and €8.2m single-A minus Class Es. The reference portfolio is composed of cashflow CDOs which are exposed to high concentrations of US sub-prime RMBS securities.

Fitch says its outlook on the US sub-prime sector is negative. This exposure, while indirect, is therefore still a risk to the performance of the transaction. However, the risk is reduced by several factors, including excess subordination at the master level, excess spread within the underlying CDOs, relatively low levels of overlapping RMBS exposures and the lack of exposure to pure second lien RMBS securities.

To fund its obligations, Series 2007-1 Mustique Segregated Portfolio and Mustique LLC - two special purpose vehicles respectively incorporated under the laws of the Cayman Islands and the State of Delaware - have issued in total €99.7m credit-linked notes. The note proceeds are invested in a GIC account with Rabobank.

Mustique provides protection on 23 tranches of cash and hybrid CDOs of mainly US prime and sub-prime RMBS through a portfolio CDS. Credit events are based on a pay-as-you go (PAUG) template, under which payments made following 'failure to pay' and 'implied write-down' floating amount events on any of the reference CDO tranches could be reimbursed within four years of the floating amount event.

When aggregate losses or write-down amounts under the terms of the CDS exceed the available credit enhancement, Series 2007 -1 Mustique Segregated Portfolio and Mustique LLC will be liable to make protection payments to the swap counterparty and the notes will be written down accordingly.

Fitch says its ratings address the payment of interest and ultimate principal in accordance with the transaction documentation. The ratings are based on the quality of the reference portfolio, available credit enhancement, the sound financial and legal structure of the transaction, and the strength of the counterparties.

Series A2A has been assigned a triple-A rating with no credit given to the remote possibility of late recoveries on the underlying CDOs.

Credit enhancement for the Class A2A notes is 18%; 15% for the Class A2Bs; 12% for the Class Bs; 10% for the Class Cs; 7% for the Class Ds; and 5.75%. for the Class Es.

MP

Structured credit hedge funds edge higher

Latest index figures indicate a reasonable April

Both gross and net monthly returns for April 2007 in the Palomar Structured Credit Hedge Fund (SC HF) Index showed a slight improvement on the previous month. The latest figures were released this week and show a gross return of just 0.40% for the month, while the net return grew by 0.25%.

Both the gross and net indices are therefore continuing to show positive cumulative returns since calculations began in January 2005, of 119.50% and 113.68% respectively. For more Index data click here.

The objective of the Palomar SC HF Index is to produce an index that represents the risk and return of investable hedge fund investments in the structured credit area. The index aims to provide a monthly measure of the performance of the universe of open, investable structured credit hedge funds. The Palomar SC HF Index is calculated in two formats - as gross asset value and as net asset value.

The Palomar SC HF Index is compiled and run by Palomar Capital Advisors and published exclusively by Structured Credit Investor. Palomar Capital Advisors is a financial advisory firm specialising in structuring, managing and placing alternative investment products, specifically credit-related securities. It is an independent firm based in Zurich, Switzerland, owned and controlled by its investment professionals.

MP

Absolute understanding

Campbell Smyth, a portfolio manager at Absolute Capital, answers SCI's questions

Campbell Smyth

Q: When, how and why did you and your firm become involved in the structured credit markets?
A: Absolute Capital has been involved in the structured credit markets since the company's inception in 2001 with the founders and senior staff being involved in the market prior to joining Absolute Capital. Structured credit has been the main focus of the firm since its inception, both as an investor and an issuer.

The rationale of a structured arbitrage transaction referencing credit makes a lot of sense. Absolute Capital has invested in the equity and debt of over 100 US and European CDO's and credit opportunity funds, along with direct investment in asset-backed securities in Australia.

Q: In your view, what has been the most significant development in the credit markets in recent years?
A: The evolution of the synthetic market, both in index and single name form, has been one of the most significant developments in the credit markets in recent years. This has allowed for more efficient pricing; greater ease of going both long and short; and allowed structured product issuance to be uncoupled from the availability of cash assets.

Q: How has this affected your business?
A: Absolute Capital, as with all participants, has had to ensure it is up to speed on the concepts around the synthetic market – in order to invest in anything one should first understand it and too often structured credit investments do not perform to expectations due to factors other than credit performance. There has been much talk in the market surrounding the CDS contracts with respect to both structured products (cash CDO tranches, for example) and individual credits (LCDS, for example). Clearly in transactions utilising synthetic technology there are additional risks that accompany the potential increase in efficiency.

Q: What are your key areas of focus today and what is your strategy going forward?
A: From an investor perspective, Absolute Capital is focussed on the same thing we have always been – levered exposure to pools of credit managed or originated and serviced by parties we consider to posses the prerequisite skills to do so throughout the whole credit cycle and who are among the top tier of firms in their respective market. Quality is the first and foremost consideration, with pricing subsequently evaluated once we are comfortable with our understanding of the potential outcomes in the transaction.

Q: What major developments do you need/expect from the market in the future?
A: As with all markets as they mature we generally expect to see greater liquidity available to investors. We expect (and in some ways have moved to anticipate) a shake up in corporate credit. The longer the extension of the cycle the greater the potential problems and therefore we would prefer to see minor corrections earlier than a wholesale blow up later. At the same time as we see future bumps in the market we are likely to see consolidation whereby only the best issuers are able to continue issuing and others are forced to get by on what they have done already or get out all together.

There is no doubt that the evolution of the LCDX and tranched versions thereof will have some impacts on the CLO market, however we will need to see decent liquidity in the index and the underlying LCDS for this to occur. Given the makeup of the LCDX it's far more likely to be useful than the ABX (other than as a short!).

Finally, one thing is for sure – when somebody says something can't happen again, it probably will.....

About Absolute Capital
Absolute Capital Limited is a specialist boutique structured credit manager, offering a range of structured credit and fixed income products to the New Zealand and Australian markets.

Established in 2001, Absolute Capital offers focused boutique capability backed by the strength of a global financial institution shareholder. Absolute Capital Group Limited is 50 per cent owned by ABN AMRO Australia, the ultimate parent of which is ABN AMRO Bank. The balance is held by Absolute Capital Management Holdings, representing the interests of staff and management.

Absolute Capital uses non-traditional investment techniques to meet the increasing demand of yield-driven products to create absolute returns. Capital preservation is achieved with risk being limited to an acceptable level sufficient for the generation of investment returns.

Deutsche makes credit trading hires

The latest company and people moves

Deutsche makes credit trading hires
Nick Pappas has joined Deutsche Bank as head of investment grade credit trading, reporting to Antoine Cornut who recently joined from Credit Suisse as head of credit flow trading (see SCI issue 39). Pappas was previously an md at Goldman Sachs in New York.

Deutsche Bank has also appointed Andrew Yeo as director, global credit trading – private equity. He will be based in Singapore and report to Sajid Javid, head of global credit trading Asia. Yeo joins from Temasek Holdings, where he was a director responsible for sourcing, leading and executing private equity investments in the global technology, media and telecommunications sectors.

ABN team exits
Ian Robinson together with three other members of his cash CDO team at ABN AMRO in London have all left the bank. Their destination, or destinations, are not yet known.

Bear grows CPPI capability
Eliav Danon, formerly a synthetic CDO structurer at Dresdner Kleinwort, has joined Bear Stearns in a similar role, reporting to European head of credit structuring and marketing Sergio Ravich. Danon is understood to have been hired to help grow and develop Bear's CPPI business.

Scott joins Calyon
Michael Scott has left ABN AMRO, where he was selling structured credit to UK investors, to join Calyon in a similar role.

Changes at UBS credit sales...
The UK credit sales to hedge funds desk at UBS, run by Tariq Islam, has made two additions. David Mcleish has been hired as a director from Morgan Stanley and Tuyen Richards joins the desk following an internal move. Meanwhile, Aneta McGuinness has left to pursue interests outside the business.

...and an addition in CDOs
Pier Cortial has recently joined UBS from Bank of America as a cash CDO structurer. He reports to Simon Perry, head of European CDOs at the bank.

Pezella goes to Goldman
Greg Pezella has left the Cash CDO structuring desk at Bear to join Norman Hardie's team at Goldman Sachs in a similar role.

Pacific Life increases structured credit focus
Pacific Life Insurance Company has formed a new asset management affiliate – Pacific Asset Management – which will focus on structured credit. Pacific Asset Management plans to leverage Pacific Life's existing expertise in this area and to become a leading third-party asset manager for a variety of structured transactions.

Future structured transactions will involve investment grade and high yield corporate bonds, leveraged loans and credit derivatives. Pacific Life will provide legal, operational and technology support, as well as capital commitments that have been identified as critical to the success of this venture.

Rex Olson will head up Pacific Asset Management and will be responsible for branding the enterprise, building the subsidiary from the ground up, and creating a team that can distinguish itself in a very competitive market. As Pacific Life's head of credit research, Olson has been responsible for overseeing high grade, high yield, private placement investments and leveraged loans. Prior to joining Pacific Life, he held portfolio management and analyst positions at GE Capital and Transamerica.

Olson will be joined by other Pacific Life investment professionals including J.P. Leasure, David Weismiller and Brian Robertson. The group has worked together at Pacific Life for over seven years.

BoA promotes White
Bank of America has named Thomas White head of global markets. White will be based in New York and will report to Gene Taylor, vice chairman of Bank of America and president of global corporate and investment banking.

White, currently global head of credit products, joined Bank of America in 1994 as a managing director in high yield trading. He replaces Mark Werner, who BoA says will be leaving the company to pursue other opportunities.

Netherlands sales moves
Amwedhkar Jethu, structured credit sales to Holland, is believed to have left Calyon, to reportedly replace Linda Koster at Merrill Lynch, who is believed to have left to join Goldman Sachs.

Husk becomes Point Nine chairman
Point Nine Financial Technologies has appointed Steve Husk as its chairman. Husk has a solid track record of growing companies from an early stage in the financial services sector.

He has held a number of senior management positions in the financial services sector and was president of SunGard Trading and Risk for five years. Point Nine supplies software and services to the buy and sell side for the pricing and risk management of complex cross-asset portfolios, particularly in the area of credit derivatives.

Wong joins Algo
Kok-Lung Wong has been appointed to lead Algorithmics' Algo Risk Service, a hosted portfolio construction and risk management offering targeted to the buy-side. He was previously the founding chief technology officer of Cube Route, a managed service provider of wireless logistics software.

Wong has also held the role of vp of research and development at Workbrain, where he was responsible for both traditional enterprise software delivery for large clients and the managed service offering for SMEs. Prior to that he was svp of the investment management and technology division of Putnam Investments, where he was responsible for all risk management systems, including enterprise risk management and portfolio construction risk analysis.

MP

Portfolio CDS trading platform launches

A round up of this week's structured credit news

Portfolio CDS trading platform launches
Q-WIXX, an electronic platform for trading large portfolios of single-name CDS (see SCI issues 20 and 23), has announced the official launch of its automated service for the global market, with broad support from major market participants.

Q-WIXX is set up to assist dealers, correlation desks, prop desks, hedge funds and other buy-side market participants to achieve improved pricing and processing efficiencies when executing OWICs or BWICs. Distributed to participating bidders via a network used actively by more than 1,000 CDS traders at the top 20 major dealers in the US and Europe, Q-WIXX allows for competitive pricing by major dealers and also provides clients with live prices, making it easier to execute trades while reducing communication and execution time.

The platform says that trades that typically took hours can be done in a matter of minutes, with reduced market risk for both the buy-side and sell-side and significantly increased assurance of execution. Through its connection to T-Zero, Q-WIXX also aims to provide significant post-trade benefits. Post-trade processing of lists is particularly cumbersome because the bulk nature of list trading currently results in very high levels of operational activity, which can dramatically increase the potential for error and backlogs in trade booking.

Major credit derivative dealers and liquidity providers have already joined the platform, signalling strong market interest and a rapid rate of adoption. Dealers include BNP Paribas, Deutsche Bank, Goldman Sachs, HSBC, JP Morgan, Merrill Lynch and UBS. Four additional dealers are in the process of joining the Q-WIXX platform.

"This is market-changing technology," says Stephen Grady, global head of trading at Fortis Investments. "Q-WIXX list execution, combined with T-Zero post-trade processing, provides buy-side institutions, such as ourselves, with the ideal end-to-end solution."

Natixis and AXA bring innovative motor synthetic
Natixis has been mandated by AXA to be sole bookrunner on its second insurance linked securitisation – Sparc Europe. The €450.8m transaction is similar to a synthetic securitisation and, unlike its predecessor, offers investors geographic diversification.

The deal offers four classes of notes, which are rated by S&P/Fitch as follows: class A €91.5m triple-A 4-year WAL; class B €220m A/A+ 4-year WAL; class C €100.1m BBB-/BBB 4-year WAL; and class D €39.2m BB/BB- 1-year WAL.

The deal is based on four portfolios of motor insurance policies that reference some 6 million vehicles through Germany (30% of the portfolio), Spain (31%), Italy (21%) and Belgium (18%), representing €2.6bn of premium income. The transaction, structured through the reinsurance subsidiary of Natixis and the FCC SPARC Europe (a mutual debt securitised fund), transfers to the capital markets €450.8m worth of loss risk for a period of four years.

The deal is roadshowing this week and next and is expected to price later in June. This new transaction follows on from the November 2005 SPARC France transaction carried out by IXIS Corporate & Investment Bank, a subsidiary of Natixis for AXA France IARD.

Hedge funds transform credit markets
Fitch Ratings says in new research that the influence of hedge funds on key segments of the credit markets continues to grow at a dramatic pace, transforming both the cash and CDS markets. Credit-oriented hedge fund assets are reported to have reached over US$300bn in 2005, a six-fold increase from five years ago, according to the IMF.

Fitch says this number excludes the 'multiplier effect' of leverage and, therefore, understates the real amount of credit risk taken and the impact on trading volumes. US$300bn of assets equates to US$1.5-1.8tr of assets deployed into the credit markets at typical leverage levels of 5-6x.

The growing role of hedge funds in the credit markets represents a true paradigm change and is evidenced in a number of ways, Fitch says. Hedge funds now account for nearly 60% of the trading volumes in the US$30tr credit default swap market, and provide significant capital flows to all areas of the cash credit markets – particularly more levered, subordinated risk exposures in pursuit of higher returns.

Given the importance of hedge funds to market liquidity and the implications of a forced selling scenario, Fitch believes liquidity risk is among the more important issues facing credit investors in the near-term. The inherent instability of hedge funds as an investor class – arising in large part from their reliance on short-term, margin-based leverage – is distinctly different from more traditional buy-and-hold institutional investors and relationship-oriented bank lenders, the agency warns.

In a market downturn, the potential for a forced unwind of credit assets can not be discounted, which in turn could lead to correlations that are different than historical expectations. For example, Amaranth was reported to have sold leveraged loans and RMBS to meet margin calls on its natural gas positions. During a period of market stress, any such forced selling of assets would be magnified by the effects of leverage.

Rare Italian synthetic marketed
Merrill Lynch and Deutsche Bank are marketing an Italian synthetic RMBS deal – the €69.6m PB Domicilio 2007-1 Ltd. The structure – which is unusual in Italy and the first of its kind rated by S&P – is similar to the German PB Domicile 2006-1 transaction, including the use of a synthetic excess spread mechanism.

The excess spread is set at 50bp a year on the performing balance of the portfolio and will accumulate until it reaches a cap of €9.2m. It will be available to cover current losses, to cure losses that resulted in note write-downs in previous periods, and to cure interest payments that were missed due to those write-downs. Because this mechanism has been introduced, there will be no additional cash reserve or unrated first-loss piece to provide further subordination.

The deal transfers the credit risk associated with a pool of 17,293 residential mortgage loans that are currently on the balance sheet of BHW Bausparkasse. The portfolio has a weighted average LTV of 57.43% and 34-month seasoning.

The deal offers six classes of rated notes, as follows: €500,000 triple-A rated Class A1+ notes; €500,000 triple-A Class A2+ notes; €38.3m double-A Class Bs; €13.4m single-A Class Cs; €11.2m triple-B Class Ds; and €5.7m double-B Class Es.

The issuance proceeds from the Class A+ to C notes are expected to be invested in pfandbriefe issued by Eurohypo, while those for the class D and E notes are expected to be invested in schuldverschreibungen issued by Deutsche Postbank.

Q1 sees only one synthetic CDO credit event
Only one new credit event was called on corporate obligors in the synthetic CDOs rated by S&P, according to its latest report for the sector.

The limited number of credit events triggered in synthetic CDOs is testament to an ongoing benign corporate credit environment. This is demonstrated by the decline in global corporate default rates since the peak in 2002, S&P says.

Since the second half of 2003, no more than three corporate obligors referenced in synthetic CDOs have triggered credit events in any single quarter. This contrasts, for example, with the more distressed period during 2002. The first quarter of 2007 was no exception, with the sole credit event recorded on Port Townsend Paper Corp.

Swiss Re issues first MedQuake notes
Swiss Re has issued the first notes under its newly established shelf programme, MedQuake Ltd. The double-B minus US$50m series 1 class A and the single-B US$50m series 1 class B principal at-risk variable-rate notes were both rated by S&P.

Through this transaction, Swiss Re, as the counterparty, has bought fully collateralised retrocession protection against high severity losses incurred from earthquakes in Greece, Turkey, Cyprus, Israel and Portugal. This is the first catastrophe bond rated by S&P that covers earthquake risks in these countries.

The issuer is a special-purpose Cayman Islands-exempted company, whose ordinary shares are held in a charitable trust. Under the programme, Swiss Re can issue additional series which may contain various classes of notes. There is no limit on the amount Swiss Re can issue under the programme.

The term of the risk period and the countries covered against earthquake events may vary between series. The first class A and B notes are due on 31 May 2010.

MedQuake invested the US$100m issuance proceeds in high-quality assets, which were placed in a collateral account. MedQuake swaps the total return of the assets with Swiss Re Financial Products Corp. in exchange for quarterly Libor-based payments.

Moody's rates Channel
Moody's Investors Service has assigned provisional ratings to Channel Capital Plc, which is the first all-European credit derivative product company rated by the agency. It has assigned a counterparty rating of (P)Aaa; and ratings of (P)Aaa to Channel's class A term and auction notes; (P)Aa2 to its class B notes; and (P)Baa2 to its senior capital notes.

Channel is incorporated in Ireland as a public company with limited liability under the Companies Act 1963-2005 of Ireland. The CDPC will invest in a diversified portfolio of eligible corporate credits and sovereigns using bespoke tranched synthetics and, in certain circumstances, single name credit default swaps. It will initially be capitalised through the issuance of notes and contributions by sponsors.

Moody's believes that the financial resources of Channel have been sized to assure that the expected loss faced by a credit default swap counterparty is consistent with the experience of investors in Aaa-rated instruments and that holders of debt obligations issued by Channel face the expected loss consistent with instruments of comparable ratings.

Moody's ratings rely on Channel's activities becoming restricted following certain trigger events, including capital adequacy tests based on a capital model. Channel's capital model examines the impacts of credit events, potential counterparty defaults, and changes in market rates on the calculation of required resources. In addition to the required resources suggested by the capital model, a buffer will be added to reflect non-quantifiable sources of risk such as operational failures.

Deutsche brings jumbo CDO
Deutsche Bank has announced the execution of one of the largest ever transactions in the CDO space with a landmark synthetic credit transfer deal between UniCredit Group (HVB) and Deutsche Bank.

The €2.2bn mezzanine and equity CDO transaction involved divesting the bespoke CDO portfolio of UniCredit Group (HVB) managed by its active credit portfolio management unit in line with the UniCredit Group's strategy to develop its investment banking business in a dedicated global division. The underlying risk was predominantly investment grade credit in the mezzanine and equity parts of the capital structure and resulted in a transfer of approximately €8bn of credit risk.

Alex Bernand, global head of credit correlation at Deutsche Bank comments: "The transaction was executed in a highly competitive environment and the coordination between Deutsche Bank and UniCredit Group (HVB) resulted in a seamless execution, without market moving significance."

MP

Trading ideas - stepping out

Tim Backshall, chief credit derivatives strategist at Credit Derivatives Research, looks at a curve steepener on RESCAP

The strong technical bid that has been evident recently in many of the homebuilders and lenders seems driven more by the spread-per-rating (or spread-per-WARF) attractiveness than by any underlying fundamental beliefs about recovery. As views on interest rates swing almost daily from cut to hike and back, the one unsettling truth remains that the US housing market continues to struggle, and with it the mortgage providers.

Delinquencies and foreclosures trend up and the while the consumer remains buoyant (based on this season's earnings), there is a clear trend down in EPS growth and further the amount of refinances and mortgage applications is dropping just as consumer debt is rising at record rates (as we switch from our HELOC-based ATM to the credit cards). RESCAP is one credit that even its parent company GMAC has concerns with but the recent debt offering was snapped up by investors, betting that the sub-prime bad news is behind us and that coupon step-ups will provide enough compensation in the event of downgrade.

When we hold a deteriorating fundamental outlook on a name, the positive carry curve steepener is our trade of choice as it pays us to be short the credit. Unfortunately, there have been very few positively-economic curve steepeners available recently. Here, we find a curve misaligned with our fundamental view of the credit and too flat to our estimate of fair value.

RESCAP has rallied recently but the differential between its 5 and 10 year levels remains flat to fair value. This appears to be due more to 5Y tightening (buying by structured credit) than 10Y widening and we believe the difference between the two tenors should widen significantly. There is also considerable uncertainty surrounding the CDS curve correct levels due to the step-up coupons in the legacy and recently issued bonds.

We take a quick-and-dirty view of the relative-value between CDS and the bonds and derive a simple framework to judge the market's expectations of downgrade which we find too low. Given the positive economics of the trade (carry and roll-down), curve flat to fair value, and CDS richness in 10Y, we believe the 5s-10s steepener is the best way to express our deteriorating fundamental outlook on RESCAP, and while concerns over tightening in long-dated spreads driven by the technical bid remain, we feel that the inevitable downgrade will stymie any ratings arbitrage very quickly.

How flat is too flat?
To estimate the fair value of the 5s-10s steepener, we model the 10 year offer level as a linear combination of the 5 year bid levels. This is undertaken across the universe of credits whose 5's are close in value to RESCAP. For each issuer, we produce an expectation of the ten-year CDS offer level. This is compared to the current market ten-year CDS bid level in Exhibit 1. Values that lie below the black line are flatter than fair value and values that lie above the line are too steep. The red square indicates RESCAP's current ten-year market level vs. fair value, which lies almost 50bp below fair value.

Exhibit 1

Differentiated differentials
RESCAP has excellent liquidity in the CDS market across the curve and now trades actively in five-, ten-, three- and seven-year maturities. Bid-offer differentials and liquidity are acceptable in the 5's and 10's. The curve differential (between five- and ten-year CDS) has risen in the very recent past, as seen in Exhibit 2, but has seen some volatility over the last 18 months.

Exhibit 2

The jump wider and steeper in Feb 2007 followed a market-wide rally in the fall of 2006 and modest sell-off through January as housing fears slopped into lenders. Notably the trend in the curve was steeper as spreads rallied in 5Y during the fall of last year and once again in the most recent past, the curve has steepened as 5Y has tightened. This behaviour seems more driven by the relative anchoring of the 10Y maturity and bid on 5Y than anything fundamental.

Mid-to-mid absolute differentials rose as spreads jumped and we saw the curve starting to steepen appreciably, followed by the flattening as credit rallied with equity, leaving the current differential at about 31bp. This differential has made the 5s-10s steepener a significantly positive carry trade. At present, RESCAP is in the middle of its differential range but not nearly as wide as we would expect.

In a slight departure from our normal analysis, Exhibit 3 shows the curve's behaviour through time relative to the 5Y CDS. This plot is typically a simple scatter-plot but we join the points in an effort to show the interesting cyclical movements in the curve relative to the spreads. The dark blue arrow shows the pre-Feb 2007 behaviour as the curve steepened (moved up the y-axis) consistently while spreads rallied (moving left) and then sold off (moving right) ending at relatively steep levels. Then the jump in Feb 2007 came and the light blue arrow shows the curve began to move flatter (after a stable period) while spreads initially widened and then tightened. The most recent move back tighter has seen the curve (red square) shifting back steeper once more.

Exhibit 3

Notably, this somewhat cyclical theme shows us that the moves in the curve are more driven by the 5Y and that there has been somewhat of a regime shift since Feb 2007 with spreads range-bound but the curve steeper. We feel that any downgrade will remove any bid pressure and leave the 5 and 10Y to float more directly with fundamental views.

Step-up thoughts
The RESCAP curve clearly trades too flat, even if one considers the likelihood of a recovery in housing then implied forward rates remain too low. It is likely that the step-up coupon features of the longer-dated bonds is responsible for much of this CDS curve flatness – maintaining an artificially low spread due to the likely jump in coupon on the downgrade, which would subsidise any default risk increase expectation.

In a perfect world, we would construct a complex ratings-migration based model of the step-up coupon bond and generate fair-values for the bond based on the current CDS curve. This process would allow us to judge whether the CDS curve is really cheap (flat) to fair-value. We prefer to take a shortcut and look for a simple yet still practical solution using market prices (rather than relying on a historically based migration matrix).

The approach is straightforward in construction. We will take the two liquid step-up coupon bonds (the 6.875% of 2015s and 6% of 2011s) and derive a series of CDS-implied Bond Prices (CDSiBP) fair-value levels for the bonds at various levels of coupon (based on the prospectus' ratings-based tables). Once completed, we can take current market prices for the bonds and solve for the market's expectation of downgrade which allows us to judge whether or not we agree.

The coupons step up by 50, 100, 150, and 200bp based on one or 'two-or-more' notch downgrades by one or 'two-or-more' agencies. There is also the interesting feature that if RESCAP is upgraded to BBB by all agencies with a stable or improving outlook then the coupon adjustments are removed for the life of the bond (even if RESCAP drops BIG again).

Our CDSiBP model generates fair-values for each potential outcome (ignoring the timing of downgrade). Based on these fair-values, and our assumption that if one agency downgrades then it is highly likely that before the next coupon another agency will follow with a downgrade, we find the market's rough expectation is a 30% chance of downgrade to junk. Furthermore, once in junk, the market seems to be pricing a 50% chance of more than one notch downgrade.

We feel that if the downgrade happened then there would likely be a trigger event. This trigger event will not only increase the cost of funding to the current bonds but will likely place more pressure on RESCAP to issue more bonds to fund the business (especially if rates fall and margins decrease). This will likely cause CDS to widen as the increased debt load will have a noticeable impact on default risk (which to us is more relevant to the synthetic curve than the bond's pricing – consider orphaned CDS at the other extreme still having value based on future issuance).

Further, we feel that the 30% chance of a downgrade is actually too low given our dour outlook for the housing and mortgage markets and RESCAP's lack of diversification from these markets leaves them prone to problems. In order for the market price of the bonds to be correct (assuming a higher chance of downgrade as we do), the CDS curve must be wider at the long end (steeper) to correctly discount the increasing coupon cashflows.

For these reasons we feel that CDS will eventually disconnect from the bond pricing as default risk rises and although the potential arbitrage is there with the CDS deliverables, once RESCAP is junk, it will likely be forced to issue more unsecured debt that potentially does not contain step-ups/downs. This disconnection is the main concern with the steepener but we have time to wait due to the attractive economics of the trade.

The real deal
While we maintain a deteriorating fundamental outlook for RESCAP, any negative view could obviously be implemented with an outright short, but, as we have stated before we prefer positive carry trades (and relative-value opportunities) that are somewhat market neutral (preferring to avoid the direct path of the technical bid as much as possible). We still prefer the steepener as our positively economic short trade (especially when our market bias is to the widening and steepening side) when we are concerned on downside and so investigating the 5s-10s relationship provides us with both relative-value and positive carry.

One more item of note with the steepener is that it buys us time (at these maturities) – if we are modestly negative on the credit but do not expect a major move soon, it does not cost us to maintain a position that benefits from over-performance. The position provides us with a cushion on our timing of any negative view and offers good upside from just rolling down the curve (if nothing happens).

Moving from the mid-to-mid (rough) curves to the bid-offer adjusted curves in Exhibit 4 paints a much clearer picture of where RESCAP's curve is trading. The differential between the standalone levels and the peer group's fair value points to an interesting divergence in 5s-10s.

Exhibit 4

The differential has been consistently negative over the past 18 months but has jumped considerably more negative since February 2007, with a slight trend narrower in the last couple of weeks. In the case of a steepener, a negative differential indicates the curve is too flat.

Since the end of February, our fair value model has indicated the RESCAP curve is too flat when compared with its peers, and given our views on fundamental deterioration leading to increased risk, relative richness of the longer-dated CDS curve to bonds, and attractive economics, the exit of the legacy steepener and initiation of the OnTR 5s-10s steepener remains a core short position for the housing/mortgage market concerns.

Risk analysis
This trade is duration-weighted to ensure positive carry as well as to reduce our exposure to absolute levels. We are therefore hedged against short-term movements in absolute spread levels, profiting only from a curve steepening between the fives and tens.

The carry cushion protects the investor from any short-term mark-to-market losses. This trade has positive roll-down thanks to the curve shape and tightness of bid-offer.

Entering and exiting any trade in these maturities carries execution risk, but this is not a major risk with this credit in these maturities as they are increasingly liquid.

Liquidity
Liquidity is a major driver of any trade – i.e. the ability to transact effectively across the bid-offer spread in the CDS markets. Our data on liquidity, created from the volume of bids, offers, and trades we see each day, provide us with significant comfort in both the ability to enter a trade in RESCAP and the bid-offer spread costs.

RESCAP has good CDS liquidity and bid-offer spreads have narrowed to around five and ten basis points in five- and ten-year CDS respectively. While we would like to see bid-offer spreads lower in 10Y, we believe we can still cross these levels and have a profitable trade.

Fundamentals
This trade is significantly impacted by the fundamentals. The technical flatness of the credit and positive economics of the trade are helped by the fundamentally deteriorating outlook for RESCAP.

Summary and trade recommendation
The housing and mortgage markets remain moribund as applications drop, delinquencies rise, foreclosures increase, and home prices and sales plummet. RESCAP has faced much of the selling and the jump in spreads in Feb07, due to the sub-prime concerns, pushed its CDS curve even flatter to fair-value.

The significant impact of the technical bid recently in many lenders and homebuilders (based on attractive spread-per-rating/WARF factors) has pushed 5Y spreads too tight and the curve even flatter in RESCAP and we reinstate our seasoned curve steepener in the on-the-run CDS with more attractive economics. The step-up coupon bonds help to maintain this artificial flatness but based on our analysis we find the expectations of downgrade are too low and that longer-dated CDS are rich relative to the bonds.

Given the relative flatness of RESCAP's 5s-10s CDS curve, we believe a positive-carry steepener trade is the safest way to express our deteriorating fundamental outlook on the company (based not only on mortgage weakness but in the Business Capital segment, which specialises in construction loans and real estate investments). Significant carry, modest roll-down, and the potential profit of a return to fair value strengthen the economics of the trade.

Sell US$10m notional Residential Capital LLC. 5 Year CDS protection at 144bp and

Buy US$6m notional Residential Capital LLC. 10 Year CDS protection at 184bp to gain 34bp of positive carry

For more information and regular updates on this trade idea go to: www.creditresearch.com

Copyright © 2007 Credit Derivatives Research LLC. All Rights Reserved.

Note: This article is intended for general information and use and does not constitute trading advice from Structured Credit Investor (see also terms and conditions, below, section 12).

Synthetic CDOs: made to measure - part 2

The second in a two part guide to valuing bespoke and non-standard tranches by Richard Huddart and Domenico Picone of the structured credit research team at Dresdner Kleinwort

Moving on from part 1, we can now think about how to use this 'liquid' market base correlation information to price a tranche of a bespoke portfolio. In this section we shall assume that we are aiming to price a regular bespoke single tranche CDO at a given maturity M, such that we can re-define our problem as: given an (interpolated) base correlation skew in the liquid market at maturity M, how do we use this information to price our bespoke tranche?

We note that this still leaves us with the problem of deciding what correlation assumptions to use at non-standard maturities. Whilst we do not touch this issue here, we aim to discuss it in a forthcoming report, where we will look at term-structure approaches to correlation modelling. All of the mappings we explain in this paper can be adapted to incorporate the term-structure of base correlation¹.

'Equivalent' tranches, same correlation
Remembering that any bespoke tranche can be priced as the difference between two equity tranches (tranche(K_L,K_U) = tranche(0,K_U) - tranche(0,K_L)), we focus on pricing (0,K) equity tranches. Our problem essentially reduces to the following:

The rationale behind this is as follows:

? The base correlation 'skew' exists because the standard Gaussian Copula model cannot fully describe the expected loss profile implied by index tranche spreads, although its shape is due to a certain extent to market supply and demand preferences, or 'technicals'.
? Tranches in related markets that are considered to be 'equivalent' should then, in theory, be subject to similar technical pressures.
? As a result, (0,K) tranches of two different portfolios in the same market should be priced with the same correlation assumption when they are 'equivalent' in some relative sense.

The mappings
The market is yet to agree on what is the 'correct' way to transform correlation assumptions from standardised to bespoke pools, although here we discuss the following common mappings in the base correlation framework:

? "Moneyness Matching": Mapping base correlation by portfolio/index expected loss
? "Tranche EL Matching": Mapping base correlation by tranche expected loss
? "Fair Spread Matching": Mapping base correlation by tranche fair spread
? "Correlation Risk Matching": Mapping base correlation by tranche correlation sensitivity

Each mapping uses a different criterion in order to decide on which (0,K) tranches are 'equivalent' across portfolios, and then prices a bespoke equity tranche with the base correlation of the 'equivalent' equity tranche in the relevant liquid market. We demonstrate the different approaches by pricing the same bespoke tranche under each mapping.

Single- vs multiple-market tranches
A final point to note before we begin is that a bespoke portfolio can be considered as being either from a single market or multiple markets. An example of a single market tranche may be a basket of European investment grade corporate names or an off-the-run iTraxx Europe tranche, in which cases the obvious liquid market to would be the on-the-run iTraxx Europe series. A CDO based on a basket of European and North American investment grade names on the other hand may be considered as a multiple market tranche, and in such a case we would look to utilise correlation information from both the iTraxx Europe and CDX.NA.IG liquid tranche markets.

We focus our discussion on mapping base correlations to price single market tranches, however we give an overview of how to apply these mappings to multiple market tranches.
Finally, we note that depending on the mapping applied, it cannot necessarily be guaranteed that a solution will be unique or arbitrage free.

"Moneyness Matching"
"Moneyness matching" is the simplest mapping to have emerged for pricing bespokes, and directly maps the base correlation skew based on the width of the (0,K) tranche relative to the index (or portfolio) expected loss (EL). Going back to our rationale that 'equivalent' (equity) tranches between baskets should be priced with the same base correlation assumption, it then follows that (0,K) tranches are considered 'equivalent' when their ratios of:

Tranche detachment point
Index (or portfolio) EL

are the same across portfolios. This mapping then assumes that the relative 'riskiness' of a (0,K) tranche can be summarised by the ratio of its detachment point to the portfolio EL. The base correlation to be used to price the two 'equivalent' tranches is given by the liquid market base correlation skew.

Remembering that

PV(6%, 10%) = PV(0%, 10%) - PV(0%, 6%)

we focus our attention on pricing tranche (0%, 6%) without loss of generality. The 'moneyness matching' criterion thus suggests that the (0%, K_iTraxx%) tranche that is 'equivalent' to our (0%,6%) bespoke solves the relationship:

As of the pricing date, the 7y iTraxx EL was 2.23%, whereas our bespoke pool EL was 2.99%, meaning that:

Our (0%, 6%) tranche is thus 'equivalent' in a moneyness matching sense to a (0%, 4.474%) 7y iTraxx tranche, and so we should use the iTraxx base correlation relating to detachment 4.474% (namely 16.52%) to price the (0%, 6%) leg of our bespoke.

The simple nature of this mapping means that if we have a base correlation skew for our relevant 'reference' liquid market (assuming a single market bespoke portfolio) then we can apply this shift to the whole base correlation skew to obtain a new skew. This new skew can then be used to price any tranches of our bespoke portfolio. This mapping of the entire skew is shown for our example in the chart below, where we assume for the moment simple linear interpolation of the skew between liquid detachment points³:

This is a constant relative shifting of detachment points across the capital structure. It should be clear that what we are adjusting here is the detachment point at which a given correlation assumption is applied – we do not scale the correlations themselves!

Applying the same mapping to a (0%, 10%) bespoke tranche and using our results to price our (6%, 10%) bespoke, we find that our bespoke is 'equivalent' to a (4.474%, 7.457%) 7y iTraxx tranche, whose fair spread is 51bp.

Doesn't capture the portfolio loss distribution
"Moneyness matching" is a simple-to-apply technique which attempts to make tranches from difference portfolios 'more comparable'. It is in a sense, however, a 'macro' mapping, in that the same (relative) transformation is applied regardless of where the tranches of interest sit in the capital structure. In moving on from this we may then look to apply a mapping which captures the shape of the loss distribution, such that more 'tranche specific' risk attributes can be used to decide on which tranches are 'equivalent'. The remaining mappings we consider in this report all capture the shape of the loss distribution in some way.

"Tranche EL Matching"
Different equity tranches, same EL, same correlation
The theory behind tranche EL matching says that two equity tranches from different portfolios should be considered 'equivalent' when they have the same tranche expected loss (EL). Equity tranches with the same EL should then be priced with the same base correlation assumption (dictated by the liquid market base correlation skew).

For example, if we want to price the (0%, 6%) equity tranche of our bespoke portfolio then we need to find the corresponding [0%, K_iTraxx%] tranche of 7y iTraxx such that:

Which liquid tranche has the same EL as our bespoke?
Of course, this is correlation trading and so the EL of any given tranche is a function of the correlation assumption used to price that tranche. Now this presents us with a situation where our aim is to find the most appropriate correlation assumption with which to price a (0,K) tranche, however in order to do this we need to calculate the tranche EL, which in turn requires a correlation input – precisely the thing we are trying to find! This means that we must solve for K*_iTraxx using a recursive procedure.

Given an initial guess forK*_iTraxx, one iteration of the recursive process would be as follows:

The recursive procedure stops when we find K*_iTraxx such that f(K*_iTraxx)= 0 (to within a pre-defined level of tolerance), and a solution can be found using a standard numerical method such as interval bisection or Newton-Raphson.

Example: pricing our 7y bespoke tranche
The following table shows the convergence of the recursive process we used to find which 7y iTraxx tranche (0%, K*_iTraxx%) has the same EL as our (0%, 6%) bespoke when priced with the iTraxx base correlation assumption corresponding to detachment K*_iTraxx%. We used an interval bisection approach with an initial estimation interval of (0%, 20%) for K*_iTraxx.

As can be seen, the tranche EL matching approach suggests that our tranche is 'equivalent' to a (0%, 4.908%) 7y iTraxx tranche, and so should be priced with the iTraxx base correlation assumption relating to detachment 4.908% (namely 17.88%).

In sum, our (6%, 10%) bespoke tranche of interest is considered 'equivalent' to the (4.908%, 8.093%) tranche of 7y iTraxx in a 'tranche EL matching' sense and as a result has a fair spread of 53bp.

"Fair Spread Matching"

A similar approach would be to match tranches based on tranche fair-spread (FS). The fair-spread essentially contains the same information as the expected loss, although it accounts for duration too. To see this, we note that a tranche is priced 'fairly' when:

Our approach to finding which liquid tranche (0%, K_iTraxx%) best corresponds to our bespoke equity tranche in this case, and hence which correlation assumption  we need to use to price our bespoke tranche (0%, K_bespoke%), then requires us to find such that:

where

denotes the fair spread of tranche [0%, K%] when priced with correlation assumption ñ. To find this solution, we apply the same recursion as with the tranche EL approach, but replacing our objective function in step 4. by:

Example: pricing our 7y bespoke tranche
As with tranche EL matching, we show how our recursive procedure converges to decide which (0%, K_bespoke%) tranche has the same fair spread as the (0%, 6%) leg of our bespoke in the following table:

Implementation of the fair spread matching method suggests that our (6%, 10%) bespoke tranche is 'equivalent' to a (4.856%, 7.915%) tranche of 7y iTraxx, and as a result it should have a fair spread of 54bp.

"Correlation Risk Matching"
"Correlation risk matching" is a slightly different approach that suggests two (0,K) tranches from different portfolios should be considered 'equivalent' when they have the same exposure to correlation risk (shifts in the tranche correlation assumption). Tranches with the same correlation sensitivity should then be priced with the same correlation assumption.

? For any given (0,K) iTraxx tranche, we know from the market base correlation skew what correlation assumption to use.
? If our chosen iTraxx equity tranche is 'equivalent' to our bespoke tranche in the sense that it has the same correlation sensitivity, then we should also use this correlation assumption to price the bespoke. But, how do we determine which iTraxx tranche has the same correlation sensitivity as our bespoke?

Which iTraxx tranche has the same correlation sensitivity as our bespoke?
We will now illustrate the process by which we can compare the correlation sensitivities of two tranches from two different baskets of 125 credits, the 7y iTraxx index and our bespoke basket of CDS with all underlying spreads at 40bp, each over the same 7y horizon.

For simplicity we will take the 0-3% tranche of 7y iTraxx and find the equity tranche of our bespoke pool that has the same correlation sensitivity.

The following chart shows how the PV of the 0-3% tranche varies with the correlation assumption used to price it⁶ - remember that equity tranches are long in correlation and so this is an increasing function. The PV of the trade is zero when correlation is 11.9%, which corresponds to the market implied base correlation at 3% detachment.

If we define tranche correlation sensitivity as the change in PV resulting from a 1% increase in the base correlation assumption, then the correlation sensitivity of the 0-3% tranche is given by:

We now need to find the corresponding (0,K) tranche of our bespoke portfolio which has the same correlation sensitivity (namely 0.78%) when priced with a base correlation assumption of 11.9% (this is important – remember that 'similar' tranches are priced with the same correlation assumption, and so to find the tranche 'similar' to 0-3% iTraxx we need to use the same correlation assumption as the iTraxx tranche).

To do this we can now plot tranche correlation sensitivity, defined as the PV gain when correlation increases from 11.9% to 12.9%, for (0,K) tranches of our bespoke portfolio, against tranche detachment point K. We also superimpose upon this chart the correlation sensitivity of the 0-3% iTraxx tranche (=0.78% as just calculated). The point K at which these lines cross tells us the (0,K) tranche of our bespoke portfolio which is 'equivalent' from a correlation sensitivity point of view to the 0-3% 7y iTraxx tranche. In this case, we note that they cross when K=2.8% and so we would price a (0%, 2.8%) equity tranche of our bespoke with a correlation assumption of 11.9%.

If we in fact wanted to price the (0%, 6%) tranche of our bespoke, then we would need to look for a more suitable, wider equity tranche of 7y iTraxx to map to. This means that we have to repeat this procedure recursively until we find an iTraxx equity tranche with the same correlation sensitivity as our bespoke when priced with the base correlation assumption from the iTraxx market.
Given an initial estimate K*_iTraxx for the detachment point of the (0,K) iTraxx tranche that most closely matches our bespoke tranche (0%, K_bespoke%), one iteration of the recursive process to be used would be as follows:

The recursive procedure stops when we find K*_iTraxx such that f(K*_iTraxx)= 0 (to within a pre-defined level of tolerance).

Example: pricing our 7y bespoke tranche
In the following table, we once again illustrate the convergence of this iterative procedure to find the K*_iTraxx such that tranches (0%, K*_iTraxx%) and (0%, 6%) of our bespoke portfolio have the same correlation sensitivity.

Applying the correlation risk matching method suggests that our (0%, 6%) bespoke tranche is 'equivalent' to a (4.785%, 7.969%) tranche of 7y iTraxx. When priced with the corresponding iTraxx base correlations, this gives us a bespoke tranche fair spread of 52bp.

The results: a comparison
In the charts on this page we show the numerical results achieved for pricing (0,K) tranches of our bespoke portfolio under the different mapping techniques. In the first case we assume that we are pricing (0,K) tranches of a 7y, 40bp homogeneous portfolio, and in the second case we assume a riskier 60bp homogenous pool. In both cases we map from the 7y iTraxx skew, which traded with an index spread of 31bp.

As can be seen, all of the mappings produce similar results when pricing tranches of the 40bp portfolio, and importantly they all infer a base correlation skew that is lower than if we were just to apply the iTraxx skew directly to our bespoke portfolio (this is due to the bespoke pool being riskier than iTraxx). This is reflected in the fact that mapped (0,K) bespoke tranche spreads are generally higher than when an unmapped skew is applied.

This effect is more pronounced when we move to the 60bp portfolio, although this is unsurprising given that the bespoke is now significantly more different from the iTraxx pool. We further note that the correlation skew inferred from the correlation risk matching method is now slightly different from the other three mappings.

A real example: off-the run index tranches
A situation in which one may wish to apply a single-market mapping would be when pricing off-the-run index tranches. Here we show how the index EL matching approach performed when pricing off-the-run iTraxx S6 and CDX S7 tranches on 16 April 2007, using the on-the-run iTraxx S7 and CDX S8 tranches as liquid prices from which to map correlations. In this case the term-structure of correlation was also incorporated and the 6 months maturity difference was also accounted for.

We note that in most cases, the tranche EL matching approach was able to produce off-the-run tranche prices inside of market bid-offer spreads.

Mapping from multiple-markets
iTraxx and CDX
Whilst the mappings we have seen are capable of reproducing reasonable market prices for bespokes and off-the-run index tranches from inside the same 'single' market, the keen reader may notice that it is not generally possible to obtain reasonable CDX tranche quotes by taking an iTraxx base correlation skew and mapping it to try and reproduce liquid CDX tranche quotes. This is something we can only attribute to underlying differences between these two different markets (Europe and North America), and as a result if we were trying to price a tranche from a bespoke portfolio consisting of both European and US credits, we would look to take information from both markets into account.

Multiple market maps – an outline
We briefly discuss one potential approach to this problem here, based on the "Tranche EL Matching" framework, although it should be clear how to adapt this for use with the other mappings.

When EL matching in a single-market (iTraxx) setting, our goal was essentially to find the liquid tranche (0%, KiTraxx%) that solved:

assuming that we are trying to price a (0%, 6%) bespoke tranche. Both liquid and bespoke tranches would then be priced with the liquid market base correlation p*(K*_iTraxx).

If we wanted to map to both iTraxx and CDX, then we would ideally like to solve both:

to find both the CDX and iTraxx tranches that are 'equivalent' to our bespoke. The problem is however, that p*(K*_iTraxx) and p*(K*_CDX)will, in general, be different. We would therefore not necessarily be any closer to deciding which correlation assumption to use in order to price our (0,K) bespoke tranche. We therefore aim to use a correlation assumption which is a weighted average of CDX and iTraxx implied base correlations and solves the following system of equations:

where c is the proportion of European credits in our bespoke portfolio. The correlation is therefore weighted more towards the region from which the highest proportion of the portfolio comes.

The use of such a multiple-market mapping would require the implementation of a multivariate solver to find a solution to this system (as opposed to the univariate interval bisection/Newton-Raphson approaches we discussed in a single-market setting).

This approach could also be easily generalised to include maps to more than two liquid indices if this was desirable.

Some comments
Whilst the approaches that we have discussed allow one to map base correlations between portfolios based on criteria such as tranche expected loss, tranche fair spread and tranche correlation sensitivity, we note that there are various factors which are not obviously accounted for. However, we do not feel that it is so realistic to be able to control for all of these different factors when mapping correlations between portfolios, and given the strong reliance of tranche prices at different places in the capital structure upon market supply and demand technicals, we think that it makes sense to focus on matching correlations across tranches that may be similarly affected by these technicals, as has been the case in this report.

Some of these factors, such as differences in geographical and industrial diversity among the underlying names, are not generally picked up on by such mappings and are often seen from a fundamental viewpoint as being key drivers of correlation. For example, if we had a portfolio consisting solely of US auto suppliers then we may expect very high correlation indeed. Other questions that need to be approached include:

Dispersion
To what extent does dispersion of spreads in the bespoke portfolio affect pricing?

Sovereign names
How do we deal with Sovereign names, or others from markets with no obvious correlation data to act as a reference?

Number of obligors
How do we account for differences in the number of obligors between a bespoke portfolio and liquid portfolios?

Short buckets etc
What if our bespoke portfolio includes a short bucket, or tranches from other CDOs/securitisations, for example?

Robustness of the mapping
Given that all of the mappings we have considered are to a certain extent ad-hoc transformations, we suggest that it may be worth applying multiple mappings when pricing a bespoke tranche and comparing the outcomes for robustness.

Notes
1. We suggest using some kind of term-structure approach when pricing any tranche (regardless of whether it has a liquid maturity or not), as it would add some consistency to prices across maturities.
2. Or, if mapping to more than one liquid tranche, some combination of the two correlation assumptions – see 'Mapping from multiple-markets' section for more.
3. We refer readers back to the previous section for a discussion of possible interpolation approaches.
4. For consistency, the same copula model should be used to calculate the liquid base correlation skew as to compute expected losses.
5 Remember that when an equity tranche gets wider, its EL (as a proportion of initial tranche notional) decreases.
6. Note we assume here that the tranche coupon is paid on an all-running basis.

© 2007 Dresdner Kleinwort. All Rights Reserved. This Research Note was first published by Dresdner Kleinwort on 14 May 2007.