08 February 2007 Fixed Income Research http://www.credit-suisse.com/researchandanalytics

Credit Derivatives Handbook Credit Strategy Contributors Ira Jersey +1 212 325 4674 [email protected] Alex Makedon +1 212 538 8340 [email protected] David Lee +1 212 325 6693 [email protected]

This is the second edition of our Credit Derivatives Handbook. With the continuous growth of the derivatives market and new participants entering daily, the Handbook has become one of our most requested publications. Our goal is to make this publication as useful and as user friendly as possible, with information to analyze instruments and unique situations arising from market action. Since we first published the Handbook, new innovations have been developed in the credit derivatives market that have gone hand in hand with its exponential growth. New information included in this edition includes CDS Orphaning, Cash Settlement of Single-Name CDS, Variance Swaps, and more. We have broken the information into several convenient sections entitled "Credit Default Swap Products and Evaluation”, “Credit Default Swaptions and Instruments with Optionality”, “Capital Structure Arbitrage”, and “Structure Products: Baskets and Index Tranches.” We hope this publication is useful for those with various levels of experience ranging from novices to long-time practitioners, and we welcome feedback on any topics of interest.

FOR IMPORTANT DISCLOSURE INFORMATION relating to analyst certification, the Firm’s rating system, and potential conflicts of interest regarding issuers that are the subject of this report, please refer to the Disclosure Appendix.

08 February 2007

Credit Default Swap Products and Evaluation Brief Overview Definition of a Credit Default Swap

3

Levered Loan CDS

5

Unwinding a CDS Transaction

5

Unwinding an Off-Market CDS

6

Points Upfront

7

Restructuring Language Differences

7

Cash Settlement for Single-Name CDS

9

Succession Rules

10

Index Products

11

Index Arbitrage

13

Corporate Bond LIBOR Pricing Methodologies

15

Adjusting Spreads on Premium-Priced Bonds

17

Basis Evaluation

18

Repo “Special”

19

Credit Suisse’s Approach to Basis Trading

21

Valuation of Step-Up Bonds

22

Steep CDS Curves Effect on Rolls

23

Steepeners and Flatteners

24

CDS Orphaning and Confusing Name Changes Credit Default Swaptions and Instruments with Optionality Credit Default Swaption Review

25 29 29

Cancelable CDS and Other Exotics

31

Implied vs. Historical

32

Where’s the Skew?

32

Options Embedded in Bonds

33

Constant-Maturity Credit Default Swaps

34

Introducing CDS Variance Swap

35

Recovery Trades. Digital CDS Capital Structure Arbitrage Debt vs. Equity. Potential to Win in the “Wings”

Credit Derivatives Handbook

3 3

37 39 39

The Trickiest Thing in Cap Arb: Hedging

40

Other Capital Arbitrage Flows to Watch

43

Trading Implied Equity vs. Implied CDS Volatility Structured Products: Baskets and Index Tranches Review of Nth to Default

46 47 47

Tranched DJ CDX Products

49

Path Dependency and Tranche Carry

50

Constant Proportional Debt Obligations (CPDOs)

52 2

08 February 2007

Credit Default Swap Products and Evaluation Brief Overview Derivatives are financial instruments that are “derived” from other base financial instruments, such as stocks, bonds, loans, currencies and commodities, and provide investors with a multitude of ways to manage risk. Since the advent of the Credit Derivatives Market over a decade ago, it has grown exponentially as participants and investors have become more sophisticated, driven by both innovation and the need to manage risk and return. Credit Default Swaps (CDS), instruments designed for investors to take or hedge a firm’s credit worthiness and default risk for the bonds and loans of companies, are now one of the main elements driving the growth of the global derivatives market. The notional amount traded in CDS now exceeds the face value of many of their reference entities and the market is often more liquid than the debt markets themselves. Exhibit 1 tracks the growth of the CDS market since 2001.

Exhibit 1: Notional Size of Credit Default Swap Market since 2001 30,000

Notional Amt (US$ billions)

25,000 20,000 15,000 10,000 5,000 0 1H01

2H01

1H02

2H02

1H03

2H03

1H04

2H04

1H05

2H05

1H06

Source: ISDA

Definition of a Credit Default Swap A Credit Default Swap (CDS) is essentially a contract struck between two parties that allows the isolation and transfer of credit risk for bonds and loans. In this agreement, the party that shorts credit risk is the protection buyer and pays fixed periodic payments to the party taking on the credit risk, the protection seller. This transaction provides the protection buyer with credit exposure similar to a short position and the protection seller with exposure similar to a levered long position in the bond. The periodic payments between the protection buyer and seller continue until either the contract expires or the “reference entity,” a third party issuer, experiences a credit event. A credit event is an occurrence where the issuer fails to fulfill its debt obligations, which includes failing to pay its coupon, defaulting, or filing for bankruptcy and, if applicable, restructuring its debt. Upon a reference entity credit event, the protection buyer gives the seller an eligible obligation of the reference entity with a par amount equal to the value of the CDS contract (most times, the buyer pays accrued interest to the date of the credit event). The seller pays the buyer par for these obligations, after which the contract terminates. So, on a credit event there is a transfer from seller to buyer equal to the

Credit Derivatives Handbook

3

08 February 2007

difference between par and the market value of the delivered obligation of the reference entity. Note that it is also possible to have a cash-settled CDS contract, where the seller pays the buyer par minus the market value of the reference obligation on a credit event.1 Due to the payout profile, CDS is commonly thought of as similar to buying insurance on a specific debt issuer. Since the nature of the contract creates residual counterparty credit risk (if the protection seller defaults before the reference entity, the buyer might be unprotected on a reference entity default), margin requirements are commonly used to minimize the counterparty exposure. If we compare the payments on a CDS contract to the payments on a leveraged position on a corporate bond, we see that the exposure is strikingly similar. As the charts below show, a long levered position in a corporate bond receives an annuity equal to the bond’s spread to LIBOR plus the LIBOR rate minus the repo rate. A seller of protection receives the fixed CDS premium (Exhibit 3). On a default, the long levered corporate bond position loses the price paid for the bond minus the recovery value of the defaulted bond. The seller of CDS loses par minus recovery. So, if the repo rate is close to LIBOR and the price paid for the bond is close to par, the payout profiles for a levered position in a bond and a CDS position are very similar.

Exhibit 2: Cash Flows for a Levered Corporate Bond Position Interest rate Swap Coupon

LIBOR + Spread Repo rate

Coupon Corporate bond

Investor Price

Repo Price

Netting the flows, the investor receives LIBOR plus Spread and pays the Repo rate. In case of a credit event, the investor can sell the defaulted bond, pay back the Repo lender and unwind the interest rate swap.

Exhibit 3: Cash Flows for CDS Transactions Par Payment in Credit Event Investor

Counterparty CDS Premium OR Par Payment in Credit Event

Investor

Counterparty CDS Premium

The investor receives/gives the CDS premium. In case of a credit event, must pay/get par in exchange for a defaulted security, which can be sold in the market. Source: Credit Suisse 1

Although not a part of the master agreement yet, cash settlement protocols are currently being considered for inclusion in ISDA agreements. The default by Dura Automotive Systems (DRRA) was the first instance of a cash settlement for single-name CDS and was highlighted in our Trading Edge published on October 16, 2006.

Credit Derivatives Handbook

4

08 February 2007

Levered Loan CDS The recent introduction of loan-only Credit Default Swaps, or LCDS, represents a new innovation in the derivatives market. LCDS are similar to vanilla CDS in most respects, but differ in several important aspects, including the deliverable obligations, the ability to cancel LCDS in certain circumstances, and the relevant obligations in the case of succession events. First, the deliverable obligations for the CDS are a certain class of levered loans, which differ from senior unsecured obligations where both loans and bonds are deliverable. For example, LCDS deliverable obligations include straight loans or revolvers, but not bonds. Seniority of liens are also specified in the contracts, so that a contract is on either First Lien or Second Lien, but typically not both. However, the contracts are typically triggered by a credit event on the loans or senior unsecured bonds. LCDS typically have HY credit event terms, or trade No R; therefore, restructuring is not usually considered a credit event. Second, unlike plain vanilla CDS, LCDS are canceled in the case that there are no longer any outstanding loans for a name. Relevant obligations can change based on which loans are outstanding; however, if there are no loans left to deliver into the contract, the contract ceases to exist after about a month and a half of being orphaned (vanilla CDS continue to exist even after being orphaned). LCDS should price based on this feature, especially on names that only have a limited number of loans outstanding, or have only short-term maturities that may not be renewed. It should be noted that investors long risk via LCDS, have somewhat more limited upside should there be less debt. There are also differences in the case of succession events, which we review under the “Succession Rules” section.

Unwinding a CDS Transaction While a CDS contract has a specified term (most commonly five years), the exposure for most contracts is closed out prior to the contract’s termination date. There are three ways for investors to close out the risk:

Credit Derivatives Handbook

1)

Entering an Offsetting Position: The CDS exposure could be offset with a position in either another CDS contract or in the underlying deliverable obligations. If the offset is in the underlying bonds, the investor will have to hedge out the residual interest rate risk separately. Offsetting with another CDS contract is most likely to result in a spread differential, allowing the investor either to collect or pay an annuity equal to the net premium until the end date of the contract or until there is a credit event (which would cause the investor to pay par for an obligation on one contract and receive par for it on the other contract).

2)

Terminating the Contract: The CDS contract can be terminated by having one counterparty pay the other the present value of the residual exposure. This is determined by figuring out the remaining annuities after an offsetting CDS is entered into and then calculating the PV. The annuity value must be adjusted for the probability that the reference entity does not survive long enough to trigger the next payment because the annuity ends early in the case of a credit event. As a result, calculating the PV of the annuity requires the use of a “risky” discount rate rather than a LIBOR discount. The most common market practice is to use the Bloomberg CDSW function to determine the PV.

3)

Assigning the Contract to Another Dealer: If investors can find a dealer that thinks the contract has a higher PV than its current dealer counterparty, they can assign the contract to that dealer in return for the PV payment. The original dealer counterparty must give permission for assignment because of the counterparty risk present in any CDS contract. Due to this counterparty exposure, we typically only see assignments when the non-dealer in the contract is replaced by another dealer.

5

08 February 2007

Unwinding an Off-Market CDS While normally choosing an unwind is just a matter of preference, the situation becomes a little trickier for contracts with an off-market strike. For example, if an investor sold $10 million of GMAC protection sometime on December 20, 2005 at 425bps, a year later, as of December 18, 2006, it would be up a little over $1.1 million in the money with 4-yr CDS trading at about 105bps. Closing this position by either assigning it to a different dealer or terminating the contract would require the counterparty to make a significant upfront payment. Therefore, the counterparty might not be willing to execute the transaction with such a large intrinsic value at market levels and would prefer to take a ‘haircut.’ For less liquid names, it is sometimes difficult to close the transaction in this manner. In our view, it might be more efficient to offset off-market CDS positions with a new contract struck at current market spreads, based on the same notional amount as the original. For example, an investor that is long GMAC risk, as in the above example, could buy $10 million of a new 4-year CDS contract at 105bps and offset the risk while creating positive cash flows for the duration of the contracts, or until there is a credit event. Investors should keep in mind that there will be no additional cash flows in the case that a credit event occurs before the contract expires. As a result, the total P&L of the trade could be less than projected by the original CDS for investors who purchased protection and unwound by selling CDS at wider levels. Additionally, the inherent duration mismatch of the two positions carries significant mark-to-market implications. For our GMAC example, the original CDS stuck at 425bps would have a DV01 of 3.2, and a year later, the same 4-year contract would have a DV01 of 3.9 as spreads rallied to 105bps. Meanwhile, a new on-the-run 4-year CDS at 105bps would have a DV01 of 3.5. Subsequently, if CDS widened by 100bps, investors would make an additional $35,000 using OTR CDS due to the difference in DV01. Another method of performing an offsetting position of the original off-market CDS is to use a DV01-weighted CDS struck at current market spreads. The main argument here is that due to a mismatch of the notional amounts, the investor is effectively long jump-todefault risk. This trade works better for investors who are concerned about a company’s future. In our GMAC example, the investor would have sold $9.0MM CDS struck at 425bps and unwound the trade by purchasing $10MM in new protection struck at 105bps. The mark-to-market profile of the DV01-weighted position is very stable; even a 75bp parallel shift of the credit curve causes only marginal P&L effects (see Exhibit 4). The changes of the curve shape, however, could have a larger effect.

Exhibit 4: The Effect of a Curve Shift on a DV01-Weighted Offsetting GMAC Position Data as of 12/18/06. Assumed recover of 40% in default. GMAC Spread as of date is 105bps

Instrument

Currently Notional

DV01

GMAC CDS struck @ 105

$10MM

3.5

GMAC CDS struck @ 425

($9MM)

3.9

Spreads Tighten 75bps

Spreads Widen 75bps

Trade’s P&L

Trade’s P&L

$10,526

$14,525

Source: Credit Suisse, Bloomberg

Credit Derivatives Handbook

6

08 February 2007

While using the OTR CDS to hedge, investors are effectively long jump-to-default risk because they would need more OTR CDS to hedge their DV01 of the original CDS struck at 425bps. Investors in such situations would prefer a credit event later rather than sooner. In the best case scenario, if a credit event occurred right before expiration, assuming recovery of 40%, the investor would receive ($10MM – $9.0MM)*(1-0.4) = $600,000 from the CDS default payments and $1,110,000 from the CDS premiums. This equates to a total of $1.71MM compared to about $1.13MM if the investor were to close the position now. To summarize, while normal unwinding of a CDS could be straightforward, situations involving off-market strike contracts could be a little trickier. We think investors who are sitting on significant unrealized gains in their CDS positions are much better off taking a view rather than simply locking in their gains via contract terminations or reassignments. We believe that investors with a more bullish view on a company’s spreads should sell the same notional amount of CDS struck at the current market spreads, effectively establishing a long duration position for the credit, while those with a more bearish view should sell a DV01-weighted amount of protection, effectively going long credit event risk. For a DV01-weighted trade, the later a credit event occurs under the terms of CDS contract, the larger the gains.

Points Upfront For highly distressed credits, trading on spread can become problematic because the P&L distribution of buying or selling protection becomes much wider than the P&L distribution of buying or selling the underlying reference entity (even though the expected values are the same). In order to keep the distributions more in line, when spreads get very wide, the market convention is for the protection seller to require a 500bps premium plus an additional upfront payment from the protection buyer. For example, assume that we have a 5-year LIBOR flat floating-rate note trading at $50 (or a discount margin of 1975bps) with a recovery rate of $40. Assuming a LIBOR funding cost, if the issue pays off at maturity, the investor makes $50, and if the issue defaults, the investor loses $10. Looking at CDS traded on a spread basis on the same credit at a spread of 1975bps, if the reference entity did not default, the protection seller would receive $19.75 for each of the next five years, for a total of $98.75. If the reference entity defaulted, the protection seller would lose $60 minus whatever premiums the seller had collected. With the issue trading at a spread of 1975bps, the possibility of near-term default is high. Therefore, the seller of protection is unlikely to collect the running premium for a long time. So trading on a spread basis would make the P&L distribution much wider (even though the expected value was the same). Instead, if the CDS is traded on a points upfront basis, the seller receives 34 points up front and a running premium of 500bps. So if the reference entity survives for the next five years, the seller collects 25 points of premium over the next five years and 34 points up front for a total of $59. On a default, the seller loses $26 (par minus $40 recovery minus $34 points up front) minus whatever premium he has collected. This P&L distribution is much closer to the $50 upside/$10 downside that the bondholder faces. In our view, shifting to trading with points up front when spreads are wide makes the P&L distribution on CDS a better match for the P&L distribution for the underlying reference entity.

Restructuring Language Differences Due to past episodes in which restructuring has caused the exposure in a CDS contract to differ from the exposure in the underlying reference obligation (Conseco is the classic example), we now commonly see CDS contracts with three types of treatment for restructuring.

Credit Derivatives Handbook

7

08 February 2007

No Restructuring (No R) In a No R contract, restructuring of debt is not defined as a credit event (only the failure to pay, bankruptcy, or default are considered credit events). High yield CDS typically trade with No Restructuring, as do the DJ CDX indexes. Modified Restructuring (Mod R) Mod R contracts create additional limits on the obligations that are eligible to be delivered on a debt restructuring (there are no additional limits on other types of credit events). The major limit is that on a restructuring credit event, the maximum maturity of the obligation delivered by the buyer is the earlier of 30 months after the restructuring date or the final maturity of the longest restructured bond or loan. The exception to this rule is that debt that matures prior to the scheduled maturity date of the original CDS contract can be delivered even if it fails the 30 months/longest restructured debt test. So one can deliver a 9-year bond into a CDS contract with 9 ½ years remaining. Modified Modified Restructuring (Mod Mod R) Mod Mod R contracts alter the deliverable limits of Modified Restructuring contracts slightly. The main difference is that the maturity limit is the later of the original CDS maturity date, 60 months after the restructuring date for restructured bonds or loans, and 30 months after the restructuring date for all other deliverable obligations. So the maturity limits are the same as Modified Restructuring except that restructured debt with a maturity of more than 30 months but less than 60 months is eligible. Determining the value of No Restructuring versus Modified Restructuring should be handled on a credit-by-credit basis because the value is dependent primarily on the possibility of a restructuring and the value of the cheapest-to-deliver option on a restructuring. In a hard credit event (bankruptcy, failure-to-pay, etc.) the price differential between different pari-passu debt tends to be small, but in a restructuring, there can be significant price discrepancies between different pari-passu bonds, depending on their maturity, coupon, currency, etc. Since the protection buyer has the option to deliver the cheapest non-subordinated eligible obligation, the cheapest-to-deliver option can be worth a significant amount on a restructuring. The value of this cheapest-to-deliver option will depend on whether the issuer has bonds with features that could generate a sizable postrestructuring price difference (like a long maturity, a currency difference, or some unusual language in the bond covenant). How expensive or cheap is Mod R CDS compared to No R? We addressed this question in our weekly publication dated August 26, 2004. As an example, we considered Shaw Communications (SRJCN). Shaw CDS was quoted for both Mod R and No R CDS. Based on mid-market CDS levels, the difference between Mod R and No R was 10bps. Assuming a constant hazard rate, we plot the probability of restructuring as a function of the recovery rate in Exhibit 5. We see that at recoveries of 50% or less, the probability of restructuring is less than 0.2% per year, compared to the probability of default: ~3%, assuming a 30% recovery rate. Is it cheap or expensive? Investors would have to make their own assessment in any specific situation, based on their estimation of the probability of a restructuring event compared to a default. In today’s market though, it is difficult to find names that trade both No R and Mod R contracts. CDS contracts on reference entities trade either No R or Mod R all the time, with No R contracts mostly limited to High Yield bonds and Mod R contracts generally limited to Investment Grade bonds. In the past, we have observed that Mod R contracts trade approximately 5% wider than No R CDS with the same level of credit risk, and the difference in levels reflects the probability of a restructuring event expected by the market participants.

Credit Derivatives Handbook

8

08 February 2007

Exhibit 5: Probability of Restructuring vs. Recovery in Restructurings

Probability of Restructuring

1.00% 0.90% 0.80% 0.70% 0.60% 0.50% 0.40% 0.30% 0.20% 0.10% 0.00% 0%

20%

40%

60%

80%

Recovery in Restructuring Source: Credit Suisse

Cash Settlement for Single-Name CDS On September 27, 2006, the International Swaps and Derivatives Association (ISDA) issued a press release in which it announced an improved protocol that would enable cash settlement for a broad range of credit derivatives, including single-name CDS. We recommend that counterparties have their legal teams review the ISDA documentation. Given the interest in the topic, we provide a brief summary of our interpretation of the cash settlement procedure. The cash settlement protocol aims to remove a major operational inefficiency arising from the current procedure, which generally requires physical delivery of a deliverable obligation of a name after a credit event (failure to pay, filing or restructuring) in order to settle single-name CDS contracts. Since the CDS notional amount outstanding might, and these days often does, exceed the amount of deliverable obligations available, the physical delivery requirement might result in an artificial inflation of the reference price, or “short squeeze”. In the case of Delphi in 2005, about $20 billion of CDS was outstanding versus $2 billion of deliverable obligations. In cash settlement, the protection seller pays the protection buyer par less the value of the reference obligation, as opposed to paying par and receiving the actual bond in a physical settlement. This way the complication resulting from the shortage of physical obligations is avoided. Under the current ISDA-determined procedures, CDX Index contracts are settled on a cash basis. A CDS Index Auction is conducted to determine the price of the defaulted reference obligation, or the “Final Price” in ISDA language. Under the new protocol, single-name CDS settlement will also be based on the result of the auction. The following is a brief description of the Auction methodology: • Participating dealers submit Inside Market Bids and Offers, which are firm quotations, expressed as percentages of the outstanding principal balance. The size of the Bids and Offers is $10 million in most cases, and the Bid / Offer spread cannot exceed two points. The “Inside Market Midpoint” is found as follows. The tradable markets (Bids higher than Offers) are excluded from the calculation. From the remaining non-tradable markets, the “Best Half” is selected as Bids / Offers with the tightest Bid / Offer spread. The Inside Market Midpoint is the arithmetic average of the Best Half market Bids and

Credit Derivatives Handbook

9

08 February 2007

Offers. Bidders whose bids and/or offers become part of the tradable market will pay an Adjustment Amount, calculated according to the protocol. Next, Open interest is found as the difference between all Buy and Sell orders. If the Open Interest is zero, the Inside Market Midpoint becomes the Final Price. • Bidders submit Limit Bids or Limit Offers, depending on the direction of the Open Interest. Open interest is then matched against the corresponding Market Orders and Limit Orders. Two outcomes are possible: 1)

All Open Interest has been matched. If the Open Interest is an offer, the Final Price becomes last market matched or Inside Market Midpoint plus 1% of par, whichever is less. If the Open Interest is a bid, the Final Price becomes last market matched or Inside Market Midpoint minus 1% of par, whichever is greater.

2)

Part of the Open Interest remains unmatched after all market and limit bids or offers have been matched. If the Open Interest is an offer, then the Final Price is set to zero; otherwise, if the Open Interest is a bid, the Final Price is set to 100%.

Under the physical delivery protocol, either the buyer or the seller of protection notifies the other party of the Credit Event by the means of Credit Event Notice or Notice of Publicly Available Information. After that the protection buyer delivers the reference obligation within 30 calendar days, the protection seller pays par in exchange. Under the cash settlement protocol, the settlement amount is par less the final price determined at the auction, which is typically held within 30 days from date of the credit event. The CDS seller pays par less recovery price on the date specified in the protocol in about 10 days following the auction date. The final pay-off of the contract ultimately depends on the recovery value of the deliverable obligation. In the case of physical settlement, it is par less the market price of the deliverable, which can be quite different from the auction price. The difference in recovery values resulting from the different settlement procedures is called the recovery basis, which the cash settlement protocol aims to minimize. Traditionally, recovery swaps have been used to hedge against recovery risk in general, and recovery basis risk in particular. (Please see more on recovery swaps in Credit Derivatives Weekly Focus - Expressing a View with Recovery Swap, dated 20 October 2005 and Credit Derivatives Weekly Focus - Valuing Recovery Swaps, dated 9 February 2006). As the new cash settlement protocol mitigates recovery basis risk and short squeeze risk, recovery swaps may experience some decline in investor interest. However, the instrument will remain a main tool to protect against inherent uncertainty associated with the recovery value of any defaulted obligation and is very useful to change singlename default swaps into fixed-default-rate contracts (also known as digital CDS).

Succession Rules Succession rules in CDS contracts following M&A, divestitures, or some major change in the corporate structure of the entity referenced in the CDS contract are important for investors considering an investment in CDS. We recommend that counterparties have their legal teams review the succession rules in any CDS contract. The standard ISDA Credit Derivatives documents can be obtained from the International Swaps and Derivative Associations, Inc. The rough rule of thumb in determining a successor is to follow the outstanding debt. The table below shows where CDS would be outstanding under the 2003 ISDA definitions. Generally, the new entity with the most debt will become the reference entity.

Credit Derivatives Handbook

10

08 February 2007

Things become more complicated when more than one successor entity has over 25% of relevant obligations assigned to it. Under this situation, firms with over 25% of the relevant obligations become successors. The notional amounts of CDS will be equally divided between the successors, with no change in spread. So even if one successor assumed 70% of the relevant obligations and another successor assumed 30%, the CDS would be split into two equal parts (a $10 million notional contract would become $5 million in each firm), even though the debt split was not proportional.

Exhibit 6: Successor Firm(s) for CDS When Reference Entity Breaks Up Based on 2003 ISDA Definitions—older contracts use “all or substantially all” language

Successor Firm assumes % Relevant Obligations of Parent

New Reference Entity

Greater than 75% Less than 75% but Greater than 25% in only 1 firm Less than 75% but Greater than 25% in 2 or more firms

All CDS to that firm All CDS to that firm All firms with 25% split CDS notional amount between number of firms equally No Change Firm with Highest percentage of Relevant Obligations becomes Successor

None Greater than 25% and Reference Entity survives None Greater than 25% and Reference Entity does NOT survive Source: Credit Suisse

LCDS Succession Event Differences US LCDS succession language differs from vanilla CDS in that for a LCDS succession, loans are only considered for a contract split after a succession event. For example, a name has $200 million in LCDS outstanding and $800 million in bonds and undergoes a credit event. Let’s say after the close of a credit event, New Corp receives 25% of the bonds, but none of the loans. The vanilla CDS that referenced the original entity would then be split in two, but LCDS would continue to reference Old Corp. Of course, it is also completely possible that half of loans are given to each name, while all the bonds remain with Old Corp. In this case, LCDS would be split, while vanilla CDS would not.

Index Products Index products of CDS are set portfolios of plain vanilla CDS traded in a standardized contract. In the US, the products are known as the Dow Jones CDX, with compositions created by polls of dealers every six months, with March 20 and September 20 contract roll dates and maturities of June 20 and December 20, i.e., maturity of the on-the-run fiveyear index can range from 4.75 to 5.25 years. The underlying reference entitles are equally weighted within the CDX products, and the index continues to exist after there is a credit event in any one of the underlying reference entitles. Credit events do not include restructuring, so only bankruptcy and failure to pay are included. Should there be no credit event, the buyer of the CDX pays a quarterly premium until maturity. Should any reference entity have a credit event, the following applies: • The buyer gives the seller deliverable obligations of defaulted entity with a principal amount equal to the notional value of the trade / # of entities. For a $100 mm position in a 30-name index, the amount would be $3.3333 mm ($100mm/30). • The seller pays the buyer the principal amount of the delivered securities (in the 30name example, it would be $3.3333 mm). • The trade continues with the notional amount reduced by the weight of the defaulted credit. In the $100 mm, 30-name example, the new notional would be $96.6667 mm ($100 - $3.3333). Number of reference entities in the index would be reduced to the 29 entities that have not had credit events.

Credit Derivatives Handbook

11

08 February 2007

When a new contract is rolled, investors wishing to participate in the roll can do so at the market spread on the roll date, which includes an upfront payment equal to the PV of the spread differential between the two contracts (roughly the yield spread multiplied by the risky DV01). Currently, there are seven IG indices and four high-yield indices. The six sub-indices of IG and three of HY include only names found in the overall index. The most liquid indices are 5yr IG, IG HiVol, XO and HY. Liquidity in the other sub-indices and indices with different maturities can be somewhat spotty.

Exhibit 7: DJ CDX Indices Number of entities as of Series 7

Investment Grade: HiVol: Financial: Energy: Industrial: TMT: Consumer:

125 diversified entities 30 from IG index, mostly BBBs 25 insurance, REIT, bank, dealer 14 from IG 30 from IG 22 from IG 34 from IG

XO: High Yield: BB: B: High Beta:

35 5B and 4B names 100 diversified entities 38 from HY rated BB when constructed 50 from HY rated B when constructed 30 from HY, wide spread names

Source: Credit Suisse, Bloomberg

While discussed later in this piece, there are a few items that investors must realize if they intend to use the index and hedge individual name exposure (either long or short credit risk). • Different Maturities: The actively traded (on-the-run) single-name CDS rolls to a new maturity every three months. The DJ CDX indices roll every six months. So there will frequently be a slight curve trade embedded into index versus underlying entities. • No R versus Mod R: Most basket products use no restructuring language. The investment grade single-name market is more liquid for Mod R language, so replicating the IG and HiVol index with a portfolio of single-name CDS usually leaves No R vs. Mod R exposure. High yield single names typically use No R. Given that the probability of a credit event is lower for No R than for Mod R, a No R index should trade tight to the sum of the Mod R underlying credits. Again, this is important for investment grade but not for high yield. • First Coupon Differential: Single-name CDS typically trades with a short first coupon. In order to improve liquidity, the index products trade with accrued. • Weighting for Survival is Necessary: Since wider spread credits are less likely to survive the full length of the CDS contract than tighter spread credits, the indices should trade tighter than the simple average of the spreads on the underlying reference entities. Let us give an example of an extreme situation. The index consists of two credits: one with the spread of 10bps and the other with the spread of 2000bps. Should the index spread be just (2000bps+10bps)/2 = 1005bps? Assuming the index trades at 1005bps, we compare the benefits of selling the index CDS ($2mn), to selling a twoname $2mn portfolio ($1mn CDS of each of the underlying credits). Judging from the issuers’ spreads, one is likely to believe that a credit trading at 10bps is unlikely to default in the near future, while the one at 2000bps is likely to experience a credit event, stopping the chunky quarterly payments. Therefore the seller would much rather receive 1005bps on a $2mn two-name index than 2000bps and 10bps from each of the credits. Even after the default of the risky credit, the index seller would collect 1005bps on the remaining $1mn, while the portfolio seller would collect just 10bp.2 2

Credit Derivatives Handbook

Marked-to-market P&L swings is the main risk of this position.

12

08 February 2007

While IG CDX is traded in unfunded form, many high yield investors cannot use derivatives, so a funded version of the HY CDX is available in addition to the HY unfunded version.

Index Arbitrage With the establishment of the DJ CDX indices as the most liquid instrument in the credit derivatives space, attempts to replicate “index arbitrage,” a popular strategy in the equity markets where investors trade the index itself versus its underlying names, have increased in recent years. Although in theory the concept sounds reasonable, the largest problem that an investor faces when trying to execute this strategy is the liquidity of the singlename CDS. For example, the CDX IG index consists of 125 credits and an investor would have to buy all of the individual CDS names simultaneously against the index, causing it to become very cumbersome; also, the existence of bids/offers often negates the opportunity. In the case of the CDX HiVol and XO, due to their smaller number of names, some investors have successfully executed the index arbitrage strategy in recent years. A more convenient way of applying the index arbitrage methodology is to use the underlying CDS spreads as a richness/cheapness indicator of the index. It is important to note that an investor would have to take into consideration two important factors listed below. Restructuring Treatments An investor must understand that there are different restructuring treatments between the indices and the underlying credits. DJ CDX indices trade with a No Restructuring provision. On the other hand, CDS protection for a majority of the investment grade credits trade with Modified Restructuring, i.e., restructuring is actually treated as a credit event. Therefore, even if an investor managed to create an “arb” position for the index against its underlying credits, the restructuring risk does not go away. A restructuring event for any of the credits triggers the payment on the single-name CDS, while the index is unaffected. There is no simple way around this problem; an easy solution is to assume that the CDS spreads of the underlying issuers are less than those observed in the marketplace by a fixed percentage. In our analysis, we assume the Mod R premium to be 5% (please remember that this is just an estimate for the Mod R vs. No R premium). Convexity Another more complicated issue is that while the average spread of all the single-name index constituents are usually priced as new contracts with strikes equal to the spreads, all the names in the CDX indices basically trade at the same strike, which is set at index series inception. This could have a significant effect on the pricing of the names and creates convexity issues with spreads being away from the strike. We correct for this effect in our fair value models. Each credit in the index receives the same annual premium (the index coupon), and as a result, we effectively have a number of CDS with off-market strikes. CDS for both TSG (at 330 bps) and TWX (at 33 bps), members of the HiVol index, receive the same annual premium of 75 bps. In this situation, TSG’s CDS PV is negative, while TWX’s is positive. To come up with a fair value assessment of the index spread, which takes into account both convexity and restructuring treatment effects, one would have to use the following procedure. (1) Initially, sum up the index-weighted PVs from all of the underlying issuers, preliminarily adjusting them for the Mod R provision (less 5%). (2) Compare the obtained PV to the PV of the HiVol. (3) Finally, determine the index spread, which equates the PVs of the index and the portfolio of credits. This spread is called the fair value spread. As we have outlined above, a pure arbitrage between the index and its underlying issuers may be difficult to realize, but we think that tracking the difference between the fair value spread/the market spread and between the fair value spread/the average spread of its underlying entities is worthwhile. In Exhibits 8 - 11, we present the historical data series for IG7 and HiVol7 since their inception.

Credit Derivatives Handbook

13

08 February 2007

ModR / NoR adjustment – 5%

ModR / NoR adjustment – 5%

0.8 0.6 0.4 0.2 11 /2 0/ 06 12 /2 0/ 06

9/ 20 /0 6 10 /2 0/ 06

0

5.0 4.0 3.0 2.0 1.0 0.0

12 /2 0/ 06

1

6.0

11 /2 0/ 06

1.2

7.0

10 /2 0/ 06

1.4

8.0

9/ 20 /0 6

Market Sprd Less Fair Sprd (bps)

Exhibit 9: HiVol7 Market Sprd vs. Fair

Average Sprd Less Fair Sprd (bps)

Exhibit 8: HiVol7 Average Sprd vs. Fair

Exhibit 10: IG7 Average Sprd vs. Fair

Exhibit 11: IG7 Market Sprd vs. Fair

ModR / NoR adjustment – 5%

ModR / NoR adjustment – 5%

0.5 0.4 0.3 0.2 0.1

9/ 20 /2 00 6 10 /2 0/ 20 06 11 /2 0/ 20 06 12 /2 0/ 20 06

0

Source: Credit Suisse as of Jan 4, 2007

4.0 3.0 2.0 1.0 0.0 -1.0 -2.0 -3.0 9/ 20 /2 00 6 10 /2 0/ 20 06 11 /2 0/ 20 06 12 /2 0/ 20 06

0.6

Market Sprd Less Fair Sprd (bps)

Source: Credit Suisse As of Jan 4, 2007

Average Sprd Less Fair Sprd (bps)

Source: Credit Suisse as of Jan 4, 2007

Source: Credit Suisse as of Jan 4, 2007

Tracking the data series for the IG7/HiVol7 since their inception in September 2006, we can make some general observations. Exhibits 8 and 10 show that the fair value spreads of the indices seem expensive compared to the average spread of the underlying CDS, but we do not think that these relationships should be used as rich/cheap indicators. These are more indicative of the pricing difference that occurs when trying to value the individual contracts that may be No R or Mod R within the index against an index that trades No R. Instead, we think that the differences between the market spreads of the CDS indices and their fair value spreads should be monitored to come up with buy/sell recommendations. Therefore, from Exhibits 9 and 11, we can see that the HiVol7 index has been relatively expensive during most of the time since its inception and is currently still somewhat expensive to fair value. The IG7 market spread has been both expensive and cheap compared to fair value in its history, and as of January 4, 20073, it appeared to be fairly valued. 3

Credit Derivatives Handbook

Past data are neither a guarantee nor indicative of future performance.

14

08 February 2007

Corporate Bond LIBOR Pricing Methodologies When valuing a bond in the corporate market, one of the main sources of confusion results from the various methods to price a bond. For example, when someone says that a bond is trading at LIBOR + 50, are they quoting this bond in reference to the I-spread, the Zspread, or the assetswap spread? When syndicate desks market a fixed-rate bond at a particular spread to LIBOR, there can some concern at the actual pricing because of the confusion about the methodology used to calculate the spread to LIBOR. We review these measures below. Each of these spread measures has advantages and disadvantages when it comes to valuing bonds, and all of them are imperfect to some degree.

Exhibit 12: LIBOR Spreads of 10 yr Corporate at T+100 Calculated using Bloomberg’s YAS and PRPL functions, as of Dec 15, 2006

Coupon

Yield

Duration

Price

ASW

Z-Spread

I-Spread

3 4 5 5.6 6 7 8 9

5.60 5.60 5.60 5.60 5.60 5.60 5.60 5.60

8.17 7.83 7.55 7.40 7.31 7.10 6.93 6.77

80.26 87.86 95.45 100.00 103.06 110.65 118.25 125.84

45.0 47.6 50.2 51.8 52.6 55.1 57.9 60.5

52.4 52.6 52.8 53.0 52.8 53.0 53.3 53.5

52.9 52.9 52.9 52.9 52.9 52.9 52.9 52.9

Source: Credit Suisse, Bloomberg

Z-spread: To calculate the Z-spread, all cash flows of a fixed-rate corporate bond are discounted by the appropriate zero-coupon swap rate. The net present value of the cash flows of the corporate bond, discounted by the swap curve, is then compared to the market price of the bond. The zero-coupon swap curve is then shifted parallel so that the calculated PV of the cash flows equals the market price. The size of the shift needed to make the PV equal to the market price is the Z-spread. The Z-spread is the closest thing to a zero-vol OAS for a bond without any embedded options. As a result, this is a consistent measure to use for making relative value decisions between various corporate bonds, as well as relative value decisions between corporate bonds and other asset classes like MBS. It provides a relatively clean way to compare the bond to the swap curve (excluding the premium adjustment), but for most bonds, it would be extremely difficult to put on a trade that would replicate the Z-spread. Accordingly, we think the Z-spread is the best analytical tool, but due to significant dislocations in Z-spreads, it cannot be cleanly arbitraged. Assetswap Spread: In an assetswap spread, the bond's cash flows are discounted using the appropriate zero-coupon swap rates, and then an annuity is added or subtracted from the bond to force the calculated bond price to be equal to the market price. The size of this annuity, spread over the floating-rate leg of the actual swap, is the assetswap spread. The key difference between Z-spread and assetswap spread involves the discount rate (zero-coupon swap rates in case of assetswaps and zero-coupon swap rates plus the Zspread in case of the Z-spread calculation) and the upfront payments for a par-assetswap. There are some very positive attributes of the assetswap spreads. There is a live and active assetswap market. One could, therefore, realize the assetswap spreads. It is not just a theoretical construct like Z-spreads. There are several problems with the assetswap spread, as well. First, by adding in an annuity, one is introducing sensitivity to parts of the curve where there should not be any.

Credit Derivatives Handbook

15

08 February 2007

For example, a 10-year zero-coupon corporate bond should not see any relative value change if the 10-year zero-coupon swap rate is unchanged but the 3-year zero-coupon swap rate rallies 10bps, but its assetswap spread would change. In addition, the upfront payment equal to the PV of the annuity is essentially a loan at LIBOR, which might be significantly different from the financing cost of the corporate instrument. In our view, the bottom line is that if one actually needs to do an assetswap to realize the value of a trade, then one ought to use the assetswap spread to make relative value judgments, its faults notwithstanding. For example, if an investor were buying a fixed-rate bond and a default swap as a package (a basis trade – see below for more), he would be much better off using the assetswap spread rather than any other measure for relative value analysis because the basis can be locked in. I-spread: The I-spread, also known as yield-yield or matched-maturity assetswap spread, is simply the difference between the corporate bond's yield-to-maturity and the swap yield to the same maturity date. An interpolated rate is used if the bond's maturity does not line up perfectly with an actively traded point on the swap curve (the 9- and 10-year swap rates would be interpolated to generate a 9 ¼-year rate). The most attractive thing about I-spreads is that they use a very straightforward and simple calculation. The I-spread works reasonably well for bonds trading near par that have a yield close to the swap yield, but if the coupon on the bond is far away from the swap yield, there will be large duration differences. For that reason, I-spread trades are typically weighted to be DV01-neutral, to account for the fact that the duration of the bond and the duration of the swap do not match. If the yield curve were perfectly flat, the duration gap would be less problematic in regard to relative value analysis, as opposed to if the yield curve were significantly steeper, where the use of maturity instead of duration can cause serious distortions. Some investors use the I-spread as their relative value measure because they do not want to make a separate premium adjustment. When the yield curve is upward sloping, the duration versus maturity issue means the I-spread on high coupon bonds should be wider, but the high dollar price of most high coupon bonds means the premium adjustment lowers the spread to LIBOR. Investors assume that these two factors will offset each other, but there is no reason to expect that the distortions will be of similar magnitudes. Accordingly, we think that investors are much better off using either the Z-spread or assetswap spread, and then making a separate premium adjustment.

Exhibit 13: LIBOR Spread Versus Price for Corporate at T+100 Different coupons at same yields are used to generate prices

Spread to LIBOR (bps)

65 60 55 50 45

12 6

11 8

11 1

10 3

10 0

95

88

80

40

Price ASW

Z-sprd

I-s prd

Source: Credit Suisse, Bloomberg

Credit Derivatives Handbook

16

08 February 2007

To summarize, the Z-spread is largely a theoretical construct and should be used mostly as an analytical tool for relative valuation rather than as a basis for executable transactions. The assetswap spread is slightly more complicated and less accurate from a relative value perspective, but more applicable since investors can actually execute trades that lock in the assetswap spread. So investors trading corporates against LIBOR or the CDS markets (which uses LIBOR by default) should stick with the assetswap calculation. The I-spread, although it is very straightforward and simple to calculate, does not discount the cash flows in a realistic manner. To complicate things further, none of these measures address the premium adjustment problem, which we will discuss in the next section.

Adjusting Spreads on Premium-Priced Bonds After a period of extreme interest rates (either high or low), corporate bonds issued during those periods tend to trade at significant premiums or discounts. This has serious implications, as premium bonds trade at a yield concession to bonds trading close to par. Relative value within and among issuers can become highly skewed due to this effect. While everybody agrees that premium bonds ought to trade at a yield concession to a par bond, the more pertinent relative value question is by how much? And is this valuation based on market data? There are many different theoretical approaches available to value cash flow differences between bonds, i.e., calculating a spread based on a Treasury spot curve. However, we believe the Credit Default Swap market provides a marketbased method of valuing the required extra yield for a premium bond or the pickup for a discount bond. With the aim of determining an appropriate yield concession, we start with the basic difference between premium and par-priced bonds – they provide different default exposures to the investor. For example, an investor can lose 80 points for a $120 bond (assuming the recovery rate is 40%), if the bond were to default immediately. The corresponding exposure for a par-priced bond is only 60 points. Our belief is that a premium-priced bond combined with default protection is the same as a par-priced bond. In the previous example, a $120 bond combined with $20 of loss (as opposed to default) protection is essentially the same credit exposure as a par-priced bond. However, as the premium on the bond is pulled toward par over time, the investor needs an amortizing amount of protection rather than a constant level of protection for the total holding period. We illustrate this with a simple example in Exhibit 14. As more CDS curves have become fully priced, we have updated the model to perform these calculations based on the actual and forward curves, rather than using only the 5-year point, but for illustrative purposes, we review them below.

Credit Derivatives Handbook

17

08 February 2007

Exhibit 14: Premium Bond Calculation XYZ 8.50% 10-year bond

Price Yield Spread to TSY CDS Spread Recovery Value Premium over par bond (Price Difference from Par) Half of Premium to Par Average Protection Needed (10/(1-0.4)) Average Cost of Protection over life of bond (16.67 * (80/100)) Par Equivalent Yield Par Equivalent Spread

120 6.03% 123 80 40% 20 10 16.7 13.3 5.9 109.7

Source: Credit Suisse

This methodology could clearly be used to price a discount bond as well. For example, a discount bond trading at $90 plus a 10-point incremental sale of exposure through default protection is the same exposure as a par-priced bond (Exhibit 15).

Exhibit 15: Discount Bond Calculation XYZ 4.40% 10-Year Bond

Price Yield Spread to TSY CDS Spread Recovery Value Discount to par bond (Price Difference from Par) Half of Discount to Par Average Protection Saved (-5/(1-.04)) Average Cost of Protection over life of bond (16.67 * (80/100)) Par Equivalent Yield Par Equivalent Spread

90 5.81 121 80 40% -10 -5 -8.33 -6.7 5.877 127.7

Source: Credit Suisse

Basis Evaluation Simple Cash/CDS The difference in spread level between a CDS and a cash bond is known as the Basis. There are many methods for evaluating the basis, and one can get very different results depending on the type of spread used – for example, an evaluation using the I-Spread will sometimes yield very different results than from using Z-Spread. For instruments with similar maturities, especially for lower-rated or higher-beta issuers, the basis is mainly positive, meaning that the CDS trades wider than cash. As the probability of default for an entity declines (e.g., for higher-rated issuers), the basis can actually turn negative, where CDS trades tighter than cash. This can be explained by looking at the default risk involved. When issuers have a very low chance of default, investors are paid on spread volatility and not for taking default risk. As the probability of default increases, the basis will increase as investors take on greater credit/default risk when purchasing a bond. Another possible reason the basis may trade positive is due to technicals. In pricing a theoretical CDS, one must take into account the cost of funding a position in a cash bond – for example to hedge a CDS position. Therefore, if a bond is trading “special” in the repo market (see next section), it is possible that the basis is being distorted due to the bond’s availability in repo.

Credit Derivatives Handbook

18

08 February 2007

In the exhibit below, one can see the historical basis behavior for DOW 5yr CDS and DOW 6.0 2012s. Over a 12 month time frame, the basis turned strongly negative due to either CDS tightening relative to the assetswap spread of the bond or the swap rate outperforming cash bonds, resulting in a widening of the assetswap spread. The times the basis went positive generally coincided with negative news regarding the company, whereupon CDS spreads widened due to the increase in default risk. The recent move towards the negative for the basis can be explained in general not only by CDS spreads tightening as outlook for the company improved, which is seen in the tighter CDS levels, but also by the recent outperformance of the swap rate, which has widened the assetswap spreads for the cash bonds. In a later section we review Credit Suisse’s preferred method of evaluating the relative value in basis trading, The Arbitrage Price Spread (APS) Model.

Exhibit 16: Basis Between DOW 5 Yr CDS and DOW 6.0 2012s Default Spread vs. Assetswap Spread as of December 8, 2006

15 10 5 0 -5 -10 -15 -20 -25 -30

12 /8 /2 00 5 1/ 8/ 20 06 2/ 8/ 20 06 3/ 8/ 20 06 4/ 8/ 20 06 5/ 8/ 20 06 6/ 8/ 20 06 7/ 8/ 20 06 8/ 8/ 20 06 9/ 8/ 20 06 10 /8 /2 00 6 11 /8 /2 00 6

Basis Points (bps)

DOW Chemical

Source: Credit Suisse

Basis Trades and LCDS With the creation of the LCDS market, the basis between levered loans and LCDS can be traded. These trades are not much different than senior unsecured bonds vs. vanilla CDS. The same issues regarding premium/discount adjustments, evaluation of swap spreads, using OAS to Swaps for callable loans, and other issues all apply. There could also be significant liquidity issues due to loan transferability in some cases. Investors considering LCDS vs. Loan basis trades also need to consider call features and LCDS cancelability, which in some cases causes a different P&L profile than vanilla CDS vs. Bonds. Since it is possible for the LCDS to cancel should it be orphaned, investors in negative basis packages are at less risk of being left with a residual short should the debt be removed.

Repo “Special” The repo market brings together bondholders who need to finance their holdings and money market investors who are looking for attractive rates. Securities dealers use repo transactions to finance their positions, while investors mainly use repos for leverage to enhance their returns.

Credit Derivatives Handbook

19

08 February 2007

Over the past several years, there have been an increasing number of market participants trading the cash versus CDS basis. As credit quality improved, the majority of these were lower risk, negative basis trades in which investors would buy a cash bond and buy CDS for the positive spread differential. If the funding was attractive enough, investors could create an arbitrage position if they held the bond until maturity. Mark-to-market and cheapest-to-deliver options in case of a credit event should be taken into account when putting on this trade. Sometimes bonds appear rich relative to CDS because they are trading “special” in the repo market.

Exhibit 17: Cash Flows in a Typical Repo Transaction Borrowing Bonds for a Short Cash Investor

Dealer Bond

Lending Bonds Out Bond Dealer

Investor Cash Minus Haircut Source: Credit Suisse

The process of shorting or lending a bond is relatively straightforward (see schematic above). An investor wishing to short calls a dealer and asks if there is any supply available to borrow for the bonds in question. If there is, the investor then pays cash to the dealer, which agrees to repurchase (repo) the bonds at a specific price in the future. Typically, the shorts receive 15-30bps below the federal funds rate (depending on the issuer), but need to pay the coupon on the bond. Most transactions for corporate bonds are done as “open” repo, in which the rate changes on a daily basis and the transaction can be terminated at any time. Other bonds (emerging markets, US Treasuries) trade on a “term” repo basis, in which the rate is locked for a specified period of time, normally from one to nine months. Corporate bond investors wishing to lend their securities are normally charged a haircut by the dealer. The haircut is the amount of capital retained in excess of the value of the collateral, i.e., for $1mn in value at a 20% haircut, the dealer would lend $800K. Haircut reductions are possible when the bond position is offset by CDS. Haircuts can normally range from as low as 2% to as high as 35% for some distressed credits and should be taken into account in P&L calculations. Trading Special On some occasions the bond can start trading “special.” In such instances, the investor shorting a bond gives cash, but receives much less than federal funds – perhaps nothing (0% repo rate) – for the loan, and still pays the coupon of the bond. This makes short trades expensive, and can explain why some positive basis trades look attractive. In the corporate bond world, the main reason bonds go “special” is the lack of supply for shorts. For smaller bond issues, if there are too many shorts already or the majority of bondholders refuse to lend their securities, bonds often “go special.” In certain extreme situations, when the bond shorts fail to deliver the bond, they can be forced to close their trades, sometimes at two to three points above the market. This is called being “bought in.” Therefore, the shorts should exercise caution and monitor the bond supply closely.

Credit Derivatives Handbook

20

08 February 2007

Sometimes it does make sense to buy the bond, trading “special” against the CDS position. For example, GM 7.2 ‘11s look rich to the 5yr CDS (data as of May 4, 2005). At first glance, it seems cost prohibitive for investors to go long cash in a basis trade (175 bps of negative carry). However, repo should be taken into account, making the trade attractive from the carry perspective (we did not take the haircut into account). Investors pay 175 bps (the difference between the CDS and Z-Spread), but get back 300 bps by lending the bonds out (fed funds rate less repo rate, i.e., 300bps - 0bps = 300bps). So effectively this trade provides 125bps of positive carry. Several risks should be taken into account. While in terms of pure carry the trade can look attractive, mark-to-market is a very important consideration. As in this GM trade, if spreads rally, one could be facing a significant loss. Another risk to keep in mind is the possible return of the “special” repo rate back to normal. Lastly, there is the loss of the voting rights on the bonds that are being repoed. Frequently, the repo market seems slow to react to company-specific news, and we believe that investors should monitor repo rate dynamics and try to lock in favorable rates for either outright cash positions or basis trades.

Credit Suisse’s Approach to Basis Trading As outright credit opportunities dwindle with ever-tighter spreads within higher quality names, relative value trading in basis packages presents a very attractive opportunity. Basis packages were very different in 2002 and 2003. The spread differentials between a cash bond and CDS were so large that any crude measure of difference in cash spread – such as I-spread or Z-spread – versus the CDS spread would have given an investor a good indication as to whether a basis package was a worthwhile investment. This is no longer the case. Since various LIBOR spreads are less than perfectly comparable to CDS spreads for risk or income, and with spread differentials at very tight levels, it is absolutely essential that we come up with a better foundation for basis trades. Credit Suisse’s quantitative strategists have devised a methodology, helping investors better assess their risks and rewards in basis trades. For a full description and the underlying theoretical justification, please see, ‘Price and probability – A new approach to trading credit,’ The Quantitative Credit Strategist, 26 Nov 2004. In simple terms, Credit Suisse’s approach consists of the following:

1) With the development of full curves going out to ten years, the CDS market contains very valuable information about probabilities of default. Using the full CDS curve, one can calculate the probabilities of default in contiguous yearly periods given survival until the start of each period.

2) At each time period when the bond pays the coupon, the issuer can either default (then no more coupons are paid and the bondholder receives the principal recovery) or all the coupons and the principal are paid as agreed. The probability of either credit event or coupon/principal payment at each time horizon is given by the survival curve from 1. By discounting each cash flow, we get the theoretical price of the risky bond. The difference between this CDS-implied price and the market price is the Arbitrage Price Difference (APD). The corresponding asset swap difference is the Arbitrage Pricing Spread (APS).

3) Statistically tracking the APD/APS changes over time provides one with much better signals than simple spread differentials for putting on and/or taking off relative value basis trades.

Credit Derivatives Handbook

21

08 February 2007

4) Finally, by computing the change in the price of the cash bond in the package relative to the change in DV01 of a CDS contract for a one-basis-point change in CDS, one can potentially compute a much better hedge ratio than simple dollar duration and DV01 comparisons. We believe this approach has several advantages: (a) it is based on market-implied default rates; (b) it fully incorporates the shape of the asset swap curve and the CDS curve; and (c) it provides insights for relative value trading of basis packages rather just the carry4 advantage of a basis package for a “buy and hold” investor. On the other hand, mark-to-market risk and the possibility of bonds going “special” on repo are the main drawbacks of any basis trade.

Valuation of Step-Up Bonds From 2000 to 2003, many telecom companies issued bonds with coupon steps linked to the firm’s rating. This was in an effort to protect investors from losses in the event of downgrade, making the deals more palatable in a turbulent market. Since then, issuance has been scattered throughout many different industries and sectors, especially with the increasing instances of leveraged buyouts that have occurred recently. This brings with it a new level of complexity for bond selection. There are a number of articles written on the subject of step-up provision estimation. We propose a simple approach that allows for estimation of step-up provisions using the rating agencies’ historical transition frequencies and issuer CDS curves. The ratings agencies regularly publish Exhibit 18: One-Year Average Transition their transition matrices, which should Probability (for Baa2 rating) give us an indication of a company’s From Moody’s Transition Matrix (1983-2005) likely rating in a year’s time. For Baa1 5.8% example, a company with a Baa2 rating Baa2 73.8% today will have a 73.8% probability of Baa3 6.7% having the same rating, 6.7% of being Ba1 1.7% downgraded one notch, 1.7% of being Source: Moody’s downgraded two notches, etc. If oneyear ratings transition frequencies are determined by the matrix A, then two-year frequencies are determined by the matrix B=A*A, three-year by the matrix C=A*A*A, etc. For each time horizon, we can determine the expected adjustment to the bond’s coupons. For example, company XYZ’s bond coupon steps up by 25 bps if each agency downgrades it by one notch below A3/ BBB+. So by multiplying the probability of reaching the rating by the corresponding bond’s coupon change, our goal is to try and determine the expected coupon step-up for a one-year horizon. This approach does not take into account the fact that a specific company’s transition profile may differ from the historical average transition matrix – there is no distinction between a stable, improving, or a deteriorating issuer. To get around this problem, we propose using the CDS curve and calibrating the transition matrix to the levels of CDS with a corresponding maturity. For example, XYZ’s 1-yr CDS trading at 55bps implies 0.78% probability of default in one year, assuming 30% recovery. Moody’s one-year transition matrix for a company with the same rating implies a 0.45% probability of default. By matching default probabilities and minimizing deviations from the original matrix, we come up with a new transition matrix for XYZ that takes into account company-specific default probabilities.

4

Credit Derivatives Handbook

Past data are neither a guarantee nor indicative of future performance.

22

08 February 2007

Another adjustment we need to take into account is the agency rating outlooks. Say for XYZ, S&P has a negative outlook, meaning that the probability of the one-notch downgrade is higher than for a company with a stable or positive outlook. To get around this problem and estimate probability adjustments due to ratings downgrades, we assume that for each main rating class the probability is normally distributed among three subclasses: positive, stable and negative. If the outlook changes (say from stable to negative), the distribution shifts with the change, implying a higher probability of downgrade than before. Calculating expected coupon changes at each time horizon for each agency and summing up its PVs, we come to the total expected change in PV due to the step-up provision. If investors have a bearish view on the company’s spreads and credit ratings, we believe they are better off going long step bonds against CDS or non-step bonds, since coupon step-ups should benefit the trade. On the other hand, if investors have a bullish view on the company, we believe they should sell step bonds against CDS or normal bonds because of the coupon step-downs.

Steep CDS Curves Effect on Rolls The roll is when the market shifts from treating one maturity (such as June) as the active benchmark to another benchmark (such as September). This three-month extension of the credit curve should have a meaningful impact on the spreads for many reference entities when credit curves are quite steep. In mid-2004, this was particularly true at the front of the curve, which tended to be quite steep for most credits. Many investors at the time felt comfortable selling 1-, 2-, and 3year CDS at that point in the economic cycle, as default rates are usually at relatively low levels in the early stages of an economic expansion before peaking when the economy moves into recession. This bias toward taking risk at the front of the curve tended to create very steep credit curves in the 1-year to 5-year sector. Since single-name CDS only extends maturities by three months versus the six-month extension we get when the DJ CDX baskets roll, the impact is unlikely to be very dramatic in most 5-year spreads (as most curves tend to be flatter beyond the 5-year point), but for 1- to 3-year CDS, we could see a significant rise in spreads. However, during some rolls when the market has extended to a new benchmark, the steepness of the credit curve may not be fully captured. With the additional risk from an extra three months of exposure from a roll tending to be undervalued, there is frequently a kink in the forward curve that offers a trading opportunity. Accordingly, we think that in these situations, buyers of shorter-dated protection are likely to find it more attractive to wait until after the roll to new benchmarks, while protection sellers (particularly at the front of steep curves) should rush to lock in pre-roll end-dates. The liquidity given up for being in an old contract in the months after a roll makes us believe that this strategy is inappropriate for investors who intend to trade out of the position very quickly. For investors who have a longer time horizon, selling off-the-run dates sometimes could be attractive relative to selling on-the-run maturities. A noteworthy tendency for credit curves to be steeper out to the 5-year point and then flatter beyond the 5-year point can clearly be seen in Radio Shack (RSH) CDS curve as of January 20075. In order to avoid the effect of a one-year maturity extension having a larger duration impact at the front of the curve than the back of the curve, the following graph shows spreads against spread durations. 5

Credit Derivatives Handbook

Past data are neither a guarantee nor indicative of future performance

23

08 February 2007

Exhibit 19: Radio Shack (RSH) Curve: Steeper Inside of 5-Years Data as of January 2, 2007

300

250

Curve beyond 5 Y ears

Spread

200 Curve out to 5 Y ears

150

100

50

0 0

1

2

3

4

5

6

7

DV01 Source: Credit Suisse, Bloomberg

The kink in the curve shape means that investors living through such situations need to be more aware of how shorter maturities roll down the curve versus longer maturities. At the June 2004 roll to September, if one assumed the curve is unchanged, 3-year CDS rolls 8.75bps per quarter, while 5-year CDS rolls 5bps per quarter and 6-year and longer CDS rolls only 1.25bps per quarter. Factoring the roll into carry calculations makes taking risk in the 5-year and under part of the curve more attractive.

Steepeners and Flatteners Many curves are very steep out to the 5-year point because a number of investors feel comfortable that default rates will be low until the economic expansion runs out of steam (so they aggressively sell short-dated protection). Beyond the 5-year point, synthetic CDOs provide support, making many 5s/7s relatively flat. This means that most DV01weighted 5s/7s steepeners are positive carry, and that many 3s/5s flatteners are also positive carry. In our view, investors who do not want to fund negative carry positions by taking risk in the wider spread names should look at curve trades to minimize carry costs while taking a view on a credit. Given the kinked curve shape, investors looking at curve trades need to be aware of the rolldown effect of the trade, as it will make many 3s/5s flatteners and 5s/7s steepeners more attractive. As an example of the impact of rolldown on 3s/5s and 5s/7s trades, we review the Clear Channel Communications (CCU) credit curve below (December 13, 2006). A 3s/5s CCU flattening trade would be DV01 weighted at just over $15 million 3s versus $10 million 5s. If we held this trade for one quarter, we would pay $42,217 for the 3-year protection ($15mn * 1.11% / 4) and receive $63,750 for the 5-year protection ($10mn * 2.55% / 4). However, the mid/mid 2s/3s CCU curve is worth 13bps per quarter, while the mid/mid 4s/5s curve is 18bps. Assuming an unchanged curve shape, the investor receives more for the rolldown in the 5-year. Accordingly, a DV01-weighted 3s/5s CCU curve trade is a 43bps positive carry flattener.

Credit Derivatives Handbook

24

08 February 2007

Investors can do a DV01-weighted 5s/7s CCU steepener that is positive carry when adjusted for the rolldown. Selling $12.5MM of 5-year CCU protection at 255bps and buying $10MM of 7-year protection at 341bps means that the investor has to pay out a net 5.4bps of additional coupon on a running basis. However, the 4s/5s curve is worth 18bps per quarter, while the 6s/7s curve is worth 11bps per quarter. So the roll down the 4s/5s curve more than offsets the higher spread paid for the 7-year protection, resulting in a 28bps positive carry steepener if calculated on a DV01-neutral basis6 7.

CDS Orphaning and Confusing Name Changes Jon Zucker +44 20 7888 8109 [email protected]

Although buying CDS on names that investors believe will significantly increase leverage is the most appropriate strategy to implement, the possibility that an entity that has issued debt in the past ceases to issue further debt or decides to take out their existing debt can cause significant problems in the market. This is especially a challenge for investors who use CDS referencing bonds to hedge M&A, LBO or leverage recapitalization risks. “CDS Orphaning” can occur when a CDS contract is left (or potentially will be left) referencing an entity that has no debt obligation. In addition, CDS orphaning can occur through other means, in particular confusing name changes, which exposes another risk in the CDS market – an operational one – that can be solved only through diligence and better information systems on the part of market participants. The main points we emphasize in regard to this section are: • The risks imposed by orphaning are real and potentially large. • Orphaning pushes CDS spreads to zero for the forward period after the longest existing bond matures. An entity that has no debt cannot default,8 so the forward credit spread – the spread covering default after the longest maturity available – must be close to zero. As a corollary, current 5-year CDS spreads move to zero as the existing maturities become shorter over time. • Orphaning has less of an effect on entities with very long-dated debt outstanding. • It is it difficult to predict orphaning activity. The typical catalysts for orphaning are M&Aprompted capital restructuring or corporate actions establishing new financing entities, together with name changes. In theory, none of these conditions are either necessary or sufficient for orphaning: only a credible promise to cease issuing further debt from a given entity (combined with no guarantees across entities) is necessary. • The confusion in the market regarding name changes can be compounded by the re-use of names. These are operational issues, which can be minimized through additional diligence and faster communication. The RED database serves as a clearing house for this information for CDS participants. Unfortunately, not all data can make it onto that database instantaneously. To that end, timely confirmations (preferably electronic, e.g., via DTCC) should prevent large value discrepancies.9

6

Past data are neither a guarantee nor indicative of future performance Transaction costs and bid-offers were not taken into account. 8 Technically, a firm can default on many obligations besides debt, e.g., loans , trade payables, or foreign exchange obligations. We avoid detailed discussion of these possibilities, except to say the spread must be “close” to zero. 9 It should be noted that even timely confirmations cannot eliminate the possibility of error. DTCC is only as good as the databases underlying it. In the Rentokil case (see below), the RED database did not catch all entity changes for many weeks, so counterparties could have confirmed on the same, erroneous data. 7

Credit Derivatives Handbook

25

08 February 2007

Orphaning: A Significant Risk to Investors Long Protection When an entity that has issued debt in the past ceases to have any outstanding debt, we deem CDS referencing that entity “orphaned.” The cessation is effectively a belief and has no legal basis. The question for analysts is whether such a belief is credible. If it is, existing CDS then references an entity with debt of limited maturity, if any at all. Absent further changes, there is little chance of a credit event beyond that maturity, and forward credit spreads should approach zero.10 A recent example from Europe illustrates one way this can occur. In 2004, ITV was formed through the merger of Granada and Carlton Communication. All existing debt in Carlton Communication was transferred to ITV, but Carlton retained guarantor status on the outstanding debt (which had maturities up to 2009). When a new 10-year ITV 2015 bond was issued, ITV indicated that Carlton’s status as a guarantor would fall away after the 2009 bond matured. Consequently, the existing Carlton CDS was orphaned. Exhibit 20 shows the spread of CDS for both Carlton Communications and ITV. Currently, the spread in 5-year CDS is 65 bps wider for ITV than for Carlton.

Exhibit 20: Orphaning Can Have a Significant Effect: the Case of Carlton Communications and ITV 5-year CDS Spread in bps

160

Carlton Com.

140

ITV PLC

120 100 80 60 40 20 0 03/2005 04/2005 05/2005 06/2005 07/2005 08/2005 09/2005 10/2005 11/2005 12/2005 01/2006 02/2006 03/2006 Source: Credit Suisse

The attempted purchase of ProSieben by Axel Springer offers another example of the power of orphaning. In its attempt to purchase ProSieben, Axel Springer announced a plan to buy back all of ProSieben’s debt. ProSieben’s CDS spread tightened from 140 bps to 30 bps until regulators quashed the deal.11 It is also important to keep in mind bond covenants that may restrict certain activities, such as change-in-control puts, or would limit a company’s ability to increase leverage in the case of certain events have the potential to create an orphaning scenario. Therefore, companies would need to take out that debt in order to lever up the balance sheet. If the new debt is secured, issued out of another entity (without a cross guarantee to the original entity), or otherwise not eligible to be a reference obligation of the CDS contract, then the CDS is orphaned. This is an issue because if the reference entity has no debt, there is almost no probability of default, so spreads tighten significantly even should the entity have a lower rating.

10 11

Credit Derivatives Handbook

See footnote 1. Here we are assuming that Axel Springer would not eliminate the ProSieben entity and create a succession. Because of the regulator’s intervention, the actual structure of the acquisition was not revealed.

26

08 February 2007

Spread (bps)

Examples of actual and possible Exhibit 21: Cendant 5-year CDS orphaning in the US market include A: Company Revises Forecasts Lower; B: Firm says it will out Holding Company debt, issuing out of Operating Cendant, Equity Office Properties, Bausch take Companies without cross guarantees & Lomb, and RJ Reynolds. In these cases, 130 CDS contracts are likely to trade 120 considerably tighter than if the entity they 110 referenced were to lever up. Cendant, a 100 B name that we’ve written on several times 90 in relation to “succession” events, is also a 80 good example of potential orphaning and 70 A we provide the relevant CDS Spread 60 history in Exhibit 21. 50 40

3/28/2006

3/14/2006

2/28/2006

2/14/2006

1/31/2006

1/17/2006

1/3/2006

Orphaning is an important consideration for investors wishing to hedge exposure to bonds, as well as going outright short. For those playing basis trades in names they believe could be involved in M&A or leveraging activities, we suggest doing a Source: Credit Suisse P&L scenario analysis in the event the bonds are taken out (say in a make whole call) and CDS is orphaned. Another option is to purchase bonds trading at a discount that would likely be taken out at par or at a significant premium to the current price, where the gain on the bond can make up for any spread tightening on the orphaned CDS. It should be noted that the spread of current 5-year CDS of an orphaned entity should steadily move toward zero as time passes because the fraction of the 5-year period with no debt outstanding increases. The basis of the orphaned entity to the replacement financing entity should widen as time passes. Orphaning and Name Changes In recent history, two instances of orphaning occurred without M&A activity. Uncertainty was compounded by the nature of the name changes. On March 15, 2005, Rentokil Initial plc created a new company under the name of Rentokil Initial 2005 plc. On June 21, 2005, Rentokil Initial plc was renamed Rentokil Initial 1927 Plc; simultaneously Rentokil Initial 2005 plc was renamed Rentokil Initial plc. The new entity became the parent holding company, owning 100% of the old entity. Thus, a new entity was established with the same name as the original entity. On December 9, 2005, the company initiated a new EMTN program. The offering circular declared that Rentokil Initial plc (the new entity) would be the primary obligor for new issuance. On December 21, 2005, the company announced that the existing debt of Rentokil Initial 1927 plc would be transferred to the new entity but with a guarantee from the old entity. The last of such notes matured on November 19, 2008, at which point the guarantee would expire. That guarantee ensured that there would be no succession12; CDS on the old company would continue to reference the old company. The CDS on Rentokil Initial 1927 plc had been orphaned.

12

Credit Derivatives Handbook

Under the terms of the ISDA 2003 Agreement

27

08 February 2007

Exhibit 22: Rentokil Name Change Process

Newco:

Rentokil Initial 2005 Plc

Rentokil Initial Plc

Rentokil Initial Plc Bond

With guarantee

Oldco:

Rentokil Initial Plc

Rentokil Initial 1927 Plc

Bond

Bond

Pre 15 Mar 2005

15 Mar 2005

21 Jun 2005

Rentokil Initial 1927 Plc

21 Dec 2005

Source: Credit Suisse

The re-use of an existing name concealed the change and eventually caused confusion in trading. Ordinarily, the trade confirmation process would rely upon reference obligations (“refobs”) to clear up discrepancies, but the new entity had no outstanding debt to reference. In addition, some parties may have felt no urgency to resolve disputes because they expected a transfer of obligation from the original entity to the new one, prompting a proper succession. Currently, 5-year CDS on Rentokil Initial plc (the new entity) trades at 62.5, 35.0 bps wider than the 27.5 bps of Rentokil Initial 1927 plc. The difference between old and new spreads is now significant – and worth fighting over in court. The WPP case is very similar. On November 9, 2005, WPP Group Plc (WPPLN) was renamed WPP 2005 Ltd (WPPGRP-2005) and a new parent company was created with the name WPP Group Plc (WPPGRP) – the same name as the earlier entity. The key difference relative to Rentokil is that the WPP Treasury had stated intent, subject to Trustee approval, to transfer all debt from the original group, WPP 2005 Ltd., to the new group, WPP Group Plc. If this occurred and the original group did not retain any guarantor status, then a succession event would occur. As this is the market expectation, the CDS curves for the two groups would be the same. However, should that not occur, the original entity’s debt would be orphaned and its credit spread should tighten. If that should happen, however, the immediate tightening would be less dramatic than for Rentokil because there exists straight debt outstanding maturing in 2014 land convertible debt with an ending maturity of 2033. Caveats in Predicting Orphaning Risk The preceding conclusions might encourage analysts to commence the hunt for orphan risk. Listed below are three factors that should moderate those vigorous urges: • The effect of orphaning is minimal on entities with long-dated debt outstanding. For those entities, the probability of default may not be changed much in the future. • Any prediction of orphaning risk is uncertain. An orphaning is usually a byproduct of another transaction (an M&A deal or a capital restructuring) for which that orphaning is a minor consideration. This was particularly true in the case of the two name changes, where the principal motivation was a restructuring to meet IFRS (International Financial Reporting Standards). In addition, any state is reversible: an intended transfer/merger of debt may never occur, and an executed transfer/merger can be reversed. • While the rewards of identifying orphaning risk are great, the probability of orphaning has been small to date.

Credit Derivatives Handbook

28

08 February 2007

Credit Default Swaptions and Instruments with Optionality Credit Default Swaption Review Credit Default Swaptions (CDSwaptions) are generically a European-style option on a plain vanilla single-name CDS or index product, such as the Dow Jones CDX and iTraxx13 products. These instruments generally trade to the next two roll dates; so, for example, in April, both June and September expiry swaptions would be available. Although dealers may quote a range of strikes (strictly speaking, any strike can be used), At-the-Money or Near-the-Money strikes are most common. Like all instruments with optionality, CDSwaptions trade with an implied volatility. Unlike bond or equity options, CDSwaptions trade using the terms Receiver and Payer in the same light as interest swaptions. In general, a Payer swaption gives the purchaser the right, but not an obligation, to buy a CDS at expiry, and a Receiver swaption gives the purchaser the right, but not an obligation, to sell a CDS at expiry. We will define these in more detail below. These instruments are used for several reasons, including expressing a directional view on a credit or the market and taking a view on volatility. First, they are used to hedge the risk of a position until after some expected event date (earnings, court case, etc). Second, investors use swaptions to take views on the credit; for example, investors selling a payer swaption if they think spread volatility should decline (selling volatility), or buying one if they think implied volatilities look cheap (buying volatility). Although the market is still maturing and liquidity might be somewhat spotty (especially in single-name CDSwaptions), it is also possible to perform most option strategies, such as straddles, strangles, or calendar spreads, but with the knowledge that unwinding the trade might be costly. In our view, the strategies work well if held until maturity. Below we go into more detail on the uses and terminology of Credit Default Swaptions. Uses: Hedged view on a credit/an index. The maximum downside in a long position in a swaption is the loss of the premium. So, for an investor who anticipates that some event will cause the market to rally/sell-off sharply but is concerned about the risk in an outright long/short position (because the event may turn out differently than anticipated), a long position in a CDS option is often an efficient way to express one’s view. Shorting volatility. Investors can sell a straddle and capture the premiums as long as spreads do not move significantly by the exercise date. In the recent low vol environment, this strategy has drawn some interest from hedge funds that fear a continuation of a low vol environment will limit their trading opportunities (so the short vol view provides some income in that environment).

13

Credit Derivatives Handbook

iTraxx is a trademark of International Index Company Limited.

29

08 February 2007

View on implied vols. With increasing liquidity and tightening bid-offer, investors can increasingly use CDS swaptions to express their views on volatility. Additionally, investors can count on a company-specific event to change current implied volatility. Alternatively, they can compare implied volatility to historical spread realized to come up with their view on cheapness/richness of swaptions. Implied spread vols versus implied equity vols. Using the Merton framework, there should be a relationship between implied spread vols and implied equity vols. This relationship should be tighter for lower credit quality names (although there tends to be a breakdown in this relationship once the credit becomes highly distressed). Terminology: Payer credit default swaption. Buying a payer swaption gives the holder the right, but not an obligation, to buy protection on the exercise date with a vanilla CDS contract at a spread equal to the strike price. So the holder has the right to pay the strike price for a short position in a given credit. If, on the exercise date, spreads on the credit are wider than the strike price, the holder will exercise the option (through physical settlement) and either retains the short in the credit or immediately unwinds protection at the market rate and takes out the differential. If spreads are tighter than the strike price on the exercise date, the holder will allow the option to expire because the holder could buy protection in the open market at a cheaper rate; hence, its intrinsic value is equal to zero. Receiver credit default swaption. Buying a receiver swaption gives the holder the right, but not an obligation, to sell protection at the strike price on the exercise date. So the holder has the right to receive the strike price for selling protection through a vanilla CDS contract. If spreads on the relevant credit are tighter than the strike price on the exercise date, the holder will exercise the option (through physical settlement) and either retains the long position in the credit or immediately unwinds protection at the market rate and takes out the differential. If spreads are wider, the holder will allow the option to expire; hence, its intrinsic value is equal to zero. Breakeven spread. The premium, or upfront fee paid to be long the option, is typically expressed in terms of basis points of the notional. The breakeven spread is how far rates on the underlying CDS need to move in favor of the option holder in order to offset the premium cost. So the breakeven is the market spread where: PV of (breakeven spread – strike price) * forward DV01 = Premium Knockout. Single-name credit default swaptions typically have a knockout feature, such that if there is a credit event prior to the exercise date, the option expires and the holder no longer has the right to buy/sell protection at the strike price. The seller of the option keeps the premium. Due to this knockout feature, using the underlying CDS to delta hedge the option leaves some exposure. Credit default swaptions on an index (CDX, iTraxx) typically do not have a knockout feature. So if there is a credit event in one of the names in the index prior to the exercise date, the holder of a long position in a payer swaption has the right on the exercise date to buy protection on the index at the strike price, and then immediately deliver the relevant obligation, and receive payment for the default.

Credit Derivatives Handbook

30

08 February 2007

If there were a credit event in one of the index members prior to the exercise date, the holder of a long position in a receiver swaption would only exercise the option if there were a dramatic tightening in the rest of the index. Because an exercise of the option would trigger a payment for the credit event, it only would make sense to exercise if the gains from taking out the strike price/market rate differential compensate the holder for the payment for default. So there is exercise only if: 1/N * (1-recovery) < (N-1)/N * DV01 * (strike – market rate) N = number of components in the index Accordingly, it is unlikely that a receiver swaption on an index would be exercised if there were a credit event prior to the exercise date. Liquidity Issues Liquidity tends to be correlated with liquidity in the underlying credit (or index) because a liquid underlying makes it less costly for a market maker to hedge positions. In addition, liquidity tends to be better for options that are reasonably close to being at the money (where the strike price is close to the forward), as the risk in a short position in far out-ofthe-money “lottery tickets” is somewhat more difficult to manage.

Cancelable CDS and Other Exotics Cancelable CDS protection gives the buyer a one-time opportunity to cancel the protection on the specified date. Effectively, buyers of cancelable CDS get into a short credit position overlaid with a slight Out-of-the-Money receiver CDS option (assuming an upward-sloping credit curve). The buyer of protection gets the flexibility to cancel the contract if spreads tighten, while the seller gets additional running premium for accepting this short-volatility position. The main risk of purchasing a CDS option is that you will lose the entire premium paid and the risk of selling an uncovered CDS option may result in losses significantly greater than the premium received. For example, an investor buys SRAC’s 5-yr CDS cancelable in three months for 180bps, while the current level is at 150bps (effectively one is paying an additional 30bps running premium for three months to get the right to cancel the contract). If SRAC’s 5yr CDS spreads tighten by more than 4bps in three months, investors would have been better off buying cancelable CDS rather than normal CDS because they can cancel the trade and reenter at the tighter spreads. We are aware of a number of exotic structures, which package CDS and CDS options to come up with various P&L profiles/features, suitable for different investors’ tastes. “GreenShoe” CDS An investor buys 5-yr protection at the regular offer level, but leaves an order to upsize the trade at a specified time (effectively, giving the seller the option to increase the trade size). In return, all CDS premiums on the original CDS and any upsized protection are waived until the option expiration. Investors effectively buy CDS protection and sell out-of-the money receiver options. The premiums are matched to provide zero carry on the CDS until option expiry. The risk of selling a CDS option may result in losses significantly greater than the premium received.

Credit Derivatives Handbook

31

08 February 2007

Cancelable CDS with Exercise Schedule This structure is mostly used to hedge bonds with call schedules that start at prices above par and decline towards par as maturity approaches (most of these bonds are in the highyield universe). Here the buyer of protection has the option to cancel the CDS on the first call date for a fee equal to the difference between the price of the bond and par, and so on to mimic the risk in the bond accurately. The risk of purchasing a CDS option is that you will lose the entire premium paid. The risk of selling an uncovered CDS option may result in losses significantly greater than the premium received.

Implied vs. Historical In the interest rate options markets, implied volatility on short-dated options tends to converge toward historical volatility over time, as volatile markets generate expectations of volatility and stable markets beget expectations of stability. With options on CDS, we would expect implied vols to be higher than historical vols on a somewhat more persistent basis because there is more gap risk in options on credit spreads (the tail is fatter than it is on interest rate options). However, for a number of names, we have seen very large gaps persist for an extended time. In particular, it appears that the more liquid credits that are frequently used to hedge portfolio risk, like the autos, tend to have large, persistent gaps between historical volatility and implied volatility. We suspect that this bias is due to investors buying options on these more liquid credits to hedge against sharp movements in market spreads. Portfolio hedging flows should push up option prices on these liquid credits, creating an upward bias in their implied volatilities.

Where’s the Skew? The market for credit default swaptions has evolved to the point where we are seeing a more developed term structure for volatility, and recently, we have begun to see a small vol skew in selected credits. In this market environment, we would expect to see less skew than normal, since many investors are looking at selling payer swaptions as a way to add carry. However, we think that the main reason skew is currently small is because the volatility surface of the swaptions market is not fully developed. As the market develops, we expect to see: • Skews that become larger with wider credit spreads; and • Skews that become larger with higher expected recovery rates on the credit. Using a Merton model framework, credit spreads are a function of the out-of-themoneyness of the “default option.” Since long positions in options have positive gamma, an equity rally and sell-off of the same magnitude will have unequal effects on spreads, with the sell-off triggering a larger spread move than the rally. Therefore, CDS swaptions vols should be skewed. Since the implied probability of default equals spread divided by loss given default (one minus recovery), we should see swaption skew increase both as spread increases and as expected recovery increases.

Credit Derivatives Handbook

32

08 February 2007

Options Embedded in Bonds In this section, we discuss a simplified approach to discover relative value in call and put embedded options. Embedded Call/Put Relationship Any callable / puttable bond can be viewed in two ways. If we consider a 15-year bond callable in five years, the owner essentially owns a package of a long position in a 15yr bullet bond and a short position in a 5yr call option. Alternatively, the callable bond can be looked at as a package of a long position in a 5yr bond and a short position in a 5yr extension (put) option for a 10yr bond. In both cases, the bond issuer decides whether to terminate the callable bond in five years or to let it mature in 15 years. Therefore, the two packages result in identical payoffs and should be priced equally: P(15 yr bullet bond) – P(call) = P(5yr bullet bond) – P(put) The same line of reasoning applies to a put bond – a 15yr bond with embedded 5yr put option can be looked upon as a 5yr bullet bond with an embedded 5yr extension (call) option, so that: P(15yr bullet bond) + P(put) = P(5yr bullet bond) + P(call) For the purposes of our analysis, we represent the bonds with embedded options as their shorter life bullet counterparts with the corresponding extension options exercisable at the bullet bond maturity. Market Price of the Option To evaluate what fraction of the dollar price of the option-bearing bond is attributable to the option, we compare the yield-to-put of the bond to a synthetic bond with the same maturity as the put date (the CDS spread plus the swap rate). The residual spread reflects what the investor actually pays / receives for the extension option. The dollar price of the option is the dollar value of the residual spread (the present value of the all future payments discounted at the bond-specific risk rate). To make a relative value call, however, we need to determine the theoretical price of the option. Pricing options on corporate bonds is difficult because the price depends on interest rate volatility, spread volatility, and the correlation between rates and spreads. But, as an input for a quick and dirty initial screen of puttable/callable bonds to find those that might be interesting, we can ignore the value of the spread option and focus on the interest rate option. So, we value a European interest rate swaption with an exercise date equal to the embedded option date, a maturity equal to the bond maturity, and a strike price equal to the bond coupon less the corresponding forward CDS spread. Example (data as of May 18, 2005) As an example, we analyze the STI 6 28s, puttable on 1/15/08 at 100. The bond is equivalent to an STI 6 01/15/08s bullet bond plus an extension (call) option expiring on the maturity date. The bond price used for this analysis is105.04. Yield to Put date = 4.54% Swap rate = 4.03 CDS to Put date = 20bps The residual spread = Yield-to-Put less Swap less CDS-to-Put = 31bps

Credit Derivatives Handbook

33

08 February 2007

The present value of the residual spread is $0.78. Since it is a positive value, this indicates that the investor is getting paid to own the extension option on this bond. Pricing on bonds with embedded options is rarely this far away from fair value, but this simple framework can be used to screen for options that may be mispriced.

Constant-Maturity Credit Default Swaps Credit default swaps are often described as being economically analogous to a levered position in a floating-rate note. Constant-maturity credit default swaps can be thought of as a levered position in a floating-rate note in which the credit spread also floats. They are useful for buying or selling protection with minimal mark-to-market pain, and for making curve trades. On the other hand, outright long/short positions in constant-maturity CDS are not appropriate for expressing one’s views on the direction of credit spreads rather than the change in the credit curve slope and one is better off trading vanilla CDS instead. A constant-maturity CDS (CMCDS) uses the same default triggers and payouts as a vanilla CDS, with the only difference being the fee that the protection buyer pays. Rather than the fixed fee that vanilla CDS uses, in its most common form, a constant-maturity CDS fee resets quarterly to a fixed percentage of the actively traded 5-year CDS rate. For example, a CMCDS on Eastman Kodak might pay 80% of EK CDS to June 2010 today (actually one-quarter of that because payments are made quarterly rather than annually), and then 80% of the September 2010 EK CDS three months later, followed by 80% of the December 2010 rate, and so on. Coming up with the fixed percentage is fairly simple, as the percentage is used to offset the expected value (or cost) of constantly rolling to the new active 5-year point. Since the payout on default in both a CDS and a CMCDS is identical ($100 minus recovery), the expected cash flows from the fees should be the same. So, the CDS curve is used to generate a string of forward 5-year CDS rates. Then a convexity adjustment is applied to it. The present value of the string of payments, assuming these forwards hold, is compared to the PV of a fixed 5-year CDS. The ratio of the PV of the vanilla CDS over the PV of the CMCDS is the fixed percentage. So, for a positively sloped credit curve, the percentage will be below 100%, and for a negatively sloped credit curve, the percentage will be above 100%. We should also mention that normally there is a cap on the underlying CDS at something like 750bps, so that the maximum CMCDS fee is the cap times the fixed percentage. This is because at very wide spreads CDS begins trading with points upfront, and translating this fee to a spread is sensitive to the assumed recovery rate (in addition, prices tend to be less stable and transparent at extremely wide spreads). The cap avoids debates about the right implied spread when an upfront fee exists. For investment grade names, the cap is normally so far out of the money that it has an extremely small impact on the CMCDS valuation. But as spreads approach the cap, and as the implied volatility in the credit increases, the cap value can become extremely important to the overall CMCDS price. As is the case with any floating-rate instrument, the mark-to-market effect on CMCDS of parallel spread changes is rather small. However, CMCDS prices are sensitive to changes in the slope of the credit curve. A steeper curve implies higher forward spreads, so the expected stream of fees for CMCDS would be higher. So, a CMCDS benefits from curve steepening, including any steepening in the curve out to the maturity of the swap plus the maturity of the reference CDS (a typical 5yr CMCDS with 5yr reference CDS would be sensitive to the CDS curve out to 10 yrs).

Credit Derivatives Handbook

34

08 February 2007

This sensitivity to curve shape makes CMCDS a useful tool in curve trades. By using CMCDS, investors avoid the two main weaknesses of curve trades using vanilla CDS – the need to rebalance the weighting on the trade constantly in order to remain DV01 neutral (trading odd-lots is rarely fun), and the residual exposure to default risk from the different nominal weightings and different maturities of the two legs. Trading Strategies Sell (buy) CMCDS if expect curve to steepen (flatten). Sell CMCDS, buy CDS. Benefits most if curve steepens and spreads widen, with no net exposure to default. Sell CMCDS rather than CDS if expect 5yr spreads to be higher than the forwards (so want spread widening without significant default risk). Buy CMCDS to hedge loan exposure, reducing mark-to-market effects. All three strategies are not appropriate for expressing one’s views on the direction of credit spreads rather than the change in the credit curve slope and one is better off trading vanilla CDS instead.

Introducing CDS Variance Swap CDS Variance Swap, referencing CDS indices in the US and Europe, have been traded by Credit Suisse since January 2006. We outline the main features and applications of the product. Product Features Variance swap is a contract in which two parties agree to exchange cash flows based on the realized variance of the specified underlying asset at contract expiration14. Long Variance Swap Position: Payout = Notional * (Realized Vol ^ 2 – Strike Vol ^ 2)15 Short Variance Swap Position: Payout = Notional * (Strike Vol ^ 2 – Realized Vol ^ 2) CDS variance swaps are similar in many ways to equity variance swaps, which have been actively traded in the marketplace for some time. Several notable peculiarities of CDS product should be noted: • CDS variance swap will specify a CDS index of corresponding maturity. To avoid doubts, the version of the index will be fixed through the transaction regardless of the occurrence of credit events. • If a new version of the index is announced following a Succession Event, the new version should become thereafter the index for the transaction. • In Europe iTraxx Variance Swap uses IIC/Markit End of Day fixing for realized volatility, while in the US Credit Suisse is the Calculation Agent for the product. The value of the product depends on realized, implied volatilities as well as the passage of time. Buyers of variance swaps are long vega – a direct function of contract’s strike volatility. Variance swaps have a non-linear P&L profile: buyers are long convexity, while sellers are short (See Exhibit 23). 14

To protect variance swaps sellers from extreme market events, contracts would have volatility caps (similar to equity product). Therefore, swap payout will reference the lower of cap or realized vol. 15 Realized Vol = SQRT [(252 / N -1) * Sum {Ln (Ci / Ci-1) ^ 2}], where Ci – daily index close, N – number of valuation days. The realized volatility is calculated upon termination of the contract. Strike Vol – volatility level specified in the contract

Credit Derivatives Handbook

35

08 February 2007

Exhibit 23: P&L Profile of Long and Short Variance Swap Positions Contract Notional: $1mln. Swap strike: 30%.

$300,000 Strike Vol

Swap P&L

$200,000 $100,000 $0 -$100,000 -$200,000 -$300,000 12%

18%

24%

30%

36%

42%

48%

54%

Realized Volatility P&L (Swap Buyer)

P&L (Swap Seller)

Source: Credit Suisse

Trades and Applications CDS variance swap is a “clean” way to express one’s view on the direction of CDS index volatility. Contrary to CDS option positions, variance swap investors should not worry about delta-hedging their positions with respect to changes in the underlying indices. Some main applications of the product are: • Directional volatility trades, including possible forward vol trades when term structure develops. • Relative value volatility trading across various indices, US vs. Europe and asset classes (debt versus equity). • Portfolio vega (volatility) hedging • Tailoring of P&L risk/return profiles and diversification. Below we provide two specific examples: Directional view Most often implied volatility is higher than historical vol over the life of the option. Therefore, investors can utilize the short variance swap potion to capture this difference (See Exhibit 24). Portfolio vega hedging In this hypothetical example, let us assume that investor’s portfolio is short volatility with a vega position of -$100,00016. Assuming variance swap is struck at 34% (current 3-mo implied vol) and the investor expects 10% volatility increase over three-month horizon. If the prediction is correct, the portfolio should lose $1mln = $100,000 * 10. To hedge it effectively, one needs to buy a variance swap with a notional of $12.82mln = $1mln / (0.44^2 – 0.34^2).

16

Credit Derivatives Handbook

Portfolio should lose $100,000 per 1% increase in volatility

36

08 February 2007

Exhibit 24: Comparison of realized, implied and variance swap vols (XO6) Realized vol – same term as implied vol. Pricing is indicative.

50.0%

Volatility (%)

45.0% 40.0% 35.0% 30.0% 25.0% 20.0%

Realized Vol (3mo)

Implied Vol

11/22/2006

11/8/2006

10/25/2006

10/11/2006

9/27/2006

9/13/2006

8/30/2006

8/16/2006

8/2/2006

7/19/2006

7/5/2006

6/21/2006

15.0%

Variance Swap Strike

Source: Credit Suisse

Potential Risks While CDS variance swaps can become an attractive instrument for CDS vol investors, several notable risks should be kept in mind: • Idiosyncratic event risk is high, possibly even higher than for equity variance swaps. This risk includes credit and/or succession events and potential LBO activity (especially for HiVol with its high concentration of LBO candidates). • Investors should be aware that liquidity of variance swaps is driven by market conditions and could somewhat dry up during stress times. • Variance swaps are relatively hard to hedge and require good liquidity for out-of-themoney CDS options. To be completely independent of the index spread moves, one needs to construct a properly weighted portfolio of payers and receivers CDS options across all strikes. Therefore, the value of variance swaps is affected by changes in volatility surface, i.e., skew flattening and steepening.

Recovery Trades. Digital CDS In a digital trade, the seller pays par on a credit event and there is no delivery of bonds. This makes it effectively like a vanilla CDS where the recovery value is equal to zero. Digital CDS is part of the broader class of customized CDS with fixed recovery, i.e., while entering the trade, the two parties specify the recovery in case of a credit event. There are two main reasons for using digital CDS in place of vanilla CDS. First, the digital offers higher leverage (assuming a par payout on the digital). Second, investors who expect the actual recovery rate to differ from the market-implied recovery rate on vanilla CDS might find the digital a cheaper way to buy or sell protection. Mark-to-market volatility for outright positions is the main risk for this trade. To examine the leverage in a digital, we use the following simplified relationship: CDS spread ≈ Expected probability of default x (1 – expected recovery rate)

Credit Derivatives Handbook

37

08 February 2007

Since the expected recovery rate for the digital is zero, its spread will be wider. The ratio of the digital spread to the vanilla spread should be close to 1 / (1 – expected recovery on the vanilla) and the spread on the digital will be more volatile. If we assume that all changes in the CDS spread are due to changes in the expected probability of default (there are no changes in the expected recovery rate), then for a 30% expected recovery rate, a 1bp movement in the vanilla CDS spread should cause a 1.43bp movement in the digital spread 1/(1-.3).

Exhibit 25: Moody's Recovery Rates by Industry from 1982 to 2003 Industry

Recovery Rate

Utility-Gas

51.5%

Transport-Air

Industry

Recovery Rate 34.3%

Oil and Oil Services

44.5%

Automotive

33.4%

Hospitality

42.5%

Healthcare

32.7%

Utility-Electric

41.4%

Consumer Goods

32.5%

Transport-Ocean

38.8%

Construction

31.9%

Media, Broadcasting, Cable

38.2%

Technology

29.5%

Transport-Surface

36.6%

Real Estate

28.8%

Finance and Banking

36.3%

Steel

27.4%

Industrial

35.4%

Telecommunications

23.2%

Retail

34.4%

Misc

39.5%

Source: Moody’s

However, it is not identical to a larger notional position because the recovery rate is not fixed for vanilla CDS. Frequently, the CDS market prices in little differentiation between expected recovery rates for different credits and for different industries, but historical default rates show significant deviations in recovery rates by industry. As the table above shows, the historical recovery rate for Gas Utilities is more than double the rate for Telecom companies. One has to keep in mind that since CDS sellers hold the option to deliver the lowest price bond in the delivery basket, the average price of a bond delivered on a CDS contract would probably be slightly below the average recovery rates listed above.

Credit Derivatives Handbook

38

08 February 2007

Capital Structure Arbitrage Capital Structure Arbitrage (CSA) is not a pure arbitrage, but a relative value strategy that attempts to capture pricing inefficiencies across a firm’s capital structure. Below, we describe, in detail, some of the most popular types of “cap arb” strategies: debt vs. equity and recovery trades (senior vs. subordinate, OpCo vs. HoldCo).

Debt vs. Equity. Potential to Win in the “Wings” With the growing number of hedge funds and other capital structure arbitrage players, wings trades have become very popular in recent years. In a typical wings trade, an investor goes long a dividend-paying equity and buys CDS to protect a long stock position. Wings trades, in our view, are ideal for investors looking for fat tail distributions where they benefit if a company’s situation dramatically improves or significantly deteriorates, providing a straddle-like payout profile. If a strong positive event takes place, the upward equity movement overshadows the losses on the CDS hedge due to the asymmetric nature of the debt-equity dependence implied by the Merton framework. On the other hand, if a negative event occurs, gains on the credit hedge should drive the trade’s P&L. Exhibit 26 presents a hypothetical wings trade P&L for two companies with characteristics similar to companies actually trading in the market, with the main difference being that one company has a wide credit spread and the other has a tight credit spread. Based on equity moves, CDS changes are projected by debt-equity relationships implied by Credit Suisse’s CUSP® model. The cost of equity financing is not taken into account, allowing for a zero P&L with no change in spread and hedging is assumed to be carry-neutral (all dividend proceeds are used to purchase CDS). From the exhibit, we see that losses on the theoretical trade occur only at small to medium stock price drops, when the CDS hedge is unable to compensate for the losses on the equity leg. In all other circumstances, the trade appears profitable. There are a number of situations where wings trades are appropriate. Companies facing litigation resolutions, expectations of positive earnings surprises, stock buybacks or other pending actions that favor shareholders over bondholders are just a few examples of when one might benefit from a wings trade. Additionally, in an LBO, investors could win on both legs of the trade on an LBO where the stock rallies and spreads widen. Recently, there have been a number of LBO situations where we believe wings trades would have proved advantageous for investors. For example, during the recent LBO of Sabre Holdings in December 2006, spreads widened from 100bps to 320bps and the stock price rallied over 12% during the three-day period when news reports were released and then confirmation of a buyout of Sabre Holdings by a group of private equity investors occurred. The main technical requirements of the trade are: • Sizable equity dividend, with little danger of being cut • Equity trading at the low end of its price range • CDS spread trading at tight levels Essentially, we are looking for cheap equity/rich debt situations where dividends finance or subsidize the CDS side of the trade. One way of gauging for such situations is by looking for dislocations in the views of equity and credit analysts where there is a positive view on the stock and a negative view on the company’s credit. Of course, these situations are not very common.

Credit Derivatives Handbook

39

08 February 2007

The main risk to these trades is a breakdown in debt-equity correlation where one asset stops moving in correlation to the other. If a wings trade is performed for a tightly traded name and equity slowly declines, CDS might not widen at all, and even if it does, the gains on the CDS might not be enough to offset the losses on the equity side. Debt-equity correlation should hold more consistently for more CUSPy credits at wider spreads, but then hedging becomes more costly. Therefore, we suggest taking a balanced, situationoriented approach.

Exhibit 26: P&L for Typical Wings Trades Theoretical Result. Assumed no funding costs and carry neutral hedge

30%

Wings P&L (%)

20%

10%

0%

-10%

-20% -50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

Equity Return (%) Low Spread Credit

High Spread Credit

Source: Credit Suisse

The Trickiest Thing in Cap Arb: Hedging There are four main ways to hedge debt versus equity trades: 1)

Default Neutral

2)

Carry Neutral

3)

Empirical DV01 Neutral

4)

Merton Model-based DV01 Neutral Estimate

Default NeutraI In a default-neutral hedge, the underlying objective is for the hedge to earn back whatever is lost on the risk side of the trade in the case of a default event. Once an investor has decided to either go long or short credit risk and hedge the exposure by buying or selling equity puts, the investor needs to figure out the hedge ratio. One assumes the residual equity price and bond recovery in a default event and calculates the potential loss on the risk side and the gain on the hedge. The ratio of the two will provide the appropriate hedge ratio for the trade.

Credit Derivatives Handbook

40

08 February 2007

Consider the following example: GM’s company’s stock had traded at $31.6, and CDS with expiration to Jan 06 was at 375 bps and the investor believed that the CDS was cheap and wanted to sell protection (data as of May 18, 2005). Assume that the residual equity price was $1 and the bond recovery in case of a default was 40%. Also, assume that the Jan 06 out-of-the money puts struck at $10 were available for $0.25. If GM defaulted the next day, the loss on the credit risk side would be $600 per $1000. On the hedge, the gains would have been $900 per put. The hedge ratio would be 600/900, or 0.667 puts per bond. The risk of purchasing a call/put is that the investor will lose the entire premium paid. Exhibit 27 provides the P&L and a simple breakeven analysis for the example trade.

Exhibit 27: GM Wings trade using default-neutral hedge ratio The risk of purchasing a call/put is that you lose the entire premium paid. Data as of May 18, 2005.

Notional / # of Shares

Ticker

Instrument

Spread (bps) /

Upfront

Maturity

Price ($)

Payment

Breakeven

maturity)

375bps $166,675

Takes 6.5 months to

$83,000

Sell

$10MM

GM

CDS

Jan 2006

Buy

6,667

GM

Put @ 10

Jan 2006

Total

P&L (at

pay for options Source: Credit Suisse

As one would expect, default-neutral hedging works satisfactorily only when the issuer either defaults or the trade is kept on until the CDS/options mature. In GM’s case we actually saw the hedge failing on a mark-to-market basis because GM’s spreads widened dramatically, while the equity changed only slightly during the same time. Therefore, a default-neutral hedge is probably more appropriate for investments in CUSPy companies rather than in investment grade (IG) or Double-B companies or for investors who prefer to hold the trade until the options expire. Carry Neutral Another popular hedging strategy is to be carry neutral. Here investors take the proceeds of one instrument over the lifetime of the trade and use them to buy a hedge. The main assumption here is that the price/spread return of the long position should outperform that of the short (hedge) position. Wings trades are probably the best example of carry-neutral trades. For example, an investor wants to execute a wings trade by purchasing equity and using the expected dividend to buy protection on the same issuer. As an example, we will use a trade idea on Reynolds American (RAI) that we suggested in a Trading Edge published on November 6, 200617. At the time, RAI’s equity price was $63.7, the dividend yield was 4.7% and 5-year protection was available at 100 basis points. With an annual dividend income of $3 per share, the income from 33,400 shares could buy protection on $10 MM of notional.

Exhibit 28: RAI Wings trade using cash neutral hedge ratio of 4.7:1 Data as of November 6, 2006

Notional / # of

Spread (bps) /

Shares

Ticker

Instrument

Maturity

Price ($)

DV01

Buy

10M

RAI

CDS

5Yr

100

4.4

Buy

33,400

RAI

Equity

63.7

Source: Credit Suisse

17

The Credit Suisse Trading Edge of November 6, 2006 can be found at http://d.research-andanalytics.csfb.com/getdoc.asp?docid=38268577

Credit Derivatives Handbook

41

08 February 2007

The biggest problem with a carry-neutral risk regime is that it is, as the name implies, carry neutral. There is no reason for the risks to be fully or substantially hedged. In the above example, if the equity price drops 5% it is quite possible that the spreads may not move at all. The investor would lose out on several hundred basis points with no cushion from the hedge whatsoever. Empirical DV01 Neutral In today’s environment an empirical DV01 neutral hedging strategy may be somewhat better. Here one looks at the relationship of debt and equity returns over the past 3m/6m/1yr and goes long the instrument that one determines to be cheap and hedges exposure with the more expensive instrument. We again take a look at the Reynolds American (RAI) trade that we highlighted earlier. To determine the empirical hedge ratio, we evaluated the correlation of equity and debt returns by calculating monthly equity and 5yr CDS spread returns over a six-month period. By running a simple regression, the hedge ratio was estimated at approximately 3.2. This results in a trade with a positive annualized carry of $46.875 for $10MM notional.

Exhibit 29: Monthly Equity vs. Spread Returns Data as of November 6, 2006, based on 6 month historical observations

20.0% 15.0% y = -3.2052x + 0.0202 Spread Returns (%)

10.0% 5.0% 0.0% -5.0% -10.0% -15.0% -4.0%

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

3.0%

4.0%

Equity Returns (%) Source: Credit Suisse

Exhibit 30: RAI Wings trade using historical hedge ratio of 3.2:1 Data as of November 6, 2006, based on 6-month historical observations

Notional / # of Shares

Ticker

Buy

$10MM

Buy

49,000

Instrument

Maturity

RAI

CDS

5Yr

RAI

Equity

Spread (bps) / Price ($)

DV01

100

4.4

63.7

Source: Credit Suisse

There are two main risks in this sort of trade that might lead to substantial mark-to-market P&L volatility. First, the relationship observed in the past might or might not remain stable going forward. Second, even if the hedge ratios do hold over the long run, they may not be consistent in the short run and/or with small price movements.

Credit Derivatives Handbook

42

08 February 2007

Merton Model-Based DV01 Neutral Estimate Finally, one can use Merton-type models, such as Credit Suisse’s CUSP, which assumes a forward-looking debt-equity relationship based on the company’s fundamentals (balance sheet data), credit spreads, stock prices and equity option-implied volatility. In CUSP, for example, we compute the Spread Widening Risk for every company in our universe and for a one-standard-deviation change in equity price; the SWR provides an estimate of the likely change in a company’s spread. By converting the equity volatility in dollar prices, one can easily come up with the appropriate hedge ratio for any equity/debt trade. A model-based hedge ratio clearly has a better theoretical foundation and presumably will provide better estimates for out-of-sample price behavior. Again, the main risk here is breakdown of debt-equity relationships for small price/spread moves. This is especially true in times when companies are increasing leverage, especially when debt financed share repurchases (i.e., levered recaps) and/LBOs are common. In conclusion, it becomes quite clear that there is not one single “good” hedging regime that is applicable to all circumstances. An investor’s challenge is further magnified by the fact that most of the time different approaches produce very different hedge ratios.

Other Capital Arbitrage Flows to Watch While traditional capital structure arbitrage activity generates a meaningful percentage of the flow in the CDS markets, it seems that slightly more complex structures are also becoming common. Traditional sell CDS/buy puts trades are likely to remain important in the “cap arb” arena, but these other structures could periodically generate enough flow to impact certain credit spreads on a selected basis. Accordingly, we review several structures below. Straddles, Strangles, Put and Call spreads, and CDS Rather than simply coupling CDS with puts, investors can couple CDS with a wide variety of equity option strategies. Strangles are likely to be particularly appropriate hedges for CDS if investors are concerned that the credit spread might widen either due to a drop in revenue (which would be likely to cause the put to be in the money), due to a takeover from a weaker credit (which would be likely to put the call in the money), or a levering event such as a recap or LBO. Put and call spreads can be coupled with CDS to create different payout profiles than would be available by using straight equity options alone. For example, if investors buy 1000 puts at $25, sell 2000 puts at $10 and sell $1MM shortdated CDS, with close to flat or positive carry on the trade, they should be protected in default (assuming 50% recovery on CDS). If equity trades below 25 they should benefit from a positive P&L. All of these option strategies can be implemented via outright or delta-hedged transactions. The latter provides the most benefit when option volatility increases/drops, affecting option pricing; therefore, it expresses one’s view on the option volatility at the time of trade execution. In many cases both hedged and unhedged trades can experience significant mark-to-market volatility. Please note that the risk of purchasing a call/put is that the investor will lose the entire premium paid. The risk of selling an uncovered put is significant and may result in losses significantly greater than the premium received. Transaction costs may be significant in option strategies calling for multiple purchases and sales of options, such as spreads and straddles. Commissions and transaction costs may be a factor in actual returns realized by the investor, and should be taken into consideration.

Credit Derivatives Handbook

43

08 February 2007

Hedging vol convergence trades with CDS The companies involved in a takeover or a merger should see the implied volatility in their equity options and their CDS converge as they become one company, assuming no large cash component of the deal. However, if the acquisition is not completed, the implied vols and the CDS could be significantly different. Investors can hedge a vol convergence trade by guarding against a breakup with CDS. These trades are most appropriate when the two companies started with significantly different vols and spreads. From the hedge ratio of options to CDS, investors have the ability to identify attractive opportunities. “Wings in a box” As we have mentioned before “Wings” trades are most appropriate for companies whose distribution of outcomes have fat tails (the equity is likely to either rally sharply or fall to a point where credit spreads are dramatically wider). Given the scarcity of sharp price swings in either the equity or the debt markets, investors may want to consider coupling a long and a short wings position as a relative value tool. In general, this strategy is appropriate for credits that have sharply different (and stable) dividend yields, similar CDS spreads, and where the equity of the firm with a higher dividend looks cheap (or its CDS looks rich). Wings trades tend to perform poorly when there is a change in capital structure that distorts the traditional debt/equity relationship (e.g., stocks sell off and CDS tightens). Debt-equity hedge for LBO candidates If one gets into a trade early enough, an LBO could present an attractive capital arbitrage opportunity. The only significant difference of an LBO debt-equity trade from a normal debt-equity hedge is that in an LBO situation, credit spreads widen and equity rises, while spreads and equity tend to move in opposite directions. Depending on whether equity or credit is leading the LBO-driven move, investors can utilize different trading strategies. If the stock price increased significantly more than spreads, investors should buy CDS and puts (or short stock). If credit is leading the move, one is better off selling CDS and hedging exposure by buying equity (or calls). The challenge is getting in early enough. Since the LBO of HCA in mid-2006, the simple mention that a name could make an attractive LBO candidate has sent spreads materially wider and equity prices higher. However, if one can stomach the mark-to-market volatility and get into trades early enough, profits can be made (see the Sabre Holdings example above). Another risk in such trades is that the purported LBO doesn’t materialize. The long credit and long equity position, in such a situation, carries a large residual market risk. Therefore, in many cases, it does make sense to abstain from putting on an “LBO trade” unless it is positive carry (possible a Wings trade as one example). While not essentially “capital arbitrage”, CDS curve steepeners are often an attractive means of getting positive or flat carry for names that have levering or LBO risk associated with them. Senior vs. Subordinated Trades A common starting point to determine relative value between senior and subordinated bonds/CDS is to plot a graph of potential senior recovery rates versus the implied subordinated recovery rates to see which senior recovery rates imply “reasonable” subordinated recovery rates. Investors who have a strong view on recovery rates should feel comfortable using this methodology to determine relative value. As a starting point, we once again return to the approximation that: Spread ≈ default probability * (1 – expected recovery)

Credit Derivatives Handbook

44

08 February 2007

Applying this equation to both the Senior and the Sub CDS and equating the default probabilities we arrive at: Sr. Spread / (1 – Sr. Recovery) = Sub. Spread / (1 – Sub. Recovery)

Exhibit 32: Implied subordinated spread at 40% senior recovery rate

100% 80%

250 230 210 190 170

Subordinated Recovery Rate

Senior Recovery Rate Source: Credit Suisse

8% 12 % 16 % 20 % 24 % 28 % 32 % 36 % 40 %

150 0%

10 0%

80 %

60 %

40 %

20 %

-80% -100%

270

4%

Implied Subordinated Spread

290

60% 40% 20% 0% -20% -40% -60%

0%

Implied Subordinated Recovery Rate

Exhibit 31: Implied subordinated recovery rate

Source: Credit Suisse

For a credit trading at 160 Senior and 285 Sub CDS, one can plot implied Sub recovery vs. Senior recovery (Exhibit 31) and find that the implied Sub recovery is negative, which suggests an existing mispricing of either Senior or Subordinated CDS. To structure a trade, one should make assumptions about the recovery levels. Assuming 20% sub recovery, the investor can conclude that the sub CDS should trade at 213bps and currently trades too wide, and can sell it against buying senior CDS. The notional amounts should be adjusted to hedge for jump-to-default risk, so that the investor buys (1Sub. recover) / (1 – assumed Sr. recovery) of senior notional. In our example, the trade results in 285 – (0.8/0.6)*213 = 72bps of positive carry. LCDS in Capital Structure Trades Capital structure trades could include trading LCDS vs. vanilla CDS or sub-CDS on names where the instruments are available. In effect, these can be considered recovery rate plays, and the evaluation is no different than Senior Unsecured vs. Subordinated. Basically, we can express this as various equivalences. As noted above: Spread ≈ default probability * (1 – expected recovery) When we re-arrange we see that: DP ≈ Spread/(1 – ER) Because there would be an arbitrage opportunity should recoveries be significantly different than what one would expect, then the expected recoveries should be similar because the default probabilities should be similar (in fact for Senior Unsecured trading No-R and LCDS the changes of a credit event are identical) due to the fact that all three contracts should be equivalent when spreads are taken as a ratio of expected recovery rates, which can be expressed as: LCDS /(1 - LCDS ER) = Sr. CDS/(1 – Sr. ER) = Sub Spread/(1 – Sub ER) An analysis of the expected recovery rates of each part of the debt capital structure would then be able to identify any anomalies, and one could perform relative value trades using LCDS against either Senior Unsecured or Subordinated CDS.

Credit Derivatives Handbook

45

08 February 2007

Trading Implied Equity vs. Implied CDS Volatility Typically, ATM implied volatilities of CDS options are significantly higher than those of equity options. CDS and equity option-implied volatilities are often used as leading indicators of spread moves: an increase in volatility is used as a bond sell signal, while a decrease is used as a buy signal. In certain cases, equity/CDS options can actually be used as trading instruments. Mark-to-market volatility is the most significant risk for these trades. As an example, assume a company that undergoes a significant diversion in the trends of its CDS option and equity option-implied volatilities. While both spreads and equity prices are under pressure, equity volatility starts going down dramatically, while CDS volatility remains near relative high levels. One way to take advantage of the situation is to sell CDS straddles and buy equity at-themoney straddles. The hedge ratio can be selected using one of the methods that was described previously. Mark-to-market volatility and illiquidity of the CDS options are the most significant risks for this trade. CDS single-name options are not overly liquid, with only a handful of names trading regularly. However, CDX index swaptions are readily available and also usually carry reasonably wide bid/offers, so investors should consider strategies that are near flat carry but have attractive P&L profiles. The exhibit below shows the CDX IG ATM vol against the VIX. As one can see, CDXimplied vols have traded higher than equity-implied vols, but the relationship and direction are occasionally different – such as in December 2006 when the VIX fell to record lows and CDX IG vols actually ticked slightly higher. These divergences can occur for any number of reasons – such as LBO concerns – but clearly the general tendency is for vols to move in the same general direction much of the time. Another interesting aspect is the relative stability of spread vol compared to equity vol, which we believe is partially a function of liquidity, but partly due to the tight overall level of spreads and low realized spread volatilities.

Exhibit 33: VIX vs. CDX IG Implied Vol 43%

25

41%

23

39%

21

37%

19

35%

17

33%

15

31%

13

29%

9

2/ 19

1/ 19

/2 0

05

05 /1 9

/2 0

/2 0 12

/1 9

/2 0 11

/1 9 10

/2 00 3/ 6 19 /2 00 6 4/ 19 /2 00 6 5/ 19 /2 00 6 6/ 19 /2 00 6 7/ 19 /2 00 6 8/ 19 /2 00 6 9/ 19 /2 00 10 6 /1 9/ 20 06 11 /1 9/ 20 06

25% 06

11

05

27%

IG Implied Vol

VIX

Source: Credit Suisse

Credit Derivatives Handbook

46

08 February 2007

Structured Products: Baskets and Index Tranches With the increased liquidity of single-name and index default swap products, investors and broker/dealers found new methods of enhancing yields. The first method is the creation of various basket products of several names. The second is slicing and dicing indices into Tranches (CDOs). In this section, some of the more common types of these Structured Products, including Nth-to-default baskets and Index Tranches, are reviewed and evaluated.

Review of Nth to Default Nth-to-default baskets tie the trigger of a CDS contract to a number of reference entities. Should a default occur for the Nth entity, the default payment is triggered. The purchaser receives a qualifying reference obligation of the defaulting issuer, and the seller pays par, after which the contract ceases to exist. So, unlike a pool of several CDS contracts where the maximum loss equals the sum of the potential losses on each contract, the seller of an Nth-to-default contract faces a maximum downside limited to the loss on the Nth reference entity to default. As such, the highest yielding (and most common) basket is the First-to-Default (FTD) Swap. FTD swaps are priced at a discount to the sum of spreads of all reference obligations – the higher the default correlation of the underlying credits, the less the likelihood of a single default, the lower the spread. In practice the spread of FTD cannot be less than the widest credit in the basket, nor can it exceed the sum of spreads (SoS) of the underlying names. Typically, FTD trades in the 65-85% of SoS range. Since default correlation is not readily observable, various models are used for pricing Nthto-Default swaps. For details on the valuation of FTD baskets using Gaussian copula and Student-t copula models, see the Portfolio Strategy Monthly: Structured Credit Products – Part II from 16 January 2004. First-to-default Baskets First-to-default baskets have become a common way for institutions to hedge, enhance yields and leverage credit exposure. Despite the fact that first-to-default basket transactions are commonly called correlation trading, the primary risk that mark-to-market investors face in FTD basket transactions is a widening or tightening of spreads. The chart below presents a simplified view of how the probability of at least one default reacts in a two-name first-to-default basket to changes in the likelihood of more than one credit defaulting (correlation changes) and changes in the probability of default of each credit driven mainly by credit spreads. We see that a decrease in default correlation should increase the possibility of at least one of the names defaulting (the main concern for FTD players), which in turn increases the spread of the basket. Additionally, increases in credit spreads of each issuer should as well increase the FTD’s spread. The first-to-default basket spread (which is a function of the cumulative default probability and recovery) tends to be more sensitive to changes in the underlying spreads. And since the reasonable range of spreads seems to be wider than the reasonable range of correlations, we think that a position in a first-to-default basket can be viewed as more of a levered play on spreads than a play on a change in implied correlation.

Credit Derivatives Handbook

47

08 February 2007

Exhibit 34: Evaluation of a Two-Name FTD Basket

High Def ault Correlation (joint area)

Low er Def ault Correlation w ith the same spreads results in the increased chance of at least one credit event.

Increased spreads w ith the same correlation could have a larger ef f ect on the FTD pricing, than compared to purely increased correlation.

Source: Credit Suisse

Accordingly, for investors trying to fully hedge a first-to-default position with the underlying CDS, movements in correlation will drive the P&L (since the impact of spread widening is at least partially hedged), but an unhedged position in a first-to-default basket is more closely related to a leveraged CDS position rather than a correlation trade. In order to give a less stylistic view of the relative impacts of spread and correlation, we determine implied spread movements in five-name FTD baskets where each underlying credit has the same spread, the same correlation, and the same recovery. The model used for calculation was one-factor Gaussian copula. As the exhibits below show, a move in spreads tends to have a much larger impact on pricing than a similar move in correlation, and while price transparency in implied correlations is somewhat limited, the standard deviation in spreads tends to be larger than the standard deviation in implied correlations. So, for unhedged transactions, spread changes should drive P&L.

Credit Derivatives Handbook

48

08 February 2007

Exhibit 35: Implied Change in FTD Spread on Similar Swings in Correlation and Spread Spread movements have much larger impact on FTD spread

Spread (bps) 90 120 150 180 210

Correlation 30% 430 570 710 849 988

40% 411 542 672 800 928

50% 386 506 623 740 856

60% 354 461 566 669 771

70% 315 408 499 588 676

Source: Credit Suisse

Exhibit 36: Implied Change in FTD Spread on Similar Swings in Correlation and Spread Spread movements have much larger impact on pricing

Spread (bps) 30 40 50 60 70

Correlation 30% 146 194 242 289 336

40% 142 188 233 278 323

50% 136 179 221 263 304

60% 128 167 205 243 280

70% 116 150 184 218 251

Source: Credit Suisse

Tranched DJ CDX Products With the DJ CDX instruments being among the most liquid credit instruments, investors can now purchase tranches of the indexes. For the IG, the table below shows the various tranches that are typically available for investors. It is possible for dealers to create custom tranches for investors, but these tend to be less liquid than the standard tranches. DJ CDX tranches are often priced using various correlation methods based on historical spread, stock market behavior or default data. Gaussian copula is the market standard. Pricing generally assumes a “delta exchange” of the underlying index, where a DV01weighted amount of the underlying DJ CDX tranche is exchanged. However, prices can be adjusted if delta is not exchanged (where the dealer is compensated for the bid/offer cost of obtaining the delta in the open market). There are many methods for theoretical pricing, and for individuals interested in the details of some of these models and our latest thinking, one can look at the Credit Suisse’s Credit Portfolio Modeling Handbook for detailed tranche-modeling information. For a detailed overview of the high yield index tranche products, please see HY CDS Index Tranches: New Product Intro from September 14, 2004.

Exhibit 37: Standard IG And High Yield CDX Tranches IG

High Yield

Tranche

Description

Tranche

Description

0-3% 3-7% 7-10% 10-15% 15-30%

Equity BBBAAA AAA+ Super Sr.

0-10% 10-15% 15-25% 25-35%

Equity Jr Mezz Mezz AAA

Source: Credit Suisse

Credit Derivatives Handbook

49

08 February 2007

The chart below is a representative structure for the HY index tranche. The typical tranche is traded under standard ISDA master agreements, but the HY tranches can often be funded. Should there be a credit event in any one name within the underlying index, payment is distributed by tranche, with the equity tranche taking the first loss, and premiums are paid on the remaining notional outstanding. The risk of selling CDO tranches may result in losses significantly greater than the premium received.

Exhibit 38: A Representative Structure for the DJ HY Index Tranche Products Credit 1 Fixed Spread

35% to 100% Credit 2 CDXNA.HY Series 7

375 bps

Dealer

Fixed Spread

….

Credit 3

25% to 35%

Fixed Spread

…..

15% to 25%

Upfro nt

10% to 15%

P remium Upfro nt P remium

Credit 100

0% to 10%

Source: Credit Suisse

In the chart below, we show the dependence on tranche losses of a HY index on attachment points (the lower bound of the loss that the tranche can endure) and tranche width (the difference between the upper and the lower bounds of the loss the tranche can endure).

Exhibit 39: Tranche Loss Varies by Attachment Point and Tranche Width 100% 90%

0% to 10% 80%

10% to 15%

Tranche seniority increases

70% Tranche Losses

15% to 25% 60%

25% to 35% Index

50% 40%

35% to 100% 30% 20% 10% 0% 0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

55%

60%

65%

70%

75%

80%

85%

90%

95%

100%

Index Losses

Source: Credit Suisse

Path Dependency and Tranche Carry Above we discussed how convexity differences in the different tranches can have significant P&L implications for delta-hedged positions. Even investors who do not delta hedge should be aware of them because of their significant effect on the majority of investors in the product who hedge their positions.

Credit Derivatives Handbook

50

08 February 2007

We show below that more frequent rebalancing reduces this convexity effect. On the other hand, it also reduces mark-to-market volatility. In addition, more frequent rebalancing makes it more likely that hedges are rebalanced close to the wides – at an extreme – causing P&L effects even if spreads don’t change at the end of the round trip; an investor who never rebalanced (static delta) would have no P&L, ignoring carry due to the passage of time (with no rehedging, P&L is no longer path-dependent).

Exhibit 40: Spread Change Impact on DJ CDX HY Tranche Deltas Assumes parallel shift in spreads of all underlying credits in the index. Correlation: 40%. Recovery: 35%

4

Minimal convexity Delta

3

Negative convexity 2

1 (50) bps

Positive convexity

0 bps

50 bps

100 bps

Spread Change 0-10%

10-15%

15-25%

25-35%

Source: Credit Suisse

Unfortunately, it is only possible to determine the basis-point impact on the value of a tranche in a static world. Because many investors will attempt to rebalance their hedges as the delta begins to deviate significantly from the original delta, a static delta will overestimate the mark-to-market impact of gamma for these investors. A dynamic delta makes P&L dependent on the path of spreads as well as spreads themselves when the trade was initiated and closed out. So the critical (and missing) element to determine the gamma impact is which strategy investors use for rebalancing their hedges. While using a different delta for each credit in the index would be a superior hedge (because the 0%-10% tranche is most sensitive to the widest spread credits, etc.), we use the market standard of hedging with the entire index. This is because there is no active market for odd-lot CDS (pieces smaller than $2 million or not in multiples of $1 million), so individually hedging each underlying credit is not cost-effective. The following tables show how P&L would behave based on various hedging strategies for investors long credit risk in different tranches. To simplify, we ignore carry effects and assume that the index widens to some rebalancing threshold and then retraces the widening back to the original starting point. The first strategy we review is rebalancing the hedge every time the underlying index moves 50bps away from the last rebalancing. So, a 120bps widening followed by a 120bps tightening would be rehedged after the initial 50bps widening, again at the 100bp widening mark from the initial spread level, then again at the 50bps widening mark from the initial spread (i.e., every 50bps of spread changes since the last rebalancing). The second strategy is rehedging every time the underlying index moves 100bps away from the last rebalancing.

Credit Derivatives Handbook

51

08 February 2007

Exhibit 41: Spreads Widen 70bps, then Retrace the Move Delta rebalanced after 50bp widening

Upfront delta

Tranche

Delta after 50bps widening

P&L from rebalancing hedge (%)

0% to 10%

1.85

1.54

0.66

10% to 15%

3.41

3.28

0.26

15% to 25%

3.71

3.75

-0.08

25% to 35%

2.23

2.43

-0.41

Source: Credit Suisse

Exhibit 42: Spreads Widen 120bps, then Retrace the Move Strategy #1 is rebalancing every 50bps. Strategy #2 is rebalancing every 100bps

Upfront

Delta after 50bps

Delta after 100bps

Strategy #1

Strategy #2

Tranche

delta

widening

widening

P&L (%)

P&L (%)

0% to 10%

1.85

1.54

1.33

1.06

2.10

10% to 15%

3.41

3.28

2.98

0.86

1.73

15% to 25%

3.71

3.75

3.77

-0.11

-0.22

25% to 35%

2.23

2.43

2.60

-0.75

-1.50

Source: Credit Suisse

As we can see from the tables above, more frequent rebalancing (Strategy 1) reduces the convexity effect and has a smaller P&L than Strategy 2. It is positive for investors long credit risk in mezzanine/AAA tranches (15%-25% and 25%-35%) and negative for equity (0%-10%) and junior mezzanine (10%-15%) tranche investors.

Constant Proportional Debt Obligations (CPDOs) Constant proportion debt obligations are a relatively new structured credit product providing dynamic exposure to high grade CDS indices, which are currently being rolled every six months. The dynamic structure is key to the distinctive characteristics of such instruments, which offer investors purportedly high coupon payments that have high and stable ratings.

Exhibit 43: Cash Flows for a Typical CPDO iTraxx

Sell Protection CS / Notional Account

Sell Protection

Coupons + Notional Account Balance at Maturity

Total Return Swap SPV Issuance Proceeds

Noteholders Issuance Proceeds

3m Euribor Ongoing Cashflows CDX

Cashflows at Inception

Source: Credit Suisse

Credit Derivatives Handbook

52

08 February 2007

A Typical CPDO Structure • Maximum leverage of 15x; • Initial leverage 8-13x (The ratio to the maximum leverage is a critical feature); • 50% iTraxx Europe, 50% CDX, rolled on the run, or potentially 100% of either; • Paying LIBOR plus 150bps - 200bps or up to 200bp of excess return; • Unwound if price hits 20 (sic: an enormous move, even leveraged: we estimate 120bp widening in the current iTraxx Europe); and • Favorably unwound if it can be done in such a way that the deal can pay out 200bp over its remaining life with no risk. CPDOs are structures that are leveraged initially and designed to “buy dips” in the credit market and so is a stabilizing instrument. It does well if spreads widen and then mean revert. A more leveraged initial structure starts to approach a pure leveraged long. In both cases, the structure is a forced seller on a very large move (indeed) in spreads. Holders of the structure will clearly have their own risk tolerance, but will be faced with initial losses if the ideal happens and spreads initially widen. We think it is safe to ignore for current purposes scenarios that call for 120bp of widening in the iTraxx Europe. However, even if the initial leverage is only eight-fold, a doubling of spreads will involve losses of approximately eight points, by our calculation. In this situation, we think that investors will be tested in the way they often are (GM springs to mind) where long-run “money-good” trades are tested to (and beyond) destruction by mark-to-market moves. Therefore CPDO trades appear to us to have the characteristics of an outright leveraged long, plus an additional deep out-of-the-money option short. The Roll: an advantage of the structure We see the following advantages to leveraging in this way (and, as we said, nothing is intrinsically wrong with it, although the presence of the zero bound raises a question). The market is generally set up to roll shorts (index payers) on the run, making the roll trade wide to the theoretical. A vehicle rolling an index long on the run might expect to benefit from this convention. We suspect that the roll, still a relatively new phenomenon, has now reached maturity and will now tend to trade at fair value. Similarly, by probably balancing the roll more effectively, these vehicles should not suffer too much from the market marking down the spread of names due to enter the index. Only six months’ default risk, but 5 years’ credit risk On the other side though, although the vehicle should experience minimal defaults, it will suffer from any credit that is deteriorating enough to leave the index and, in efficient markets, will experience losses that equate to the market’s perception of increased default risk over the next five years. The vehicles are not taking five years’ default risk, but are taking five years’ credit risk. Transitions out of the iTraxx Europe have been fairly limited, reflecting the fact that the indices have developed during a quiet period. Still, ratings transition is a risk, albeit one that is fully reflected in the current index price, and is broadly covered by the Gaussian approximation referred to above.

Credit Derivatives Handbook

53

08 February 2007

Jumps to Default from the iTraxx are not Impossible Credit returns are known not to be Gaussian, as we have analyzed on numerous occasions: it drives our strategic “Underweight” recommendation. Defaults most certainly are not Gaussian. Our CUSP® model applies a “jump-diffusion” assumption, among other more sophisticated methodologies, to appraise relative value within and between equity and credit markets and their derivatives. We are not applying a full CUSP analysis here, but we point out that jumps to default from the iTraxx are unknown, but not impossible. It is an improvable counter-factual assumption, but we strongly suspect that Parmalat would have been in the iTraxx at least until September 2003 and quite possibly into default: the worst of the news flow was in November 2003. Jump to default from investment grade is rare, but it does happen. Exhibit 44 shows the Moody’s average cumulative default rate for initial Baa-bucket credits over the 1980-2005 period.

Exhibit 44: Jump to default is rare, but it does happen Moody’s cumulative 5-year default rate by year for initially Baa credits, percent by year

1.4 1.2 1 0.8 0.6 0.4 0.2 0 1

2

3

4

5

Source: Credit Suisse

Parmalat should be the exception. However, the first-year default rate for BBBs is 0.1% . This is 0.05%, or one in 2,000 over the six-month 18 contract run. This needs to be considered in light of the fact that there are 46 BBB iTraxx members. The current ratings distribution of the iTraxx Europe members is shown in Exhibit 45.

Exhibit 45: The iTraxx Europe contains 46 Baas iTraxx S6 members by Moody’s rating

Moody’s rating A1 A2 A3 Aa1 Aa2 Aa3

Total 12 17 16 4 2 10

Moody’s rating Ba1 Baa1 Baa2 Baa3 N/R Grand Total Baas

Total 1 17 25 9 12 125 46

Source: Credit Suisse

18

Credit Derivatives Handbook

We consider jumps to be homogenous over these periods.

54

GLOBAL HIGH GRADE CREDIT RESEARCH AND STRATEGY Bunt Ghosh, Managing Director, Global Head of Fixed Income Research William Porter, Managing Director, Head of European High Grade Credit Research and Strategy Robert Schiffman, Managing Director, Head of North America High Grade Credit Research Jamie Nicholson-Leener, Managing Director, Head of Latin America Credit Research Damien Wood, Director, Head of Non-Japan Asia Credit Research and Strategy US CREDIT Auto & Auto Parts Mark Altherr, CFA Lauren Taylor

Director Analyst

212 538 4082 212 538 4453

Basic Industries, Healthcare, Manufacturing Gene Cheon Vice President John Nordstrom Analyst

212 325 9372 212 325 0852

Consumer Products Robert Schiffman Christine Dillon

Managing Director Associate

212 538 3972 212 325 8983

Credit Strategy (US) Ira Jersey Jason Rayman, CFA Alex Makedon David Crick David Lee

Vice President Vice President Vice President Associate Associate

212 325 4674 212 325 4908 212 538 8340 212 325 3070 212 325 6693

EUROPEAN CREDIT Basic Materials Fouad Fenianos

Associate

44 20 7883 4188

Consumer, Retail, Tobacco Clark McPherson

Director

44 20 7883 8571

Emerging Markets Yuri Poliakov

Vice President

44 20 7883 3738

Credit Strategy (Europe) William Porter Jon Zucker Has Tank Andrea Cicione Nicolas Caille

Managing Director Vice President Vice President Associate Analyst

44 20 7888 1207 44 20 7888 8109 44 20 7883 5822 44 20 7888 0757 44 20 7883 6422

Financial Institutions Leslie Mapondera Stephane Suchet

Director Associate

212 538 0410 44 20 7883 4278

Crossovers, Supermarkets, Tobacco Mark Altherr, CFA Director Lauren Taylor Analyst

212 538 4082 212 538 4453

Industrials Haroon Hassan

Associate

44 20 7883 7178

Electric Utilities Paula White-Lavitt Lora Branz

212 325 4853 212 325 4893

Telecom, Media, Technology Robert Lambert Hayley Walker

Director Associate

44 20 7883 5455 44 20 7883 5569

Energy, Pipelines, Project Finance Andrew Brooks Director Christine Dillon Associate

212 325 3014 212 325 8983

Atif Ali Nicholas Harrison Neel Shah

Analyst Analyst Analyst

44 20 7883 7978 44 20 7883 7992 44 20 7883 0894

Financial Institutions Leslie Mapondera Natasha Tsiouris

212 538 0410 212 538 0560

Real Estate, Insurance, Homebuilders Matthew Lynch Vice President John Nordstrom Analyst

LATIN AMERICA Consumer, Industrials, Transportation Jamie Nicholson-Leener Managing Director Marianela Najera Associate

212 325 6456 212 325 0852

Media, Telecom Carlos Gomez-Ramos

Associate

212 325 0112

Retail Ee Lin See

Associate

212 538 5785

Supermarkets, Tobacco Mark Altherr, CFA Lauren Taylor

Energy, Utilities Eduardo Vieira Carlos Gomez-Ramos

Director Associate

212 325 5999 212 325 0112

Director Analyst

212 538 4082 212 538 4453

Banks, Metals, Mining Jamie Nicholson-Leener Daniela Aslan

Managing Director Analyst

212 538 6769 212 538 0490

212 538 3972 212 538 2675

NON-JAPAN ASIA Banks, Industrials, Credit Strategy (NJA) Damien Wood Director

65 6212 3195

Oil, Telecom, Media Sandra Chow, CFA

Vice President

65 6212 3411

Industrials, Utilities, Resources Erly Witoyo Vice President

65 6212 3275

Managing Director Associate

Director Analyst

Telecom, Media, Technology Robert Schiffman Managing Director Christina Leonetti Associate

Property Vivien Gui

..

Associate

212 538 6769 44 20 7888 1229

852 2101 6455

Disclosure Appendix Analyst Certification The analysts identified in this report each certify, with respect to the companies or securities that the individual analyzes, that (1) the views expressed in this report accurately reflect his or her personal views about all of the subject companies and securities and (2) no part of his or her compensation was, is or will be directly or indirectly related to the specific recommendations or views expressed in this report. CUSP® Certification and Definition With respect to the analysis in this report based on the CUSP® methodology, Credit Suisse certifies that (1) the views expressed in this report accurately reflect the CUSP® methodology and (2) no part of the firm's compensation was, is, or will be directly related to the specific views disclosed in this report. CUSP® is an analytical model that relates to an issuer's capital structure, stock price, and the option-implied volatility of its shares to credit risk. CUSP® provides systematic monitoring of credit risk from forward-looking, market-based measures (one standard deviation spread widening, SWR) and relative-value tracking that incorporates both risk and return (probability weighted return, PWR). Details on the CUSP® methodology can be found on http://www.credit-suisse.com/cusp or by contacting the Credit Suisse Fixed Income Portfolio Strategy Group. Important Disclosures Credit Suisse's policy is only to publish investment research that is impartial, independent, clear, fair and not misleading. For more detail, please refer to Credit Suisse's Policies for Managing Conflicts of Interest in connection with Investment Research: http://www.csfb.com/research-and-analytics/disclaimer/managing_conflicts_disclaimer.html Credit Suisse’s policy is to publish research reports as it deems appropriate, based on developments with the subject issuer, the sector or the market that may have a material impact on the research views or opinions stated herein. The analyst(s) involved in the preparation of this research report received compensation that is based upon various factors, including Credit Suisse's total revenues, a portion of which are generated by Credit Suisse's Investment Banking and Fixed Income Divisions. Credit Suisse may trade as principal in the securities or derivatives of the issuers that are the subject of this report. At any point in time, Credit Suisse is likely to have significant holdings in the securities mentioned in this report. As at the date of this report, Credit Suisse acts as a market maker or liquidity provider in the debt securities of the subject issuer(s) mentioned in this report. For important disclosure information on securities recommended in this report, please call +1-212-538-7625. For the history of any relative value trade ideas suggested by the Fixed Income research department over the previous 12 months, please view the document at http://researchand-analytics.csfb.com/docpopup.asp?docid=35321113&type=pdf. Credit Suisse clients with access to the Locus website may refer to http://www.credit-suisse.com/locus. For the history of recommendations provided by Technical Analysis, please visit the website at http://www.credit-suisse.com/techanalysis. Credit Suisse does not provide any tax advice. Any statement herein regarding any US federal tax is not intended or written to be used, and cannot be used, by any taxpayer for the purposes of avoiding any penalties. Emerging Markets Bond Recommendation Definitions Buy: Indicates a recommended buy on our expectation that the issue will deliver a return higher than the risk-free rate. Sell: Indicates a recommended sell on our expectation that the issue will deliver a return lower than the risk-free rate. Corporate Bond Fundamental Recommendation Definitions Buy: Indicates a recommended buy on our expectation that the issue will be a top performer in its sector. Outperform: Indicates an above-average total return performer within its sector. Bonds in this category have stable or improving credit profiles and are undervalued, or they may be weaker credits that, we believe, are cheap relative to the sector and are expected to outperform on a total-return basis. These bonds may possess price risk in a volatile environment. Market Perform: Indicates a bond that is expected to return average performance in its sector. Underperform: Indicates a below-average total-return performer within its sector. Bonds in this category have weak or worsening credit trends, or they may be stable credits that, we believe, are overvalued or rich relative to the sector. Sell: Indicates a recommended sell on the expectation that the issue will be among the poor performers in its sector. Restricted: In certain circumstances, Credit Suisse policy and/or applicable law and regulations preclude certain types of communications, including an investment recommendation, during the course of Credit Suisse's engagement in an investment banking transaction and in certain other circumstances. Corporate Bond Risk Category Definitions In addition to the recommendation, each issue may have a risk category indicating that it is an appropriate holding for an "average" high yield investor, designated as Market, or that it has a higher or lower risk profile, designated as Speculative and Conservative, respectively. Options Disclaimer Structured securities, derivatives and options are complex instruments that are not suitable for every investor, may involve a high degree of risk, and may be appropriate investments only for sophisticated investors who are capable of understanding and assuming the risks involved. Supporting documentation for any claims, comparisons, recommendations, statistics or other technical data will be supplied upon request. Any trade information is preliminary and not intended as an official transaction confirmation. Use the following links to read the Options Clearing Corporation's disclosure document: http://www.cboe.com/LearnCenter/pdf/characteristicsandrisks.pdf Because of the importance of tax considerations to many option transactions, the investor considering options should consult with his/her tax advisor as to how taxes affect the outcome of contemplated options transactions. Transaction costs may be significant in option strategies calling for multiple purchases and sales of options, such as spreads and straddles. Commissions and transaction costs may be a factor in actual returns realized by the investor and should be taken into consideration. HOLT™ With respect to the analysis in this report based on the Credit Suisse HOLT methodology, Credit Suisse certifies that (1) the views expressed in this report accurately reflect the Credit Suisse HOLT methodology and (2) no part of the Firm’s compensation was, is, or will be directly related to the specific views disclosed in this report. The Credit Suisse HOLT methodology does not assign recommendations to a security. It is an analytical tool that involves use of a set of proprietary quantitative algorithms and warranted value calculations, collectively called the Credit Suisse HOLT valuation model, that are consistently applied to all the companies included in its database. Third-party data (including consensus earnings estimates) are systematically translated into a number of default variables and incorporated into the algorithms available in the Credit Suisse HOLT valuation model. The source financial statement, pricing, and earnings data provided by outside data vendors are subject to quality control and may also be adjusted to more closely measure the underlying economics of firm performance. These adjustments provide consistency when analyzing a single company across time, or analyzing multiple companies across industries or national borders. The default scenario that is produced by the Credit Suisse HOLT valuation model establishes the baseline valuation for a security, and a user then may adjust the default variables to produce alternative scenarios, any of which could occur. Additional information about the Credit Suisse HOLT methodology is available on request. The Credit Suisse HOLT methodology does not assign a price target to a security. The default scenario that is produced by the Credit Suisse HOLT valuation model establishes a warranted price for a security, and as the third-party data are updated, the warranted price may also change. The default variables may also be adjusted to produce alternative warranted prices, any of which could occur. Additional information about the Credit Suisse HOLT methodology is available on request.

..

References in this report to Credit Suisse include all of the subsidiaries and affiliates of Credit Suisse operating under its investment banking division. For more information on our structure, please use the following link: http://www.credit-suisse.com/en/who_we_are/ourstructure.html. This report is not directed to, or intended for distribution to or use by, any person or entity who is a citizen or resident of or located in any locality, state, country or other jurisdiction where such distribution, publication, availability or use would be contrary to law or regulation or which would subject Credit Suisse or its affiliates (“CS”) to any registration or licensing requirement within such jurisdiction. All material presented in this report, unless specifically indicated otherwise, is under copyright to CS. None of the material, nor its content, nor any copy of it, may be altered in any way, transmitted to, copied or distributed to any other party, without the prior express written permission of CS. All trademarks, service marks and logos used in this report are trademarks or service marks or registered trademarks or service marks of CS or its affiliates. The information, tools and material presented in this report are provided to you for information purposes only and are not to be used or considered as an offer or the solicitation of an offer to sell or to buy or subscribe for securities or other financial instruments. CS may not have taken any steps to ensure that the securities referred to in this report are suitable for any particular investor. CS will not treat recipients of this report as its customers by virtue of their receiving this report. The investments and services contained or referred to in this report may not be suitable for you and it is recommended that you consult an independent investment advisor if you are in doubt about such investments or investment services. Nothing in this report constitutes investment, legal, accounting or tax advice, or a representation that any investment or strategy is suitable or appropriate to your individual circumstances, or otherwise constitutes a personal recommendation to you. CS does not advise on the tax consequences of investments and you are advised to contact an independent tax adviser. Please note in particular that the bases and levels of taxation may change. Information and opinions presented in this report have been obtained or derived from sources believed by CS to be reliable, but CS makes no representation as to their accuracy or completeness. CS accepts no liability for loss arising from the use of the material presented in this report, except that this exclusion of liability does not apply to the extent that such liability arises under specific statutes or regulations applicable to CS. This report is not to be relied upon in substitution for the exercise of independent judgment. CS may have issued, and may in the future issue, other reports that are inconsistent with, and reach different conclusions from, the information presented in this report. Those reports reflect the different assumptions, views and analytical methods of the analysts who prepared them and CS is under no obligation to ensure that such other reports are brought to the attention of any recipient of this report. CS may, to the extent permitted by law, participate or invest in financing transactions with the issuer(s) of the securities referred to in this report, perform services for or solicit business from such issuers, and/or have a position or holding, or other material interest, or effect transactions, in such securities or options thereon, or other investments related thereto. In addition, it may make markets in the securities mentioned in the material presented in this report. CS may have, within the last three years, served as manager or co-manager of a public offering of securities for, or currently may make a primary market in issues of, any or all of the entities mentioned in this report or may be providing, or have provided within the previous 12 months, significant advice or investment services in relation to the investment concerned or a related investment. Additional information is, subject to duties of confidentiality, available on request. Some investments referred to in this report will be offered solely by a single entity and in the case of some investments solely by CS, or an associate of CS or CS may be the only market maker in such investments. Past performance should not be taken as an indication or guarantee of future performance, and no representation or warranty, express or implied, is made regarding future performance. Information, opinions and estimates contained in this report reflect a judgement at its original date of publication by CS and are subject to change without notice. The price, value of and income from any of the securities or financial instruments mentioned in this report can fall as well as rise. The value of securities and financial instruments is subject to exchange rate fluctuation that may have a positive or adverse effect on the price or income of such securities or financial instruments. Investors in securities such as ADR’s, the values of which are influenced by currency volatility, effectively assume this risk. Structured securities are complex instruments, typically involve a high degree of risk and are intended for sale only to sophisticated investors who are capable of understanding and assuming the risks involved. The market value of any structured security may be affected by changes in economic, financial and political factors (including, but not limited to, spot and forward interest and exchange rates), time to maturity, market conditions and volatility, and the credit quality of any issuer or reference issuer. Any investor interested in purchasing a structured product should conduct their own investigation and analysis of the product and consult with their own professional advisers as to the risks involved in making such a purchase. Some investments discussed in this report may have a high level of volatility. High volatility investments may experience sudden and large falls in their value causing losses when that investment is realised. Those losses may equal your original investment. Indeed, in the case of some investments the potential losses may exceed the amount of initial investment and, in such circumstances, you may be required to pay more money to support those losses. Income yields from investments may fluctuate and, in consequence, initial capital paid to make the investment may be used as part of that income yield. Some investments may not be readily realisable and it may be difficult to sell or realise those investments, similarly it may prove difficult for you to obtain reliable information about the value, or risks, to which such an investment is exposed. This report may provide the addresses of, or contain hyperlinks to, websites. Except to the extent to which the report refers to website material of CS, CS has not reviewed any such site and takes no responsibility for the content contained therein. Such address or hyperlink (including addresses or hyperlinks to CS’s own website material) is provided solely for your convenience and information and the content of any such website does not in any way form part of this document. Accessing such website or following such link through this report or CS’s website shall be at your own risk. This report is issued and distributed in Europe (except Switzerland) by Credit Suisse Securities (Europe) Limited, One Cabot Square, London E14 4QJ, England, which is regulated in the United Kingdom by The Financial Services Authority (“FSA”). This report is being distributed in Germany by Credit Suisse Securities (Europe) Limited Niederlassung Frankfurt am Main regulated by the Bundesanstalt fuer Finanzdienstleistungsaufsicht ("BaFin"). This report is being distributed in the United States and Canada by Credit Suisse Securities (USA) LLC; in Switzerland by Credit Suisse; in Brazil by Banco de Investimentos Credit Suisse (Brasil) S.A; in Japan by Credit Suisse Securities (Japan) Limited; elsewhere in Asia/ Pacific by whichever of the following is the appropriately authorised entity in the relevant jurisdiction: Credit Suisse (Hong Kong) Limited, Credit Suisse Equities (Australia) Limited, Credit Suisse Securities (Thailand) Limited, Credit Suisse Securities (Malaysia) Sdn Bhd, Credit Suisse Singapore Branch, and elsewhere in the world by the relevant authorised affiliate of the above. Research on Taiwanese securities produced by Credit Suisse, Taipei Branch has been prepared by a registered Senior Business Person. Research provided to residents of Malaysia is authorised by the Head of Research for Credit Suisse Securities (Malaysia) Sdn Bhd, to whom they should direct any queries on +603 2723 2020. This research may not conform to Canadian disclosure requirements. In jurisdictions where CS is not already registered or licensed to trade in securities, transactions will only be effected in accordance with applicable securities legislation, which will vary from jurisdiction to jurisdiction and may require that the trade be made in accordance with applicable exemptions from registration or licensing requirements. Non-U.S. customers wishing to effect a transaction should contact a CS entity in their local jurisdiction unless governing law permits otherwise. U.S. customers wishing to effect a transaction should do so only by contacting a representative at Credit Suisse Securities (USA) LLC in the U.S. Please note that this research was originally prepared and issued by CS for distribution to their market professional and institutional investor customers. Recipients who are not market professional or institutional investor customers of CS should seek the advice of their independent financial advisor prior to taking any investment decision based on this report or for any necessary explanation of its contents. This research may relate to investments or services of a person outside of the UK or to other matters which are not regulated by the FSA or in respect of which the protections of the FSA for private customers and/or the UK compensation scheme may not be available, and further details as to where this may be the case are available upon request in respect of this report. Copyright © 2007 CREDIT SUISSE GROUP and/or its affiliates. All rights reserved.