Contents

Japanese CDOs boom

Investor appetite for synthetics grows in Tokyo

Synthetic CDOs are attracting a rapidly growing number and range of Japanese structured credit investors, who are looking at a wider variety of products. At the same time, cash CDOs are also benefiting from strong investor interest.

Investors both old and new are now tapping the Japanese synthetic CDO market, according to Dean Rostrom, head of CDOs Japan at Deutsche Securities in Tokyo. "Those who have previously invested are now coming back with a stronger interest for more deals and larger size. And there are also those investing in synthetic CDOs for the first time, ranging from large institutional investors - such as banks and insurance companies - to smaller regional banks, credit cooperatives, corporates and so on."

Such recent synthetic CDO activity is driven by a number of factors, Rostrom says. "First, there is the realisation that the tightening in the market is something we will have to live with for some time - people are losing out on the hope that spreads are going to widen significantly for one reason or another over a 6-12 month horizon."

Second is the clarification of the accounting rules in Japan that took effect from last April. Their introduction means that investors in certain circumstances no longer need to bifurcate the derivative from the cash instrument in a structured credit product and mark it to market.

"Now that the complexity of the accounting regime for structured notes has been clarified, a number of investors have been able to get their internal approvals and accounting standards established, so they are now starting to invest," explains Rostrom.

In addition, says Aaron MacDougall, head of credit trading at Deutsche Securities in Tokyo: "A lot of the institutional investors in Japan are banks, so a big theme is the roll-out of Basel II. Certainly, a lot of the clients we are talking to are focused on optimising their portfolios in this respect."

Basel II will impact banks in different ways, depending on whether their focus is global, regional or local, but a shift is emerging away from non-rated pieces and lower-rated tranches. Increasingly, investors are looking for high grade or rated assets that have a lower risk weighting under the new regime.

Whatever the driver, Tokyo-based dealers have in recent months witnessed a move away from international names and an increased use of more complex structures. "We are starting to see Japanese investors come back to Japanese portfolios and adapting generic structures to their needs - in some cases getting simpler. For example, there is a lot more interest in simple basket trades and, in some cases, interest in more complex structures: step-ups and callable notes," explains Rostrom.

The number of CDO market participants in Japan has been rising along with the continued global expansion of the synthetic CDO market, confirms a new report from Moody's. The agency says its Tokyo office has seen an increase in enquiries from these players about the stability of synthetic CDO tranche ratings.

According to some sample analysis described in its report, "Understanding Rating Stability of Moody's Rated Synthetic CDOs - Commentary for Japanese CDO Market," the maximum change in a tranche rating is limited to one or two notches, even if a single obligor is downgraded severely. In this sample case, the tranche would theoretically not suffer a one-notch downgrade until 33 reference credits experience one-notch downgrades.

"At the same time, the passage of time has a decidedly positive impact on a CDO tranche's rating", says Yusuke Seki, Moody's senior vice president and an author of the report. It provides sample analysis using Moody's CDOROM Model of how a synthetic CDO tranche can be affected by the passage of time, as well as by rating changes in its reference credits.

Meanwhile, there is also plenty of activity in the country's cash CDO sector. "We do see very strong interest in both the cash and synthetic markets in Japan – it is not just a synthetic market alone, as some people seem to think," says Rostrom.

He continues: "On the cash side, we are seeing continuing strong interest in US and European leveraged loans. The types of products are changing a little bit. For example, we are seeing a lot more interest and buying in credit opportunity funds, which tend to be lower-leveraged flexible investment vehicles with liquidity after one-year instead of liquidity after 12 years like a typical CLO. Further, unlike traditional CDOs with the leverage fixed over the course of the transaction, credit opportunity funds allow the asset manager substantial flexibility to adjust leverage and the asset mix in accordance with market conditions."

MP

All about equity

Rated equity set to be investor focus

The quietest week of the European summer - sandwiched between a UK and a US public holiday - is traditionally a time for preparing for the busier periods ahead. This year, the omens for volumes are good and equity tranches look set to be the key driver behind growth.

"We expect a pick-up in the structured credit pipeline in September," confirms Dirk Muench in the credit derivatives quantitative credit research team at JP Morgan. "This is based on our own market insight and the increasing activity levels of structured products being marketed elsewhere."

In the absence of any macro shocks, such a pipeline should lead to well-supported absolute spread levels, Muench suggests. "We don't have any indication that a certain part of the capital structure will receive more interest than the other and therefore would expect relative tranche valuations to remain broadly stable," he adds.

Equity looks set to continue to be the most attractive sector, at least in the near term. "The printing of leveraged super senior deals in Q1 slowly transferred value back into equity, which reached year-to-date highs at the end of Q2," explains Mehernosh Engineer, senior credit strategist at BNP Paribas.

To exploit the value in the equity tranche, banks have primarily repackaged the risk and transferred it to hedge funds in the form of zero-coupon structured products. However, many real money investors are still unable to take advantage of such value because these instruments were unrated - which led structurers to create rated equity products, the first of which was launched in July.

Equity is by definition an unrated entity, but by placing the equity into a structured note whose coupons are guaranteed and whose principal is partially guaranteed, it is possible to obtain similar default dynamics and returns as a portfolio of similarly rated issuers. "In essence, getting the equity rated is a de-leveraging exercise with the extent of de-leveraging dependent upon the rating required," says Engineer.

He continues: "Depending upon the structure that investors choose, they may also be able to immunise themselves from the timing of the defaults. This may be attractive to those investors who are concerned about the slowdown in the US. A second advantage could also be gained by having the credit selection process dynamically managed by the investor himself or by a credit manager, as opposed to having a static portfolio. As with any other credit investment, a lower default rate will lead to higher returns and vice versa - and the key to maximizing IRR lies in the credit selection process."

JP Morgan's Muench says that his firm sees an increasing demand of equity risk too, particularly in the context of rated equity. "Nevertheless, zero coupon structures haven't lost all their appeal. In fact, they might become more interesting as spreads widen due to the maturity extension brought about by the September 20 iTraxx index roll and upfront premia on zero coupon equity subsequently fall."

MP

"As the deal gets more complex, the transparency disappears"

Buy-side analytics provider CDO2 is the subject of this week's Provider Profile.

Fittingly for a firm that deals in the specifics of pricing, CDO² was created with very specific aims. It was founded in early 2004 to provide a pricing service targeting structured credit investors and focussing on complex structured deals. Furthermore, CDO² was set up with the express aim of plugging the gap between the price that investment banks put on the deals that they market, and what their clients need: an independent validation of that price.

Gary Kendall

"A typical CDO² client might be a mid-sized European bank. They are a 'buy and hold' type client that has an appetite for more exotic structured credit deals. We work with many German banks in particular, as there is a regulatory need for an independent price for deals with these types of structures in Germany. We provide them with that additional level of comfort by supplying them with an independent pricing mechanism" says Gary Kendall, md and founder of CDO².

Perhaps surprisingly, the hedge fund market has, so far at least, not made so many enquiries about the CDO² service. Kendall explains: "Our clients are not in and out of the market quickly, they will look very carefully at a CDO, analyse its complexity and price it, and if they invest they will be in there for the medium term at least."

Hedge funds, in Kendall's view, adopt a different strategy. "The hedge fund sector trades simpler credit derivatives deals, such as single tranche and zero coupon CDOs; it's easier for them to do that, as they can get out of a deal quickly. It's very difficult, however, to model structured credit. We don't see them getting into the highly exotic CDOs," he says.

So when does an investor look for help from CDO²? A typical deal often includes an unequal weighting of names across a CDO or an unusual payoff. "Once you do this you find that you fall off the scope of traditional software to price the deal," claims Kendall. "CDS and index tranche trades are relatively straight-forward, with enough transparency and liquidity, but as the deal gets more complex, the transparency disappears."

Investors' desire to regain that transparency through CDO²'s service is driven by two key factors, according to Kendall. First, a deal structure, having usually first been agency rated, will need to be marked-to-market to comply with IAS39 accountancy regulations. Second, investors question whether the price they've been given by the issuing dealer is fair.

He explains: "Basically these guys see that they might be getting a poor deal and so need a third party to provide an accurate price on their behalf. The banks give them spreadsheets when marketing a deal, but such tools only go so far. These investors don't employ many quants so we fill that analysis gap. The more complex the deal, the more the investor wants an independent view."

Investors' desire for third-party input could potentially antagonise the issuing banks and raise concerns that their own pricing services are being replaced, but the banks appear comfortable with the situation. "The sell side is happy for their clients to get an independent view, if it helps them invest. We even have investment banks using CDO², to get an additional view on a price," claims Kendall.

A typical client for CDO² might well have only a few deals on the go, explains Kendall, but naturally as the bank gains confidence and invests more in exotic structures, they always want to do better with each additional deal, and so more is demanded of the software tools used to do their pricing. "There are some additional tools out there besides what the investment banks offer. Bloomberg, for example, lets you price two or three deals, but sophisticated investors need a complete system," Kendall observes.

Pricing the highly complex structures now on offer to such investors requires more computing power than a desktop pc can provide. So CDO² has partnered with Sun Microsystems, to give its clients remote access to Sun's network of servers, which delivers grid computing.

Grid computing provides the vast processing power needed to run the thousands of Monte Carlo simulations required to value a structure; having Sun as a business partner has therefore benefited all concerned. "This approach works well for the banks as they pay for what they use on Sun. And because the system is networked, a work-group can be formed and they can see each other's deals," explains Kendall.

The use of a complex set-up such as this to price deals will serve the structured credit market well going forward as structural innovation will inevitably continue. At present, the explosion in the number and variety of CPPI products is a key focus for CDO².

"This is definitely the next big challenge for CDO². Credit CPPI today is where CDO squared deals were a few years ago. We hear that there's a lack of transparency out there, so it's a natural area for us to consider," says Kendall.

JW

Markit acquires Chasen Enterprises

The latest company and people moves

Markit acquires Chasen Enterprises
Markit Group continues its aggressive expansion with the acquisition of Chasen Enterprises, a provider of structured securities modeling systems for the agency and non-agency collateralised mortgage obligation (CMO) and asset-backed securities (ABS) sectors.

Markit will exploit Chasen's structured finance know-how to expand its reference cashflow database (RCD), which is expected to launch toward the end of this year. The database will form the central monitoring and settlement platform for the CDS of ABS market. Markit's ABX calculator, which is used for settlement of the ABX index trades and is free to market participants, will also benefit from enhanced functionality.

"We are very excited about becoming a part of Markit," says Andrew Chasen, founder and president of Chasen. "Markit has helped to build liquidity in the sub-prime residential mortgage market through ABX. We look forward to contributing our modeling expertise in order to provide the investment community with the tools needed to analyse synthetic ABS."

Kevin Gould, executive vice president at Markit, adds: "The acquisition of Chasen demonstrates Markit's commitment to support the structured finance markets, particularly with respect to the valuation of synthetic ABS instruments. The CDS of ABS market is growing rapidly and we are seeing an ever-increasing demand for benchmark data and independent valuations for this product."

Chasen has been offering software and services to the structured finance sector since 1989, having modelled over 18,000 CMO and ABS structures to date. The company is based in Peekskill, New York.

Deloitte hires CDO analytics expert
Deloitte & Touche has hired former director of European operations at Atlantic Information Services, Bryan Roberts. He will join Deloitte's securitisation services practice based in the US as part of the team that markets and supports its CDO system offering, CDO Suite. In his new role, Roberts will report to Hillel Caplan, global product leader for CDO Suite, and ultimately to Mark Scherer, the firm's head of securitisation technology.

Bear brings in new credit derivatives md
Bear Stearns has appointed Brad Mazur as managing director of credit derivatives structuring and marketing. He was previously a vice president in structured credit for JP Morgan in New York.

Send all your people moves to: John Donnelly

Moody's highlights LCDS template pros and cons

A round up of this week's structured credit news

Moody's highlights LCDS template pros and cons
Moody's has published its initial study of ISDA's June documentation template for credit default swaps that reference US syndicated loans (LCDS). The template - which was created to meet the specific needs associated with syndicated loans - is widely expected to increase the size of the LCDS market and to facilitate the inclusion of LCDS in CLOs.

However, while the LCDS template is meant to create a synthetic position that mimics a cash position in a secured loan, some of the features allowable within the terms of the document may prevent a simple pass-through of the default and recovery assumptions used to assess cash secured loans, according to Moody's analyst Rudolph Bunja.

One key area of concern for the rating agency is in connection with default probability assessments because, under the terms of the template, 'failure to pay' triggers can incorporate any of the reference entity's other obligations. Unless the obligation is redefined as a 'reference obligation only', there is a risk that other obligations could trigger a credit event.

One way to alleviate this concern, says Moody's, is to make sure that the LCDS is physically settled and that the delivered obligation is a senior secured loan that meets the CDO's collateral debt security eligibility criteria. As such, even if a credit event is triggered by a default of the reference entity's unsecured obligations, it can be held by the CDO as a performing collateral debt security, providing the delivered senior secured loan is performing.

CDO issuance continues to grow
Barclays Capital's report on CDO activity in the first half of 2006 finds that, in spite of compressed CDO risk premia, issuance in the sector continues unabated - as demonstrated by both the large number of new issues and the strong demand for credit that the issues embody. The bank says 720 transactions were completed in H1 06, compared to 371 for H1 05 and 224 in H1 04.

"When we examine delta-adjusted volumes we find that bespoke CDO issuance amounted to US$233bn in H1 06, compared to US$137bn for the same period for 2005, which is a significant increase, while volumes are currently similar to the whole of 2005," the report adds.

Meanwhile, the main change in the composition of bespoke non-equity issuance is a shift towards CDOs that attach lower down in the capital structure. A breakdown of the delta-adjusted volume of non-equity protection bought by dealers shows that for H1 06 junior mezzanine transactions accounted for 61% of issuance volume, compared with only 36% in 2005.

"This is even more significant, given that we have seen a larger share of transactions with longer maturities, which have higher attachment points compared to the shorter maturities," the report says.

Structured credit reduces volatility
A report released by the Bank for International Settlements last week acknowledges the positive benefits of structured credit instruments. The paper - which studies the behaviour of financial market volatility since 1970 - notes that in the last few years new products, such as CDS and CDOs that allow investors to transfer credit risk, have become popular.

The BIS report says: "This significant growth in risk transfer instruments may indirectly enhance market liquidity and hence reduce volatility, in that it allows investors to take on or unwind exposures in a short period of time without having to trade in the cash market. However, the effect may depend on conditions in financial markets (Ferguson (2005). In general, by allowing scope for the transfer and dispersion of market and credit risk, the new instruments create a more resilient financial system, thereby reducing volatility. However, it has been argued that these new instruments may at times increase asset price volatility."

Send all your news to: John Donnelly

Smoothing tranche risk

Matthias Neugebauer and Cheiron Osako of Fitch Ratings look at understanding and hedging risks in synthetic CDO tranches

A synthetic CDO tranche is subject to mark-to-market movements as underlying risk drivers move over the life of the transaction. For example, an increase in the credit spread of an underlying credit causes a loss in the value of a long equity tranche position because it implies that the risk of the portfolio has increased and therefore the expected loss of the portfolio has increased. A key element in hedging against unfavourable movements in CDO values is the proper understanding of the sensitivities to risk factor shifts such as changes in:
- credit spreads of the underlying credit default swap ("CDS") contracts;
- credit quality of the underlying CDS contracts;
- default correlation between the underlying CDS contracts; and
- time to maturity

This article details how to measure these sensitivities and shows how this information can be used for risk management purposes. For example, an investor wanting to know their exposure to General Motors would use DV01 ("Dollar Value of a Basis Point") and Value on Default to assess the likely impact of spread widening on GM CDSs and a default, respectively. An investor more worried about systemic risk, which could be seen through a general widening of credit spreads, would use S.DV01 ("Systematic DV01") to see the sensitivity of their instrument.

1. Credit Spread Sensitivities

DV01 and S.DV01
One of the most important factors that drive the risk of a synthetic CDO is the credit quality of the instruments in the underlying pool of CDSs. The market view on their creditworthiness is mainly reflected in credit spreads: the wider the spreads the greater the perceived riskiness of the credit. However, not only the absolute level of credit spreads is relevant but also the risk of changing credit spreads.

Deltas and DV01s are associated with the risk of changing spreads for a particular reference entity, whereas S.DV01 is associated with measuring the impact of a market-wide increase in spreads that may occur with an increase in systemic risk.

DV01 is a straightforward measure showing the impact of increasing credit spreads on the value of the tranche.

DV01 with respect to a specific underlying credit is defined as the change in present value ("PV") of a synthetic CDO tranche as a result of a 1bp increase in the CDS spread of the underlying credit.

Considered below is the tranched DJ iTraxx Europe Series 5 with its five traded tranches. The data is as of 24 May 2006 (CDS spreads and market consensus recovery rates are from Valuspread. Base correlations are implied from observable market prices of iTraxx tranches provided by GFI).The calculations are done using Fitch Ratings' Risk Analytics Platform for Credit Derivatives ("RAP CD").

Chart 1 below shows the DV01 for the same credit for different subordination levels (represented by the attachment point). It can be observed that a credit's DV01 is higher for a tranche with less subordination. This is usually true; however, as noted in the Time to Maturity section, DV01s may not be ordered by tranche seniority for risky or long dated tranches. Usually, equity tranches are much more sensitive to spread movements of individual names than senior or mezzanine tranches.

Furthermore, Chart 2 shows the DV01 of the equity tranche for a number of credits with different spread levels. This allows the riskiest names in the portfolio to be identified.

In addition to DV01 for a specific credit, investors also look at S.DV01, the so-called systemic DV01.

S.DV01 measures the change in the PV of the CDO tranche for a 1bp increase in the spread of all names in the reference portfolio at all maturities.

The above chart 3 shows the S.DV01 for all tranches of the iTraxx EUR IG index.

As was the case with DV01, the S.DV01 is highest for the equity tranche and reduces as the attachment point of the tranche increases. Therefore, the equity tranche is the most sensitive to single-name and portfolio-wide spread moves.

Delta
Although the DV01 is a good measure for the sensitivity of the value of a CDO to changes in CDS spreads, the concept of DV01 is not sufficient to define a hedging strategy against spread changes. This requires a delta.

Tranche deltas can be used by tranche owners to hedge their exposure to movements in the spread of this credit, i.e. they define the percentage of the notional amount of the credit that needs to be sold/bought to hedge a short/long synthetic CDO tranche position.

However, even buy and hold investors, who do not intend to hedge their exposures, may find the delta to be a useful measure. For example, for managed transactions, delta gives an indication of the cost of substituting underlying credits. Deltas can range from 0% to 100%, but are usually quoted in terms of notional amount needed to hedge the tranche against spread movements of the particular name.

The tranche delta for a specific underlying credit is defined as the ratio of (i) the PV change of the tranche with respect to a 1bp shift in the CDS spread of the underlying credit over (ii) the PV change in the single name CDS on the same credit. The following chart shows the delta hedge notional for the tranches of the iTraxx EUR IG S5 index for one of the underlying credits. The delta is highest for the equity tranche and declines further up in the capital structure.

To immunise the equity tranche position against small movements in the spread of the credit, protection has to be purchased via a single-name CDS on that name with a swap notional equal to the delta hedge notional. The delta hedge only works for small spread movements in the underlying credit and fails to protect the investor against large movements. The general assumption is that spread movements will be small over short time intervals. This is clearly an assumption that is relied upon with care, i.e. it is important to also consider scenarios with idiosyncratic or systemic large spread moves. To remain fully hedged through time, the delta hedge has to be adjusted by buying or selling more protection in the underlying credit. However, due to transaction costs and a lack of liquidity for certain names, this adjustment might be too expensive or not possible at all.

An investor can delta-hedge all the names of a synthetic CDO. However, implementing and managing such a trade can be very expensive so investors will typically only hedge the exposures they are most concerned about.

Alternatively, index tranches could be hedged by taking a delta position in the underlying index portfolio. In this example, the equity investor could buy a certain amount (delta) of protection on the iTraxx EUR IG index. However, this would only be an average delta-hedge. In the case of the equity tranche, the tighter names (with higher-than-average deltas) would be over-hedged, the wider names (with lower than average deltas) would be under-hedged. For senior tranches it would be the other way round.

Tranche deltas depend on several parameters:
- attachment point and thickness of the synthetic CDO tranche;
- CDS spreads of underlying CDS contracts;
- time to maturity;
- default correlation; and
- recovery rates of the reference credits.

To illustrate the relationship of the deltas to each one of these parameters, Fitch employs a simplified example CDO structure with the following characteristics:
- reference portfolio of 125 names;
- flat credit spread of 30bp for each of the names;
- recovery rate of 40%;
- three tranches;
- senior: 9%-12%; 14bp;
- mezzanine 2: 6%-9%; 45bp;
- mezzanine 1: 3%-6%; 160bp; and
- equity: 0%-3%; 950bps; no upfront.

Subordination and Tranche Thickness

Changes in the subordination level, change the risk profile of the tranche. To visualise the dependency of deltas on the subordination level, tranche deltas for the example structure above were calculated. The smaller the subordination level, the higher the tranche delta for a specific credit. This is intuitive because tranches that are lower in the capital structure are more risky as there is less protection against default risk.

For a fixed level of subordination, a wider tranche is exposed to a larger band of losses and is therefore more risky. The increase in tranche delta flattens out once the limit of possible losses is reached.

Dependence on CDS Spreads
Changes in the credit spreads of the underlying reference instruments affect the risk profile of the reference pool and the tranche and therefore also the tranche delta. For the equity tranche, a higher credit spread of a specific name generally means a higher tranche delta for that credit. (For high spread levels, a decrease in delta is seen due to default correlation effects. High spread levels result in an increase in default correlation. For equity tranches, this reduces the expected losses and hence a reduction in the delta.)

For the senior tranche, a higher credit spread means a lower delta. This is caused by the different behaviour of tranche PV changes relative to CDS PV changes for equity and senior tranches. For mezzanine tranches, the relationship also depends on other factors (such as the correlation level).

Time to Maturity
When Fitch examined the five-year iTraxx tranches it was noted that DV01, S.DV01 and Delta all decrease as the seniority of a tranche increases. This is not necessarily the case.

Charts 7a and 7b below show obligor DV01s for the five-year and 10-year iTraxx tranches. For the 10 year example it can be seen that DV01s are no longer ordered by seniority – the 3-6 tranche has a higher DV01 than the equity tranche. This chart is based off real CDS credit spread and tranche implied spread term structure data, so the loss of ordering derives from a combination of maturity, dependence on correlation, and on CDS and index tranche term structure differences between the five-year and 10-year maturities.

One way to conceptualise this effect is that the tranche with the highest DV01s indicates where the expected pool loss will fall. Fitch's data indicates that for the five-year example the pool expected loss is less than 3% (the equity detachment point) and for the 10-year example the pool expected loss lies between 3% and 6% (i.e. in the first mezzanine tranche).

As time to maturity decreases, tranche deltas change. In general, the delta of the equity tranche converges to 100% of the credits reference notional and the deltas of all other tranches converge to 0% as time to maturity tends to zero (all other parameters assumed to be equal). The equity tranche becomes more risky in relation to the other tranches as there is less time for defaults to hit mezzanine or senior tranches.

Dependence on Correlation
Tranche deltas are also influenced by the changes in the underlying default correlations. Considering the simple implied correlation framework, if correlations increase, more of the risk is shifted to the senior tranche as higher correlations mean that there is a higher probability of joint defaults in the reference pool. Therefore, the senior tranche delta increases whereas the equity tranche delta decreases.

It is also interesting to examine how deltas change when switching between two alternative implied correlation frameworks (compound and base). Chart 8a shows DV01s for the 3-6 tranche of the 5 year iTraxx based on real market data and calibrated correlations in the base and compound correlation frameworks. Both frameworks were calibrated such that the 3-6 tranche was priced at the market level. However, it can be seen that although the modelled tranche spread is the same in both frameworks (73bp), the DV01s are dramatically different. DV01s in the base correlation framework are significantly lower than in the tranche correlation framework, and have the opposite curvature.

Dependence of Delta on Recovery Rates
Varying the recovery rate while keeping the spread fixed, amounts to changing the default probability of the credit, much like it changes when spreads move for fixed recoveries. Therefore, the dependencies between delta and recovery rates are very similar and are not discussed further here.

2. Instantaneous Default Risk or Value on Default

Value on Default
Deltas and DV01s assess the impact of a fall in the credit quality of a reference entity symbolised by a widening of its spread, but do not address the effect of a default on the value of a CDO tranche. Investors should look at Value on Default ("VOD") in combination with DV01 to get a fuller picture of the risk involved, especially if they are worried about the long-term health of an entity rather than a cyclical lull.

Fitch defines the VOD sensitivity as the net profit and loss of a tranche position resulting from an instantaneous default of a credit, keeping all other credit spreads in the portfolio unchanged.

For unhedged tranches, the default risk is high, for delta-hedged tranches, lower, but it can still be significant. Ideally, when hedging a synthetic CDO tranche, it is best to be delta- as well as default-neutral. A typical approach to combine delta- with default- hedging is the following: Apply delta-hedging as described above, but also trade short-term CDSs to be default-neutral too (if the maturity of the CDS contracts are very short, the delta is close to zero).

The following chart shows the equity, mezzanine and senior VOD for several names in the iTraxx EUR IG S5 index. It can be seen that the VOD is largest for the equity tranche, which actually includes the settlement payment of the underlying name. The VOD also strongly corresponds to the spreads of the underlying names and wider spread names have higher VODs. This makes sense intuitively as removing wider spread names will have a larger impact on the PV and VOD of the tranche than for tighter spread names.

The default of a high spread name will cause a larger improvement in the credit quality of the remaining performing portfolio compared to a low spread name. Assuming the net loss amount is the same in both cases (same recovery and same notional) the tranche PV following default of a high spread name will be greater compared to a low spread name. Of course the PV change is usually dominated by the reduced AP following the default of a name, which reduces the PV for the protection seller, since the spread remains the same. However the improved credit quality of the remaining portfolio compensates for some of the effect of the lower attachment points. As a result the VOD for a high spread name is higher compared to the VOD of a low spread name.

3. Correlation Sensitivity

Corr01
A synthetic CDO is a correlation product, which means the expected loss and therefore the value of a tranche depends on the correlation of the underlying reference pool. To determine whether an investor is getting a fair return for this risk, they must be able to measure the correlation risk. As with spreads, correlation may change over time, although the base correlation skew is reasonably stable most of the time. However, base correlation is also a measure for the relative supply and demand across the capital structure, so shifts over time may be caused by technical factors rather than a fundamental change in the market's view of correlation. Technical factors were a key driver behind the turmoil in the correlation market during May 2005. Fitch defines Corr01 as the change in the PV of a synthetic CDO tranche for a one percentage point increase in asset correlation.

Working within the base correlation framework, this means that the base correlation for the attachment and detachment points increase by 1%. In the compound correlation framework, this means that the compound correlation of a specific tranche increases by 1%.

Corr01 depends on several parameters:
- attachment point and thickness of a synthetic CDO tranche;
- average level of CDS spreads of underlying CDS contracts;
- time to maturity; and
- level of correlation.

Using the standard iTraxx EUR IG S5 example, Fitch calculated Corr01 for all tranches.
If a long correlation position is defined as having a positive Corr01 and a short correlation position a negative Corr01, then a long equity position (protection seller) has a long correlation position (positive Corr01) and a long senior position a short correlation position (negative Corr01). This means that for an investor in the equity tranche, an increase in correlation increases the PV of the tranche, because the expected PV of the default leg decreases and the expected PV of the premium leg increases.

As correlation increases, the portfolio loss distribution changes so that there are more extreme outcomes. An increase in correlation decreases the expected loss of the equity tranche and increases the expected loss of senior tranches. Therefore, Corr01 changes sign from negative to positive in a shift from junior to senior tranches. This allows one to easily create a correlation hedged tranche by choosing an appropriate attachment and detachment point around the point where the correlation sensitivity changes its sign. More generally, any desired correlation exposure can be obtained by choosing an appropriate tranche (i.e. attachment and detachment point) or combinations of tranches of the same CDO.

Note that the value of a single name CDS and the index portfolio have zero sensitivity to correlation. This is obvious for a single-name CDS. The index value only depends on the expected loss of the portfolio, which is not affected by correlation. Therefore, delta hedging (whether with a single-name CDS or the index) does not change the correlation sensitivity of the tranche.

4. The Effect of Time Decay

Traded index tranches are typically traded to have zero value at closing. This means the spread is set such that the expected value of the premium leg equals the expected value of the loss leg. The notable exceptions are equity tranches in the index, which are traded with an upfront payment.

Investors need to be aware that as time elapses the value of their CDO investment will change even if spreads and correlation remain the same. In this respect, synthetic CDOs behave very similarly to traditional bond investments, which can trade below or above par over time, even if valued at par at issuance.

The value of a synthetic CDO is determined by the value of the loss and premium leg in the underlying swap. The expected loss leg depends on the timing of defaults over time, which is driven by the term structure of credit spreads and is portfolio specific. For example, defaults in one portfolio could be more back-loaded while in another the default timing could be relatively even over time. Therefore, as time passes, the value of the loss leg and premium leg will change by different amounts and hence the tranche value may go up or down with the passage of time. The following chart shows the expected loss and the expected premium payments per quarter for the iTraxx EUR IG S5 3-6% tranche.

Losses in tranches with subordination are generally back-loaded because of the effect of the subordination preventing the first losses from impacting the tranche. The expected premium payment per quarter declines over time as expected losses accumulate, which causes an erosion of the tranche notional.

Chart 12 shows the change in the value for an investor selling protection in the iTraxx EUR IG S5 3-6% tranche between quarters. For the first 12 quarters, the change in value is negative, meaning the PV of the investment declines as time passes.

From quarter 14 onwards, the change in value becomes positive for the protection seller. In other words, the investment will gain value, because a shorter term removes expected losses at an increasing rate, which exceeds the reduction in premium payment. This behaviour is very similar to the "pull to par" for a traditional bond investment.

The change in value is often measured by theta, which is defined as the change in value of the tranche when the valuation date is moved on one day.

Base correlation and the spreads of the underlying names are assumed to remain the same.

Conclusion

Investors can use the sensitivities described in this report to gain a more complete understanding of the risks of a CDO. These sensitivities can be used to determine the impact of the main factors driving the value of a CDO, that is: spread movements, defaults, changes in correlation and time-decay.

Copyright 2006 by Fitch, Inc., Fitch Ratings Ltd. and its subsidiaries. This Research Note, originally published on 4 August 2006, is being republished with additional information on time to maturity and DV01s.