Cover.qxd
30/04/2004
11:48
Page 1
The of The credit guide to collateralised debt obligations Published in association with
Title page.qxd
28/04/2004
11:00
Page 1
Published in association with
The credit guide to collateralised debt obligations
Editor’s letter.qxd
27/04/2004
15:43
Cover illustration Tan Doan
Head office Incisive Financial Publishing Ltd Haymarket House 28-29 Haymarket London, SW1Y 4RX United Kingdom Tel +44 (0) 20 7484 9700 Fax +44 (0) 20 7930 2348 Editor, Credit David Watts Author Philip Moore Director of Editorial Services Celia Mather Subeditors Rebecca Geldard, Gary Fox Designer Matt Hadfield Associate Publisher Simon Crabb Publisher Sean O’Callaghan © Incisive Media Investments Ltd, 2004. All rights reserved. No part of this publication may be reproduced or introduced into any retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the copyright owners.
02
credit The ABC of CDS
Page 2
Editor’s letter.qxd
27/04/2004
15:43
Page 3
editor’s letter
Editor’s letter n the past few years, CDOs have experienced a baptism of fire. Just as ‘CDO’ was beginning to be a recognisable term in the financial markets the structure was hit by the record downgrades and defaults of 2001 and 2002. The natural leverage of the subordinated tranches caused many such tranches to be entirely wiped out. At that time it was difficult to see CDOs being rehabilitated back into mainstream finance, and especially not within the space of one year. But a succession of developments to better protect investors and the 2003 credit rally turned CDO from being a dirty term to practically a household phrase. The trend for CDOs to replace static CDOs is helping them to become a real tool for quick and easy diversification compared with the more accident-prone issue-and-forget structure of a static CDO. The idea of trading the correlation between credit defaults is becoming a widely accepted development, a possibility that has been driven by the tranching of debt. And while banks disagree on the exact methodology to use to price that correlation, the introduction of standardised CDO baskets in the form of tradable credit default swap indices is allowing the market to set the price for itself. In fact, the innovation that is doing more to make CDOs a recognisable and accepted product is the introduction of tradable credit default swap indices such as iBoxx and Trac-x, which allow investors to properly trade a diversified basket of credits. It is difficult to imagine that the market can continue the innovation seen over the past five years in the next half decade. But with so much innovation behind us, it is now time for the CDO to settle down and take its place as an accepted part of the credit market.
I
David Watts Credit
www.creditmag.com
credit The ABC of CDO
03
Contents.qxd
27/04/2004
15:53
Page 2
contents
Contents 06 Introduction
13 Static versus managed CDOs
The start of the market
Turning the tanker
Collateralised debt instruments were launched
CDOs were originally static portfolios where the
initially for mortgages in the 1980s, but the
underlying names rarely changed. But the fallout
market has been adopted over the years to
in credit since 2000 increased the attractiveness
cover practically every type of debt.
of portfolios watched over by fund managers.
17 Tranching
08 Market evolution The reason to issue
Structuring by seniority
Since the mid-1990s CDOs have become a way
CDOs provide senior creditors with greater
for originating firms to arbitrage rating
protection and allow subordinated investors to
inefficiencies.
lever up their investments.
12 Cashflow vs synthetic CDOs
20 Investor demand
The synthetic solution
Out of the ordinary
The development of the CDS
Investors have frequently questioned
market has allowed originators to issue CDOs
originating firms’ motivation for issuing
where the exposure is referenced via a CDS
CDOs. But the 2003 credit rally has largely
rather than cash assets.
rehabilitated the structure.
Winner! Risk Magazine’s 2004 Trading technology product of the year Trading - Trading Support - Risk Management 04
credit The ABC of CDS
Contents.qxd
27/04/2004
15:53
Page 3
contents
credit www.creditmag.com
21 Recovery rates
29 Rating agencies
Severity of default is the key
Rating the product
Investors in CDOs must feel confident about
The complexity and opaqueness of many CDOs
the circumstances under which they will win
mean rating agencies have carved out a
and lose. Here the recovery rate is the key to
niche in the market, helping investors
performance.
understand the risks.
30 Conclusion
23 Product innovation Diverse and sophisticated
The CDO contribution
CDO originators and arrangers have recently
Five years ago, CDOs were just a relatively
ignored no debt or proxy for debt to act as the
small part of the overall credit derivatives
underlying assets in a collateralisation.
market.Today they are still small in comparison with the overall financial market, but their
26 Market information
contribution is impossible to ignore.
Understanding the product
32 Index
Critics of CDOs express concern about transparency. In the past few
Index of terms and deals
years a succession of products and initiatives
A list of some major CDO transactions and a
has been produced to change this.
rundown of important terminology.
JRisk
TM
flexible credit trading solutions rapidly deployed www.application-networks.com www.creditmag.com
credit The ABC of CDS
05
1intro.qxd
27/04/2004
15:50
Page 2
introduction
The start of the market Since the collateralisation of debt instruments was first developed and applied to mortgage bonds in the 1980s, the market has grown at an exponential rate and quickly been adopted to cover practically every type of debt
C
lare Island, Galway Bay, Copernicus, Concerto, Thunderbird. The names are more likely to conjure up images of the racecard on Grand National Day than of a menu of financial products. These, and hundreds of other equally colourfully named instruments, are examples of collateralised debt obligations (CDOs), which in recent years have been among the fastest-developing investment vehicles in the financial services industry. As the rating agency Moody’s noted in its Review of activity in the structured finance arena in 2003: “The dramatic growth in the number of deals means that this market is now 100 times the size it was in 1998.” The breakneck speed of the market’s development has brought with it an apparently bewildering amount of jargon, introducing new issuers, intermediaries and investors to a world of combo-notes and repacks, of CDOs squared and rating agency drill-downs. The rapid emergence of much of this jargon, which this guide will aim to demystify, is in large measure a by-product of the dynamism, flexibility and adaptability of an instrument that has built upon a blindingly obvious concept. This is that the advancement of any form of credit should be based upon the ability of the borrower to repay – or on the collateral, security or compensation in the event of default that a borrower is able to provide. That concept is pivotal to instruments that can be generically described as ‘collateralised obligations’ (COs), one of the most straightforward definitions of which is the one given by Dutch asset management company Robeco to its investors in a description of one of its CDOs. This is that the instrument is simply a “promissory note backed by collateral or security”. In the market for COs, that security can be taken from a very wide spectrum of alternative financial instruments, such as bonds (collateralised bond obligations, or CBOs), loans (collateralised loan obligations, or CLOs), funds (collateralised fund obligations, or CFOs), mortgages (collateralised
00 06
credit The ABC of CDO
mortgage obligations, or CMOs) and others. They can (and frequently do) source their collateral from a combination of two or more of these asset classes. Collectively, these instruments are popularly referred to as CDOs, which are bond-like instruments that use the cashflows from their assets to pass coupon payments on to their investors. In a technique known as tranching (slicing up), those payments are made on a sequential basis, depending on the seniority of investors within the capital structure of the CDO.
The history of collateralisation The market for CDOs is generally believed to date back to the late 1980s and the repackaging and redistribution in the US by houses such as Drexel Burnham Lambert and Kidder Peabody of portfolios of high-yield bonds and loans. Prior to those transactions, however, a market for CMOs – the forerunner of modern CDOs – was taking shape in the US market by the early 1980s. In 1983, for example, the Federal Home Loan Mortgage Corporation pioneered structures that built on existing mortgage securitisation templates by creating so-called paythrough structures. These divided cashflows up into a number of tranches to suit investor preferences, and by the late 1980s these securities remained the only form of COs familiar to the market cognoscenti. In their 1988 book Securitization, S G Warburg’s John Henderson and Jonathan Scott make no mention of COs other than CMOs, which are described at length. And COs of any kind are the subject of even less attention in Securitization of Credit, published in the same year under the auspices of the McKinsey Securitization Project. That book makes no more than a single, passing reference to CMOs. Clearly, therefore, although portfolios of securities were being re-parcelled into COs by the late 1980s, the rapid evolution of CDOs is very much a story of the 1990s. In the US market, the degree to which that story was being backed by highly liquid,
1intro.qxd
27/04/2004
15:50
Page 3
introduction jumbo issuance was becoming self-evident by the middle of the 1990s, with volume breaking through the $10 billion mark for the first time in 1996. It was September 1997, however, that provided a key landmark for the US market when NationsBank launched a $4 billion CLO repackaging just over 1,000 commercial loans valued at about $6 billion. That made it the largest CLO the US market had ever seen, prompting an excited Barron’s, a US financial weekly newspaper, to describe the bond as “the type of transaction that bond geeks live for”. Unusually, however, by 1997 the US market had already been eclipsed by Europe in terms of its capacity to structure and distribute jumbo COs. A year before the NationsBank deal, the UK’s NatWest had thrown down the gauntlet in the market with the launch of Rose Funding, a $5 billion CLO backed by more than 200 loans to corporates in 17 different countries. By the late 1990s, the structure of the international market for COs of all kinds was becoming characterised by a number of conspicuous and interrelated trends. First, issuance volume was rising exponentially, as was understanding and acceptance of the CDO technique. That process led Duff & Phelps (now part of Fitch Ratings) to describe 1999 as a year that was marked by “considerable maturation and change in the CDO market”, and as the landmark period in which CDOs crossed what the agency described as “the line from ‘esoteric’ to ‘mainstream’ assets”. Second, cross-border investment flows into CDOs were rising steeply. For example, in November 1998, when BankBoston launched a $2.18 billion CLO, more than 25% of the notes were sold to investors outside the US, which was viewed at the time as an unusually high proportion. Third, by now more and more asset classes were being used as security for COs. In February 1998, for example, Credit Suisse First Boston (CSFB) launched the first CLO backed entirely by project finance loans, with a $617 million transaction collateralised against 41 fully secured loans, the majority of which were accounted for by power projects. Another trend that had become conspicuous by 1999 and, more strikingly, 2000 was the speed with which the concept of the CO was being popularised across continental Europe. As Europe’s largest economy, Germany was an important source of new deal www.creditmag.com
flow in the CDO market. Elsewhere in Europe, in November 1999, Banca Commerciale Italiana (BCI) became the first Italian bank to issue a public CLO backed by its corporate loan book in an €170m transaction dubbed Scala 1 Ltd. The following summer, meanwhile, saw the first securitisation of European loans by a Belgian bank (KBC Bank), as well as the first CLO in Switzerland for UBS (named the Helvetic Asset Trust). By the late 1990s, changes to legislation and regulation were emerging as important sources of support for new issuance in the CDO market in some pockets of Europe. In May 1999, Spain passed a law that specifically allowed for the provision of government guarantees for securitisations of loans originated by banks under a special line of credit for small and medium-sized enterprises (pequeñas y medianas empresas) from Instituto de Credito Oficial, the state funding agency. That allowed for
The breakneck speed of the market’s development has brought with it an apparently bewildering amount of jargon the issuance of CLOs backed by tranches effectively offering investors quasi-government risk. By 2000 and 2001, the most important determinant of increasing volumes in the CDO market globally, however, was the explosive growth in the market for credit derivatives in general, and for credit default swaps (CDS) in particular. That growth paved the way for an equally explosive expansion of the market for synthetic CDOs, which had made their first appearance in Europe at the end of 1997. In 2003, 92% of all European CDOs rated by Moody’s were accounted for by synthetic structures, up from 88% in 2002. Following its explosive growth between 1997 and 2002, the expansion in the European market for CDOs paused for breath in 2003. Moody’s reported in January 2004 that the total market volume of rated CDO issuance in the Europe, Middle East and Africa region declined by 8% to $82 billion (€71 billion) in 2003. ■ credit The ABC of CDO
07
2recentperf.qxd
27/04/2004
16:01
Page 2
market evolution
The reason to issue The initial reason for collateralisation of debt was the same as that for any securitisation, to free up more of the banks’ balance sheets. But since the mid-1990s CDOs have become a way for originating firms to arbitrage rating inefficiencies
A
n apparent paradox associated with the rapid expansion of the CDO market is that it has come against a backdrop of a performance track record that can hardly be said to have been distinguished. A study published by Moody’s in February 2003 found that between 1991 and 2002, CDOs had what the agency described as an “extremely high downgrade rate” (of 10.9%) and a very low upgrade rate (of 0.6%). That, Moody’s explained, was “primarily due to the extraordinary number of downgrades and defaults in the corporate bonds that underlie these securities”. Nevertheless, support for the CDO market from mainstream institutional investors continued to grow. Federal Reserve chairman Alan Greenspan observed in January 2004 that “insurance compa-
Part of the complexity associated with the CDO market arises from the fact that originators have different reasons for issuing CDOs nies, especially those in reinsurance, pension funds and hedge funds, continue to be willing, at a price, to supply this credit protection, despite the significant losses on such products that some of these investors experienced during the past three years”. There is a simple enough explanation for the persistence of this apparently irrational demand, which is that against a backdrop of collapsing returns from government, supranational, agency and other toprated assets, the structure of the CDO market is such that it can offer much more attractive yields for comparably rated securities. Returns in the CDO
08
credit The ABC of CDO
market are also considerably higher than in other securitised instruments with longer track records. According to research published in 2003 by Barclays Capital, triple-A rated tranches of CDOs have spreads as high as three times wider than credit card-backed deals.
Credit quality in 2003 The decline in the credit quality of CDOs slowed dramatically in 2003. Moody’s downgraded the tranches of 279 CDOs in 2003, compared with 339 in 2002, and the majority of those downgrades came during the first half of the year. Additionally, the number of Moody’s upgrades almost tripled from 10 in 2002 to 28 in 2003. That fed through into a recovery in demand for CDOs in 2003. As Dresdner Kleinwort Wasserstein (DrKW) explained in its review of activity in the market in 2003: “Stabilising credit fundamentals, structural innovation and a resurgence in investor demand marked a more positive year for European CDOs in 2003.” Balance-sheet and arbitrage CDOs Part of the complexity associated with the CDO market arises from the fact that originators can and do have different – even antithetical – reasons for issuing CDOs. The very different motives for issuing is one pivotal difference between the two fundamental forms of all COs, which are known either as balance-sheet or arbitrage instruments. In very crude terms, a balance-sheet CDO is issued by a bank almost out of necessity, or as a means of addressing an existing problem or challenge. An arbitrage obligation, by contrast, will generally be issued by an asset management company as a strategic means of exploiting latent opportunities arising from perceived market inefficiencies. On either side of the Atlantic, it was the issuance of balance-sheet CLOs by commercial banks that formed the basis for the growth of the broader
2recentperf.qxd
27/04/2004
16:01
Page 3
market evolution A second, related influence on the growth of the CLO market in the 1990s was the increasing emphasis 50 10 that banks were forced to CDO equity (left-hand axis) 40 8 place on the delivery of Baa US credit - (right-hand axis) 30 6 enhanced shareholder value 20 4 and improved return on equi10 2 ty (ROE) , which in conti0 0 nental Europe were especially -10 -2 low. By 1996, publicly quot-20 -4 ed banks in the UK were 1991 1993 1995 1997 1999 2001 2003 enjoying average ROE levels Source: Lehman Brothers Equity performance has been calculated for CDOs containing triple-B rated credits, in excess of 20% (with Lloyds so comparison is with Baa section of the Lehman’s US credit index TSB leading the way with an ROE of over 30%). By conCDO market in the mid to late 1990s. A balance- trast, the average in continental Europe was a little sheet CLO is a form of securitisation in which assets over 11%, with banks in the largest markets posting (in this instance, loans) are removed from a bank’s ROEs well below even this modest average. In balance sheet and repackaged into marketable secu- Germany in 1996, the average was 7.5%, in France rities that are then sold on a private placement basis it was 6.9% and in the notoriously inefficient Italian banking sector it struggled to reach 3%. to investors. For a variety of reasons, feeble ROEs in the In the 1990s, at least three influences combined to fuel the expansion in banks’ issuance of banking sector in Europe had not much mattered in balance-sheet CLOs. The first and comfortably the 1980s and early 1990s. In many countries, the most important of these can be traced back to banking industries remained sheltered within a the Basel Capital Accord of 1988, which laid cocoon of local protectionism and regulation that down the first, universally accepted framework for discouraged competition among largely statecalculating bank capital for regulatory purposes. owned banks and was unwelcoming to foreign Securitisation in general, and the use of the fast- financial services providers. But as privatisation, libdeveloping CLO market in particular, allowed eralisation, deregulation and – ultimately – consolibanks to transfer the risk associated with their dation all gathered momentum in Europe in the loan portfolios to third parties via the capital mar- 1990s, banks were forced to focus more intensely ket, which in turn qualified them for regulatory on their profitability. Securitisation rapidly emerged capital relief and removed or reduced constraints as one important tool to help in the process. A third, albeit less important, driver for increased on fresh lending capacities. Rating agencies believe that the new Basel II issuance in the CLO market in the 1990s was the guidelines will reduce the attractiveness of securiti- recognition among commercial lenders that institusation as a means of achieving regulatory capital tional investors were probably much more profirelief and more efficient capital allocation. “In our cient at pricing and managing credit risk than the opinion,” notes Fitch in a report published in banks were. In Europe (as it had been many years February 2004, “Basel II, as it is now proposed, will before in the US), the development of securitisation indeed serve to make origination of securitisation was part of a broader transformation in the corpoless attractive to banks, given the stricter (in com- rate finance landscape. Among UK clearing banks, Barclays had pioparison with Basel I) capital requirements.” That may well be. But it is clear that as far as the CDO neered the concept of securitisation in the early world is concerned, Basel has already played a criti- 1990s, with the launch of a £280 million securitisation of personal loans in October 1993. But the cal role in kick-starting activity in the market. www.creditmag.com
Performance of Baa US credits (%)
Performance of CDO equity (%)
Performance of average equity tranche versus Lehman’s US Credit index
credit The ABC of CDO
09
2recentperf.qxd
27/04/2004
16:01
Page 4
market evolution market for CLOs took its most significant step forward in September 1996 with the jumbo, $5 billion securitisation by NatWest of an international portfolio of more than 200 corporate loans. In the first half of 1996, NatWest had posted an ROE of just 13.4%, which was well below the average for the UK banking sector, making the bank’s shareholders increasingly agitated. Issuance of CLOs in the US market was also motivated in large measure by banks’ need to make more efficient use of capital and to bolster ROE levels. When NationsBank launched its record-breaking $4 billion CLO deal in September 1997, its publicly stated motive was to bring about a reverse in the decline of its ROE, which had fallen from 18.5% in 1996 to 14.6% in the first half of 1997. In the late 1990s, the CDO market was also viewed by the Japanese banks as being an appealing means of transferring the risk on large international portfolios of loans. The year 1998 saw transactions such as Industrial Bank of Japan’s $1.165 billion CLO of triple-A loans, swiftly followed by an even larger $2.4 billion deal for Sumitomo Bank. Sumitomo was also the originator in 1998 of the £479 million Aurora deal, which was the first occasion on which a Japanese bank had launched a CLO backed by a portfolio of UK corporate loans. Among continental European banks, meanwhile, Deutsche Bank in particular was involved in an all-out campaign to bolster its ROE in the late 1990s. In July 1998, it launched the first of its Core (corporate and real estate) securitisations – a landmark securitisation of more than 5,000 loans worth $2.4 billion (equivalent) to some 4,000 companies and carrying ratings of between triple-A and double-B. As much of the portfolio backing Core was accounted for by loans to medium-sized German companies, for many investors this represented the first occasion on which they were offered exposure to loans to the so-called Mittelstand companies often accredited for being the motor of Europe’s largest economy. The following year, Deutsche began to securitise loans from its global banking division. The first transaction under its Globe programme, securitising loans extended by its global banking operation to large corporates, was a CLO launched in
10
credit The ABC of CDO
May 1999 backed by 144 loans to 75 companies rated from as high as A+ to as low as C–. While these transactions were so-called traditional or ‘true sale’ securitisations, in 1999 Deutsche launched its first synthetic CLO (see page 12), which was named Blue Stripe 1999-1 (after the Deutsche logo) and referenced 330 loans to 240 companies. The second Globe transaction that followed soon afterwards was also a synthetic deal, as was the first Cast transaction launched by Deutsche Bank in 2000, which parcelled just over €4.5 billion of loans to small and medium-sized companies and broke new ground at the time for being the largest public synthetic CLO ever issued. For Deutsche, it was also an innovative structure in that it parcelled short-term loans denominated in 10 currencies. Another German bank that helped to spearhead the development of the CDO market in Continental Europe was Bayerische Hypotheken- und Vereinsbank, which in February 1999 established a landmark when it launched its €2.2 billion Guarantees of Euro-Loan Debt in Luxembourg (Geldilux) transaction, which was the first public CLO denominated in euros. While the majority of balance-sheet CDOs originated by commercial banks have been motivated by regulatory capital considerations, some issuers have had alternative objectives. In the case of the Caesar Finance CDOs launched by Banco di Roma in 2000, for example, the bank’s stated goal was to support the development of the broader Italian corporate bond market rather than to achieve capital relief. Thanks in no small measure to transactions such as Core, Geldilux and others, by the late 1990s an active, liquid and increasingly wellunderstood market for balance-sheet CDOs originated by banks had taken shape in Europe, but it was not until 1999 that the other basic form of CO, the arbitrage CDO, took its bow in the European market. The motives for issuing arbitrage CDOs are radically different from those prompting the launch of balance-sheet instruments, with the market driven by asset management companies using the CDO technique principally as a means of increasing their assets under management. In a nutshell, in an arbitrage CDO, the originator’s objective is to take advantage of
2recentperf.qxd
27/04/2004
16:01
Page 5
market evolution the yield differential between the assets within a CDO portfolio and the cost of funding the CDO through the sale to investors of securities. A key difference between a balance-sheet and an arbitrage CDO is that in the case of a balance-sheet instrument the issuer will be securitising assets that it already owns, whereas in an arbitrage CDO the issuer will generally buy new assets in the market earmarked as collateral for a new CDO issue. This explains why, in the US capital market in particular, arbitrage CDOs are often identified as important sources of demand in the new issue market for investment-grade and high-yield bonds as well as in the syndicated loans market. Arbitrage deals are now the main motor of expansion in the CDO market in Europe. According to DrKW’s review of market activity in 2003: “Arbitrage was the prevailing motive for European CDOs in 2003, with only around a quarter of public transactions done for balance-sheet purposes and the majority of these backed by loans to small and mediumsized enterprises.” That represents a very rapid growth rate for the arbitrage CDO market, given that Europe’s first deal was launched in August 1999 by the Intermediate Capital Group (ICG). Eurocredit I was a €400m CDO arranged by Morgan Stanley that securitised a portfolio of high-yield bonds, senior debt and mezzanine loans. Strength of investor demand for its inaugural arbitrage CDO encouraged ICG to launch Eurocredit II just over a year later, by which time a number of other asset management companies in Europe were also scrutinising the potential of the market for arbitrage CDOs. In Continental Europe, Deutsche Bank’s asset management arm, DWS, was at the forefront in terms of issuance of arbitrage CBOs, launching its first deal in May 2000 with a €318 million transaction backed by a mixture of European and US investment-grade and high-yield bonds. Among other asset management companies, one of the most regular, successful and innovative players in the CDO market has been Axa Investment Managers. In February 2002, it followed up on its previous ‘classical’ CDO (named Concerto) with the launch of the actively managed synthetic €1.05 billion Jazz CDO 1. As the Credit Guide to CDOs published in 2002 commented: “Jazz rewrites the rules, allowing www.creditmag.com
its managers to do their own thing. This being a CDO, there are still some strict rules for the portfolio manager to follow, but Jazz has achieved a degree of flexibility never seen before in a cashflow CDO.” Aside from being something of a landmark for the market for so-called managed CDOs (see page 15), another notable characteristic of the Jazz product was that it represented the first time that an arbitrage CDO was able to take short positions. A more recent contribution made by Axa to innovation in the CDO market was its launch of Overture, which at $3.5 billion was the largest CDO ever structured in the European market. More important than its size, however, was the way in which it was distributed. In that sense, the Overture deal was distributed like a conventional public bond, rather than as a private placement,
CDOs are often identified as important sources of demand in the new issue market which is very good for liquidity. The fact that the deal was bought by 96 institutions – more than in any previous CDO – attests to the success of the syndication method. The growth of the market for arbitrage CDOs in Europe was characterised in 2001 and 2002 by the emergence of increasingly specialised and liquid CDOs. Duke Street Capital’s Duchess CDO, which was launched in June 2001, was an example of a specialised arbitrage product that raised the bar in terms of its size. Initially an €750 million CDO, Duchess was the largest European product specialising in leveraged buyouts, which in February 2002 was tapped for an additional €250 million, bringing its total size to the €1 billion threshold. Transactions of this size demonstrated very clearly the power of the CDO to add to originators’ total assets under management. So too did the success with which Axa, for example, rolled out its series of innovative arbitrage CDOs in 2002. In that year, Axa reported that its institutional business division had attracted total inflows of €4.1 billion, of which three CDOs alone (Jazz I, Jazz II and Khaleej) accounted for €3.5 billion. ■ credit The ABC of CDO
11
3cashvsynth.qxd
27/04/2004
16:05
Page 2
cashflow versus synthethic CDOs
The synthetic solution The development of the credit default swap market has allowed originators to issue CDOs where the exposure to underlying credits is referenced via a CDS rather than directly to the funded asset
Balance-sheet and arbitrage CDOs can be structured as cashflow or synthetic instruments, although an increasingly popular formula among originators is to combine the two into so-called hybrid CDOs. The cashflow CDO, which formed the bread and butter of the market in its formative years, is a structure in which CDO notes are collateralised by a portfolio of cash assets purchased by the originator. In other words, in this classical structure the CDO owns the physical bond, loan or other security referenced by the instrument. The volume of traditional cashflow CDOs has been eclipsed in recent years by synthetic products, sometimes referred to as collateralised synthetic obligations. In a synthetic CDO, no legal or economic transfer of bonds or loans take place, with the underlying reference pool of assets remaining on the balance sheet of the originator. Instead, the CDO gains exposure to credit risk by selling protection to others through a CDS, which functions very much like an insurance contract. In other words, the CDO is still being paid for bearing credit risk, just as it
would do if it physically owned a bond or loan. From the perspective of originators, there are a number of clear benefits associated with synthetic CDOs. One of these is that risk transfer via synthetic structures allows bank originators in the CDO market to ensure that client relationships are not jeopardised. That is an especially relevant consideration in the market for CLOs, given that deal documentation in the syndicated lending market often prevents the transfer of loan ownership. Even where loan transfer is permitted, CDOs would often need, in theory, to secure the written permission of each borrower in order to construct a cashflow, which would amount to an impractical burden. Synthetic structures are also attractive for originators securitising multi-jurisdictional portfolios or loans made in countries where the local legal framework either does not allow for the so-called ‘true sale’ of assets or, more probably, where the local tax system makes the transfer of legal title of assets uneconomic. Until recently, it was the tax-related
Cash collateralised loan obligation Step 1 Bank sells loans to SPV Loan bank
Loan portfolio
Step 2 SPV securitises loan portfolio
credit The ABC of CDO
Capital markets
SPV
Loan portfolio
* The equity tranche is frequently retained by the loan bank
12
Step 3 Credit risk is tranched and sold on to bondholders
Triple-A notes
Investors
Double-A notes
Investors
Triple - B notes
Investors
Equity / first loss
Investors*
3cashvsynth.qxd
27/04/2004
16:05
Page 3
cashflow versus synthethic CDOs
Synthetic collateralised loan obligation Step 1
Step 2
Loan bank buys CDS protection on least risky ~90% of loan portfolio from big bank
Loan bank
Big bank* Exposure (via CDS) to least risky ~90% of loan portfolio
CDS protection
CDS premium
Step 3
Loan bank buys CDS protection on most risky ~10% of from SPV
Loan portfolio
SPV sells loan portfolio exposure as credit linked notes to investors
SPV CDS protection
CDS premium
Exposure (via CDS) to riskiest ~10% of loan portfolio
supersenior swap
* This is typically referred to as a bank in an OECD country
disadvantages with true sales that made synthetics the main source of securitisations in Germany. The market for synthetic CDOs owes its dramatic growth in recent years to the explosive expansion in the market for CDS. A CDS is a privately negotiated bilateral agreement in which one party, variously known as the protection buyer or risk shedder, pays a premium to another, generally referred to as the protection seller or risk taker, in order to secure protection against any losses that may be incurred through exposure to an investment as a result of an unforeseen development (or ‘credit event’). A Deutsche Bank report on synthetic CDOs traces the strong growth in investment-grade CDOs back to 2000, by which time – notes the Deutsche report – “the credit default swap market was expanding at a seemingly exponential rate. We estimate the outstanding notional amount was growing at about 75% per annum and that the market totalled about €800 billion. Between the US and Europe, about 150–200 names were actively traded.” Since then, liquidity in the CDS market has continued to grow at breakneck speed, with some estimates suggesting that by the end of 2004, the CDS market will be worth some $4,800 billion. For investors there are a number of important attractions associated with exposure to the CDS market rather than to cash bonds. CDOs made up of CDS allow investors to buy ‘pure’ credit because the structure separates the credit risk component of www.creditmag.com
triple -A CLNs (CDS exposure + Pfandbrief )
Investors
double -B CLNs (CDS exposure + Pfandbrief )
Investors
triple -B CLNs (CDS exposure + Pfandbrief )
Investors
Unrated equity tranche (first loss tranche) †
Investors †
The equity tranche is frequently retained by the Loan Bank
from the other asset’s risks, such as interest rate and currency risk. The explosion of liquidity in the CDS market has had a beneficial knock-on effect on the market for synthetic CDOs at a number of levels. For one thing, it has allowed for portfolios of default swaps to be assembled (or ramped up) much more quickly than those of cash instruments. As a report published by Citigroup analysis explains: “This is especially important in such markets as European investment-grade cash, where the liquidity in the corporate bond market does not permit the ramp-up of a diversified portfolio within such a short time.” Synthetic CDOs began to appear for the first time in the European market in 1997, with JP Morgan’s Bistro (Broad Index Secured Trust Offering), launched in December of that year, one of the first instruments of its kind to transfer the risks embedded in a portfolio of loans to the capital market, and hence reduce regulatory capital requirements. Synthetic CDOs were much slower to catch on in the Asian market, which was attributed by some market commentators to a reluctance among Asian investors to buy bonds that are not backed by physical, tangible assets, which in turn explained why the broader CDS market was slower to develop in Asia than in Europe. Since 2001, however, issuance of synthetic products has been rapidly gaining in popularity in Asia. credit The ABC of CDO
13
3cashvsynth.qxd
27/04/2004
16:05
Page 4
cashflow versus synthethic CDOs Recently the market has seen the development of standardised synthetic CDOs in the form of tradable CDS indices – most notably iBoxx and Trac-x (see page 25). These products allow investors to buy and sell a proxy for credit market risk and individual subsectors, quickly.
The structure of a synthetic CDO In the so-called unfunded portion of a synthetic CDO, the risk embedded in a portfolio of assets (as opposed to the assets themselves) is transferred directly to a ‘super-senior counterparty’ via a supersenior CDS. In this instance, the CDO acts as the protection buyer, by agreeing to pay a premium to the counterparty (the protection seller) in return for a commitment from the counterparty to pay compensation to the CDO in the event of any defaults in the reference portfolio. The super-senior swap is a vital driver behind the economics of a synthetic CLO and the key reason underlying the compelling cost benefits of these structures for originators. Within a synthetic structure, the super-senior swap will typically account for at least 80% of the CLO’s capital structure, and will generally be provided by a highly rated bank or insurance company. Those super-senior buyers or sellers of credit protection are attracted by the security of the instrument, which is often referred to as a ‘quasi quadruple-A’ or triple-A-plus tranche, and is therefore, presumably a more solid credit than the US government or the World Bank, which of course is not possible. Nevertheless, it is broadly accepted that the risk embedded in the super-senior tranche of a syn-
Growth in cashflow and synthetic CDOs 70 Synthetic (balance sheet)
billions €
60
Synthetic (arbitrage CDO)
50
Cash (balance sheet CDO)
40
Cash (arbitrage CDO)
30 20 10 0
1998
1999
2000
Source: Lehman Brothers
14
credit The ABC of CDO
2001
2002
2003
thetic CLO referencing a pool of investment-grade assets is remote in the extreme. This is because even if a super-senior swap accounts for as much as 90% of a CDO’s capital structure (which is not uncommon), more than 10% of the assets within the reference pool would have to default for losses to be sustained by the most senior tranche – an improbably high default rate for investment-grade bonds when the historical norm has been for a default rate of about 0.3%. Because the perceived risk associated with the super-senior swap is so low, the investor in this tranche is typically paid a premium that is no more than 8 basis points to 10bp of the CDO’s notional size, which is considerably
The market for synthetic CDOs owes its dramatic growth in recent years to the explosive expansion in the market for CDS below what an investor in the most senior tranche of a cash CDO would demand. The very compelling cost benefits associated with synthetic CDOs compared with their traditional cash counterparts are neatly outlined in a primer on the market published in 2002 by Bear Stearns, which analyses the liability structure of a hypothetical CDO with a notional value of $1 billion compared with that of a cash product. The collateral pool for both CDOs is investment-grade credits. But in the synthetic transaction, no cash is paid upfront for physical bonds, whereas in the cash CDO the entire liability structure is used to fund the physical purchase of the collateral. In the synthetic CDO, 89% of the capital structure is accounted for by the super-senior swap, for which the CDO is paying just 8bp. The result is that the weighted average cost of liabilities for the whole capital structure is 20bp. That compares with a weighted average cost of 66bp for a comparable cash CDO in which 85% of the capital structure is accounted for by triple-A Class A notes, 10% by A3 Class B notes and the remaining 5% by junior-most equity. ■
4statvmancdos.qxd
27/04/2004
16:07
Page 1
static versus managed CDOs
Turning the tanker CDOs were originally static portfolios where the underlying names rarely changed. The fallout in credit in the first years of the new millennium changed this, and increased the attractiveness of portfolios actively watched over by fund managers
I
t can be helpful to think of a CDO like an ordinary business. The management of a CDO is given the task of investing in debt instruments and the better it is at doing that, the more the equity holders will receive. Rather than selling a product the CDO manager is buying debt. The CDO’s selling of bonds is analogous to any company using the bond markets to finance its capital structure. In the formative stages of the development of the market, CDOs were generally categorised as ‘static’ or ‘passive’ products, meaning that the original composition of their underlying portfolios remains unchanged. The obvious advantage associated with this structure is that it calls for minimal resources in terms of management expertise and time, and reduces the costs involved in trading or ‘churning’ a portfolio. The drawback with the static structure is that it has all the manoeuvrability of an oil tanker. As long as credit quality was improving or remaining stable, that did not matter. But as credit quality began to deteriorate sharply in the downturn between 2000 and 2002, many investors in static CDOs found themselves holding instruments that were declining in value and – worse – were unable to do anything to reverse that decline. The result was that by 2001 and 2002, actively managed CDOs – in particular managed synthetic products – were rapidly gaining in popularity. The growth of managed products, however, was also helped by the growing maturity of the CDO market, and by the increasing number of managers with proven experience in managing credit in general and credit derivatives in particular.
The asset manager With the expansion of managed CDOs at the expense of more traditional static products, asset managers (or collateral managers) have become increasingly important protagonists in the CDO www.creditmag.com
world. They will generally hold some of the equity in their CDO so as to ensure that they have a vested interest in the success of the business. However, the growing responsibilities of CDO asset managers in Europe has caused disquiet among some observers who have suggested that the majority of European managers remain highly inexperienced relative to their US counterparts. Although the overall European market for managed CDOs may be in its infancy relative to the US, some sub-sectors of the market are maturing rapidly. In its 2003 year-end review, DrKW points out that of the 35 managed CDOs launched during the year, more than half were by repeat issuers, suggesting that a handful of the best CDO managers are now developing enhanced credibility based on established track records. Credit’s April 2004 salary survey has shown that complexity remains profitable and trading the more exotic structured instruments pays. Base salaries for traders of such products as single-tranche CDOs are 25% up on a year ago, and bonuses are 50% higher. One analyst that moved between American banks last year received $500,000 in total compensation and a structured product sales professional at a leading US investment bank is understood to have received $3 million. According to Napier Scott’s salary survey, London-based tier-one banks paid managing directors in exotic trading £125,000 basic plus a £1.5 million bonus. In credit structuring, managing directors received £125,000 and a £1.1 million bonus. As asset managers will generally hold some of the equity in their CDOs, their size can be an important consideration, given that they will need to have the resources to hold (and, in some cases, to replenish) the equity. Size can also be an important consideration if it gives the CDO manager the resources it needs to maintain an in-house research credit The ABC of CDO
15
4statvmancdos.qxd
27/04/2004
16:07
Page 2
static versus managed CDOs department. Nevertheless, analysts appear to agree that there are no hard and fast rules about the benefit of the size of a collateral manager. As a JP Morgan report observes: “Some CDO investors prefer a small management company that will be focused on the CDO. Others prefer a large manager that has available back-up personnel and clout in gaining access to allocations.” While size may therefore not be relevant, proven competence and management skills most assuredly are. Those skills will feed directly through to the management company’s bottom line, given that the CDO manager payment is typically made up of a base management fee twinned with an incentive management fee that is strongly performance-related. That additional incentive fee will generally only be paid if certain predetermined targets (known as the manager’s hurdle rate) are met. The market is becoming an increasingly efficient arbiter of the competence of collateral managers in the CDO market. By 2001 and 2002, an increasingly conspicuous process of price-tiering had emerged in the market, whereby those managers with poor track records were penalised by investors who would demand a premium of between 3bp and 5bp for CDOs managed by those with less impressive performance histories. Rating agencies are also important referees of the quality of collateral managers. Although much of the rating agencies’ analytical work is quantita-
tive, all the leading agencies also publish detailed qualitative research on the performance, strengths and weaknesses of collateral managers. Fitch assigns scores ranging from CAM1 to CAM5 to the asset managers it rates, with CAM1 representing the top rating. Those ratings are subdivided into scores based on a range of criteria, including: company and management experience; financial condition; staffing; procedures and controls; credit underwriting and asset selection; portfolio management; CDO administration; technology; and portfolio performance. Among the other top agencies, Standard & Poor’s (S&P) describes its CDO Manager Focus as “a comprehensive report of a CDO manager’s capabilities and track record developed through in-depth site visits and evaluation of past transactions.” In March 2003, meanwhile, Moody’s announced the release of a new CDO performance report series aimed at helping market participants track and compare the credit performance of US CDO transactions over time. According to the agency: “Moody’s Deal Score Reports provide investors and other market participants with independent, objective criteria by which they can measure an individual deal’s performance. Rather than focusing on equity performance or total return, these reports measure and compare Moody’s rating performance on the deal and the manager level.” ■
Proportion of managed synthetic CDOs to static synthetic CDOs 100%
80% static managed 60% 100%
91%
92%
9% 2001
8% 2002
78%
40%
20% 22% 0%
2000
Source: Lehman Brothers 'European CDOs: review & outlook'
16
credit The ABC of CDO
2003
5tranching.qxd
27/04/2004
16:08
Page 1
tranching
Structuring by seniority CDOs use classical securitisation methodology to provide senior creditors with greater protection and allow subordinated investors to lever up their investments
A
common characteristic of all securisations, including CDO collateralisations, is socalled tranching – the structuring of the product into a number of different classes of notes ranked by the seniority of investors’ claims on the instrument’s assets and cashflows. As with any business, alongside the ‘owners’ (the holders of the equity tranche), a CDO has creditors with varying degrees of seniority. The more senior the creditor, the less risky the investment and hence the less they will be paid in interest. The way it works is frequently referred to as a ‘waterfall’ or cascade of
cashflows, because in bankruptcy the proceeds from liquidating a CDO assets will first be used to repay the most senior creditors, the senior debt tranche, and only then, if there is remaining money, the next most senior tranche. And so on. The most senior note is rated triple-A, with the tranche below this generally referred to as the mezzanine notes, which are usually rated from high triple-B to low single-B. These can be in fixed- or floating-rate form and pay note-holders a regular coupon. As such each individual tranche is very much like a bond.
Over-collateralisation Over-collateralisation (OC) is one of a broader range of structural features of CDOs – collectively known as credit enhancement – that allows for higher-quality debt to be issued relative to lower-rated underlying collateral. The concept of over-collateralisation is pivotal to all forms of securitisation, and refers to the excess of the par amount of collateral available to secure one or more of the note classes over the par amount of those notes. To illustrate how the level of OC is determined, consider this example described by Standard & Poor’s, which is a cashflow transaction involving the issuance of $80 million of rated senior debt supported by a collateral pool with a total par value of $100 million. This is therefore known as an ‘80/20’ liability structure consisting of 80% of rated senior debt and 20% of unrated supporting debt or equity. The level of over-collateralisation is 125%, which equals the ratio of assets over liabilities.
www.creditmag.com
Tests to ensure that the OC level is maintained (OC tests) fall into two categories. The first is the par value test, which requires that the value of the rated notes is equal to a minimum percentage of the underlying collateral. The higher the ranking of the note in the capital structure, the higher this is required to be. In other words, the par value test may call for 115% coverage in the case of the senior notes and for 105% in the case of the mezzanine tranche. The second OC test is known as the interest coverage test. This is designed to ensure that interest income earned by the collateral is sufficient to cover potential losses and to maintain interest payments to senior note-holders, with the difference between the two referred to as the excess spread. In the event of a breach of the OC test, managers will need to remedy the situation usually within two to 10 days by, for example, purchasing additional collateral with any available excess interest.
credit The ABC of CDO
17
5tranching.qxd
27/04/2004
16:08
Page 2
tranching
The cashflows of the waterfall To illustrate how the cashflows of the CDO waterfall function, consider this simple example of a CDO of a portfolio of $300 million of triple-B rated asset-backed securities (ABS) described in a Bear Stearns primer to the CDO market published in 2002. In this instance, the portfolio generates an annual cashflow of 4.25% after the deduction of the CDO’s expenses and hedging costs, which equates to a total of approximately $12.7 million. Deducted from that total is the $7.3 million paid to the holders of three tranches of rated notes. The most senior tranche is accounted for by the $240 million of Aaa/AAA rated Class A-1 notes, paying Libor plus 54bp, a total of $5.7 million. The next claim is from the holders of the $26 million Aa3/AA– rated Class A-2 notes, paying Libor + 79bp, or $700,000, with the holders of the more deeply subordinated $20 million tranche of Baa2/BBB rated Class B paper bringing up the rear in terms of claims on the rated paper. The depth of that subordination, however, is compensated for by the spread of Libor plus 275bp that they are paid. Assuming that all these payments are
In addition to being senior to the subordinated debt and the equity holders, the most senior tranches can be given an added degree of protection in the form of guarantees from monoline insurance companies. As Duke Street Capital explains in a report on its Duchess I CDO, in this instance “Financial Security Assurance (FSA), an insurance company, ‘wrapped’ the €865 million of A/A2 notes to an AAA/Aaa rating. In the event of severe underperformance... rendering the Fund incapable of paying the interest and principal due to the AAA/Aaa noteholders, these note-holders would then have their interest and principal paid by FSA.” The final tranche within the CDO structure, in terms of seniority of sequential payment claims, is the equity portion, and it is this junior position in the capital structure that explains why the equity is also
18
credit The ABC of CDO
made in full, with no defaults on the portfolio, the residual $5.4 million flows to the holders of the $14 million tranche of income notes, or equity – a very healthy return of almost 40%. This equity represents credit enhancement for the notes ranking above it in the capital structure, because any defaults on the underlying portfolio will be borne by holders of this ‘first-loss piece’. In other words, if the underlying portfolio suffers losses of, say, $3 million, the cashflows due to the rated note-holders remain the same, with those due to the equity holders shrinking to $2.4 million and reducing their return to 17% – still within the 15% to 20% range that investors in the most junior tranche of a CDO will expect to earn. What of the risks to the most senior claimants? In this example, because the $240 million of Class A-1 notes are backed by $300 million in collateral, the portfolio would need to suffer a loss of $60 million before these first priority notes would suffer any loss. But because of the OC triggers, the excess spread (of $5.4 million) provides additional protection for the senior note-holders.
described as the ‘first-loss’ piece. Also sometimes known as junior subordinated notes, preferred stock and secured income notes, the equity tranche is generally unrated and can account for anything between 2% and 15% of a CDO’s total capital structure. Unsurprisingly, the equity tranches of CDOs have historically delivered the highest returns but also exposed investors to the highest risks, in just the same way as investing in the equity of any public company is associated with higher risks and rewards than investing in its debt. In the early 1990s, the equity tranches of some US high-yield CDOs provided returns of as high as 50% or more, but in the downturn of 2000 and 2001 investors in equity tranches of CDOs sustained some very heavy losses. There are no predetermined parameters dictating how many tranches an individual CDO can contain,
5tranching.qxd
27/04/2004
16:08
Page 3
tranching although there is usually a minimum of three. Nor are there any governing the optimum weighting of any class of note within the overall structure. Indeed, one of the principal attractions of CDOs is the flexibility of their capital structure, which can create scores of different risk profiles by adjusting the structure of the instrument and the credit quality of its collateral.
Single tranche versus whole capital structure CDOs A notable recent trend within the CDO market has been the growing popularity of single-tranche products, which are generally bespoke transactions structured on a reverse enquiry basis – in other words, to cater to the specific requirements of individual investors. As Fitch explains in a recent report: “The ability of the investor to have a higher degree of input to the characteristics of the transaction is a common element of the single-tranche synthetic. Portfolio composition, tranche size, desired spread,
management/substitution rights, and, importantly, rating can all be selected to a greater or lesser extent by the protection seller [CDO investor].” In a single-tranche structure, only a specific level of the portfolio credit risk is transferred to the investor, with the remaining risk dynamically or delta-hedged by the dealer. “In other words,” Fitch explains, “risk transference is achieved using a derivatives model in the case of single-tranche synthetics versus a securitisation model in the case of traditional synthetics.” Given that single-tranche trades are arranged on a bilateral basis and not generally disclosed to the market, it is very difficult to gauge precise issuance volumes. But as DrKW’s 2003 review of activity in the CDO market notes: “Based on figures from Creditflux, we estimate that the total notional amount referenced by single tranche CDOs in 2003 was in excess of US$400 billion, although the amount of protection sold to hedge these tranches was a fraction of that number.” ■
Tranching of an average CDO
87.50%
88.00% Super senior tranche ('AAA+')
Senior tranche (Aaa) Mezzanine tranche (Aa2) Mezzanine tranche (Baa2)
3.75% 2.25% 2.75%
7.00%
Equity
3.75%
Synthetic CDO
www.creditmag.com
2.25% 2.75%
Cash CDO
credit The ABC of CDO
19
6investordemand.qxd
27/04/2004
16:09
Page 2
investor demand
Out of the ordinary Investors have frequently been sceptical of CDOs, questioning the originating firm’s motivation for issuing. But with credit's startling performance in 2003, CDOs’ failings were quickly forgotten as investors enjoyed the natural leverage such instruments could provide
T
he investor base for collateralised notes of all kinds has broadened considerably over the past eight years. Securitised products are no longer the exclusive preserve of a narrow clique of investors, but have been increasingly appealing to pension funds and insurance companies as well as hedge funds and banks’ proprietary trading desks. Most recently, even private high-net-worth individuals have also become active in the market for CDOs. So why do investors buy CDOs? After all, it is the investor who ultimately bears the costs of the infrastructure and services associated with the establishment and management of a CDO. So why should investors buy CDOs rather than participate directly in the market for investment-grade or high-yield bonds, leveraged loans, mortgage-backed bonds or whatever other assets are referenced by the CDO? At a fundamental level, buying into a CDO can give investors exposure to a well-diversified range of credits, industries, geographical regions or structures that they may have been unable to access independently. An additional attraction of the market for collateralised instruments is that they generally provide investors with exposure to an asset with low correlation to other securities such as vanilla bonds and equities. Since the earliest days of the market for securitised notes of any kind, however, the principal attraction for investors has been the greater yield they offer compared to similarly rated corporate issues. Describing the emergence of the US market for collateralised mortgage obligations in their 1988 book, Securitization, Henderson and Scott explained that: “CMO bonds (most of which have been rated triple-A) have tended to yield 30–40bp per annum more than corporate bonds of similar maturity.” More than 15 years later, collateralised products continue to offer a spread pick-up. According to research published at the start of 2004 by DrKW: “Triple-A CDO tranches price at similar
20
credit The ABC of CDO
levels to triple-B corporate credit and offer a pickup of 20bp over single-A names.” Within the CDO universe itself, there is a wide range of spreads available to investors, which vary depending on the structure of the product. Cash transactions generally have lower spreads than their synthetic counterparts, and this trend applies for all types of collateral. Spreads on synthetic transactions, meanwhile, tend to be broadly similar, although static deals appear to trade very slightly wider. The broadest diversity of returns in the CDO market, is the product of the tranching structure of individual transactions. Historically, the most challenging tranche of CDO deals to market to investors has been the junior equity, or ‘first-loss’ piece. According to a Fitch report published in February 2004: “As markets have developed and the number of potential investors has grown, they have become more capable of absorbing the riskier pieces of an issue. However, originating banks still often find that they have no choice but to retain the first-loss piece of securitisations or part of it, because of lack of market appetite at a given spread.” Analysts believe that investors weighing up the value to be found in the equity tranche of CDOs would be well rewarded. In a report published in October 2002, JP Morgan argued that the equity tranches of certain CDOs should warrant consideration as an alternative asset class (alongside investments such as hedge funds and private equity) for pension funds. More specifically, JP Morgan calculated that “investing in 10% equity tranches of fiveyear Baa3 collateral pools would have returned an average of 14% annually since 1984, with a 13% standard deviation”. The same report calculated that this portfolio of CDO equity exhibited an extremely low correlation with both public equity and fixed-income markets – a correlation of just 9% with the S&P 500 benchmark and of 11% with the Merrill Lynch investment-grade bond index. ■
7perf&recovrates.qxd
27/04/2004
16:13
Page 1
recovery rates
Severity of default is key For investors to feel comfortable with investing in CDOs they must feel confident about the circumstances under which they will win and lose. Here the recovery rate of the underlying assets is the key to performance
F
or investors in the CDO market, it is important to distinguish between default and recovery rates. A default is defined as occurring at the moment that a promised payment on a bond is missed by the issuer, or the time at which an announcement of a missed payment is made regardless of the allowable grace period. For example, the way that Moody’s rates bonds means: “If issuer ABC misses an interest payment on the due date but makes the payment during the grace period, Moody’s treats ABC as a defaulted issuer at the time of the missed payment.” For CDO investors, therefore, the severity of loss rather than the severity of default is the key. Clearly, recovery rates in the event of liquidation of assets will vary widely across various claims in the capital structure. S&P’s recovery assumptions, for example, range from highs of 50% to 60% in the case of senior secured bank loans through to lows of 15% to 28% for subordinated debt and just 15% for emerging market corporates. For individual distressed credits, therefore, recovery rates can also vary dramatically. For example, according to figures published by S&P, recovery rates have varied from as low as 9–12% for
WorldCom and 11–24% for Enron through to as high as 78–90% in the case of Railtrack in the UK. Aside from the specific circumstances of default, a number of other factors can and do influence recovery rates. For example, on average, the longer collateral managers hold on to defaulted securities, the greater their recovery values become. That does not necessarily mean that collateral managers will usually aim to retain ownership of defaulted securities, because in most cases the terms of their contracts will make them forced sellers in a default scenario. The important differences between default and recovery rates mean that calculating historical recovery levels and therefore extrapolating likely future trends is far from straightforward. A complication in the European market is that information on recovery rates has historically been kept private by banks in the loan market, forcing rating agencies and other analysts to apply a so-called ‘haircut’ to recovery rate data from the US, where much broader information on recovery rates for bonds and loans is available. An added complication, especially for CDO investors tutored in the bankruptcy laws that apply in
CDO repackaging (repacks) The repackaging of CDOs (known as CDO repacks) is a relatively recent phenomenon arising from the poor performance of a number of CDOs in 2002 and 2003, and another good example of the flexibility and adaptability of the market to respond to fluctuations in credit quality and economic volatility. Repacks are considered to be ‘first derivatives’ of CDOs and, as Moody’s explains: “In a typical repack, the terms of the existing CDO are restructured, with changes in seniority, notional amount, coupon, maturity
www.creditmag.com
and waterfall priority. The cashflows of the existing debt are used to support restructured debt securities to achieve the desired ratings.” Moody’s adds that repackaged structures, 45 of which were rated by the agency in 2003 compared with just 11 in 2002, will be able to achieve a higher rating due to an increased subordination and the support of extra interest.“After the restructuring of the existing CDO structure, the new bond will be more appealing to the investors who are seeking higher credit quality,” Moody’s notes.
credit The ABC of CDO
21
7perf&recovrates.qxd
27/04/2004
16:13
Page 2
recovery rates
The role of the SPV ABS (including CDOs) are generally issued by SPVs or special-purpose entities set up to allow for the transfer of risk from the originator to an entity that is generally thinly capitalised, bankruptcy-remote and isolated from any credit risk associated with the originator. According to a JP Morgan handbook:“To limit the universe of an SPV’s potential creditors, it is usually a newly established entity, with no operating history that could
the US, is that insolvency regimes continue to differ throughout Europe. France, for example, is notorious for being highly protective of borrowers while Germany is regarded as being much more pro-secured creditors. Furthermore, as the Credit Guide to CDOs published in 2002 observed: “In many European jurisdictions, bond investors have no control over any work-out process: this is in stark contrast to the situation in the US, where both loan and bond investors get a seat at the table. As a result of these structural features, European high-yield bonds are proving to have abysmal recovery rates.” Those poor European recovery rates, however, are not confined to the high-yield market. According to Moody’s, while the default rate in the European corporate bond market plunged from 20.1% in 2002 to 6.9% in 2003, the average recovery rate was almost unchanged at 19.9% in 2003 compared with 20% the previous year. Recovery rates in Europe, Moody’s advises, continue to be roughly half the North American average.
Structuring and constructing a CDO The financial press will often make its first mention of a ‘new’ CDO on or around the time of its closing – with the closing date generally the day on which the CDO issues tranches of debt and equity to investors. Prior to that day, however, there will have been a socalled pre-closing or ‘warehousing’ period, typically lasting between three and six months. During that period the asset manager will have acquired (or ‘warehoused’) assets to act as collateral for the securities to be issued by the CDO via a special-purpose
22
credit The ABC of CDO
give rise to prior liabilities. The SPV’s business purpose and activities are limited to only those necessary to effect the particular transaction for which the SPV has been established (for example, issuing its securities and purchasing and holding its assets), thereby reducing the likelihood of the SPV incurring post-closing liabilities that are in addition or unrelated to those anticipated by rating agencies and investors.”
vehicle (SPV – see box) on the closing day. Closing of a fund usually occurs when a CDO has acquired between 40% and 60% of its targeted assets. Clearly, however, given that the proceeds of the CDO notes will only become available following their sale on closing day, CDO managers will often need a bridging loan facility (or ‘warehouse facility’) during the warehousing period. Following the issuance of notes on closing day, the CDO will have a period usually lasting between 60 and 360 days – although the period can also be much longer – in which to complete the process of buying the assets backing the CDO. This important phase is commonly known as the ramp-up period, and the year in which the ramping-up takes place is referred to as the CDO’s vintage. The final investment amount amassed following the ramp-up is sometimes known as the target par amount, which is the total size of the fund less its start-up costs. A portfolio that has been ramped-up with a relatively large number of small exposures is described as being granular, whereas a more concentrated portfolio with a small number of exposures is known as a lumpy fund. After completion of the ramp-up, there is usually a reinvestment (or revolving) phase lasting up to five years, during which any cashflows arising from amortisation, maturity, prepayment or the sale of assets can be reinvested, as long as a number of basic performance objectives have been maintained. Finally, during the amortisation phase, which can last for between five and 30 years depending on the underlying assets of the CDO, cashflows earned by the fund are used to pay down its liabilities. ■
8productinnov.qxd
27/04/2004
16:15
Page 1
product innovation
Diverse and sophisticated In the past few years, originators and arrangers of CDOs have ignored no debt or proxy for debt to act as the underlying assets in a collateralisation, including other CDOs
“T
he process of increasing sophistication in the CDO market that began last year cannot be reversed,” noted Moody’s in its review of activity in the CDO universe in 2003, noting that much of the new issuance was driven by transactions with a more innovative collateral mix. Much of the recent product innovation has been aimed largely at increasing the diversification of collateral underlying CDOs, with resecuritisations – or CDOs of ABS – one example of a product that has
become increasingly popular. While the diversity embedded in these products is to be welcomed, it is also clear that in the case of resecuritisations, investors are advised to scrutinise each transaction very carefully on a case-by-case basis. This is because, as a Barclays Capital analysis of the market published at the start of 2003 notes: “Spreads on resecuritised transactions can differ significantly from one transaction to another, which is unsurprising given that the collateral underlying these transactions can also vary
CDOs of EDS Another variety on an existing theme that was first developed in 2003 was the CDO of equity default swaps (EDS). That did not quite represent the first time that CDOs emerged giving investors exposure to equity risk. Prior to the arrival of CDOs of EDS, there were, for example, CDOs with exposure to convertible bonds that were typically more volatile than their counterparts exposed to pure debt. EDS are modelled on the template established so successfully in the CDS market, and are contracts in which payments are made to buyers of protection if certain trigger events cause the equity price of a reference entity to fall below a certain predetermined threshold. In other words, as Moody’s expresses it:“An EDS is a derivative product which attempts to mimic a CDS by replacing a ‘credit event’ with an event on the stock price of the reference entity.” Moody’s adds:“The trigger event definition, being objective and verifiable by all the parties, benefits from a considerable degree of simplicity and clarity.”That is in obvious contrast to the definition of credit events in the early
www.creditmag.com
days of the CDS market, which left substantial scope for interpretation and controversy. The first EDS-backed transaction to be rated by Moody’s was the hybrid Odysseus deal arranged by JP Morgan, which consisted of a portfolio of 100 reference entities, although only 10% of these were EDS, with the remaining 90% accounted for by CDS. All the default swaps in the portfolio were referenced to blue-chip corporate names with a weighted average rating of Baa. In February 2004, Daiwa Securities SMBC took the EDS concept a step further when it launched the first publicly rated arbitrage CDO 100% collateralised by EDS. Zest Investments V issued ¥31.5 billion of notes in five different classes, backed by EDS on a portfolio of 30 quoted blue-chip corporates with an aggregate notional amount of ¥45 billion. Payment to the protection buyer will be triggered if the share price of any of the companies referenced in the EDS portfolio falls by more than 70% from its initial price and if the share price fails to recover to the initial level by December 2008.
credit The ABC of CDO
23
8productinnov.qxd
27/04/2004
16:15
Page 2
product innovation
CDOs by underlying collateral
High-yield loans 26% High-yield bonds 25% Structured finance 21% Investment grade 6% Credit derivatives 5% Other 4% Trust preferred 3% Emerging markets 3% Euro high-yield bonds 2% Middle market loans 2% Euro high yield 1% CBO 1% Euro investment grade 1%
Source: Morgan Stanley
considerably.” The same report adds: “Let’s consider, for instance, a transaction backed only by residential mortgage-backed securities and another one backed by whole business securitisations; there could not be two more different deals.” Much the same could be said of other COs packaging a pool of diversified assets, such as CDOs of funds. The first CDO of hedge funds – or collateralised fund obligation (CFO) – appeared in April 2002. This was a $250 million, five-tranche, fiveyear transaction managed by Investcorp, named the Diversified Strategies CFO and lead managed by CSFB. “Funds of hedge funds have relatively low volatility and are uncorrelated to other asset classes such as debt and equities,” Bahrain-based Investcorp announced at the time of the CFO’s launch. Another relatively recent innovation designed to increase diversification has been the advent of CDOs of CDOs, known as CDOs squared, in which
24
credit The ABC of CDO
CDOs’ underlying assets are other CDOs. The concept of the CDO squared was pioneered by Christian Zugel, a former JP Morgan bond trader who established the Zais Group, which originally focused on investment opportunities in the equity tranches of CDOs, in July 1997. In September 1999, Zais launched its first CDO of CDOs, the $343 million Zais Investment Grade (Zing I) product, which has since been followed by a number of other Zing instruments. Another early entrant into the world of CDOs squared was Triton Partners, which launched its first CDO of CDOs, the $307 million Triton Opportunities Fund, in July 2001. The rationale behind actively managed CDOs squared such as the Zing and Triton products is that they are able to offer an increased level of diversification on the one hand and more highly leveraged, equity-like returns across all tranches of the structure on the other. ■
8productinnov.qxd
27/04/2004
16:15
Page 3
product innovation
Standardising the market The introduction of tranched tradable credit indices that reference CDS as the underlying is bringing synthetic CDOs to the mainstream in a liquid format In the past few years investment banks have developed new tradable credit indices that reference exposure to the CDS market, the largest of which are iBoxx and Trac-x. These products are similar to static synthetic CDOs. They provide investors with exposure to a standardised static (changing either quarterly or half-yearly) list of names via the CDS market. Credit investing at its purest – ignoring maturities, currencies and underlying government interest rates – is primarily about selecting individual names depending on their fundamental credit quality and secondly taking a view on the direction of the credit market as a whole. With this in mind, the idea of index-linked products is to allow investors to take a view on the credit market in one transaction. An investor wanting to reduce exposure to the whole credit market previously had to either sell stacks of bonds or buy protection on those names in the credit default swap market – a costly and onerous task. But basket- and index-linked notes allow investors to separate individual name risk from the market risks that affect all bonds, loans, convertibles or credit default swaps to some degree. At the moment the biggest enthusiasts are hedge funds, fixed-income prop desks and loan portfolio managers, who use these products as a proxy to buy or hedge the market. Where benchmarked investors find basketlinked notes useful is to ramp up a fund quickly. If a fund manager is launching a corporate bond fund for retail investors, the fund may stay open for several months and money will accumulate in dribs and drabs. The money can be left in cash, invested in low risk assets such as government bonds, or invested in the corporate bond market. And since the fund is a corporate bond fund, the
www.creditmag.com
latter course of action is probably the best. But buying individual corporate bonds every time more money is received is expensive in transaction costs and will take considerable time. So a product that allows investors to buy the market cheaply and quickly every week will satisfy a need. Once the fund reaches its target size, the individual bonds can be bought and the basket-linked note sold down. Finally, basket-linked notes can also be used by corporates. If a company has a bond
Tradable indices are bringing synthetic CDOs to a mainstream investor audience maturing in 18 months, it is unlikely to refinance much before that. But the company may believe that current interest rates are low and spreads are tight. The company can tie in the interest rates using derivatives and the same is now true for spreads. Although corporates only really care about their own spreads, they cannot buy credit default swap protection on themselves because questions would be asked in the market. By buying a basket-linked note they can lock in the tight market spreads until they are ready to refinance. What makes these indices so similar to synthetic CDOs is that the banks behind them allow investors to buy tranched versions of the indices. This means that investors now have a way to trade standardised CDO tranches. This in turn is making complicated strategies such as trading how correlated credits are defaulting a real possibility, and is bringing synthetic CDOs to a mainstream investor audience.
credit The ABC of CDO
25
9marketinfo.qxd
27/04/2004
16:21
Page 2
market information
Understanding the product Since their inception, CDO critics have expressed concern about the transparency of the product. In the past few years a succession of products and initiatives has been produced to help investors understand what a CDO contains and how it is performing
A
s volumes in the CDO market began to expand rapidly in the mid to late 1990s on both sides of the Atlantic, misgivings were frequently expressed about the level of transparency (or lack of it) in the primary and secondary market. As one analyst was quoted as telling the Wall Street Journal in September 1997, when NationsBank launched its record-breaking $4 billion transaction: “You’re blind to the underlying obligors.” Misgivings about the transparency of the CDO market may have receded since the late 1990s, but they have certainly not disappeared altogether. Indeed, the emergence since then of increasingly sophisticated second-generation products has prompted concerns that investors and even originators may be unable to identify or monitor the precise exposure of portfolios underlying individual CDOs.
Theoretically, the problem of lack of transparency in the CDO market is addressed by trustees Theoretically, the problem of lack of transparency in the CDO market is addressed by trustees, which distribute monthly reports to investors detailing the composition of CDO portfolios along with updated information on any changes to the ratings of underlying assets. But as the 2002 Credit Guide to CDOs pointed out, trustee reports are “notoriously difficult to use”, with investors complaining that “they are dense, inaccessible and frequently riddled with basic errors”. The news is not all bad. Recent years have seen a proliferation of new initiatives aimed at enhancing the credibility of the CDO market by promoting its increased transparency, many of them initiated by
26
credit The ABC of CDO
some of the banks that are more active in the market. In May 2002, for example, Goldman Sachs announced that it would be making key data on most of its cashflow CDOs more broadly available to all qualified subscribers of Intex’s Analytical Products, which is part of the Intex Solutions software family and which maintains extensive deal model databases of structured finance securities. Goldman Sachs announced at the time that “CDO investors have long been concerned about the lack of transparency in the market. We believe broader dissemination of our CDO information sets an important precedent. We welcome closer scrutiny of CDOs by a broader pool of investors. A more informed and efficient market-place should lead to a larger and more robust CDO market over time.” Similar measures in other asset classes, added Goldman Sachs, had “engendered more accurate valuations and tighter bid/offer spreads. They have also produced deeper pools of capital in the secondary market.” Goldman Sachs was by no means alone. By the end of 2003, eight underwriters had agreed to release information on their deals to Intex’s subscribers, with Morgan Stanley, CSFB, Lehman Brothers, JP Morgan, Merrill Lynch, Citigroup and Wachovia all having joined Goldman Sachs in the initiative. Those banks account for about two-thirds of the 650 deals modelled by the Massachusettsbased Intex, and more seem likely to sign up to the initiative. Delegates from Bear Stearns and UBS, for example, both indicated at a Bond Market Association (BMA) conference in New York in September 2003 that they were considering posting deal documentation on Intex. Other initiatives aimed at improving transparency in the CDO market have included the agreement reached in 2002 between JP Morgan and RiskMetrics to develop and license custom-designed web-based tools for CDO market participants. Using JP Morgan’s deal data and powered by
9marketinfo.qxd
27/04/2004
16:21
Page 3
market information RiskMetrics’ CDO manager software, the result of the joint venture is a website, CDOVantage, aimed at providing CDO market participants with “a consistent and comprehensive source for CDO market data”. At the time of its launch, JP Morgan announced: “CDOVantage will provide users with unprecedented control and insight over their CDO transactions. CDOVantage will help to develop a market for these products by dispelling misperceptions about the CDO market-place, and demonstrating how open and liquid the business can be.” Wall Street Analytics (WSA) has developed CDOnet, which it describes as “the first structuring and analysis product designed specifically for CDOs with the ability to run complex default scenarios, including Monte Carlo simulations, to value CDOs”. WSA has also used CDOnet to develop CDOcalc, providing investors with access to a library of more than 500 CDOs which “allows investors to run price-yield analytics, project cashflows and coverage tests, download collateral information [and] track exposure across portfolios”. One of the most active champions of increased transparency (and hence liquidity) in the CDO market has been the BMA itself. For example, the Association has built up a comprehensive library containing information on specific CDO transactions that had previously not been available to any market participant not directly involved in the transaction. The library posted details for each deal on swap agreements, offering memoranda, indenture documents and monthly trustee reports. Adding updates such as amendments to the original documentation is optional, with all transaction information supplied by investment banks on a voluntary basis. The information is searchable on www.cdolibrary.com by deal name, underwriter, collateral manager and/or deal date. At the start of March 2004, the BMA announced the new library open to qualified institutional buyers and dealers had gone live, with 13 banks having agreed to provide deal details to the library: Banc of America Securities, Bear Stearns, Citigroup, CSFB, Deutsche Bank, Goldman Sachs, Greenwich Capital Markets, JP Morgan, Lehman Brothers, Merrill Lynch, Morgan Stanley, UBS and Wachovia. A more recent initiative developed by the BMA has been its recommendation that participants in www.creditmag.com
the CDO market consider using a standardised template when creating monthly CDO trustee reports. The aim with this project, says the BMA, is to provide “a standardised format for monthly CDO trustee reports that can be easily read, searched and sorted”. The Association adds: “With a standardised report, the same information categories will always be included in the same location (that is, rating history is always included in the ‘structured-ratings-payment history’ section). Widespread adoption of this standardised format, therefore, would allow the user to quickly and efficiently find information in and reference the report as necessary. The end result is a much more manageable and functional trustee report.” The BMA is hoping to replicate in Europe much of what it has achieved in the US CDO market. In March 2004, it established a new CDO Committee in Europe dedicated to the promotion of standardisation, transparency of information and increased liquidity in the market. Fritz Thomas, head of CDO structuring and distribution at Deutsche Bank, was appointed chairman of the committee, and Oldrich Masek, co-head of global structured credit origination at JP Morgan, as vice-chairman.
The application of technology for improved transparency The increasing complexity of the CDO market, and the ever-growing range of products being made available to investors in the CDO world, are inevitably creating challenges in terms of information assembly, management and dissemination. In the UK, the Financial Services Authority drew attention to the concerns arising from these challenges in its Financial Risk Outlook, published in January 2004. “One of the ways in which credit risk is being transferred from the banking sector to the insurance sector is through the use of portfolio credit default swaps or notes whose performance is linked to a basket of credits,” this notes. “Many of these transactions are booked in banks’ trading books, and as a result must be marked-to-market each day. The market in portfolio trades is still new and relatively illiquid, so banks usually rely on models to revalue and risk-manage the transactions on a day-to-day basis. Valuing and risk-managing complex and illiquid structures like the portfolio trades credit The ABC of CDO
27
9marketinfo.qxd
27/04/2004
16:21
Page 4
market information described above presents challenges for even the largest and most sophisticated of banks. We are currently reviewing firms’ valuation of such positions to assess systems and controls in this area.” It is the development of hybrid second-generation credit derivatives, however, that is potentially presenting the most formidable challenges for dealers’ IT infrastructure in the CDO arena. In a deliberately oversimplified illustration of these challenges, take the various ways in which an investor’s exposure to a highly liquid credit such as Ford may be expressed in today’s market. At the most basic level, the investor may hold a Ford bond, the value of which is well known, allowing for very straightforward marking-to-market of the investor’s exposure. At a more complex level, he may have exposure to Ford through a first-generation static cashflow CDO, which again is relatively straightforward to quantify and value. The same exposure in a managed synthetic CDO will be marginally more complicated to monitor. But in the event of a CDO of CDOs, it will become increasingly complicated – and even more so if CDOs of CDOs themselves buy CDOs squared, in effect creating CDOs cubed. In a report published in the September 2003 edi-
tion of Risk magazine, software company Application Networks spelt out the challenges. “In the case of CDOs of CDOs, for example, the bank can be faced with extremely large underlying portfolios to track and risk-manage. The same piece argued that dependence by banks on a traditional approach to new product development, or on bolting on new functions to old systems, are no longer an adequate way of dealing with the rapid emergence of new, increasingly hybrid products. In the past, quantitative analysts would price a new product, they would then pass it to the technology department in order to incorporate the new structure within the existing software infrastructure, to capture and process the trade and to manage its risk. The key drawback with this design was that they were initially built with a single asset class in mind, which made them inflexible, cumbersome and difficult – in many cases impossible – to extend to new product lines. Given the requirements to fit new products into existing systems, such an approach could delay product rollout by weeks or months, This problem is particularly acute today, given the speed with which quants can generate new structures, suggests Application Networks. ■
Model risk A key point in CDO analysis is calculating the degree of asset correlation in the underlying portfolio – the probability that if one asset defaults, a second will default. But this is subject to a variety of methodologies. Moody’s bases its correlation analysis on the assumption that two assets in different industries are fully independent, and that two assets in the same industry are quite highly correlated, say 25–30%. The rating agency then issues a diversity score for the CDO, as distinguished from its rating. While Moody’s admits that this is a simplistic model, it says that in reality two assets in two different industries can be correlated. But while assuming zero inter-industry correlation may be generous, the intra-
28
credit The ABC of CDO
industry correlation assumptions may be quite conservative. S&P evaluates the credit quality of a portfolio of CDO assets using its CDO Evaluator platform. This tool uses similar assumptions to the Moody’s analysis: a correlation coefficient of 0.3 is used within ABS sectors and 0.1 between ABS sectors. For corporate sectors, it uses 0.3 within a given industry and zero between two separate sectors. Fitch is the only rating agency that has progressed to modelling inter-industry correlation. While the paucity of data in the European corporate bond market complicates such a task, Fitch’s recently launched Vector method uses equity market correlations to arrive at a relatively accurate model.
11conclusion.qxd
27/04/2004
16:24
Page 2
Conclusion
The CDO contribution Even five years ago, CDOs were just a relatively small part of the overall credit derivatives market, and in terms of the overall financial market they were barely noticeable. Today they are still small in comparison with the overall financial market, but their contribution and importance is impossible to ignore
C
redit derivatives as well as COs have not been without critics, who have bandied around words such as “toxic” or expressions like “weapons of mass financial destruction” to describe them. In his May/June 2003 newsletter, for example, Pimco’s Bill Gross warned that “unregulated hedge funds, collateralised debt obligations and poorly structured derivatives of all kinds that redistribute risk but do not eliminate it portend the likelihood of another LTCM debacle at some point.” Gross is by no means the only market commentator who has expressed concerns that an era of rising interest rates (which the CDO market has yet to experience in its current, highly liquid form) will wreak the same sort of devastation that affected the structured finance market in 1994 when the US interest rate cycle turned. Many share the belief that the credit derivatives market in general and CDOs in particular are making a dangerously unsustainable contribution to a global credit bubble. The April 2003 edition of Marc Faber’s Gloom, Doom and Boom Report carries this withering analysis of the collateralised obligations market: CDOs, it argues “are a collection agency of every debt owed by anyone that the lender is willing to sell. Investment banks corral
thousands of these debt claims and turn them into CDOs, a bond. The CDOs are impossible to understand in detail, so they are mathematically modelled to predict how they will behave in aggregate.” It is not just Jeremiahs of the Marc Faber mould who have suggested that there is a dangerous ignorance about credit derivatives within the global financial community. The Economist chose The swaps emperor’s new clothes as the title of a piece it published on credit derivatives in February 2001. And in the same year Kenneth Chenault, chairman and chief executive of American Express, was quoted as saying that his bank “did not comprehend the risk” of the CDOs that it had acquired in the late 1990s leading to losses amounting to hundreds of millions of dollars. Aside from fuelling an unsustainable credit bubble possibly constructed on the basis of ignorance, sceptics such as Warren Buffet have argued that the credit derivatives market has led to an equally dangerous concentration of risk. A survey published by Fitch in September 2003 went some way towards corroborating these concerns, finding that the 10 largest global banks and broker-dealers accounted for 70% of the credit derivatives market. An important counterforce to this apparent concentration, however, is that the growth of the overall market for
Growth in credit derivatives and CDOs 5.0
$ trillion
4.0
Size of CDO market Size of total credit derivatives market (excluding CDOs)
26%
3.0 2.0 22% 1.0 0.0
18% 1999* Source: British Bankers' Association
30
credit The ABC of CDO
2001† * 2002 survey † 2002 survey
2004†
11conclusion.qxd
27/04/2004
16:24
Page 3
Conclusion credit derivatives is a positive development as it supports diversification of credit risk and results in improved liquidity in underlying credit markets. Those with a more upbeat view than Marc Faber or Warren Buffet of the contribution made by credit derivatives to the global financial system argue that far from encouraging the irresponsible build-up of an ever-proliferating stock of credit, the CDS market in particular acts as an extremely efficient arbiter of value and hence of risk. Many also argue that rather than concentrate risk, the credit derivatives market has made a remarkable contribution to diffusing it. Notably, Federal Reserve chairman Alan Greenspan has lent his voice to this particular argument, describing financial derivatives in January 2004 as “the new instruments of risk dispersion” that “have enabled the largest and most sophisticated banks in their credit-granting role to divest themselves of much credit risk by passing it to institutions with far less leverage”. Those instruments, Greenspan added, had contributed to the development of a “far more flexible, efficient and hence resilient financial system than existed just a quarter-century ago”. Tangible proof of this was provided, he said, by the telecommunications sector, which on a worldwide basis borrowed more than $1 trillion between 1998 and 2001. Although a substantial amount of this debt defaulted, the capacity of banks to pass their exposure on to institutions and hence to diffuse the risk associated with the telecoms boom meant that “no major financial institution defaulted, and the world economy was not threatened. Thus, in stark contrast to many previous episodes, the global financial
system exhibited a remarkable ability to absorb and recover from shocks.” Disasters waiting to happen or a benign means of dispersing and hence reducing risk? The truth probably lies somewhere in-between, as Fitch appeared to conclude from a study published in September 2003, entitled Global Credit Derivatives: A Qualified Success. “On balance,” this commented, “Fitch believes that credit derivatives have been a positive development for the global financial system, facilitating enhanced risk transfer and dispersion.
“CDOs are impossible to understand in detail, so they are mathematically modelled to predict how they will behave in aggregate” Growth in the credit derivatives market has significantly increased liquidity in the secondary markets and allowed the efficient transfer of risk to other sectors that lack direct origination capabilities.” Nevertheless, Fitch added that there are risks that bear watching. These include “low financial transparency, informational asymmetries (which create the potential for unanticipated, incorrectly priced and poorly managed concentrations of risk), and the possible promotion of new forms of moral hazard within the banking system as the linkage between origination and management of credit risk becomes more attenuated.” ■
Credit derivatives as a proportion of all derivatives as of December 2002
trillions $
100
102
80 60 40 20
18.5 Interest rate
Foreign exchange
19.3 Other
2.31
2.10
Equity
Credit
Source: Bank for International Settlements 'OTC derivatives market activity 2002'
www.creditmag.com
credit The ABC of CDO
31
index.qxd
27/04/2004
16:26
Page 2
index of terms & deals Terms
Deals
A Arbitrage (CDO) 8, 10, 11, 12, 23
A Aurora Sumitomo Bank 10
B Balance-Sheet (CDO) 8, 9, 10, 11, 12 Basel Capital Accord 9 Basel II 9
B Blue Strip Deutsche Bank 10 Bistro (Broad Index Secured Trust Offering) JPMorgan 13
C Cashflow (CDO) 11, 12, 17, 26, 28, 29 CDO squared 24 Churning 15 Credit event 13 , 23
C Caesar Finance Banco di Roma 10 Cast Deutsche Bank 10, 29 Concerto Axa 6, 11 Core (Corporate and Real Estate) Deutsche Bank 10
D Diversity scores 28, 29
D Duchess Duke Street Capital 11
F First-loss (tranche) 18, 20
E Eurocredit I, II Intermediate Capital Group 11
H Hybrid 12, 23, 28 M Managed (CDO) 11, 15. 24, 28
G Geldilux (Guarantees of Euro-Loan Debt in Luxembourg) Bayerische Hypotheken- und Vereinsbank 10 Globe Deutsche Bank 10
O Over-collateralisation (OC) 17, 28, 29
H Helvetic Asset Trust UBS 7
P Passive 15
I iBoxx 14, 25
R Ramp up 13, 22, 25 Reinvestment phase 22 Return on equity (ROE) 9, 10 Repacks 6, 21
J Jazz I & II Axa 11
S Special purpose vehicle (SPV) 22
O Overture Axa 11
T Tranching 6, 17, 20 True sale 10, 12, 13
R Rose Funding NatWest 7
S Single-tranche (CDO) 19 Static 15, 20, 25, 28 Synthetic 7, 10, 11, 12, 13, 14, 15, 19, 20, 25, 28, 29 Super-senior 14 V Vintage 22 W Warehouse (facility) 22 Waterfall 17, 18, 21
32
credit The ABC of CDO
K Khaleej Axa 11
S Scala 1 Banca Commerciale Italiana 7 T Trac-x 14, 25 Triton Opportunities Fund Triton Partners 24 Z Zing I (Zais Investment Grade) Zais Group 24
Outside back cover.qxd
30/04/2004
11:54
Page 1