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 Page 2

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

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

I 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.

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
The start of the market
Collateralised debt instruments were launched
initially for mortgages in the 1980s, but the
market has been adopted over the years to
cover practically every type of debt.
13 Static versus managed CDOs
Turning the tanker
CDOs were originally static portfolios where the
underlying names rarely changed. But the fallout
in credit since 2000 increased the attractiveness
of portfolios watched over by fund managers.

08 Market evolution 17 Tranching

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
The development of the CDS
market has allowed originators to issue CDOs
where the exposure is referenced via a CDS
rather than cash assets.
Out of the ordinary
Investors have frequently questioned
originating firms’ motivation for issuing
CDOs. But the 2003 credit rally has largely
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
Investors in CDOs must feel confident about
the circumstances under which they will win
and lose. Here the recovery rate is the key to
performance.
Rating the product
The complexity and opaqueness of many CDOs
mean rating agencies have carved out a
niche in the market, helping investors
understand the risks.

23 Product innovation 30 Conclusion

Diverse and sophisticated
CDO originators and arrangers have recently
ignored no debt or proxy for debt to act as the
underlying assets in a collateralisation.
26 Market information
Understanding the product
Critics of CDOs express concern
about transparency. In the past few
years a succession of products and initiatives
has been produced to change this.
The CDO contribution
Five years ago, CDOs were just a relatively
small part of the overall credit derivatives
market.Today they are still small in comparison
with the overall financial market, but their
contribution is impossible to ignore.
32 Index
Index of terms and deals
A list of some major CDO transactions and a
rundown of important terminology.

TM
JRisk
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 col-
lateralised debt obligations (CDOs), which in recent
years have been among the fastest-developing invest-
ment 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 dra-
matic 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 redis-
tribution in the US by houses such as Drexel
Burnham Lambert and Kidder Peabody of portfolios
of high-yield bonds and loans. Prior to those transac-
tions, however, a market for CMOs – the forerunner
of modern CDOs – was taking shape in the US mar-
ket by the early 1980s. In 1983, for example, the
Federal Home Loan Mortgage Corporation pio-
neered structures that built on existing mortgage
securitisation templates by creating so-called pay-
through 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 men-
tion 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, pub-
lished 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 securi-
ties 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
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 mar-
ket 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 internation-
al market for COs of all kinds was becoming charac-
terised by a number of conspicuous and interrelated
trends. First, issuance volume was rising exponential-
ly, 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 col-
lateralised against 41 fully secured loans, the major-
ity 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 econ-
omy, Germany was an important source of new deal
www.creditmag.com
introduction
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 sum-
mer, 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 reg-
ulation were emerging as important sources of sup-
port 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 govern-
ment guarantees for securitisations of loans origi-
nated 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 determi-
nant 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 struc-
tures, 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 vol-
ume 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 corpo-
rate 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 signifi-
cant losses on such products that some of these
investors experienced during the past three years”.
There is a simple enough explanation for the per-
sistence of this apparently irrational demand, which
is that against a backdrop of collapsing returns from
government, supranational, agency and other top-
rated 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 cred-
it 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 innova-
tion 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 issu-
ing is one pivotal difference between the two fun-
damental 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 strate-
gic 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 influ-

Performance of average equity tranche versus ence on the growth of the
Lehman’s US Credit index CLO market in the 1990s
50
was the increasing emphasis
Performance of CDO equity (%)

Performance of Baa US credits (%)

10
CDO equity (left-hand axis) that banks were forced to
40 8
Baa US credit - (right-hand axis) place on the delivery of
30 6
enhanced shareholder value
20 4
and improved return on equi-
10 2 ty (ROE) , which in conti-
0 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
Source: Lehman Brothers enjoying average ROE levels
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 con-
CDO 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
to investors. banking sector it struggled to reach 3%.
In the 1990s, at least three influences com- For a variety of reasons, feeble ROEs in the
bined 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 state-
calculating 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, lib-
developing CLO market in particular, allowed eralisation, deregulation and – ultimately – consoli-
banks 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.
on fresh lending capacities. A third, albeit less important, driver for increased
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 institu-
sation as a means of achieving regulatory capital tional investors were probably much more profi-
relief 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 corpo-
less attractive to banks, given the stricter (in com- rate finance landscape.
parison with Basel I) capital requirements.” That Among UK clearing banks, Barclays had pio-
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 securitisa-
cal role in kick-starting activity in the market. tion of personal loans in October 1993. But the

www.creditmag.com 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 for-
ward in September 1996 with the jumbo, $5 billion
securitisation by NatWest of an international port-
folio 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 lev-
els. When NationsBank launched its record-break-
ing $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 appeal-
ing means of transferring the risk on large interna-
tional portfolios of loans. The year 1998 saw trans-
actions 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 cor-
porate loans.
Among continental European banks, mean-
while, 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 repre-
sented 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 securi-
tise loans from its global banking division. The
first transaction under its Globe programme, secu-
ritising loans extended by its global banking oper-
ation 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 tradi-
tional 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 trans-
action that followed soon afterwards was also a
synthetic deal, as was the first Cast transaction
launched by Deutsche Bank in 2000, which par-
celled 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 origi-
nated 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 well-
understood market for balance-sheet CDOs origi-
nated 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 arbi-
trage CDOs are radically different from those
prompting the launch of balance-sheet instru-
ments, with the market driven by asset manage-
ment companies using the CDO technique princi-
pally 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
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 expan-
sion 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 transac-
tions done for balance-sheet purposes and the major-
ity of these backed by loans to small and medium-
sized 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 mez-
zanine 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 transac-
tion 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 ‘clas-
sical’ 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
market evolution
its managers to do their own thing. This being a
CDO, there are still some strict rules for the portfo-
lio 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 specialis-
ing 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 clear-
ly 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 struc-
tured 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 struc-
ture 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 ref-
erenced 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 eco-
nomic 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
Cash collateralised loan obligation
Step 1 Step 2
SPV securitises
Bank sells loans to SPV loan portfolio
SPV
Loan bank
Loan Loan
portfolio portfolio
* The equity tranche is frequently retained by the loan bank
12 credit The ABC of CDO
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 syn-
thetic 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 owner-
ship. 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 origi-
nators securitising multi-jurisdictional portfolios or
loans made in countries where the local legal frame-
work 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
Step 3
Credit risk is tranched and
sold on to bondholders
Capital
markets
Triple-A
notes Investors
Double-A
notes Investors
Triple - B Investors
notes
Equity / Investors*
first loss
3cashvsynth.qxd 27/04/2004 16:05 Page 3

cashflow versus synthethic CDOs

Synthetic collateralised loan obligation

Step 1 Step 2 Step 3
Loan bank buys CDS Loan bank buys CDS SPV sells loan portfolio
protection on least risky ~90% of protection on most risky exposure as credit
loan portfolio from big bank ~10% of from SPV linked notes to investors

Loan
Big bank SPV
bank* triple -A CLNs
Investors
(CDS exposure + Pfandbrief )
Exposure CDS CDS Exposure
protection Loan protection
(via CDS) (via CDS)
to least
portfolio to riskiest double -B CLNs
~10% of Investors
risky CDS (CDS exposure + Pfandbrief )
CDS
~90% of premium premium
loan
loan portfolio
triple -B CLNs
portfolio (CDS exposure + Pfandbrief )
Investors
super-
senior swap
Unrated equity tranche
Investors †
(first loss tranche)

* This is typically referred to as a † The equity tranche is frequently retained by the Loan Bank
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 dramat-
ic growth in recent years to the explosive expansion
in the market for CDS. A CDS is a privately negoti-
ated bilateral agreement in which one party, vari-
ously 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 expand-
ing 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 con-
tinued to grow at breakneck speed, with some esti-
mates 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
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 quick-
ly than those of cash instruments. As a report pub-
lished by Citigroup analysis explains: “This is espe-
cially important in such markets as European invest-
ment-grade cash, where the liquidity in the corpo-
rate 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 phys-
ical, 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 pop-
ularity in Asia.

www.creditmag.com credit The ABC of CDO 13

3cashvsynth.qxd 27/04/2004 16:05
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 sub-
sectors, 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 super-
senior 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 com-
pensation 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 struc-
ture, 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)
60
Synthetic (arbitrage CDO)
50
billions €
Cash (balance sheet CDO)
40
Cash (arbitrage CDO)
30
20
10
0
1998 1999 2000 2001
Source: Lehman Brothers
14 credit The ABC of CDO
Page 4
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 pre-
mium 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 tradi-
tional 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 bil-
lion 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 struc-
ture 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 compa-
rable cash CDO in which 85% of the capital struc-
2002 2003 ture is accounted for by triple-A Class A notes, 10%
by A3 Class B notes and the remaining 5% by jun-
ior-most equity. ■
4statvmancdos.qxd 27/04/2004 16:07 Page 1
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 ordi-
nary 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 equi-
ty holders will receive. Rather than selling a product
the CDO manager is buying debt. The CDO’s sell-
ing 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 orig-
inal composition of their underlying portfolios
remains unchanged. The obvious advantage associ-
ated 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 mar-
ket, 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
static versus managed CDOs
world. They will generally hold some of the equity
in their CDO so as to ensure that they have a vest-
ed 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 man-
aged 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 dur-
ing the year, more than half were by repeat issuers,
suggesting that a handful of the best CDO man-
agers 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 com-
pensation and a structured product sales profes-
sional at a leading US investment bank is under-
stood 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, manag-
ing directors received £125,000 and a £1.1 mil-
lion 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 impor-
tant 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 ben-
efit 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 manag-
er 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-relat-
ed. 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 increas-
ingly 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 impres-
sive performance histories.
Rating agencies are also important referees of
the quality of collateral managers. Although much
of the rating agencies’ analytical work is quantita-
Proportion of managed synthetic CDOs to static synthetic CDOs
100%
80%
static
managed
60%
100% 91%
40%
20%
9%
0%
2000 2001
Source: Lehman Brothers 'European CDOs: review & outlook'
16 credit The ABC of CDO
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 represent-
ing 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; cred-
it underwriting and asset selection; portfolio man-
agement; 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 transac-
tions.” In March 2003, meanwhile, Moody’s
announced the release of a new CDO perform-
ance report series aimed at helping market partic-
ipants 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 inde-
pendent, 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.” ■
92% 78%
22%
8%
2002 2003
5tranching.qxd 27/04/2004 16:08 Page 1
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 so-
called 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 busi-
ness, 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 fre-
quently referred to as a ‘waterfall’ or cascade of
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
tranching
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 mez-
zanine 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.
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 tranch-
es can be given an added degree of protection in the
form of guarantees from monoline insurance com-
panies. 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 underper-
formance... rendering the Fund incapable of paying
the interest and principal due to the AAA/Aaa note-
holders, 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 gen-
erally 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 pro-
vided 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 qual-
ity 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 prod-
ucts, which are generally bespoke transactions struc-
tured 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 deriv-
atives model in the case of single-tranche synthetics
versus a securitisation model in the case of tradi-
tional 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+')

3.75% Senior tranche (Aaa)

2.25% Mezzanine tranche (Aa2) 2.25%
2.75% Mezzanine tranche (Baa2) 2.75%

3.75% Equity 7.00%

Synthetic CDO Cash CDO

www.creditmag.com 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 individu-
als 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 infra-
structure and services associated with the establish-
ment 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 struc-
tures that they may have been unable to access
independently. An additional attraction of the mar-
ket for collateralised instruments is that they gener-
ally 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 securi-
tised 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 pick-
up 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 chal-
lenging 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 mar-
ket 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 consider-
ation as an alternative asset class (alongside invest-
ments such as hedge funds and private equity) for
pension funds. More specifically, JP Morgan calcu-
lated that “investing in 10% equity tranches of five-
year 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
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 allow-
able 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 dur-
ing the grace period, Moody’s treats ABC as a default-
ed 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
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
recovery rates
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 recov-
ery rates. For example, on average, the longer collat-
eral 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 recov-
ery 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
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 situa-
tion 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 recov-
ery 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 so-
called 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 securi-
ties 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, where-
as a more concentrated portfolio with a small number
of exposures is known as a lumpy fund.
After completion of the ramp-up, there is usual-
ly 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
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 col-
lateral underlying CDOs, with resecuritisations – or
CDOs of ABS – one example of a product that has
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
product innovation
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 transac-
tion to another, which is unsurprising given that the
collateral underlying these transactions can also vary
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
considerably.” The same report adds: “Let’s consid-
er, for instance, a transaction backed only by resi-
dential 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 pack-
aging a pool of diversified assets, such as CDOs of
funds. The first CDO of hedge funds – or collater-
alised fund obligation (CFO) – appeared in April
2002. This was a $250 million, five-tranche, five-
year 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
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
CDOs’ underlying assets are other CDOs. The con-
cept 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) prod-
uct, 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
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 basket-
linked 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
product innovation
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 main-
stream 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 origina-
tors 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 transparen-
cy in the CDO market is addressed by trustees,
which distribute monthly reports to investors detail-
ing 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 “notori-
ously difficult to use”, with investors complaining
that “they are dense, inaccessible and frequently rid-
dled 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 mar-
ket-place should lead to a larger and more robust
CDO market over time.” Similar measures in other
asset classes, added Goldman Sachs, had “engen-
dered 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 sub-
scribers, 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 Massachusetts-
based 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 transparen-
cy 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
RiskMetrics’ CDO manager software, the result of
the joint venture is a website, CDOVantage, aimed
at providing CDO market participants with “a con-
sistent 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 mispercep-
tions about the CDO market-place, and demon-
strating 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 cash-
flows and coverage tests, download collateral infor-
mation [and] track exposure across portfolios”.
One of the most active champions of increased
transparency (and hence liquidity) in the CDO mar-
ket has been the BMA itself. For example, the
Association has built up a comprehensive library
containing information on specific CDO transac-
tions that had previously not been available to any
market participant not directly involved in the trans-
action. 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 docu-
mentation is optional, with all transaction informa-
tion supplied by investment banks on a voluntary
basis. The information is searchable on
www.cdolibrary.com by deal name, underwriter, col-
lateral manager and/or deal date.
At the start of March 2004, the BMA announced
the new library open to qualified institutional buy-
ers 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
market information
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 cate-
gories 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 effi-
ciently find information in and reference the report
as necessary. The end result is a much more man-
ageable 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 standard-
isation, 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 origina-
tion 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 informa-
tion assembly, management and dissemination. In
the UK, the Financial Services Authority drew
attention to the concerns arising from these chal-
lenges 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 mod-
els to revalue and risk-manage the transactions on a
day-to-day basis. Valuing and risk-managing com-
plex 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 cur-
rently reviewing firms’ valuation of such positions to
assess systems and controls in this area.”
It is the development of hybrid second-genera-
tion 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-genera-
tion 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-
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
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 port-
folios 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 deal-
ing 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 incorpo-
rate 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 sin-
gle asset class in mind, which made them inflexible,
cumbersome and difficult – in many cases impossi-
ble – to extend to new product lines. Given the
requirements to fit new products into existing sys-
tems, such an approach could delay product roll-
out by weeks or months, This problem is particu-
larly acute today, given the speed with which
quants can generate new structures, suggests
Application Networks. ■
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 exam-
ple, Pimco’s Bill Gross warned that “unregulated
hedge funds, collateralised debt obligations and
poorly structured derivatives of all kinds that redis-
tribute risk but do not eliminate it portend the like-
lihood 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 mak-
ing 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 obliga-
tions market: CDOs, it argues “are a collection
agency of every debt owed by anyone that the
lender is willing to sell. Investment banks corral
Growth in credit derivatives and CDOs
5.0
Size of CDO market
Size of total credit derivatives market (excluding CDOs)
4.0
$ trillion
3.0
2.0
1.0 18%
0.0
1999*
Source: British Bankers' Association
30 credit The ABC of CDO
thousands of these debt claims and turn them into
CDOs, a bond. The CDOs are impossible to under-
stand 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 igno-
rance about credit derivatives within the global finan-
cial community. The Economist chose The swaps emper-
or’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 loss-
es amounting to hundreds of millions of dollars.
Aside from fuelling an unsustainable credit bub-
ble possibly constructed on the basis of ignorance,
sceptics such as Warren Buffet have argued that the
credit derivatives market has led to an equally dan-
gerous 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 impor-
tant counterforce to this apparent concentration,
however, is that the growth of the overall market for
26%
22%
2001† 2004†
* 2002 survey † 2002 survey
11conclusion.qxd 27/04/2004 16:24
credit derivatives is a positive development as it sup-
ports 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 cred-
it 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 telecom-
munications sector, which on a worldwide basis bor-
rowed more than $1 trillion between 1998 and
2001. Although a substantial amount of this debt
defaulted, the capacity of banks to pass their expo-
sure 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 con-
trast to many previous episodes, the global financial
Credit derivatives as a proportion of all derivatives as of December 2002
100
102
80
trillions $
60
40
20
18.5
Interest rate Foreign exchange
Source: Bank for International Settlements 'OTC derivatives market activity 2002'
www.creditmag.com
Page 3
Conclusion
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 prob-
ably 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 signifi-
cantly 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 trans-
parency, 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.” ■
19.3
2.31 2.10
Other Equity Credit
credit The ABC of CDO 31
index.qxd 27/04/2004 16:26 Page 2

index of terms & deals

Terms Deals
A A
Arbitrage (CDO) 8, 10, 11, 12, 23 Aurora Sumitomo Bank 10

B B
Balance-Sheet (CDO) 8, 9, 10, 11, 12 Blue Strip Deutsche Bank 10
Basel Capital Accord 9 Bistro (Broad Index Secured Trust Offering)
Basel II 9 JPMorgan 13

C C
Cashflow (CDO) 11, 12, 17, 26, 28, 29 Caesar Finance Banco di Roma 10
CDO squared 24 Cast Deutsche Bank 10, 29
Churning 15 Concerto Axa 6, 11
Credit event 13 , 23 Core (Corporate and Real Estate) Deutsche Bank 10

D D
Diversity scores 28, 29 Duchess Duke Street Capital 11

F E
First-loss (tranche) 18, 20 Eurocredit I, II Intermediate Capital Group 11

H G
Hybrid 12, 23, 28 Geldilux (Guarantees of Euro-Loan Debt in
Luxembourg) Bayerische Hypotheken- und
M Vereinsbank 10
Managed (CDO) 11, 15. 24, 28 Globe Deutsche Bank 10

O H
Over-collateralisation (OC) 17, 28, 29 Helvetic Asset Trust UBS 7

P I
Passive 15 iBoxx 14, 25

R J
Ramp up 13, 22, 25 Jazz I & II Axa 11
Reinvestment phase 22
Return on equity (ROE) 9, 10 K
Repacks 6, 21 Khaleej Axa 11

S O
Special purpose vehicle (SPV) 22 Overture Axa 11

T R
Tranching 6, 17, 20 Rose Funding NatWest 7
True sale 10, 12, 13
S
S Scala 1 Banca Commerciale Italiana 7
Single-tranche (CDO) 19
Static 15, 20, 25, 28 T
Synthetic 7, 10, 11, 12, 13, 14, 15, 19, 20, 25, 28, 29 Trac-x 14, 25
Super-senior 14 Triton Opportunities Fund Triton Partners 24

V Z
Vintage 22 Zing I (Zais Investment Grade) Zais Group 24

W
Warehouse (facility) 22
Waterfall 17, 18, 21

32 credit The ABC of CDO

Outside back cover.qxd 30/04/2004 11:54 Page 1