ASSET-BACKED

16 July 2003

Alexander Batchvarov, PhD, CFA

[email protected]
ABS/MBS/CMBS/CDO 101
Jenna Collins
[email protected] A Compendium of Pieces for New and Not So New Investors
William Davies
[email protected]
London Europe
(44) 20 7995 2685
Asset-Backed
Highlights of This Issue
In the course of the last several years the Merrill Lynch International
Structured Finance Research Team has published numerous introductory
reports to the international securitisation and structured credit markets. Until
the beginning of this year we thought that they have fulfilled their role and
can be put to rest. Since then, however, we have received numerous
requests for introductory reports and commentaries, driven by the wide entry
of new investors into the sector. Hence, the need to dust out some of the old
reports, update them and publish them in a new binding.
Here are the key topics covered in this publication:
• Investors’ approach to building a pan-European structured credit and
securitisation portfolio
• Introduction to securitisation
• ABS and CDO rating transitions as a key investor consideration
• Analyzing subordinated tranches of ABS
• Master Trust structures on the example of credit card ABS
• Investor analysis of MBS
• Real estate and CMBS considerations
• Key features of synthetic and cash CLOs
• Static and dynamic synthetic CDOs

Refer to important disclosures at the end of this report. Investors should assume that Merrill Lynch is
Analyst Certification on page 1. seeking or will seek investment banking or other
business relationships with the companies in
Merrill Lynch Global Securities Research & Economics Group
this report.
Global Fundamental Equity Research Department
RC#61419701
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

CONTENTS
n Section Page

Section 1 European Structured Credit Portfolios 3

Section 2 Securitisation Fundamentals 11

2.1 What is Structured Finance and What is Securitisation? 12

2.2 Structural and Legal Aspects of ABS and MBS 17

2.3 Asset Classes and Basic Securitisation Structures 26

2.4 Benefits and Potential Drawbacks of Secuitisation 33

2.5 ABS Investors and Their Considerations 38

Section 3 Rating Stability of Structured Finance Bonds 41

3.1 Moody’s ABS Rating Transition Study 2002 – Many Things to Write 42
Home About!

3.2 Moody’s CDO Rating Migration Study 1996-2002 – More Doom Than 50
Gloom!

Section 4 Assessing Subordinated Tranches in ABS Capital Structure 55

4.1 Framework for Subordinated Tranches Analysis 56

4.2 Break-Even and Sensitivity Analysis 60

Section 5 Key Consideration for Master Trust Structures 65

Section 6 Commercial Real Estate and Securitisation 73

6.1 Investment Landscape 74

6.2 CMBS: Securitised Property Portfolios 76

6.3 Legal Aspects of European Commercial Property 81

Section 7 Basics of Synthetic Collateralised Debt Obligation 85

7.1 Synthetic CLOs Explained 86

7.2 Synthetic Static and Dynamic CDOs – Structural Variations 99

7.3 Investor Protection and Strategies 104

Section 8 MBS Investor Considerations 107

8.1 Analytical Framework 108

8.2 Calculus for MBS 119

Refer to important disclosures at the end of this report. 2

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

1. European Structured
Credit Portfolios

Refer to important disclosures at the end of this report. 3

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Structured finance and securitisation are usually used as substitute terms, although
the former is by definition broader and includes traditional asset backed
securitisations, as well as more recent, innovative funding and risk transfer
approaches. Such innovations include the future flow (or revenue-based) financing
and operating assets (or whole business) securitisation. More recently, a
combination of structured finance techniques and credit derivatives has come to
create synthetic credit exposures and credit risk transfers, embraced under the
generic term synthetic securitisations.
In this article we introduce an analytical credit framework for structured finance
bonds in order to make them easier to understand and position in an investment
portfolio. It is important to emphasize that securitisation’s most ‘notorious’
feature stems from the level of separation of the underlying pool of assets from
their originator. For that separation to occur, however, third parties may be
introduced and other credit linkages may be created. Hence, the need for a two-
dimensional credit plane to populate with structured finance and securitisation
credits.

Defining the Credit Space of Securitisation

All fixed income credits can be placed on a credit continuum:
• From credits fully dependent on the operating and payment ability, and hence
bankruptcy risk, of the company issuer/originator within the respective
industry-specific and market environment.
• Through credits gradually removed from those risks to a point where they are
fully separated from the originator and they rely entirely on the performance
of the ring-fenced assets, transferred and pledged as bond collateral.
The former, are usually referred to as plain vanilla unsecured bonds, while the
latter have come to be known as asset-backed securities. Between these two well-
understood extremes lies a series of credits, which occupy an intermediate position
in terms of credit dependence, structural enhancements, legal underpinnings,
operating exposure – they have come to be referred to interchangeably as
structured financings or securitisations.
The different types of structured credits occupy a different spot along that credit
continuum. The degree of their autonomy from the underlying originator is
usually reflected in the gap between the rating assigned to the structured credit and
the rating assigned to the related asset originator. The level of dependence on the
credit quality of the underlying assets is reflected in the level of credit
enhancement, which partially bridges that gap. In an attempt to mitigate the
influence of the originator of the assets or the risks associated with the assets
themselves on the structured bonds and achieve a higher rating for the latter,
additional parties with their own credit risk profiles are introduced. We refer to
the latter collectively from here on as third parties or third parties risk. Such third
parties may include servicer, insurance provider, swap provider, portfolio
manager, property manager, off-takers, letter of credit provider, etc. Hence, a
two-dimensional credit plane is needed to position the universe of securitisations
and structured finance credits. (See Chart 1)
Understanding the position of any given structured credit within the frame of the
credit plane has serious implications for:
• Originators, as they determine the role they play during the life of a given
structured credit and related effects on their operating and financial results,
and ;
• Investors, as they determine the role of a given structured credit in the overall
portfolio management - the portfolio’s credit diversification, performance,
monitoring.

Refer to important disclosures at the end of this report. 4

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 1: Credit Independence of Securitisation Credits from Originator (horizontal axis)

and from Third Party (vertical axis)

General
HIGH insurer
Monoline
wrap Single asset/manager
Few assets portfolio
De-linked Insured Single industry portfolio
Securitisations
Synthetic Securitisations ABS/MBS /CMBS
THIRD Portfolio of CLO/CBO
PARTY CLN insured
assets Diversified
multiple asset
Future
portfolio
Flow Multiple assets
Operating Assets
LOW Single asset

Corp Bond
HIGH ORGINATOR/ISSUER LOW
Ideal Securitised
Bond
Source: Merrill Lynch

Source: Merrill Lynch

Populating the Credit Space of Securitisation

Structured finance or securitisation comprises an array of approximations to debt
financing. With a range of options to use, companies can implement variations of
any one basic design, depending on their particular needs. With a range of options
to choose from, investors can find the best match for their portfolios and portfolio
management objectives.
The terms structured finance and securitisation are applied freely to include:
• Asset-backed securities (ABS) and Mortgage-backed securities (MBS),
Collateralised Debt Obligations (CDOs) and Commercial Mortgage Backed
Securities (CMBS). This is the most fundamental format of securitisation and
involves the issuance of off-balance-sheet debt, backed by a homogenous
pool of assets. Securitised assets include bank and finance company assets
such as credit cards, auto loans, mortgages, leases, corporate loans, real estate
loans, etc., or corporate assets like trade receivables, vendor financing, real
estate, etc. Securitisation in its narrow traditional sense is usually associated
with ABS and MBS only.
• Future flow (or revenue-based) securitisations. This type of financing is
usually associated with emerging markets and project finance. Assets include
revenues generated in the normal course of business of a given company, be it
a manufacturer or a property manager - export receivables, fee-based
revenues, rental flows, etc.
• Operating assets (or whole business) securitisations. Viewed as, but not
limited to, exit strategies for principal finance and mergers and acquisitions
activities, the financing in this case is backed by the core operating assets of a
company and the revenues they generate, and typically has a strong real estate
element, which helps put some of these deals in the commercial real estate
securitisation and CMBS field.

Refer to important disclosures at the end of this report. 5

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Insured securitisations. Any of the above securitisation formats can be

subject to a ‘wrap’ – insurance protection against the credit and cash flow
risks of a securitised transaction. It allows for the substitution of the credit
risk of the assets for the credit risk of the insurance provider.
• Synthetic securitisations. It is aimed at achieving a transfer of the credit risks
associated with given assets, while the originator retains them on its balance
sheet. Full or partial financing against those assets may or may not be
realised. Synthetic securitisations may actually involve many of the
securitisation listed so far and it should be viewed more as a form of
execution rather than independent securitisation type.
n Asset-Backed and Mortgage-Backed Securities
The legal separation of assets from the originator (i.e., a ‘true sale’) ensures that
investors in the asset-backed bonds are not directly exposed to credit deterioration
or bankruptcy of the originator. This feature fundamentally distinguishes asset-
backed securities from unsecured debt issued by the same originator or from other
forms of securitisation. Bondholders have no recourse to the originator, except
under limited circumstances.
In most cases, the originator remains the servicer of the pool of assets, with a
back-up servicer identified at the outset of the transaction in case the originator
underperforms or is declared bankrupt during the term of the transaction. Cash
flows generated by the underlying assets are used to service the bonds. The credit
quality of the asset-backed debt is, therefore, based in the first instance on the
credit quality of the legally segregated, ring-fenced pool of assets. However, the
ratings assigned to the asset-backed bonds will reflect the extent of internal or
external credit support, i.e. their credit enhancement.
From investors’ point of view ‘true’ securitisation provides credit exposures
mitigated through structural, legal and credit enhancements, generally unavailable
in the plain vanilla bond market. The exposures can vary widely and include lots
of combinations: numerous obligors – numerous industries portfolio, numerous
obligors - single industry portfolio, numerous obligors – single asset, single
manager – diversified portfolio, single asset - single manager, etc.

n Future Flow (Revenue-Based) Securitisation

In case of the revenue-based securitisation, the well being of the company
generating the assets-revenue is of paramount importance for the performance of
the securities. The bonds are usually backed by a defined and specific revenue
stream generated in the normal course of business of the respective company.
These could include fees for services provided to customers; payments due for
exported goods, utilisation fees, settlement payments, money transfers, etc.
Bondholders’ rights are not limited only to those revenues: in case of default or
bankruptcy of the originator, the bondholders have a right to recourse and become
general unsecured creditor to the originator.
The credit quality of the revenue-based bonds are generally in line with the credit
quality (the unsecured senior debt rating) of the respective originator (representing
generation risk for the assets) and the related obligors (representing payment risk
for the assets, off-take contacts). Credit support in the structure is achieved
through a desired level of debt service coverage, and excess debt service flows are
remitted back to the originator or retained in case of adverse events. Covenants
may also play an important role.
Investors take full credit exposure to the company and a set of its customers, yet
certain other risks (which are inherent in the company’s plain-vanilla bonds) may
be well mitigated (e.g. currency and sovereign risks in case of emerging markets
export based deals). They can also have access to areas of credit traditionally
reserved for the banks and bank loan market, as is the case with project finance.
In the worst case they may be treated as a general non-secured creditor of the
company. Investors’ exposure can be defined as multiple obligors – single
originator or single obligor-single originator.
Refer to important disclosures at the end of this report. 6
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Operating Asset Securitisation

• An operating asset securitisation quite simply encompasses debt financing
backed by a company’s operating assets, typically its core assets generating
highly predictable income streams given the company’s unique position.
Under the specific legal structure, known as ‘secured bond structure’, a
combination of fixed and floating charges over the assets of the company
gives investors the legal right to determine whether to operate or liquidate the
company in question in case of default on their debt.
• The transaction can be structured with credit enhancement (insurance or credit
tranching, or both) and structural covenants to protect investors, just like in
traditional ABS. Unlike them, though, operating asset-based securitisations
retain a certain level of operating risk. The assets are not typically
transferred, only the income from such assets is assigned for the benefit of
investors. However, in case of a bond default, investors can choose
collectively to run the company or liquidate it (and apply the liquidation
proceeds to redeem their principal).
Credit quality of operating asset securitisations does not need to be capped at the
level of the operating company, particularly, if the assets can be shown to generate
income reliably with little input from the operator and/or to preserve their value in
case of company’s bankruptcy. Obviously, the credit quality of the underlying
obligors (i.e. paying customers of the company, in some cases government or
quasi-government bodies) is also an essential feature of the bonds’ credit.
Investors gain exposure to companies which otherwise may not issue debt and
face full credit exposure to the company in question. They are entitled to a full
priority claim against the company with the potential downside in the value of the
assets (decline in the market value of the assets) and the timing of the realisation
of that value (length of period needed to realise those assets). Credit exposure
may take the forms: pool of assets - single industry, single asset – multiple
obligors.

n Insured Securitisations
Insurance can play different roles in securitisations – it can be applied to the assets
in the securitised portfolio, or it can be applied to the securitisation bond in order
to enhance its credit quality. Insurance can be provided on the bond level by both
general and specialised bond (monoline) insurance companies and on the
collateral level by both general and specialised (primary mortgage) insurers.
Through the insurance the risk of the assets or the bonds is shifted to the insurance
provider. And the insurers’ credit substitutes the credit of the bond or the assets.
It is prudent, though, to look through the insurance cover or wrap and determine
the risk, to which the insurance provider is exposed and the investor or issuer
could be exposed in case of insurers’ default.
The application of the ‘look through’ principal require understanding of the
process of insurance underwriting, the minimum credit level requirement for the
insurance to be applied, the general business of the insurer and the management of
the insurers’ risk portfolio and overall business. In that respect, the levels of risk
should be different and analysed differently for
• Insurance providers on assets level (e.g. insurers specialising in mortgage
insurance who provide coverage on an individual or pool basis) where the
securitisation bond is backed by a pool of multiple assets insured on
individual or pool basis by one or several insurers;
• Insurance providers on a bond level through specialised bond (monoline)
insurers, whereby such insurers take only investment grade risk, manage their
risks on a portfolio basis and are tightly monitored by the rating agencies in
terms of individual and portfolio risks;
• General insurance providers, for whom bond insurance is just one of the
numerous lines of business.
Refer to important disclosures at the end of this report. 7
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

For investors, this is a way of shifting the exposure from the assets and the
originator away to another party, yet such exposure to the third party is only to the
extent of the credit performance of the underlying bond or pool of assets. In
addition, depending on the type of insurance as discussed above, investors
exposure to the third party may be viewed as exposures to: a pool of assets, a
managed portfolio of risks or a general insurance business.

n Synthetic Securitisations
In the case of synthetic securitisation, the assets, usually used for credit reference
only, remain on the balance sheet of the respective originator/issuer. Through a
number of mechanisms the originator seeks to shift the risk of those assets to other
parties: through a credit default swap to the swap provider or through the purchase
of protection from the market, or a combination thereof. Raising financing is of
secondary importance, if any, for the seeker of credit protection.
The reference portfolio of assets is clearly defined but remains on the sponsor’s
balance sheet – the portfolio comprises residential or commercial mortgages,
unsecured or secured consumer and corporate loans, real estate. The risks
associated with this portfolio and subject to the transfer are defined through ‘credit
events’. If bonds are issued, in the case of partially funded or fully funded
structures, their credit performance is fully linked to the portfolio and its credit
enhancement. In an effort to de-link the rating of the bonds from the rating of the
issuing bank, the note proceeds may be used to purchase a portfolio of government
bonds, mortgage bonds or MTNs, as collateral for the bonds.
Hence, the credit performance of the bonds depends on the credit performance of
the referenced portfolio, the enforcement of credit events, while their cash
performance is linked to the cash performance of the collateral portfolio. Unlike
conventional ABS or CLN structures, the amortization proceeds of the referenced
assets are not used to make payments under the structured bonds. Investors only
have synthetic exposure to the referenced portfolio, whilst debt service is met by
the yield of the collateral portfolio supplemented by insurance premiums paid by
the sponsor in return for credit protection for the referenced portfolio. The
purchased collateral is reduced to the extent necessary in order to compensate the
sponsor for losses incurred on the referenced portfolio. Ultimately, the collateral is
liquidated to meet note holders’ principal payments.
Credit exposures are related to the reference portfolio, its credit quality and credit
performance related to the definition of credit events and the determination of
their occurrence (where the trustee’s role is broader than usual). Others stem from
the collateral portfolio and its performance, mechanics to protect against its
market risk. Credit default swap counterparty is another element in the credit
picture. The originator plays a role by assuming the negative carry on the
collateral through regular payments made in the form of insurance premiums.
Investors’ levels and nature of credit exposure depends on the particular type of
synthetic securitisation. In the case of credit-linked bonds, investors’ exposure is
to both the credit quality of the asset portfolio and the issuer’s credit standing. In
the case of unfunded or partially funded structures, the exposure is to a diversified
assets pool, highly leveraged for the latter. Investors are gaining exposure to the
credit risk as defined by the credit events of a diversified and enhanced pool, while
the cash flows for debt service are derived from a non-related highly liquid, high
credit quality bond portfolio.

n Securitisation Credit Space and Investment Portfolio

The stated initial purpose of securitisation was to achieve structured bonds whose
credit risk lies with the transferred assets and is entirely de-linked from the credit
risk of the originator of the assets. The developing the structured techniques and
their application to different asset classes and company needs have increasingly
lead to introduction of third party risks or to leaving residual originator risk in the
securitisation bonds.

Refer to important disclosures at the end of this report. 8

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Hence, the need for a credit framework within which to position the securitisation
bonds. The proposed framework includes two dimensions to reflect the level of
credit (in)dependence of the securitisation bond from the originator and the third
parties. The credit risk of the asset itself is a reflection of the risk of the originator
and the third party in combination with the credit enhancement as determined by
the rating agencies.
In the language of our schematic increasing credit dependence of the securitisation
bond from the originator and the third parties would be reflected in moving left on
the horizontal axis and up on the vertical axis, respectively. By determining the
level of credit dependence of each bond on the originator and the third parties, we
can position it in our securitisation credit plane:
• We believe the ideal starting point for full credit independence would be a
securitisation bond based on a multiple asset, well geographically and
industry diversified amortising portfolio with easily replaceable servicer. In
the low right corner of our credit plane credit dependence is de minimum.
Such a securitisaton bond should pretty well fit in any portfolio with minimal
surveillance.
• Moving left (backwards) on the horizontal axis, the securitisation bonds have
an increasing degree of originator risk, i.e. risk specific to a given company
and a given industry. At its extreme, future flow securitisation bonds offer
almost full company risk along with mitigated through structural
enhancements certain particularly unpalatable risks. In this case, credit
surveillance would require primarily focus on the originator. Intermediary
points will include operating assets securitisations based on portfolio of
businesses units or a single business.
• Moving up the vertical axis from the ideal starting point, securitisation bonds
have an increasing degree of a third party risk to the point of extreme of, say,
fully insured by a general insurer single asset transaction with some risk
mitigants on the underlying asset. Hence, surveillance is required for the
insurance company in question and the associated insurance industry segment.
Intermediary points would include on the insurance side: bonds based on a
broad portfolio of insured assets, monoline insured bonds, and on the
traditional ABS/ MBS side multiple industry portfolios, single industry
portfolios, decreasing number of assets in a pool down to a single asset/ single
manager situation.
• Moving away from the axes and towards the middle of the credit plane,
securitisation bonds would represent different combinations of originator –
third parties credit risks. Different level of monitoring of all the parties
involved would then be necessary.
The above three extreme points and the relevant intermediary points should be
used as points of reference when building and re-balancing the credit risk of a
securitisation bond portfolio.

Refer to important disclosures at the end of this report. 9

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Refer to important disclosures at the end of this report. 10

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

2. Securitisation Fundamentals

Refer to important disclosures at the end of this report. 11

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Structured finance and securitisation are usually used as substitute

expressions, although the former is by definition broader and includes
traditional asset-backed securities, as well as more creative and innovative
approaches, such as future flow (or revenue-based) financing and operating
assets (or whole business) securitisation, and most recently synthetic
securitisation.

2.1 What is Structured Finance and What is

Securitisation?
Securitisation is a form of debt financing technology, developed two decades ago
and actively used in a variety of forms to raise off-balance and alternative
financing for companies and banks. In its most recent modification, synthetic
securitisation is used as a risk transfer rather than financing mechanism.
In its most generic form, securitisation involves the sale, transfer or pledge of the
specified assets to a special purpose, bankruptcy-remote vehicle or trust (SPV),
which in turn issues notes or certificates to investors. Investors (banks, insurance
companies and specialised funds) generally rely on those assets and associated
pledges for the redemption of their bonds, either from the cash flows generated by
the assets or from the assets’ sale/ liquidation under adverse conditions.
Securitisation and structured finance are generic terms, which are applied
interchangeably to include:
• Asset and mortgage-backed securities (ABS and MBS). This is the most
fundamental format of securitisation and involves the issuance of debt, or
asset-backed securities, secured by a homogenous pool of assets. Securitised
assets could include bank and finance company assets such as credit cards,
auto loans, mortgages, real estate, equipment leases and corporate loans, or
corporate assets like trade receivables, vendor financing and real estate.
Initially, the term securitisation was used to mean ABS and MBS only.
• Future flow (or revenue-based) financing. In this case, the financing is
backed by specified revenues generated in the normal course of business of a
given company - export receivables, settlement and utilisation fees, workers
remittances, etc.
• Operating assets (or whole business) securitisation. The financing in this
case is backed by the core operating assets of a company and typically has a
strong real estate element.
• Synthetic securitisation. A combination of structured finance and credit
derivatives techniques, synthetic securitisation has recently come to address
banks’ need for transfer of risk associated with given assets without the
transfer of the assets themselves. It can be executed as partially or fully
funded securitisation.

n Asset and Mortgage-Backed Securities

The legal separation of assets from the originator (i.e., a ‘true sale’) ensures that
investors in the asset-backed notes are not directly exposed to credit deterioration
or bankruptcy of the originator. This feature fundamentally distinguishes these
securities from unsecured debt issued by the same originator or from other forms
of securitisation.

Refer to important disclosures at the end of this report. 12

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 2: Traditional Securitisation – ABS/MBS

❏ Legal segregation of the assets backing the ABS/MBS from the assets originator
❏ Bankruptcy- remote issuing vehicle
❏ No investor recourse to assets originator in case of losses in the assets pool
❏ Rating on senior notes exceeds rating of asset originator
❏ Off-balance sheet financing for asset originator
❏ Higher yield and more protections for investors

Source: Merrill Lynch

In most cases, the originator remains the servicer of the pool of assets, with a
back-up servicer identified at the outset of the transaction in case the originator
underperforms or is declared bankrupt during the term of the transaction. Cash
flows generated by the underlying assets are used to service the notes issued. The
credit quality of the debt is therefore, based in the first instance on the credit
quality of the legally segregated, ring-fenced pool of assets. However, the ratings
assigned to the asset-backed notes will take into account not just the credit quality
of the underlying pool, but also the extent of internal or external credit support, i.e.
the credit enhancement for the notes, as well as other structural and legal features.
The assets backing the transaction are sold or transferred to the special purpose,
bankruptcy remote entity, issuing the asset-backed securities. The sale or transfer
is ‘absolute’, so that the creditors of the originator of the assets have no claim
against those assets, whereas the investors in the asset-backed securities have no
recourse to the originator of those assets in case of losses under the latter.
The benefits of securitisation to issuers is in the off balance sheet treatment
achieved, as well as the capital relief gained to the extent that the underlying assets
attract regulatory capital charges. Another essential benefit is the diversification
of funding sources. The cost of funding through securitisation is sometimes more
competitive for certain issuers.
Investors in asset-backed securities may be able to pick up yield over comparable
plain vanilla bonds. To a large degree, this incremental spread reflects the relative
complexity of structures and – sometimes – the lower liquidity of securities in the
secondary marketplace.

n Future Flow (Revenue-Based) Financing

In case of the revenue-based securitisation, the wellbeing of the company
generating the assets is of paramount importance to the performance of the
securities. The bonds are backed by the revenue stream generated in the normal
course of business of the respective company. These could include fees for
services provided to customers, payments due for exported goods, utilisation fees
(charges for the use of a pipeline or other distribution networks), settlement
payments for telecom services or credit card usage, etc.

Refer to important disclosures at the end of this report. 13

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 3: Future Flow Securitisation

❏ Defined revenue stream pledged to noteholders

❏ Revenue stream depends on generation capabilities of originator and paying capabilities of
its clients
❏ Originator circumvents restrictive covenants and sovereign ceiling
❏ Investors gain access to borrowers in emerging markets without the corresponding
sovereign currency exposure
❏ Rating of notes in line with rating of originator

Source: Merrill Lynch

A defined revenue stream is pledged to the bondholders for the purposes of the
securitisation deal. However, bondholders’ rights are not limited only to those
revenues. In case of default or bankruptcy of the originator, the bondholders have
a right to recourse and become general unsecured creditors to the originator.
Many of the future flow transactions are structured to mitigate certain sovereign
risks (e.g. currency risk) and allow a highly rated corporate or bank borrower from
a given low rated country to raise financing at more advantageous conditions than
the country’s sovereign ceiling would otherwise permit.
The ratings of the revenue-based bonds are generally in line with the unsecured
senior debt ratings of the respective originator. Credit support in the structure is
achieved through a desired level of debt service coverage, and excess debt service
cash flows are remitted back to the originator or retained in case of adverse events.
As in some other debt financing transactions, covenants play an important role to
protect investors.
Revenue-based securitisation is often used when the company faces restrictive
covenants on its other debt preventing it from pledging assets. The sale of future-
generated revenue circumvents such restrictions. In addition, in emerging markets
it helps accommodate higher credit quality companies located in lower credit
quality sovereigns, and allow the former to raise funds at more advantageous
terms than the sovereign ceiling would otherwise allow them. Project finance
bonds are yet another user of this securitisation structure.

n Operating Asset Securitisation

An operating asset securitisation simply encompasses debt financing backed by
the operating assets of a company, typically the core revenue generating assets of
the company. Such assets should generate highly predictable income streams
given the company’s superior competitive position as a result of, say, a monopoly
or a particular business niche.
Many operating asset securitisations have been employed as a mainstream
corporate finance exercise by companies involved in acquisition finance,
especially by principal finance groups and companies involved in operations
related to government tendered service contracts.
In operating assets-based securitisation, the core assets of a company are
specifically ear-marked to secure the debt issued. Such securitisations have been
executed mainly in the UK, where legal aspects of the transaction typically ensure
that investors have first recourse to the assets in case of originator’s bankruptcy

Refer to important disclosures at the end of this report. 14

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

through the so-called ‘secured loan structure’: investors are granted fixed and
floating charges over the assets of the company, which allow them to take over
those assets in case of default and determine whether to operate or liquidate them
to satisfy debt payments before potential creditors.

Slide 4: Operating Asset Securitisation

❏ Assets backing the transaction are the core operating assets of a company or entity
❏ ‘Secured loan structure’ allowing investors to take over the company in case of default on
the notes
❏ Operating risk
❏ Reliance on operating cash flows and liquidation value
❏ Widely applicable in acquisition, privatisation and principal finance
❏ Exposure to sectors which otherwise may not issue debt
❏ Risks of changes in the value of the assets and the timing of its realisation

Source: Merrill Lynch

Operating assets securitisation is based on operating cash flows (EBITDA) and
can be structured with credit enhancement (insurance or credit tranching, or both)
and structural covenants to protect investors.
However, operating assets-based securitisations are fundamentally different in that
a certain level of operating risk remains and is assumed by investors. The assets
are not typically transferred. The income from such assets is assigned to the SPV
in benefit of investors. Generation of revenue used to service the debt depends on
the continued operational use of these assets by the company.
From issuers’ perspective, operating asset securitisations can be employed as yet
another corporate finance tool especially by companies involved in acquisition and
principal finance, companies involved in operations related to government
tendered service contracts.

Refer to important disclosures at the end of this report. 15

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Credit ratings for operating asset securitisations do not need to be capped at the
ratings of the company, particularly, if the assets can be shown to generate income
reliably with little input from the operator and to preserve their value in case of
originator bankruptcy. Obviously, the credit quality of the underlying obligors
(i.e. paying customers of the company) is also an essential feature of the rating
analysis. In addition, some operating assets-based securitisations benefit from
reliance on payments from the government or quasi-public bodies.
From issuers’ perspective, operating asset securitisations can be employed as yet
another corporate finance tool especially by companies involved in acquisition and
principal finance, as well as companies involved in operations related to
government tendered service contracts.
Investors gain exposure to companies which otherwise may not issue debt and
face full credit exposure to the company in question. They are entitled to a full
priority claim against the company with the potential downside in the value of the
assets (decline in the market value of the assets) and the timing of the realisation
of that value (length of period needed to realise those assets).

n Synthetic Securitisation

Slide 5: Synthetic Securitisation

❏ Non-funded, partially funded or fully funded structures

❏ Main goal credit risk transfer (and not necessarily asset transfer or financing)
❏ Referenced pool of assets remain on balance sheet
❏ Collateral portfolio of highly rated bonds assures rating de-linkage
❏ Bank pays for the protection - ‘insurance premium’ makes up the difference between the
bond coupon and the yield of collateral portfolio
❏ Applicable to a variety of asset classes on bank’s balance sheet
❏ Ease of execution for investors
❏ Wide range of exposures for investors-sellers of credit protection

Source: Merrill Lynch

In the case of synthetic securitisation, the assets, usually used for credit reference
only, remain on the balance sheet of the respective originator/issuer. Through a
number of mechanisms, the originator seeks to shift the risk of those assets to
other parties: through a credit default swap to the swap provider or through the
purchase of protection from the market, or a combination thereof. Raising
financing is of secondary importance, if any, for the seeker of credit protection.
The reference portfolio of assets is clearly defined but remains on the sponsor’s
balance sheet – the portfolio comprises residential or commercial mortgages,
unsecured or secured consumer and corporate loans, as well as real estate. The
risks associated with this portfolio and subject to the transfer are defined through
‘credit events’. If bonds are issued, in the case of partially funded or fully funded
structures, their credit performance is fully linked to the portfolio and its credit
enhancement. In an effort to de-link the rating of the bonds from the rating of the
issuing bank, the note proceeds may be used to purchase a portfolio of government
bonds, mortgage bonds or MTNs, as collateral for the bonds.

Refer to important disclosures at the end of this report. 16

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Hence, the credit performance of the bonds depends on the credit performance of
the referenced portfolio and the enforcement of credit events, while their cash
performance is linked to the cash performance of the collateral portfolio. Unlike
conventional ABS or CLN structures, the amortization proceeds of the referenced
assets are not used to make payments under the structured bonds. Investors only
have synthetic exposure to the referenced portfolio, whilst debt service is met by
the yield of the collateral portfolio supplemented by insurance premiums paid by
the sponsor in return for credit protection for the referenced portfolio. The
purchased collateral is reduced to the extent necessary in order to compensate the
sponsor for losses incurred on the referenced portfolio. Ultimately, the collateral is
liquidated to meet note holders’ principal payments.
Credit exposures are related to the reference portfolio, its credit quality and credit
performance related to the definition of credit events and the determination of
their occurrence (where the trustee’s role is broader than usual). Others stem from
the collateral portfolio and its performance, mechanics to protect against its
market risk. Credit default swap counterparty is another element in the credit
picture. The originator plays a role by assuming the negative carry on the
collateral through regular payments made in the form of insurance premiums.
From issuer’s point of view, synthetic securitisation is a simpler and easier way of
transferring risk especially when financing is not a primary goal, and can be
applied to much larger (than for traditional ABS) portfolios.
Investors’ exposure levels and the nature of credit exposure depend on the
particular type of synthetic securitisation. In the case of credit-linked bonds,
investors’ exposure is to both the credit quality of the asset portfolio and the
issuer’s credit standing. In the case of unfunded or partially funded structures, the
exposure is to a diversified assets pool, highly leveraged for the latter. Investors
are gaining exposure to the credit risk as defined by the credit events of a
diversified and enhanced pool, while the cash flows for debt service are derived
from a non-related highly liquid, high credit quality bond portfolio.

2.2 Structural and Legal Aspects ABS and MBS

The securitisation process features many legal and structural aspects with the main
objective of separation of credit risk of the underlying assets from the originator’s
credit risk. That separation is the necessary prerequisite for achieving a higher
credit quality of the asset-backed securities than that of the originator. Other
prerequisites include credit and liquidity risk mitigants.

Refer to important disclosures at the end of this report. 17

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n The Securitisation Process

Slide 6: Generic Securitisation Chart

Originator
Originator Assets
Assets

Servicer
Servicer Transfer of Asset Pool
(Sale or Assignment)

Credit
Credit SPV
SPV // Trust
Trust -- Issuer
Issuer Swap
Swap
Enhancer
Enhancer

AAA rated bonds

Structural/
Structural/ Legal
Legal
Elements
Elements Liquidity
Liquidity Provider
Provider
Investors
Investors

Source: Merrill Lynch

Let us look at a very simple schematic diagram that outlines the main generic
structure of securitisation transaction (Slide 9).
The originator, a company or bank originates assets in the normal course of its
business and retains them on its balance sheet. It needs financing, and one way of
acquiring it, is by monetising the assets it has, selling them to another entity, (an
SPV) established solely for the purpose of that financing. The SPV (the Issuer)
issues securitisation bonds to investors and applies the issuance proceeds to
purchase the assets from the asset originator. The SPV’s balance sheet now
consists of assets – the assets acquired from the originator (as they are no longer
on the originator’s balance sheet) and liabilities in the form of the bonds issued.
The SPV is a shell company, which holds the assets for the benefit of the bond
investors. Hence, the need for a company to look after those assets (the servicer).
In order to achieve desired credit quality of the bonds issued based on the assets,
there is a need for additional supports: credit, structural and legal enhancements.
These are put in place to mitigate the credit, legal, liquidity, interest rate, currency
or other risks associated with the assets and the transaction. Hence, the roles of
credit enhancer, liquidity provider and swap counterparty. The securitisation
transaction is structured within a given legal framework in order to achieve the
legal separation of the assets from the originator.
The most senior tranches of a securitisation bonds, generally achieve AAA rating.
This is an essential aspect of securitisation. Securitisation allows originators with
lower credit quality to issue AAA rated bonds. In other words, the originator
receives better funding irrespective of its own credit worthiness on a stand-alone
basis.
The legal separation of the assets from the originator, the credit and structural
enhancement in the securitisation structure altogether provide for a securitisation
bond for highest credit quality (rating).

n Key Securitisation Parties

Let us analyse the parties and their role in a securitisation structure.

Refer to important disclosures at the end of this report. 18

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 7: Key Parties to ABS/MBS Financing

❏ Assets Originator
➢ entity with funding needs and with assets which can be used as collateral for ABS/MBS funding
❏ Issuer of ABS/ MBS
➢ entity specifically created for the purposes of the securitisation - Special Purpose Vehicle (SPV) or
Special purpose Company (SPC)
❏ Outside Credit Enhancer
➢ entity providing credit enhancement through insurance, guarantee or reserve account
❏ Servicer
➢ entity which collects and distributes the cash flows from the assets
❏ Liquidity Provider
➢ entity that addresses the timing mismatches between the collected cash flows from the assets and the
cash flows to be distributed under the structured bonds
❏ Rating Agencies
➢ determine the credit strength of an ABS and size the credit enhancement level necessary to achieve
that credit strength

Source: Merrill Lynch

Originator The originator is an entity which has funding needs and holds assets that are used
as collateral in an asset-backed structure to achieve higher rating and, hence, better
funding. The originator could be any entity, which has well defined assets on its
balance sheet. These assets should generate predictable and stable future cash
flows. They should not only be clearly defined, but also be legally transferable by
the originator to the issuer.
Issuer The issuer is a specially created entity for the purposes of the securitisation,
known as a special purpose vehicle (SPV) or special purpose company (SPC), or a
trust under some Anglo Saxon jurisdictions. We have to stress the fact that this
entity is established only for the purposes of the respective funding and expected
issuance of asset-backed securities. Its obligations, hence, should be limited to the
satisfaction of this particular purpose. It is an entity structured to be legally
different and independent from the originator of the assets. The functions of that
entity are limited to the issuance of the bonds (its main creditors are the
bondholders) and the acquisition of the assets (thus its main assets are the
securitised asset pool). The SPV is not meant to have any other substantial
obligations or incur other debt, which could jeopardise its status of a bankruptcy –
remote entity.
Outside credit enhancer Outside credit enhancer could be an insurance company or monoline insurer, or in
some cases could simply be a reserve account which is funded with a letter of
credit. For most securitisation transactions, however, the credit enhancement is
internal to the deal, i.e. through subordination or overcollateralisation.
Servicer The servicer could be an outside party or a party in the structure such as the
originator. This is a very important point to understand when analysing asset-
backed securities. The servicer is responsible for the assets: the generation of the
cash, its collection, pursuing delinquent and defaulted accounts. This role is
crucial because investors rely for their repayment only on the cash generated from
the assets, as there is no recourse back to the originator of those assets. If the
servicer is the originator, any negative changes that the originator could encounter
could affect his role as servicer, influencing the performance and rating of the
respective asset-backed securities. As we mentioned earlier, one of the objectives
of the securitisation process is to de-link the rating of the asset-backed securities
from the rating of the pool and the rating of the originator. However, the rating of
the originator may continue to have a bearing over the structure and the
performance of the deal in its role as a servicer.

Refer to important disclosures at the end of this report. 19

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

The liquidity provider addresses only timing mismatches between the cash flows
Liquidity provider
generated in the asset pool and cash flows needed to be paid under the asset-
backed security. These timing mismatches may arise because of delays in
transferring the money, rising delinquencies or some technical glitches. It is
important to remember that the liquidity provider only covers timing mismatches
in the cash flows, not cash shortfalls due to losses in the asset pool.
Rating agencies Important players in the securitisation process are the rating agencies. The rating
agency in the asset-backed securities and mortgage-backed securities market
establishes the credit enhancement levels, i.e. the cushion that investors receive to
protect them against losses, according to the desired rating levels. The rating
agency has the role of determining the size of the cushion, which ultimately
depends on the credit quality of the asset pool and the desired rating of the bonds.

n Key Legal Concepts in ABS Securitisation

Slide 8: Key Legal Concepts in Securitisation

Separation of Asset Risk from Originator Risk

❏ Assets are transferred from the originator to the SPC

➢ assets are removed from the bankruptcy estate of the originator who retains no legal
and equitable interest in those assets - absolute removal, true sale
➢ absolute (true) sale tests: degree of transfer of ‘risk of loss’ away from the originator;
degree of retention of ownership benefits by originator; degree of control over the assets
after the transfer; accounting treatment of asset transfer by originator; intent of sale
➢ asset transfer results in monetising the assets for cash in consideration of their value - no
fraudulent conveyance
❏ SPC is structured to be a bankruptcy remote (not bankruptcy-proof) entity
➢ its assets are unlikely to be considered part of the assets of the originator in case of the
originator’s bankruptcy - no substantive consolidation
➢ limited risk of bankruptcy filing - voluntary or involuntary

Source: Merrill Lynch

It is vital to achieve a full separation between the assets and the originator of those
assets to obtain AAA rating for the asset-backed securities, so that the assets (and
subsequently the bond investors) are not affected negatively by the originator’s
bankruptcy, which could be particularly severe in the case of low-rated
originators. There must be a particular structure that allows ‘absolute’ separation
in the ownership of these assets from the originator in order to avoid the threat of
the assets being consolidated back into the bankruptcy estate of the originator.
One of the aspects of the separation is the so called ‘true’ sale of the assets, i.e.
ensuring that no creditors of the originator have any claims against the sold assets,
and those assets cannot be consolidated in the bankruptcy estate in case of
insolvency proceedings against the originator. Furthermore, the sale of the assets
is subject to a special agreement between the originator and the SPV.
The agreement must be an arms-length agreement and the payment must represent
fair value of the assets. The originator normally does not retain any control over
the SPV nor over the transferred assets. In general, the aim is to fully transfer the
assets with their underlying benefits and risks to the SPV.

SPV is structured to be a Another element in the structure is the SPV (Issuer), a bankruptcy remote, not
bankruptcy proof, entity. The SPV could be structured as a corporate entity or a
bankruptcy remote, not
trust, in all cases independent from the originator. Arrangements are made so that
bankruptcy-proof the risk of involuntary or voluntary bankruptcy of the SVP is remote. One way of

Refer to important disclosures at the end of this report. 20

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

achieving this is defining explicitly the purpose of the SPV and its obligations
towards outside parties. These obligations could be related to the payment of
servicing fee of the servicer, fees under the letter of credit provider, amongst
others, but its core obligations remain those to the securitisation bond investors.
Normally, the SPV should not incur any additional debt.
If there is a swap in the structure, it is important to determine what is the potential
termination payment that the SVP could owe to the swap counterparty, and what
priority this payment has in the deal’s cash flow waterfall. In case of currency
swaps, the termination payments could be large and the issuer may face a problem
making such a termination payment.

n Credit Enhancement – A Key Feature of All Securitisations

A common feature of all types of securitisation transactions is the use of credit
enhancement, i.e. a cushion put in place to protect investors against expected
losses. The credit enhancement is sized to reflect an expected loss level
determined under a series of adverse scenarios that could affect the asset pool
during its life.
The credit enhancement for a specific deal is usually a combination of several
forms of credit enhancement mechanisms and is a reflection of the specific
characteristics of the securitised assets, the goals of the securitisation sponsor and
the requirements of the rating agencies.
Credit enhancement is normally sized by the rating agencies to help attain the
desired ratings for the securitisation notes. Senior asset-backed notes are usually
assigned the highest rating (triple-A).
Internal or external credit enhancement for asset-backed notes is typically
complemented with structural investor protection built into the securitisation
transaction.
The credit enhancement and the other structural enhancements should amongst
them address most of the adverse eventualities that could affect the asset pool and
the securitisation bonds. Along with the legal protections they help create bonds
which are fundamentally different and more resilient than other fixed income
instruments.

n Forms of Credit Enhancement (CE)

One of the most essential elements in structuring an asset-backed security is
establishing adequate credit enhancement levels. The role of credit enhancement
The role of credit enhancement is to bridge the credit quality of the assets, which may be B or BB, to the level of
the desired rating of the asset-backed security, generally AAA. The credit
is to protect against the credit
enhancement is sized to absorb the expected losses that the pool could experience
risk of the assets and help during the life of the asset-backed security, down to a residual level corresponding
achieve the required rating of with the expected losses under the required rating level.
the asset-backed security
The credit enhancement can be structured in a number of different ways.
External - Provided by an Outside Party
In the earlier stages of the development of the asset-backed market, the external
In the earliest stages of the
credit enhancement prevailed. It is called external because it is provided by an
market development, the outside party, a bank opening a letter of credit (LOC) or an insurance company
external credit enhancement and a monoline insurer providing a surety bond, or a company giving some other
prevailed form of guarantee. It would also take the form of a loan provided by a third party
and subordinated to the senior asset-backed bonds sold to investors.

Refer to important disclosures at the end of this report. 21

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 9: Forms of Credit Enhancement

❏ External - provided by an outside party

➢ bank letter of credit
➢ insurance company surety bond
➢ financial assurance company guarantee
➢ subordinated loans from third party
❏ Internal - provided by originator or within the deal structure
➢ reserve account/refunded or build up from excess spread
➢ originators guarantee
➢ senior-subordinated structure
➢ excess spread
➢ overcollateralisation
➢ minimum required debt service coverage ratio (DSCR)
❏ Trigger events

Source: Merrill Lynch

However, it is important to remember that in the case of external credit

enhancement the rating of the security has a direct link with the rating of the credit
enhancer, whether it is a bank providing an LOC or an insurance company
providing a surety bond. Any rating downgrade, any bad news, any volatility in
the quality or performance of the respective credit enhancer will have a direct
impact on the performance or the rating of the insured securitisation bond.
Internal - Provided by Originator or Within the Deal Structure
As the market developed, a new type of credit enhancement emerged. This credit
support is provided within the structure by the originator or through mechanisms
internal to the deal structure (subordination, overcollateralisation, etc). The
originator for instance, could provide some kind of a corporate guarantee for the
asset-backed securities issued. Such a guarantee is typically attached to the most
junior tranches of an asset-backed security for the purposes of improving their
rating and improving their distribution in the market place.
The most common form of internal credit enhancement is subordination.
Subordination or ‘credit-tranching’, means that the cash flows generated by the
assets are allocated with different priority to the different classes of notes in order
of their seniority. In case of subordination, the face value of the bonds is equal to
the value of the assets. The subordinate structures are known as a
senior/subordinated structure, or also as a senior/mezzanine/subordinated
structure. In a simple senior/subordinated structure the senior tranche is usually
rated AAA and it receives the cash flow generated by the assets first, for the
purposes of interest and principal payment while the subordinated piece (also
called equity piece), receives cash flows second and absorbs the losses first. The
priority of the cash flow distribution comes from the top, waterfall like, while the
distribution of the losses rises from the bottom.
In case of overcollateralisation, the value of the assets exceeds the face value of
the notes. For instance, the assets can be purchased at a discount. The level of the
discount reflects the level of expected losses to be potentially incurred by the
assets, as well as the level of deal expenses.
Another form of internal credit enhancement is the requirement that the cash flows
generated by the assets over a specified period of time exceed the debt service
requirement of the bonds over the same period by a predetermined factor. The

Refer to important disclosures at the end of this report. 22

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

minimum required debt service coverage ratio (DSCR) it that factor. The revenue
from the assets must exceed the debt service several times, thus allowing for
monitoring performance and applying excess debt service to accelerate bond
amortisation in case of adverse events affecting the transaction.

n Excess Spread Capture Mechanics

Slide 10: Calculation of Excess Spread

❏ Yield- revenue generated by the asset pool from interest payments and other charges -
penalties, annual fees, etc.
Minus
❏ Coupon (interest paid under the securitisation bonds) Base Rate
❏ Servicing fees (payments made to the servicer)
Net Yield
Minus
❏ Losses (or charge-offs, lost revenue of the asset pool due to defaults, shortfalls in asset
liquidation proceeds, etc)
❏ Excess Spread

❏ Excess spread capture mechanisms

Source: Merrill Lynch

There is one additional aspect of credit enhancement, which rating agencies

frequently do not take into consideration when they establish the required credit
protection level for a given rating. This is the excess spread. The assets in the
securitised pool generate certain revenue, called revenue yield or gross yield. The
assets generate revenues, associated with the interest charges on the respective
debt obligations in that pool. The revenue is used to cover the expenses of the
SPV. The expenses are related to the coupon payments under the asset-backed
securities to investors and payments to the other parties in the deal (like the
servicing fee, the swap counterparty fee or letter of credit fee). Generally, the
combination of a coupon payment and servicing fee payment is known as the base
rate in a structure. The difference between the revenue yield and the base rate or
base expenses is known as net yield.
The net yield absorbs one of the main ‘expenses’ in a securitisation, which are the
The excess spread reflects the
pool losses. The excess spread is an indicator of the credit health of the asset pool
health of the credit since the excess spread moves in the opposite direction of losses and
performance of the asset pool delinquencies. By deducting the losses from the net yield, we get the excess
spread, also known as excess servicing fee. The excess spread generated in each
period could be used to cover losses produced in that period or in the next one, if
there is any way to capture this excess spread and use it for in future.
Generally, the excess spread is remitted back to the originator when not needed.
However, many structures have mechanisms in place to build up reserve accounts
from excess to be used in case of pool performance deterioration. Thus, the
excess spread in each period, or excess spread captured through respective
mechanics in a reserve account, provide additional credit support for the asset-
backed security. This additional credit enhancement is of particular importance to
investors who are focusing on the mezzanine, subordinated and equity tranches.
The trigger events in an asset-backed security are structural enhancements in the
transaction and are often linked to certain credit events. Occurrence of specified

Refer to important disclosures at the end of this report. 23

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

events such as insolvency of originator or deterioration in pool credit quality

expressed in increase in delinquency or loss levels or decrease in excess spread,
triggers ‘early amortisation’, leading to an accelerated repayment of the notes.
n Credit Enhancement

Slide 11: Credit Enhancement

❏ Credit enhancement sized to protect against prolonged reductions in cash flow

➢ modelled on severe stress scenarios, conservative assumptions

➢ based on a multiple of historical losses

➢ adjusted for certain structural risks (e.g., commingling, set-off, servicer transfer, cash
transfer delays)

➢ level of enhancement commensurate with desired rating

❏ Differences in rating approaches

➢ ‘weak link’
➢ ‘expected loss’

Source: Merrill Lynch

The credit enhancement is sized We can recall that the credit enhancement is established to absorb expected losses
to absorb expected losses in the in the performance of the underlying pool of assets. It acts like a loss cushion that
asset pool, backing the ABS is there to absorb the expected losses that the pool could accumulate during the
life of the asset-backed security. Consequently, to determine the credit
enhancement the rating agencies try to predict the performance of the pool. Its
future performance depends on the initial pool quality, general development of the
economy, the performance of the originator and the servicer, etc. In other words,
certain assumptions about pool characteristics and how they can be affected by the
future developments should be made to size the credit enhancement. One of the
best predictive tools available is the scenario analysis based on either probability
distribution of scenarios or on certain extreme (stress) scenarios.
The stress scenarios are generally related to the bankruptcy of the originator and
assume that the asset pool incurs excessive losses. The bankruptcy of the
originator is usually the stating assumption in sizing the credit enhancement; if the
originator is also the servicer, the bankruptcy of the former would require the
transfer of the servicing function, and this would increase the pool losses. These
losses then accumulate during the life of the pool beyond historical averages or
extremes. The new line should reflect particular risks at a given pool, for
example, the risk of set off, commingling and others. In case of credit cards the
originator may issue a credit card to a physical person who also holds deposits
with the bank. In case the bank goes bankrupt, these deposits form part of its
bankruptcy estate, yet the individual can claim that his debt under the credit card
(subject to securitisation) is off-set (extinguished) by the deposit he/she holds with
the bank (and now held in the bankruptcy estate); thus the credit card
securitisation pool incurs a loss equal to that deposit.
Furthermore, if the servicer is the originator, when accumulating the cash flows
Credit enhancement is sized to from the assets transferred to the SPV, these collections, may be co-mingled, or
protect from certain risks: mixed with collections from other assets of the originator. If the originator goes
commingling, set-off, service bankrupt it may be difficult to differentiate between the two cash flows, the ones
transfer, cash transfer delays that belong and should be remitted to the asset-backed pool and the others that
belong to the bank (and other creditors).

Refer to important disclosures at the end of this report. 24

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

The level of credit enhancement Finally, the credit enhancement level, or the cushion sized to absorb the asset pool
losses, should correspond to the required or expected credit rating under the asset-
corresponds to the level of back securities: the respective rating requires different levels of credit
desired rating enhancement: highest for AAA, lower for A and even lower for BBB or BB
security.

n Differences in Rating Approaches

We emphasise that the credit enhancement is determined by the rating agencies,
after evaluating the current, historical and future performance of the pool of assets
backing the securitisation bond. To establish its size though, the rating agencies
use different approaches.

Expected loss approach Some of them apply the expected loss approach, where they determine the
expected losses in the pool under various scenarios. Under this approach, the
expected severity of loss to investors as well as their frequency or probability of
occurrence is determined. The rating agencies simulate the expected cash flows
that the pool could generate, determining the potential losses that it could
accumulate. Some go further and link the expected loss to the level of reduction
of the internal rate of return (IRR) of the bond: the higher the IRR reduction, the
lower the bond rating.
Weak link approach Other rating agencies base their assessment on the so-called weak link approach.
Under this methodology, the rating agencies look at the confluence of different
entities and assets in the structure and determine where the structure could ‘break’,
that is, the weakest link in the chain of assets, counterparties and entities. The
final rating of the security cannot be higher than the weakest link in the structure.
On that basis, the rating agency could determine the probability of first dollar loss.
‘Probability of first dollar loss’ is another rating approach, which can be
contrasted to the ‘expected loss’ approach. Market convention differentiates
First dollar loss between the two by stating that the expected loss approach involves a
consideration of both probability of default and severity of loss on the liability side
of the securitisation, while first dollar loss approach considers only the probability
of default, or other words, a missed dollar payment is considered a default
regardless of the recovery (may be even equal to 100%) that could follow.
It is also important to understand that more often than not an asset-backed security
is rated by at least two rating agencies. Each of them may have a different
approach and may focus on different factors or weigh differently the same factors
to determine the performance under stress scenarios and related expected loss.
The credit enhancement which the respective asset-backed security carries is the
highest required by any one of the rating agencies in order to achieve the desired
bond rating. In this respect, it is worth investigating any split ratings that exist
especially on lower-rated tranches of the securitisation bonds.

Refer to important disclosures at the end of this report. 25

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

2.3 Asset Classes and Basic Securitisation

Structures
The different classes of assets determine the different securitisation structures that
are available in the market. Almost any asset type can be securitised if it meets
some basic conditions. The broad variety of assets in general fits into two
securitisation bond structures: revolving and amortising.

n Asset Classes

Slide 12: Selected Asset Classes

❏ Auto Loans ❏ Auto Leases

❏ Wholesale Auto Receivables ❏ Aircraft and Computer Leases

❏ Credit Card Receivables ❏ Reinsurance
❏ Home Equity Loans ❏ Tax Liens
❏ Manufactured Housing Contracts ❏ CBO’s/ CLO’s
❏ Recreational Vehicles
❏ Commercial Real Estate
❏ Boat/Truck Loans
❏ Credit Card Voucher Receivables
❏ Agricultural Equipment Loans
❏ Workers’ Remittances
❏ Student Loans
❏ Oil and Other Raw Materials Exports
❏ Trade Receivables
❏ Project Finance
❏ Stranded Costs
❏ Sports and Entertainment (Sales, TV
Rights)

Source: Merrill Lynch

Almost any asset type can be Let us briefly talk about the asset types that can be securitised. On slide 15, one
can see a panoply of asset classes ranging from auto loans or credit cards all the
securitised, if it meets some
way through TV rights, sports and entertainment contracts, exports and workers
basic conditions remittances. We generally talk about traditional and more esoteric, or non-
traditional, asset types. But the range of different asset classes that have been
securitised basically proves that pretty much any asset type can be securitised if it
meets a few basic conditions.
Some of these conditions are related to the ability to transfer the legal ownership
or entitlement to the benefits and losses of these assets by the originator to another
party for the benefit of the asset-backed investors. These assets should also
generate predictable and stable cash flows. Further, rating agencies usually
require some historical performance data and performance track record of the
respective assets in order to be able to quantify their credit risk and to size the
credit enhancement for the asset-backed deals. To illustrate these points we look
at the nature of corporate asset securitisation.

n Nature of the Assets Backing Corporate Securitisations

Corporate assets (as opposed to bank ones) present a particular challenge in
determining the type of securitisation they are subject to. Corporates have assets
of different nature:
• Some of them are existing assets, such as real estate or export receivables,
trade receivables, vendor financing, generated based on the performance of
certain, often contractual, obligations (goods exported, services performed,
products shipped, etc.).

Refer to important disclosures at the end of this report. 26

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Others are future assets, such as export receivables, trade receivables, vendor
financing, to be generated in the future based on the performance of some,
often contracted today, obligations.
The key differentiating factor is whether the obligation has been performed, so
that the assets are existing now (current assets), or will be performed, so that the
assets will be existing some time in the future (future assets). Hence:
• The current assets will be less dependent on the corporate originator, because
it has already performed its obligations and the assets are in existence – now,
it is a matter of collection or obligation performance by the counterpart
(importer, buyer). So, the linkage with the originator is limited if the
originator is also the servicer and to the degree a real servicer replacement is
available.
• The future assets will be more dependent on the corporate originator, because
it has to perform its obligations in the future in order to generate the assets
(export the products, ship the goods, perform the services such as deliver
electricity for example). If the originator does not exist in the future, there
will be no assets backing the securitisation. So, the linkage to the
creditworthiness expressed through the credit rating of the originators is
almost 100%. We say ‘almost’ because the credit rating expresses the
probability of default on a financial obligation, and not necessarily on a
contractual obligation – a company may be in a financial default, but still
continue to perform services to clients and thus generate receivables.
Hence, the linkage of the securitisation bonds’ rating with the corporate rating of
the originator in the second case is much stronger than in the first case. Case in
point, EDF transaction in comparison to Cremonini deal (Italian trade receivables
deal priced last week).
Another Issue is the Value of the Assets Backing the Corporate Securitisation.
Deals may be structured on the basis of existing receivables (Cremonini), future
receivables (EDF) or a mixture of both (Chargeur). A revolving feature added to
the existing receivables securitisation does not change their nature – every new
purchase of receivables involves existing receivables.
We suggest a simple calculation to determine corporate linkage and exposure: a
review whether the corporate receivables’ face value is higher, equal or less than
the bonds’ face value:
• If the existing receivables’ value is higher (bond is structured with over-
collateralisation) or equal (bond is structured with subordination) than the
bond’s face value, then the corporate linkage is low – securitisation bond
rating will stand alone dependent more on the credit quality of the
receivables, than on the credit quality of the originator.
• If the existing receivables value is less than the bond’s face value or the bond
is backed by future receivables whose generation depends on the performance
of the originator, then the corporate linkage is high – the securitisation bond
rating will be almost fully linked to that of the originator.
The above has credit analysis, rating and pricing implications:
• In the first case, the focus is on the credit quality of the assets and linkage to
servicer, hence the pricing should be more independent of the pricing of the
originator’s stand-alone bonds.
• In the second case, the focus is on the credit quality of the originator and
secondarily on the credit quality of the assets, hence the pricing of the
securitisation bonds should be closely linked to the pricing of the corporate’s
stand-alone bonds.
• Furthermore, the rating of the securitisation bonds in the first case should be
fairly independent from the rating of the corporate sponsor, while in the
second case it will not only be highly dependent on day one, but will be as
volatile during the life of the transaction as that of the corporate.
Refer to important disclosures at the end of this report. 27
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Basic Structure Types

The different asset classes give rise to different ABS and MBS structures. The
important element in these structures is to match the cash flows generated by the
assets with the cash flows required to service the asset-backed security. It is
almost like asset and liability matching in a balance sheet situation.

Slide 13: Basic Structure Types

ABS
ABS

REVOLVING
REVOLVING PASSTHROUGH
PASSTHROUGH
HYBRID
HYBRID
Bullet/Cont.Am.
Bullet/Cont.Am. Amortising
Amortising

Credit
Credit Cards
Cards MBS
MBS
Floorplan
Floorplan HELOCs
HELOCs HELs
HELs
Trade
Trade Receivables
Receivables Trade
Trade Receivables
Receivables Auto
Auto Loans
Loans
CLO/CBOs
CLO/CBOs Auto
Auto Leases
Leases

Source: Merrill Lynch

There are two basic asset-backed securities structures: the revolving structures,
and the pass through structure.
Revolving Structures
In most general terms, the revolving structure is applied when a pool of short-term
Collected principal is used to assets is used to back a longer-term asset-backed security. Let us say, a pool of
purchase new assets, which credit cards or trade receivables, which have a life of between 60 and 120 days, is
extends the maturity of the used to back five-year credit card or trade receivables asset-backed notes. The
underlying pool and the notes it pool of assets generates interest and principal, and how and when the principal is
supports distributed to investors is the main differentiating feature of the respective
structure applied. In the case of a revolving structure, the principal generated is
used to purchase new receivables during a specified period of time (the revolving
period) and is applied afterwards to repay the bonds.

Refer to important disclosures at the end of this report. 28

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 14: Revolving ABS Cashflows

Interest Only (Revolving)

12

10 Controlled Amortisation Bullet

Payments
8

0
1 2 3 4 5
Year
Bullet – One payment at the end of last year.
Controlled Amortisation – 12 equal monthly payments over the last year.

Source: Merrill Lynch

Generally, these revolving structures incorporate two periods. The first period is
the revolving period. During the revolving period the principal under the asset-
backed security remains outstanding in full and the investors receive only interest
payments. This reflects the cash flows generated by the assets. The revenue or
the yield generated by the assets is used to pay interest to investors (coupon) and
all the other expenses we talked about – servicing fee, LOC provider fees, losses.
The principal is used to purchase new receivables. This could continue for a
number of years, and in our example on slide 17, it continues for four years.
The question now is, once the revolving period is over, how do we apply
principal? One way of using principal is to deposit it into a reserve account, and
generally, we refer to this as principal accumulation period, an accumulation
period. We accumulate principal in a reserve account in order to pay the asset-
backed security at its maturity through a soft bullet payment. In other words, we
have a revolving/accumulation/soft bullet structure.
In other cases, instead of accumulating the principal in a reserve account, we can
start paying it out in regular instalments to investors. In other words, a revolving
period is followed by controlled amortisation and investors receive their principal
back in several regular equal instalments. This is referred to as controlled
amortisation period. Alternatively, we have a revolving/controlled amortisation
structure.

n Early Amortisation Triggers

Essential features in these structures are the early amortisation events or triggers.
During the revolving period, the principal is applied to buy new receivables. If the
originator of those receivables goes bankrupt, there will be no receivables to
generate and there will be no receivables to be purchased with the principal
collected. The solution then is to start passing through the principal to investors.
In other words, certain events (or the trigger event), ‘trigger’ the amortisation of
the bonds on a pass through basis. All principal, as collected, is passed through to
investors to pay down the asset-backed notes. The early amortisation event could
occur at any point in time during the revolving period, the accumulation period or
the controlled amortisation period.

Refer to important disclosures at the end of this report. 29

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Examples of typical amortisation events are shown on Slide 15. Some of them are
related to the originator or the servicer in the transaction, but others are more
‘economic’ in nature, for example, the deterioration of the excess spread below a
certain level or the increase of pool losses above a given level.

Slide 15: Typical Amortisation (Payout) Events

❏ Deterioration of net portfolio yield / net excess yield triggers

❏ Originator / servicer insolvency

❏ Failure to pay principal on expected payment date

❏ Minimum seller interest triggers

❏ Bank breach of representations and warranties

❏ Failure to transfer funds

❏ Failure to replace interest rate cap provider in the event of rating reduction below
required level

Source: Merrill Lynch

Early amortisation triggers are As we discussed earlier, the difference between the revenue generated by the
assets and the expenses needed under the asset-backed security provides excess
established to protect investors
spread. If this excess spread goes below a certain level, this indicates deterioration
against further deterioration of in the pool performance. In order to protect investors against further deterioration,
the underlying pool an early amortisation of the bond is triggered. The revolving or accumulation
period is over, and now principal is passed through directly to investors and they
can still benefit from the remaining excess spread generated in the structure.
Just to illustrate this with numbers, let’s assume that the trigger is set at 3%. If
during the revolving period the excess spread varies between 4% and 6%, the
structure is OK. If the excess spread falls below 4%, it signals deterioration and
indicates a risk of the excess spread trigger being breached. When the excess
spread reaches 3% the trigger is breached and the early amortisation begins. The
excess spread trigger is usually set on a three-month rolling average basis, which
means that if in any one month excess spread is below 3% and in the other two
months of a three-month period it is above 3%, so that the average is above 3%,
there will be no early amortisation. This protects against one-off adverse events
and seasonal fluctuations.
Pass-Through Structures
On the other side of the structural spectrum is the pass through structure where the
principal, instead of applied to purchase new receivables, is passed through to
investors to repay gradually the outstanding amount of the asset or mortgage-
backed securities. Assets with longer maturities include mortgages, auto loans and
home equity loans. Because they have a longer amortisation horizon, generally
they are used to structure longer-term pass-through securities. Again, this is in
most generic terms.
Now let us briefly look at the pass through structures. In the case of pass through
structures, the principal collected from the assets is passed directly through to
investors to pay down the bonds. Slide 19 shows the repayment schedule of a
typical pass-through structure (or a typical loan underlying such structure).

Refer to important disclosures at the end of this report. 30

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

You have probably seen this graph in any textbook dealing with mortgage or auto
loans, where it shows how such a loan amortises.

Slide 16: Pass-Through Securities

Stylised Repayments Schedule for Pass-Through Security

2000

1800

1600
Principal
1400

1200

Monthly 1000
Payments 800 Interest

600

400

200

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Time

Source: Merrill Lynch

In the case of a level pay loan, the borrower makes an equal payment every month,
which is split between interest and principal. Usually, in the earlier stages of the
life of the loan, more of this payment is allocated to interest and less to principal,
and as the loan amortises more payment is allocated to principal and less to
interest. From the point of view of the borrower, the borrower pays more interest
at the beginning and more principal later on, which means that later on the
borrower acquires more equity in the asset and has higher motivation to continue
paying the loan.
Constant (Conditional) Prepayment Rate (CPR)
Many of the auto loans and mortgage loans allow borrowers to prepay. One of the
main reasons borrowers prepay is because they can obtain a cheaper financing for
the respective house or car, which usually happens when interest rates go down or
price competition among lenders increases. So, at any stage in the life of a
mortgage, the borrower could take a new mortgage and prepay the outstanding
amount of the old ones. This is probably all good and clean from the point of
view of the borrower. However, from the point of view of the investor in a
mortgage-backed security, this means that the investor receives back the principal
much earlier than expected. If this happened while interest rates are falling, the
investor in a mortgage-backed security would be exposed to higher reinvestment
risk and convexity risk.
An essential element when Hence, a key element in determining the payment profile of such pass-through
determining the payment profile securities is the determination of the constant prepayment speed for the purposes
is to establish the CPR of pricing the security. The CPR states the average speed at which an investor
should expect to receive principal back.
The CPR is established as an average over a certain horizon, while in real life the
prepayment rate varies from month to month. The monthly prepayment rate could
vary due to changes in the interest rate environment, the industry, as well as due to
seasonal developments. For example, competition in mortgage lending, could
force mortgage lenders to lower the price of a mortgage loan, and could also make
borrowers more aware of their refinancing options. That would speed up

Refer to important disclosures at the end of this report. 31

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

refinancing of the loans. Specific industry events or tax considerations could

introduce definite pattern of seasonal changes. Another factor would be the
existence of prepayment penalties and their enforceabilities. Generally, borrowers
facing high prepayment penalties are less willing to refinance their loans.
However, if the interest rates drop sharply and there are other legal or structural
elements in the mortgage, (say, reset dates when they are allowed to prepay
without penalty) the borrower may consider refinancing the mortgage even if
she/he has to pay a prepayment penalty.
Other elements driving prepayments could also be related to the demographics of a
given country or people moving from one city to another: in the US from the East
Coast to the West Coast; in the Euro-zone, relocating for jobs or other purposes,
which has not happened on a large scale yet. In other words, when investors consider
buying pass through amortising bonds, (especially if they have fixed rate coupons)
investors are facing prepayments risk and have to understand clearly what factors
drive prepayment in the underlying assets. To sum up, prepayments are important in
determining the maturity of a floating rate pass-thorough securitisation bond, and the
convexity and maturity of a fixed rate pass-through securitisation bond.

Refer to important disclosures at the end of this report. 32

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

2.4 Benefits and Potential Drawbacks of Securitisation

The most often touted benefits of securitisation are for issuers the ability to raise
off-balance sheet financing and to diversify funding sources, and for investors –
the ability to pick up yield while buying bonds based on credit enhanced, well-
diversified assets pools. However, securitisation can have drawbacks for both
issuers and investors if not properly executed or not fully understood.

n Securitisation Sponsors

Slide 17: Securitisation Sponsors

❏ Banks
➢ assets on their balance sheet, risk transfers
❏ Companies
➢ trade receivables, exports
❏ Project Finance
➢ cash flow stream generated post completion
❏ Municipalities
➢ tax liens, social security contributions, parking tickets
➢ revenues from a specified entity - toll road, bridge, hospital, etc.
❏ Real estate developers
➢ commercial real estate (offices, shopping malls, hotels, etc.)
❏ Countries
➢ privatisation (PFI in the UK)
➢ export credits, external credits, etc.

Source: Merrill Lynch

Banks - Consumer and Corporate Loans, Mortgages, Real estate
Banks are the most obvious candidates for securitisation. They have assets on
their balance sheet and want to transfer them to raise funding. In addition, they
can transfer the risk associated with these assets through their sale or other means.
By doing that, banks release some of the regulatory capital they hold against those
assets.
Companies - Trade Receivables, Exports
Companies have assets in the form of trade receivables, export receivables or any
other assets (real estate) that could be used in securitisation transactions, for the
purposes of financing and streamlining their balance sheet.
Project Finance - Cash Flow Stream Generated Post Completion
After the completion, a project generates sizeable stable cash flows that can be
used for the purposes of securitisation. It is also possible early in the construction
phase to leverage the cash flows to be generated post completion.
Income from Public Entities - Tax Liens, Social Security Contributions and
Tolls
Municipalities generate stable cash flows and possess certain assets that can be
securitised, such as tax liens, social security contributions, parking tickets and taxi
medallions. Public institutions may also sponsor specific projects such as toll
roads, bridges or hospitals. Municipal assets also can be subjected of
securitisation whether they are in the form of existing assets and related cash
flows (like tax liens, social security contributions, government reimbursements for
taxes collected), or assets which will be generated in the future (like fees from
crossing a bridge or for using a toll road).

Refer to important disclosures at the end of this report. 33

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Real Estate Developers - Commercial Real Estate

Real estate developers have commercial real estates, offices, shopping malls,
hotels that generate rental and capital income. Many banks also hold on their
balance sheets respective real estate assets. Such assets can be subject to
commercial mortgage-backed securities or commercial real estate securitisations.
Countries - Privatisations, Export Credits, External Credits, etc.
Countries have export credits, different assets that they may want to privatise (as
was the case in the UK with the Private Finance Initiative [PFI]). These assets
could be securitised or the purchase of these assets by another entity, privatisation
could be financed through securitisation.

n Incentives for Securitisation – The Originator’s Perspective

Slide 18: Securitisation – Benefits for Originators

Why securitise?

❏ Off-balance sheet financing

❏ Regulatory treatment (regulatory arbitrage) for banks
➢ off-balance sheet treatment allowed by regulators - by transferring risk capital is
‘freed up’ (equity relief) - “cost of securitisation < cost of capital”
➢ asset transfer off-balance sheet reduces leverage and improves ROE
❏ Alternative source of liquidity
➢ diversified funding mix
➢ reduced correlation between the financial performance of the issuer and the risk
of the assets
❏ Exposure management
➢ transferring loans from credit lines where exposure limits have been reached
❏ Removal of illiquid assets from loan book
❏ Transfer uncertainties related to loan prepayments to investors (improve maturity
management)
❏ Achieve better pricing (through higher debt rating) or longer term financing especially for
cross-border securitisations

Source: Merrill Lynch

Off-balance sheet financing securitisation allows companies and banks to raise

financing off-balance sheet, by monetising their currently existing or future assets.
Such financing, especially given its non recourse nature most of the time, is not
recorded as debt and does not affect their financial ratios.
Regulatory Treatment (Regulatory Arbitrage) for Banks
We already briefly referred to the banks’ need for regulatory capital relief. When
Off-balance sheet treatment selling assets off-balance sheet the banks gain off-balance sheet funding (funding
allowed by regulators - by against their assets or monetisation of their assets). Also by transferring these
transferring risk, capital is assets or risk associated with them away from their balance sheets, they free up
‘freed up’ (equity relief) regulatory capital, receiving equity relief. The asset transfer off the balance sheet
reduces the leverage and improves the return on equity. In order for securitisation
to make economic sense, the cost of securitisation should be less than the cost of
equity for the bank. That may or may not be the case, yet there are other benefits
from securitisation for the banks.
Alternative Sources of Liquidity and Diversified Funding Mix
Through securitisation banks and companies can access alternative sources of
Dependence on traditional liquidity and diversify their funding mix reducing the correlation between their
sources of funding is reduced own financial and the risks of the assets – a point we made earlier. A lower credit
quality bank or company, say B or BB, faces very high funding spreads. This
issuer may reduce such funding spreads by issuing asset-backed securities, which
featuring a AAA rating would be priced much tighter.
Refer to important disclosures at the end of this report. 34
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Exposure Management
Banks and companies have certain exposure limits or credit lines, specified limits
Transfer of loans from credit
for their credit exposure to different entities. By transferring loans off its balance
lines where exposure limits sheet, or transferring risk associated with some of their loans and other assets, the
have been reached banks could free their credit lines and avoid breaching their credit limits. They
can free up the lines for further lending to the respective sectors or clients.
Removal of Illiquid Assets from Loan Book
Banks and companies can remove illiquid assets from their balance sheets, as is
Increased liquidity the case with non-performing or sub-performing loans (NPL), and sub-performing
real estate. Securitisation is often the solution for banks burdened by NPL and an
acute need to sanitise their balance sheet. Companies that have accumulated large
amounts of real estate may want to focus on their core businesses, but retain use of
their real estate assets without burdening their balance sheets. They can transfer
those assets through securitisation by selling real estate and leasing it back to their
own rise.
Transfer Uncertainties Related to Loan Prepayments to Investors
Improvement of risk and asset- When finance companies and banks lend for mortgages, they are exposed to the
liability management interest risk of those loans, expressed in their prepayment behaviour. Normally,
when interest rates fall mortgage borrowers tend to refinance their mortgages, i.e.
taking a new mortgage with a lower interest rate, and using its proceeds to repay
in full the old mortgage loan. This creates prepayment risk for the banks on their
balance sheets, and requires more active management of their assets. By
transferring such assets away, the banks transfer the related prepayments risk,
simplify, and improve the maturity management of their balance sheet.
Achieve Better Pricing (Through Higher Debt Rating)
With certain types of structures, like future flow deals (say in the case of exports
Better pricing and longer term
from emerging markets), highly rated companies in lower rated countries can raise
financing especially for cross- debt at better terms than the respective sovereign ceiling would allow them to do
border securitisation on a straight corporate bond basis. This is achieved by structuring a deal, where
the cash flows are generated by exports denominated in hard currency and
captured offshore. The deal can be structured with a rating above the sovereign
ceiling of that country, usually at the level of the credit rating of the exporter on a
stand-alone basis, and the bond may have much longer maturity that their
respective country or exporter would normally get under straight bond funding.

Refer to important disclosures at the end of this report. 35

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Slide 19: Bank Securitisation – a Numerical Example

Assets Bank Funded ABS Funded

AA
AA rated
rated Senior
Senior AAA
AAA rated
rated
Debt
Debt Senior
Senior Debt
Debt
(92)
(92) (91.5)
(91.5)
Portfolio
Portfolio (100)
(100)
VS L+11bp
L+11bp
L+15bp
L+15bp
Yield:
Yield: L+35bp
L+35bp
Losses:6bp
Losses:6bp pa.
pa.
Junior
Junior Debt
Debt (7.5)
(7.5)
Tier
Tier 22 (4)
(4) L+40bp
L+40bp
L+60bp
L+60bp

Tier
Tier 22 (0.5)
(0.5) L+60
L+60
Equity
Equity (4)
(4)
Equity
Equity (0.5)
(0.5)

Assumes LIBOR = 5% Bank Funded ABS Funded

Funding Cost = 4.96% Funding Cost = 5.12%
Net Income = 5.29% Profit = 0.33% Profit = 0.17%
ROE = 8.25% ROE = 34%

Source: Merrill Lynch

We discussed the different motivations different types of issuers may have to

resort to asset-backed financing. We conclude this discussion with a numerical
example, which illustrates the effect on securitisation on bank ROE, to describe
the benefits for the banks. We show in Slide 24 a pool of assets that a bank can
fund in a traditional way - through a bank loan, or innovatively - through ABS.
There are two important points to note in the table: the funding cost and the
resulting ROE under the two different funding alternatives.
It is obvious that the asset-backed funding may be slightly more expensive than
Securitisation helps to improve the bank funding. On the other hand, it results in a much higher return on equity
return on equity than the bank funding. In this general example, the asset-backed funding achieves
a ROE of 34% versus 8.25% in the case of traditional funding sources, i.e. ROE
increases roughly four times using securitisation. This is a very important issue, in
the context of merger and acquisition activities, and growing shareholder
demands.

Refer to important disclosures at the end of this report. 36

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Potential Drawbacks of Securitisation

We looked at the incentives banks and companies have to securitise as well as the
related derived benefits. Securitisation also has some potential shortcomings and
negative aspects.
Retaining the First Loss Position
In many securitisations the banks or the companies retain the most junior piece in
the bond structure. As we discussed earlier, cash flows generated by the pool are
distributed from the top, from the senior tranche downwards, whereas the losses
are absorbed from the bottom, from the junior tranche upwards. A bank or a
company retaining the junior equity piece is in a first loss position and in one
small piece, retains the concentrated losses of the large pool of assets it has sold.
When an entity relies heavily on securitisation by retaining more equity tranches it
increases its effective leverage of the balance sheet. One of the best ways to avoid
that is to find a way to sell the equity piece to investors willing to assume the
associated risks.

Slide 20: Bank Securitisation

Why not to securitise? Or…how not to securitise?

❏ Company retains its first loss position

➢ increased effective leverage of the firm to the detriment of unsecured creditors
➢ by transferring assets off-balance sheet leverage declines even though no material
change has occurred in the capital structure and no material transfer of risk has
taken place
➢ company transfer of its best assets results in a deterioration of its balance sheet and
risk profile
❏ Company’s commitment to support securitisation
❏ Over-reliance on securitisation
❏ Asset risk transference can be achieved through other means - monoline wraps, synthetic
structures

Source: Merrill Lynch

Over-reliance on Securitisation
If a company heavily relies on securitisation, it will be determined to maintain
high levels of performance of the securitised assets at any cost. Performance of
any given asset pool securitisation bond below expectations would make it more
difficult for that company to access the securitisation market in the future. To
avoid that the originator may be willing to take some residual risk in the asset pool
or to step in to support the asset pool performance. In other words, the originator
may indirectly allow for partial recourse, which is not really the stated purpose of
securitisation and may provoke regulators’ objections.
Asset Risk Transfer Can Be Achieved Through Other Means
Finally, the transfer of risk associated with a given pool of assets could be
achieved without selling the assets, but simply by the transfer of that risk through
credit default swaps in credit structures, or insuring that risk with an insurance
policy for an outside insurer or guarantor. We are referring here to securitisation
methods – synthetic and insured structures, different from the traditional
securitisation based on asset sale.

Refer to important disclosures at the end of this report. 37

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

2.5. ABS/MBS Investors and Their Considerations

It is important to underline the broad range of investors in the securitisation
market and the dominant role of banks in that context. The benefits of
securitisation relative to other fixed income papers undoubtedly attract ever-
increasing number of ABS investors.

n Investors in Non-US ABS/MBS

Chart 21: Tentative ABS Investors Distribution by Country Chart 22: Tentative Investor Distribution by Investor Type
Scandinavia Rest of the World Sovereign /
3% Other Supranationals /
4% 8% Govt Agencies
UK 3%
Ireland 30% Banks
6% 43%
Italy
6% Asset Managers
20%

Iberia
4%
Germany / Austria
Benelux 18% Corporates SIV
14% 1% Pension Funds Insurance Co.s
France 10% 3% 12%
15%
Source: Merrill Lynch Source: Merrill Lynch

Chart 23: Tentative Corporate Bond Investors by Country Chart 24: Title Tentative Corporate Bond Investors Distribution
by Investor Type

Scandinavia Rest of the World Other

2% 2% Banks
5% UK 15%
18% Insurance Co.s
18%
Ireland Germany / Austria
1% 27%
Italy
12%
Pension Funds
6%
Iberia Asset Managers
10% 57%
Corporates
Benelux 2%
8% France SIV
17% 0%
Source: Merrill Lynch Source: Merrill Lynch

Slide 21-24 presents indicative investor distribution for non-US ABS and MBS in
Banks continue to be the 2002. Two features clearly stand out. Firstly, banks continue to be the biggest
biggest investor in ABS/MBS investors in asset and mortgage-backed securities. Nevertheless, the role of
investment advisers and commercial paper (CP) conduits has been growing,
representing 20% and 10% of the investor base for new issuance. We have to
stress the role of the CP conduits, as in CP conduits we include two different types
of entities: the asset-backed commercial paper conduits, which focus exclusively
on purchasing bond portfolios and funding them with CP, and the stand-alone
securities finance companies, the likes of Beta, Sigma, Centauri, K2, Links, and so
on, which also purchase substantial amounts of asset-backed securities.

The share of insurance Two other entities on this pie chart are worth mentioning, the insurance companies
and the pension funds. Their share of the purchases of new issuance of asset-
companies and the pension
backed securities has been gradually growing over the years, and it may also have
funds has been gradually influenced the structures of the securities issued. Traditionally, asset-backed
growing over the years
Refer to important disclosures at the end of this report. 38
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

securities in Europe are issued as floating rate pass-through notes. With the
demand of the insurance companies and pension funds for ABS/MBS increasing,
more of them are issued as fixed rate bullet notes, as these are the structures which
insurance companies and pension funds tend to buy, given their asset/liability
structure.
In terms of geographic distribution of new issuance of non-US asset and
mortgage-backed securities, the UK and Irish investors prevail and account for
UK and Irish investors prevail roughly 40% of the placements. The role of German, French and Italian investors,
with roughly 40% of the though, has been growing fast in recent years, while the Benelux countries already
placements have a group of well established sophisticated investors for this type of securities.
We have to emphasise the rapid development of the investor base since the
introduction of the Euro. As the currencies were eliminated and 11 countries in
Europe started functioning as one Eurozone market, investors who traditionally
focused on their domestic bonds due to currency considerations are now able to
purchase Euro denominated asset-backed securities from issuers domiciled in any
country in the Eurozone. In addition, with the rapid development of securitisation
within the Eurozone and the affirmation of the Euro as a reserve currency, there is
increasing demand for Eurozone ABS and MBS from non-European investors.

n Investor Considerations

Slide 25: Investor Considerations

❏ Attractiveness of ABS investments

➢ a diversified pool of consumer or corporate assets
➢ built-in structural and legal protections
➢ credit enhancement reflecting multiple performance scenarios
➢ limitations for sovereign risk exposure
❏ Diversification tool
❏ Defensive investment
❏ Higher risk-adjusted returns
❏ Rating Stability

Source: Merrill Lynch

Attractiveness of ABS Investments

Asset-backed securities are attractive to investors because they give them
exposure to a diversified pool of consumer or corporate assets. Instead of buying
loans or bonds in a number of industries or geographic areas and building up such
a portfolio alone, investors could simply buy into a CLO, a collateralised loan
obligation, which is already such a portfolio. They could also buy into a
diversified pool of consumer loans, be it through a credit card, mortgage or auto
loan-backed security.
In addition, each of the securities has built-in structural and legal protections we
discussed earlier, and are also structured to withstand levels of losses sized under
negative future scenarios. This certainly limits the risks, especially for the senior
investors in the asset-backed structure. Investors buying into asset-backed or
future flow securities from emerging markets are acquiring bonds with reduced
currency and sovereign risks.

Refer to important disclosures at the end of this report. 39

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Diversification Tool
Buying into an asset or mortgage-backed security is definitely a diversification
tool when managing a broad portfolio of fixed income instruments. In addition,
asset-backed securities generally price wider than comparable corporate bonds,
which gives them attractive higher risk adjusted returns.
Rating Stability
Historically asset-backed securities ratings have been demonstrated to be generally
stable. There is lower rating volatility in the asset-backed bonds world than in the
corporate or bank bonds world. This is to a large degree because these securities
already have built in credit and structural enhancements, as discussed earlier.

Refer to important disclosures at the end of this report. 40

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

3. Rating Stability of Structured

Finance Bonds

Refer to important disclosures at the end of this report. 41

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Rating agencies rating transition studies have long been a key investment
decision making tool. In this regard, every new study is anticipated with great
interest – even more so, in difficult years for the market. The catch, however, is
that even if the results presented by the studies are positive, investors, and
particularly those who have experienced problems, remain sceptical. And that is
natural – even when the averages are good, there are extremes with both
positive and negative signs, and the individual investor’s perception depends on
under which sign s/he happens to be.

3.1 Moody’s ABS Rating Transition Study 2002 –

Many Things to Write Home About!
This year’s Moody’s rating structured finance transition study ‘Structured Finance
Rating Transitions: 1983 – 2002’ published in January 2003 is an exception in
several respects. First, it is much more detailed and comprehensive than ever
before. Secondly, it brings much needed good news (contrary to the expectations,
we would say) for structured finance investors. Thirdly, it deepens the
comparative analysis between structured finance and corporate bond ratings, as
well as the comparisons within the structured finance sector itself. These are all
strong points! Where the study is a bit weak is on the interpretations of the
observed rating behaviors. To fully understand them, one has to sift through other
Moody’s reports – alas, they exist!
While we encourage everyone to read the study for oneself, we undertake to share
our read of the study. We follow the transition report as it unfolds and provide
some of the conclusions and our associated comments.
Given the volume of the study, we plan to provide a series of commentaries on its
different aspects. In this first instalment we focus on the study’s data pool and
more general rating transitions findings.
Let’s start with the study methodology and the data pool. Here one familiar with
the performance of the structured finance market to-date can detect the early signs
of the results to come. Let’s look at the data (see Tables 1 and 2):
• Structured finance ratings are predominantly – close to 90% - investment
grade, with predominance of Aaa‘s –about a third, while in corporate finance
slightly above half are investment grade – a quarter of them Baa, and more
than 40% are speculative grade.
- This reflects the nature of the two different sectors: structured finance
deals are ‘structured’ to achieve a certain rating – preferably the highest,
while corporate finance ratings reflect the creditworthiness of the
respective company as a going concern.
- It also raises questions about the validity of comparisons of the two
sectors, particularly in the sub-investment grade area – too few structured
finance ratings to derive viable comparisons with corporate high yield
ratings.
• The number of ratings differs substantially between structured finance and
corporate finance. The former are related to the number of securities
(approximation for tranches) and the latter – to number of the corporate
issuers. This, along with the growth of the structured finance market in recent
years, explains the significantly larger (by a factor of almost 5) rating
universe in structured finance compared to corporate finance. If we compare
the number of structured finance deals to the number of corporate issuers, the
differential is much smaller (5136 vs. 2950).
- A question remains unanswered here and that is – how are the ratings of
different categories of debt of a given corporate issuer treated? The
capital structured of a corporate issuer has numerous layers of debt, at a
minimum three – senior secured, senior unsecured and subordinated debt,
and each has a different rating level with a differential as high as 3 - 5
notches.
Refer to important disclosures at the end of this report. 42
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

- The answer to the above question above should be used in the

interpretation of the contemporaneous rating change dependency
discussed in the report with regards to structured finance, but not
analyzed for corporate finance – that is, the likelihood that if one tranche
of a securitization transaction is downgraded other tranches will be
downgraded as well. That should be true for the different rated
categories of debt for corporates, but it is unclear how it is reflected in the
study, if at all.
• The composition of the structured finance ratings across sectors has
dramatically changed over time – from highly weighted in favor of MBS to a
more balanced distribution across sectors.
- It is worth noting, however, that the overwhelming majority of structured
finance ratings are assigned to transactions based on pools of secured
loans (RMBS, HEL, CMBS).
- The composition also shows the much higher weight of the traditional
securitization (ABS, RMBS and one can argue CMBS) vs. the newer
categories (CDOs and ‘others’). One should relate to that the
considerably different structures and motivations underlying the
execution of the traditional and non-traditional structured financings –
food for thought when interpreting rating performance.
- Structured finance ratings include securities whose ratings can be
influenced by the ratings of two or more securities or issuers, even if one
plays a dominant role. In practical terms that means that ratings highly
dependent on corporate or sovereign entity are included in the study and
may to some degree distort the results – example, rating transitions of
tranches guaranteed by Green Tree/ Conseco have a serious impact on the
Baa rating transition levels; another example – structured ratings subject
to sovereign ceiling in emerging markets.

Table 1: Distribution of Structured Finance and Corporate Finance Ratings

Ratings Structured Finance Corporate Finance
1/1/85 1/1/90 1/1/95 1/1/00 1/1/02 1/1/85 1/1/90 1/1/95 1/1/00 1/1/02
Aaa 91.7% 36.4% 41.5% 35.4% 33.2% 3.5% 3.7% 2.5% 1.4% 1.5%
Aa 0.0% 58.2% 28.2% 16.9% 16.2% 17.4% 10.0% 10.1% 8.3% 10.3%
A 8.3% 1.8% 13.7% 20.0% 20.3% 31.2% 24.4% 27.6% 21.9% 22.4%
Baa 0.0% 2.9% 10.9% 16.2% 17.4% 17.1% 15.1% 18.6% 22.1% 24.3%
Ba 0.0% 0.5% 3.7% 6.2% 7.7% 18.0% 22.1% 17.0% 11.9% 11.5%
B 0.0% 0.0% 1.7% 3.7% 3.7% 9.8% 19.6% 18.8% 25.2% 19.6%
Caa-C 0.0% 0.2% 0.3% 1.6% 1.6% 3.0% 5.2% 5.4% 9.3% 10.4%
Investment 100% 99.4% 94.3% 88.6% 87.1% 69.2% 53.1% 58.8% 53.7% 58.5%
Grade
Speculative 0.0% 0.6% 5.7% 11.4% 12.9% 30.8% 46.9% 41.2% 46.3% 41.5%
Grade
Total ratings 12 649 3325 8201 12296 1585 2119 2200 3149 2950
outstanding
Source: Moody’s Investors Service

Refer to important disclosures at the end of this report. 43

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Table 2: Distribution of Structured Finance Ratings and Deals by Sector

Ratings Structured Finance Ratings Structured Finance Deals
1/1/85 1/1/90 1/1/95 1/1/00 1/1/02 1/1/85 1/1/90 1/1/95 1/1/00 1/1/02
ABS 8.3% 18.6% 23.9% 36.7% 35.6% 8.3% 18.3% 25.7% 38.6% 38.8%
CDO 0.0% 0.0% 2.0% 9.3% 14.5% 0.0% 0.0% 2.5% 8.0% 12.3%
CMBS 0.0% 1.4% 5.1% 13.4% 16.5% 0.0% 1.5% 3.0% 5.7% 6.9%
RMBS 91.7% 79.8% 68.1% 36.3% 29.4% 91.7% 80.1% 67.5% 39.1% 33.2%
OTHERS 0.0% 0.2% 0.9% 4.3% 4.0% 0.0% 0.2% 1.2% 8.6% 8.8%
Total 12 649 3325 8201 12296 12 617 2072 3788 5136
ratings/deals
outstanding
Source: Moody’s Investors Service

Now that we better understand the data, let’s look at the study’s results and its key
summary – the annual rating transition matrix, 1983 – 2002 (see Tables 3, 4 and
5). We particularly focus on the comparison between structured finance and
corporate finance rating transitions:
• Structured finance ratings are very stable with five of seven structured finance
(and only one of seven corporate) broad rating categories experiencing no
rating change in more than 90% of the cases.
- Structured finance ratings are more stable in both investment and sub-
investment grade categories.
- While in the sub-investment grade category structured finance ratings are
less likely than corporate ratings to be downgraded, they are nonetheless
less likely to be upgraded – this may be a reflection of the going concern
nature of a corporate entity and the liquidating nature of the assets
backing a structured finance rating.
• In addition, structured finance ratings are particularly less likely to be
downgraded than similarly rated corporates.
- Downgrade to upgrade ratio is much lower for structured finance than for
corporate ratings historically and remained lower in 2002 despite the
significant increases in downgrades experienced by both sectors.
- Structured finance ratings, however, appear less likely to be upgraded
than corporate ratings judging by the upgrade ratios. We believe this
conclusion while correct based on the numbers presented, is somewhat
misleading when considering existing market practices. It is a well-
known fact that as the senior tranches of a structured finance security
amortize the credit strength of the junior tranches improves all other
conditions being equal – however, this is rarely reflected in a rating
upgrade for a number of reasons.
• Structured finance securities appear more likely (almost twice) to be
downgraded to Caa and below category from all broad rating categories
(except single B) in comparison to corporate finance ratings. This is a
somewhat worrisome observation. A better review of the instances of
structured finance ratings moving into the Caa and below category is needed
in order to derive the reasons behind such transition results.
- A possible explanation is that once a significant deterioration takes place
in a securitized pool, a reversal is unlikely and the ratings are likely to
continue to slide to the lowest possible category.
- The above finding, on the other hand, may contradict some assertions that
structured finance ratings are stable because of intervention of originators
to support troubled transactions, a limited instances of which have been
observed in the past.

Refer to important disclosures at the end of this report. 44

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Table 3: Moody’s All Structured Finance Annual Rating Transition

Matrix, 1983-2002 (adjusted for Withdrawn Ratings)
Moody’s Structured Finance Rating Transitions 1983-2002
To
From Aaa Aa A Baa Ba B Caa or
below
Aaa 98.90% 0.89% 0.13% 0.04% 0.00% 0.00% 0.03%
Aa 5.45% 91.46% 2.28% 0.63% 0.09% 0.03% 0.06%
A 1.13% 2.74% 93.54% 1.82% 0.52% 0.07% 0.18%
Baa 0.53% 0.65% 2.25% 90.40% 3.83% 1.26% 1.08%
Ba 0.14% 0.06% 0.78% 3.99% 86.33% 3.24% 5.46%
B 0.00% 0.06% 0.06% 0.46% 0.85% 88.95% 9.62%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.17% 0.34% 99.49%
Source: Moody’s Investors Service

Table 4: Moody’s All Structured Finance Annual Rating Transition

Matrix, 1983-2002 (adjusted for Withdrawn Ratings)
Moody’s Corporate Finance Rating Transitions 1983-2002
To
From Aaa Aa A Baa Ba B Caa or
below
Aaa 89.93% 9.17% 1.00% 0.00% 0.00% 0.00% 0.00%
Aa 0.79% 89.66% 9.04% 0.37% 0.09% 0.02% 0.03%
A 0.05% 2.53% 90.68% 5.77% 0.70% 0.22% 0.04%
Baa 0.05% 0.28% 5.94% 86.95% 5.25% 1.12% 0.41%
Ba 0.01% 0.04% 0.61% 5.50% 82.59% 9.01% 2.23%
B 0.01% 0.06% 0.23% 0.61% 6.19% 81.22% 11.68%
Caa or below 0.00% 0.00% 0.00% 1.01% 2.57% 6.53% 89.88%
Source: Moody’s Investors Service

Table 5: Comparison of Aggregate Average Downgrade and Upgrade

Rates between Structured Finance and Corporate Finance, 1983-2002
(Broad-Rating-Based, Adjusted for Withdrawn Ratings)
Downgrade Upgrade Unchanged Downgrade/Upgrade
Rate Rate Rate ratio
Structured Finance, 1983-2002 3.21% 2.70% 94.09% 1.2
Corporate Finance, 1983-2002 9.42% 4.14% 86.44% 2.3
Structured Finance, 2002 only 6.46% 1.41% 92.13% 4.6
Corporate Finance, 2002 only 14.42% 2.61% 82.97% 5.5
Source: Moody’s Investors Service

In our weekly commentary of the same title from Feb 20, 2003 we drew attention
to the latest Moody’s structured finance rating transition study capturing the
period 1983-2002. Here we expand our commentary to include several additional
aspects of the study.
We emphasised Moody’s conclusion about the higher rating stability of structured
finance as compared to corporate ratings.
• Structured finance better rating stability is confirmed over a horizon longer
than one year. For example, triple-As retain their ratings in more than 95% of
the cases over even a five-year period.
Reviewing the multi-year transition matrices we note the following:
• The incidence of ratings withdrawn for investment grade structured credit
ratings is much higher for the higher investment grade ratings than for lower
investment grade and sub-investment grade by a factor of two or three. This is
the exact opposite to corporate ratings – they have a much higher incidence of

Refer to important disclosures at the end of this report. 45

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

withdrawn ratings at sub-investment grade level. This difference should be

distorting somewhat other conclusions made on the basis of rating transitions
for corporate and structured finance bonds.
• The longer the period, the higher the incidence of ratings withdrawn for the
lower rated classes in structured finance – this is logical given the
amortisation structure of many structured finance bonds and the longer
average life of the junior classes compared to the senior classes in a given
structure.
• The cumulative frequency of transition from investment grade into the lowest
rating category is higher for structured finance bonds, while the cumulative
frequency of transition from sub-investment grade to the lowest rating
category is higher for corporate bonds.
While these results may be considered intuitive by many, they should be viewed in
the light of the background data. It shows that investment grade ratings dominate
in structured finance and sub-investment grade ratings are significantly higher
percentage of all ratings in the corporate world as well as much higher percentage
compared to the structured finance world. In addition, the prevailing amortising
nature of structured finance bonds (although that depends on the asset class) as
compared to predominantly bullet corporate bonds must also have an impact on
the results. The much higher cumulative frequency of transition of sub-investment
grade corporate to the lowest rating category in comparison to structured finance
ratings refutes strongly the claims by some market participants that junior tranches
of structured financings are riskier than comparatively rated corporate bonds. The
data does not support such a claim, or, at worst, is inconclusive.
• Given the nature of a structured finance rating, i.e. it is a transaction specific
rating and remains outstanding for the life of the transaction, it is worth
tracking rating transition performance according to the outstanding age profile
of the ratings. Given the issuer specific nature of corporate ratings and the
going concern nature of the issuers, such ageing profile will be less
meaningful, except maybe in high yield world.
Structured finance ratings ageing seem to confirm what investors have long know
about the ageing, that is seasoning, of underlying pools – their credit quality may
deteriorate with time initially, then stabilise after a peak. So it seems to be true for
the ratings as well – the rating volatility increases up to about year four, then
decreases and stabilises. Such rating volatility, though, while having an
interesting pattern – both upgrades and downgrades climb up initially, with
downgrades exceeding upgrades, and then both climb down after year 3d – to - 5th
with upgrades exceeding downgrades.
This confirms that it is not only that the pool credit quality is age dependent, but
also that the rating transition frequencies may also be age-dependent, as Moody’s
says. The two may not be necessarily correlated, though, although some
relationship must exist. Such conclusions may argue in favour of maintaining a
weighed average seasoning for the deals in a portfolio in addition to the weighting
average seasoning for the underling pools.
• Rating drift is defined by Moody’s as the difference between the weighted
average upgrade frequency and the weighted average downgrade frequency,
weighted by notches per rating change and adjusted for withdrawn ratings.
Rating volatility, on the other hand, is the sum of the upgrade and downgrade
rates. While the first shows the prevailing direction of the rating changes, the
second indicates the relative volume of rating changes.
Historically, rating volatility was about 20% with no particular rating drift, that is
rating changes in both directions traditionally offset each other over time. The
exception is year 2002, which recorded a significantly higher negative rating drift
and rating volatility. It is certainly a question whether 2002 was an aberration
from historical averages or a beginning of a new trend. There are arguments in
favour of both views: aberration induced by the negative performance of the CDO
sector or a new trend driven by the credit deterioration in consumer credit in the
US!

Refer to important disclosures at the end of this report. 46

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

We could not find a similar data for the corporate bond world in order to derive
meaningful comparisons, maybe because the corporate world is more preoccupied
with defaults – that should be another point in favour of structured finance.
However, comparison of the average number of notches per rating move shows a
much more stable corporate sector (approximately 1.6) versus a more dynamic
structured finance sector (more than 2.6 notches in recent years and trending
upward).
This leads us to conclude that while structured finance ratings are more stable,
but once they start changing their rating changes tend to be much more dramatic
in either direction.
• The question arises then as to whether a given rating change is predicated on a
preceding rating change of the same security – Moody’s calls this feature a
rating change momentum or path dependency. Structured finance ratings
appear to have stronger downgrade and upgrade momentum than corporates –
this is to some extent confirmed by the conditional annual rating transition
matrices.
When talking about rating dependency, it is worth noting that when one tranche in
a structured finance transaction experiences a rating change, there is a high degree
of certainty that one or more tranches of the same transaction will experience a
rating change in the same direction. That is particularly notable in the sub-
investment grade tranches. This confirms what investors have always know
intuitively that there is a common denominator to the ratings of all tranches in a
given structured financing – that of the credit performance of the underlying pool,
among others.

Table 6: Rating Change Momentum 1983-2002

Structured Finance Ratings
Downgraded in t Upgraded in t Unchanged in t
Downgraded in t-1 36.36% 3.55% 60.10%
Upgraded in t-1 3.39% 9.97% 86.65%
No change in t-1 3.52% 4.32% 92.16%
Unconditional 4.15% 3.71% 92.14%
Corporate Finance Ratings
Downgraded in t Upgraded in t Unchanged in t
Downgraded in t-1 25.78% 7.11% 67.11%
Upgraded in t-1 6.68% 16.15% 77.17%
No change in t-1 14.01% 8.53% 77.46%
Unconditional 14.70% 8.84% 76.46%
Source: Moody’s Investors Service
Note: t is the current year and t-1 is the previous year, and downgrades include transitions iinto default
(Adjusted for Withdrawn ratings, Refined-Rating Based, Unadjusted for the Number of Notches Per Rating Move)

Refer to important disclosures at the end of this report. 47

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Table 7: Conditional Annual Rating Transition Matrices in Structured Finance,

1983-2002
Structured Finance Annual Rating Transition Matrix, Conditional on previously downgraded
Aaa Aa A Baa Ba B Caa or below
Aaa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%
Aa 0.36% 83.27% 10.32% 5.34% 0.71% 0.00% 0.00%
A 0.30% 0.30% 80.91% 13.33% 3.03% 0.00% 2.12%
Baa 0.48% 0.00% 0.48% 62.32% 22.22% 5.80% 8.70%
Ba 0.00% 0.00% 0.00% 15.96% 52.84% 13.83% 17.38%
B 0.00% 0.00% 0.00% 0.00% 0.00% 44.06% 55.94%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 100.00%
Structured Finance Annual Rating Transition Matrix, Conditional on previously upgraded
Aaa Aa A Baa Ba B Caa or below
Aaa 100.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%
Aa 13.44% 86.32% 0.24% 0.00% 0.00% 0.00% 0.00%
A 2.14% 8.93% 88.21% 0.71% 0.00% 0.00% 0.00%
Baa 1.12% 5.03% 7.82% 59.78% 26.26% 0.00% 0.00%
Ba 0.00% 0.00% 3.03% 21.21% 75.76% 0.00% 0.00%
B 0.00% 0.00% 0.00% 0.00% 22.22% 77.78% 0.00%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%
Source: Moody’s Investors Service

Question arises, however, about the corporate ratings, given that in the corporate
world there are both issuer and issue specific ratings, and the range of issue
specific ratings are dependent on the issuer rating level and it changes, read
notching. In other words as corporate rating transitions track only issuer specific
ratings, the magnitude of rating change and particularly of contemporaneous rating
changes is not reflected in the corporate rating transition study, while it affects,
apparently quite severely, the structured finance rating transition study.
Finally, as rating changes due to the ratings of Conseco/ Green Tree were reflected
in the structured finance study, we believe this fact distorts the rating transition
conclusions made with regards to Baa and Ba – both in terms of general and
conditional (upgrade conditional on a downgrade) rating transitions.

Table 8: Contemporaneous Rating Change Dependency across Tranches in the

Same Deal, 1983-2002
Downgraded Upgraded Unchanged
Conditional on Another Tranche Being Downgraded 70.53% 0.60% 28.87%
Conditional on Another Tranche Being Upgraded 0.57% 66.62% 32.81%
Conditional on Another Tranche Sustaining No Rating Change 1.69% 2.29% 96.02%
Unconditional 4.15% 3.71% 92.14%
Source: Moody’s Investors Service
(Adjusted for Withdrawn Ratings, Refined-Rating Based, Unadjusted for the Number of Notches Per Rating Move)

• Finally, while the conclusions from the rating transition study are generally in
favour of structured finance ratings in comparison to the corporate ratings, the
structured finance world is not homogeneous and comprises several distinctly
different sub-sectors. While they demonstrate similar rating stability levels
(excluding, naturally CDOs), they are distinctly different in terms of their
downgrade and upgrade frequencies and rating withdrawn frequency.
The tables below demonstrate the significant differences in terms of rating
stability and transition among the four major sub-sectors of the structured finance
market.

Refer to important disclosures at the end of this report. 48

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Table 9: Weighted Average Annual Transition Matrices in the

Structured Finance Sectors, 1991-2002
ABS
Aaa Aa A Baa Ba B Caa or below
Aaa 99.05% 0.82% 0.04% 0.01% 0.00% 0.00% 0.08%
Aa 2.57% 94.62% 1.82% 0.67% 0.08% 0.00% 0.24%
A 0.63% 1.07% 96.34% 1.15% 0.63% 0.08% 0.10%
Baa 0.59% 0.46% 0.97% 89.59% 6.47% 1.22% 0.71%
Ba 0.28% 0.14% 0.42% 7.45% 74.40% 5.63% 11.67%
B 0.00% 0.00% 0.00% 0.60% 0.00% 76.65% 22.75%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 100.00%
CDO
Aaa Aa A Baa Ba B Caa or below
Aaa 95.37% 3.09% 0.93% 0.51% 0.10% 0.00% 0.00%
Aa 0.58% 89.04% 5.48% 3.50% 1.05% 0.23% 0.12%
A 0.16% 0.98% 89.59% 6.02% 1.95% 0.33% 0.98%
Baa 0.00% 0.09% 0.51% 86.19% 6.17% 3.69% 3.34%
Ba 0.00% 0.00% 0.00% 1.37% 81.48% 5.83% 11.32%
B 0.00% 0.00% 0.00% 0.00% 0.00% 68.98% 31.02%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 100.00%
CMBS
Aaa Aa A Baa Ba B Caa or below
Aaa 98.49% 1.51% 0.00% 0.00% 0.00% 0.00% 0.00%
Aa 5.58% 93.35% 0.63% 0.18% 0.00% 0.18% 0.09%
A 1.45% 3.23% 93.87% 1.45% 0.00% 0.00% 0.00%
Baa 0.62% 1.24% 3.03% 93.18% 1.45% 0.28% 0.21%
Ba 0.00% 0.00% 0.59% 2.52% 94.67% 1.78% 0.44%
B 0.00% 0.00% 0.16% 0.65% 1.79% 94.95% 2.44%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 2.70% 97.30%
RMBS
Aaa Aa A Baa Ba B Caa or below
Aaa 99.34% 0.49% 0.14% 0.03% 0.00% 0.00% 0.00%
Aa 8.03% 89.85% 1.70% 0.43% 0.00% 0.00% 0.00%
A 2.07% 4.28% 91.06% 2.15% 0.21% 0.03% 0.21%
Baa 0.67% 0.76% 3.37% 91.58% 1.88% 0.85% 0.88%
Ba 0.21% 0.07% 1.41% 4.08% 90.01% 1.76% 2.46%
B 0.00% 0.14% 0.00% 0.42% 0.56% 92.31% 6.57%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.28% 0.00% 99.72%
Source: Moody’s Investors Service

As becomes clear from the comparison among sub-sectors, CDO has the worst
track record in structured finance world. While this is true, investors should
consider the CDO rating transition study for better understanding of the CDO
sector performance – as we have stated on many occasions this is not a
homogeneous sector and it owes its poor name primarily to US high yield
arbitrage CDOs.
In addition, the different type of structures predominant in the CDO and other
sectors of the structured finance market, along with the different levels of stress
experienced by the underlying should be taken into consideration when
interpreting the data. We note the strong performance of the real estate based
securitisations, RMBS and CMBS sectors – there is something to be said about the
strength of the collateral! While the CDO sector tracked the performance shadows
the performance of the corporate sector, RMBS and CMBS mirror the real estate
one. However, we again emphasise the importance of the deal structures in the
respective sectors to counter or connive with the collateral performance.
Considering the performance of the sub-investment grade area, again there is
evidence of significant difference among sectors – CDOs and ABS being much
more volatile than CMBS and RMBS. Furthermore, the frequency of emerging
from the Caa and below is very low in the structured finance world – and much
higher in the corporate world – there seems to be little cure for terminal illness!
Refer to important disclosures at the end of this report. 49
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Comparison of rating transitions between the US and international structured

finance sectors yield remarkably similar results. A closer look at the numbers
indicates a marginally higher downgrade frequency for international
structured finance, which Moody’s attributes to the higher share of the CDO
sector in international structured finance compared to US structured finance.

Table 10: International Structured Finance Annual Rating Transition

Matrix, 1989-2002
Caa or
Aaa Aa A Baa Ba B below
Aaa 97.43% 2.18% 0.32% 0.07% 0.00% 0.00% 0.00%
Aa 1.48% 92.65% 4.18% 1.28% 0.27% 0.07% 0.07%
A 0.55% 4.79% 90.96% 2.81% 0.68% 0.00% 0.21%
Baa 0.00% 0.24% 1.71% 88.20% 5.13% 2.44% 2.28%
Ba 0.00% 0.00% 0.20% 2.00% 85.60% 4.40% 7.80%
B 0.00% 0.00% 0.00% 0.00% 0.00% 77.51% 22.49%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 100.00%
Source: Moody’s Investors Service

Table 11: U.S. Structured Finance Annual Rating Transition Matrix,

1989-2002
Caa or
Aaa Aa A Baa Ba B below
Aaa 99.12% 0.70% 0.10% 0.04% 0.01% 0.00% 0.03%
Aa 6.00% 91.26% 2.04% 0.55% 0.06% 0.03% 0.06%
A 1.23% 2.41% 93.96% 1.65% 0.49% 0.08% 0.18%
Baa 0.62% 0.72% 2.35% 90.77% 3.61% 1.06% 0.87%
Ba 0.17% 0.07% 0.88% 4.32% 86.46% 3.04% 5.07%
B 0.00% 0.06% 0.06% 0.50% 0.95% 90.22% 8.20%
Caa or below 0.00% 0.00% 0.00% 0.00% 0.18% 0.37% 99.45%
Source: Moody’s Investors Service

3.2 Moody’s CDO Rating Migration Study 1996-2002:

More Doom than Gloom!
The long-awaited Moody’s update on the performance of the CDO sector globally
is finally out (see Credit Migration of CDO Notes, 1996-2002, for US and
European Transactions, April 15, 2003).
Although it is a mandatory read for any investor, it is hardly telling the
experienced CDO veteran anything new. It may further enhance the position – be
it negative or positive – regarding CDOs, a specific investor has adopted over
time.
But, maybe it is a time for another look! Why so? To answer this question, we
will take a look at some of the data and conclusions of the study:
First, the credit trends of the various sectors of the CDO market exhibited
during 2001 continued in 2002. That is, some sectors continued to suffer, others
continued to be stable or mildly affected. However, these trends highlight again
the existence of significant differentiation and related diversification possibilities
within the sector.
Here comes a market reality, which has so far been largely disregarded by
European investors: the CDO market has various sectors and they have performed
credit-wise differently, often significantly differently, from one another:
• Moody’s counts 12 such sectors. An additional one can be added – B/S Cash
Flow non-US$. Furthermore, one of the sectors is obviously a composite of
several more – resecuritisations may include CDO of CDO, CDO of ABS,
CDO of Real Estate, etc.
Refer to important disclosures at the end of this report. 50
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Moody’s has increased over time the number of CDO sectors – from 4 in the
2000 study, to 10 in 2001 to 12 at present – this reflects the market
development and data availability, and is akin in some respects to adding new
industries to the corporate sector.
As the data shows arbitrage cash flow CBOs have been consistently the worst
performers in the CDO world in terms of downgrades – say, half of all
downgrades, followed well behind by the arbitrage synthetic US$ and non-US$
transactions, and even further behind, by Balance Sheet Synthetic US$ and non-
US$ transactions. It is interesting, that while arbitrage cash flow CBOs have
performed fairly poorly across the board, the other two transaction types have had
a distinctly different performance: arbitrage synthetic non-US$ deals performed
worse than the US$ ones, while balance sheet synthetic non-US$ deals performed
better than the US$ ones. Overall, more than half of the CDO types have had a
fairly good performance, all considered and particularly taking into account the
most stressful conditions the corporate and other fixed income markets
experienced in the last several years.

Table 12: Moody’s CDO Downgrade by Sector for US and European Market (Tranches)
(% of total CDO downgrades from January 1, 2002 to December 31, 2002 (Row 2002)
and from January 1, 2001 to December 1, 2001 (Row 2001))
Deal Type/ Vintage 1995 1996 1997 1998 1999 2000 2001 2002 Total
ACF CBO 2001 3.45 14.94 14.37 13.22 1.15 47.13
2002 4.0 6.8 10.8 23.6 8.5 0.8 54.6
Arb Synth non-US$ 2001 0.57 2.87 10.34 13.79
2002 0.2 3.6 8.9 12.7
Arb Synth US$ 2001 1.15 3.45 4.6
2002 0.4 0.8 5.5 0.4 7.2
B/S Synthetic US$ 2001 1.15 10.92 2.3 4.6 0.57 19.54
2002 0.8 2.1 0.4 2.1 1.3 6.8
ACF CLO 2001 1.15 0.57 4.02 5.75
2002 1.1 1.5 2.8 0.4 5.7
ACF IG CBO US$ 2001
2002 0.2 0.2 0.6 3.6 4.7
B/S Synthetic non-US$ 2001 1.15 5.75 6.9
2002 0.2 2.5 1.7 4.5
Resecuritisations 2001
2002 0.2 0.4 0.6 1.3
Market Value 2001
2002 1.1 1.1
Emerging Markets 2001 1.15 0.57 1.72
2002 0.4 0.2 0.2 0.8
B/S Cash Flow US$ 2001 0.57 0.57
2002 0.4 0.4
ACF IG CBO non-US$ 2001
2002 0.2 0.2
Totals 2001 4.6 17.8 29.9 19 17.8 10.9 174
2002 0.4 5.3 9.1 16.1 26.8 19.1 22.7 0.4 471

Note: CDOP Downgrades by vintage, by tranches as % of tal CDO Downgrades by tranche

Source: Moody’s, Credit Migration of CDO Notes, 1996-2002, for US and European Transactions, April, 15, 2003 and
Credit Migration of CDO Notes, 1996-2001, February 27, 2002

A review of this table coupled with the experienced accumulated on the CDO
market would suggest that:
• It is incorrect to talk about the CDO market performance in general, as much
as it is incorrect to talk about the corporate market performance in general.
• Different credit performance may be suggesting that there is a potential
diversification benefit from investing in different CDO types. This is not,
however, the same as saying that the underlying CDO exposures present

Refer to important disclosures at the end of this report. 51

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

diversification opportunities from the perspective of a corporate bond investor

– they may (as in the case of emerging markets or structured finance
instruments) or they may not (as in the case of corporate bonds and loans).
• There is also a timing aspect of investing in CDOs as performance by vintage
indicates, but there may well be a seasoning effect at least in some of the sub-
sectors.
• CLOs perform better than CBOs, primarily related to the inherent differences
between loans and bonds as it affects arbitrage transactions and the different
level of lender’s involvement in each case.
Second, comparing corporate and CDO rating transitions could be quite
challenging. Data for 2002 indicates wide differences by credit category and by
CDO type. It also confirms some of the experiences of 2001 and refutes others.
Yet, on an ‘apples with apples’ basis the performance of CDOs and corporates
looks remarkably similar, as one would expect them to.

Table 13: Corporate vs. CDO Transitions in 2002 and 2001, Probability for Downgrade
for Selected Credit Ratings (adjusted for withdrawn ratings)
Synth Arb Synth Arb
2002 ACF-CBO ACF-CLO (US$) (Non US$) Em. Mkts All CDOs Corporates
AAA 19.4 1.0 31.8 55.0 0.0 10.7 8.2
Aa2 47.4 12.5 80.0 100.0 12.5 23.9 14.7
Baa2 44.3 4.3 66.7 66.7 16.7 26.4 24.1
Baa3 55.6 30.8 nm 85.7 12.5 35.9 21.6
Ba2 59.6 5.6 71.4 nm nm 30.6 33.6
Ba3 46.2 3.3 nm nm nm 26.1 27.7

Synth B/S Synth B/S

2001 ACF-CBO ACF-CLO (US$) (Non US$) Corporates
AAA 1.7 0.0 10.0 0.0 1.0
Aa2 25.6 8.0 0.0 7.7 16.3
Baa2 12.7 2.8 33.3 11.1 14.8
Baa3 34.8 7.7 nm nm 10.8
Ba2 12.1 15.4 50.0 30.0 23.4
Ba3 14.7 9.1 nm nm 18.2
Source: Moody’s, Credit Migration of CDO Notes, 1996-2002, for US and European Transactions, April, 15, 2003 and
Credit Migration of CDO Notes, 1996-2001, February 27, 2002
It is worth noting that:
• CDOs on average have performed more or less in line with corporates,
although the differences among CDO types relative to corporates are clearly
quite significant. We believe that such significant differences will exist
among the different industries within the corporate sector, as well. We still
wonder, though, why the former difference is viewed so negatively by the
market participants, while the latter is widely accepted as normal.
To expand on this point: CDOs on average have performed more or less in line
with corporates, although the deviation from the corporate performance may differ
significantly by CDO type. Similarly, in the corporate world alone – different
industries perform differently and have different risks. So deviations from the
averages can be also significant for different industries. This is widely accepted as
normal, and necessitates diversification and a view on a particular industry. The
fact that different CDO asset classes perform differently and have different risk is
not, in contrast, viewed by the market as an opportunity to diversify and take a
view on different CDO types, but, perversely, used as an excuse to stay out of the
sector as a whole.
• From downgrade point of view, investment grade rating categories for CDOs
have exhibited higher downgrade risk than similarly rated corporates, while
sub-investment grade CDO tranches have demonstrated lower downgrade risk
than similarly rated high yield corporates. Such observations go against the

Refer to important disclosures at the end of this report. 52

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

widely accepted views, which have ultimately found their way into investor
and regulator stance towards CDOs.
• These numbers are actual and are not adjusted for the significantly different
numbers of securities in each rating category for corporates and CDOs, that’s
there is sample size error inevitably distorting the results. Said differently, the
comparison is between ‘apples and oranges’ – you choose which is which.
It is a pity that Moody’s has not presented a similar analysis of the different
corporate sector – we are fairly certain that different corporate sectors downgrade
performance will look significantly different among themselves and against the
composite over a given period of time.
While the actual downgrade experience in given years is enlightening, it is longer-
term averages on which investment portfolios are structured. In that regard
Moody’s creates a weighted-average one-year transition matrix for CDOs and
theoretical one-year transition rate for Corporates (weighting each year’s one-year
corporate transition rates by the number of CDO ratings outstanding at each rating
level as of the beginning of that year) to allow for proper comparison. On such
basis the comparison is between ‘apples and apples’, and it looks increasingly
similar, as most apples do, except for those belonging to top brands.
Table 14: Average One-Year Downgrade Risk.
Comparison of 1-Year Average Rating Transition, 1996-2002, Probability of
Downgrade*
1yr Av ACF-CBO ACF-CLO Synth B/S (US$)** Em. Mkts All CDOs Corporates***
AAA 7.0 0.4 8.8 0.0 5.2 6.0
Aa2 16.3 6.0 6.7 9.3 13.4 13.5
Baa2 20.5 3.6 26.7 3.7 14.3 17.1
Baa3 21.8 10.7 37.5 17.2 17.5 14.3
Ba2 36.6 10.5 31.3 9.1 20.4 26.2
Ba3 19.8 5.8 50.0 14.3 16.8 24.1
*Probability of downgrade is adjusted for withdrawn ratings
**For the period 1998-2002
***Figures are from ’theoretical’ corporate transition matrix, developed by Moody’s
Source: Moody’s, Credit Migration of CDO Notes, 1996-2002, for US and European Transactions, April, 15, 2003
Third, near-term CDO sector performance trends, observed during the last
several years, look set to continue. That means that sub-sectors that under-
performed in the past should continue to do so, maybe to a lesser extent due to
deal de-leveraging and hopefully some improvement in corporate credit, while the
CDO sectors, which performed reasonably well in the past may continue to do so
with some exceptions, which may see slight deterioration.
An indicator of future performance could be the CDO watchlist. Below we
compare Moody’s CDO watchlists in early 2002 and early 2003. This watchlist
from early 2002 can be compared with the performance in 2002 shown earlier in
this report to determine its predictive power. The comparison seems to indicate a
fairly strong predictive power.
In the early 2002 watchlist the sectors ranked from worst-to-best in the following
order: ACF CBO, ACF CLO, B/S Synthetic USD, Arbitrage Synthetic USD, etc.
The actual performance in 2002 ranks the sectors from worst-to-best in the
following order: ACF CBO, Arbitrage Synthetic non-USD, Arbitrage Synthetic
US$, B/S synthetic US$, ACF CLO.
Judging by the watchlist from early 2003, the 2003 ranking of CDO sector by
performance in worst-to-best order could be as follows: ACF CBO, ACF IG CBO,
ACF CLO, Arbitrage Synthetic US$, Resecuritisation, etc. And again ACF CBO
are well ahead of the pack – a watchlist ratio of 4 to 1, at least – and will continue
casting a long shadow on the CDO market.

Refer to important disclosures at the end of this report. 53

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Table 15: Moody’s CDO Downgrade Watchlist by Sector for US Market and European
Market As of March 18, 2003 for Row 2003 and as of February 8, 2002 for Row 2002
Deal Type/ Vintage 1995 1996 1997 1998 1999 2000 2001 2002 Total
ACF CBO 2002 4 12 20 36 8 80
2003 2 8 27 37 16 90
Arb Synth non-US$ 2002 2 2
2003 4 2 6
Arb Synth US$ 2002 7 7
2003 1 17 1 19
B/S Synthetic US$ 2002 6 2 8
2003 1 4 5
ACF CLO 2002 2 6 6 14
2003 1 2 6 9 2 20
ACF IG CBO US$ 2002
2003 2 13 6 21
B/S Synthetic non-US$ 2002
2003 1 1 2
Resecuritisations 2002
2003 10 5 15
Market Value 2002
2003 4 2 6
Emerging Markets 2002 2 1 3
2003
B/S Cash Flow non-US$ 2002
2003 3 3
B/S Cash Flow US$ 2002 2 2
2003 3
ACF IG CBO non-US$ 2002
2003
Totals 2002 2 6 19 32 38 8 11 116
2003 1 4 14 44 74 46 4 187
Source: Moody’s, Credit Migration of CDO Notes, 1996-2002, for US and European Transactions, April, 15, 2003 and
Credit Migration of CDO Notes, 1996-2001, February 27, 2002

In conclusion, CDO market performance apparently owes its bad name to the
arbitrage cash flow CBO sector, which in fact represents 36.56% of all CDO
tranches rated by Moody’s. It is unfortunate that one third, the bad third, of the
market has such a strong negative bearing on the rest of it in the eyes of European
investors. It may be time to re-examine the current attitude and adopt a stance
much more relevant to the market realities.

Refer to important disclosures at the end of this report. 54

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

4. Assessing Subordinated
Tranches in ABS Capital Structure

Refer to important disclosures at the end of this report. 55

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

As subordinated tranches continue to increase in availability, investors need a

framework for their analysis. Regardless of whether it is a true sale or
synthetic transaction, subordinated tranche investments require a more
intensive initial analysis and more active monitoring. The lower investors
move in the ABS capital structure, the more intensive the credit analysis
should become. Analytical methods such as sensitivity and break even
analysis come handy when reviewing payment rates, charge-offs and excess
spread.

4.1 Framework for Subordinated Tranche Analysis

The analysis of the subordinated tranches of unsecured consumer assets ABS
follows the general principles and guidelines of the investment analysis for ABS
Analysis of senior and and their senior tranches. It is or should be, however, more credit intensive - the
subordinated tranches share a lower in the ABS capital structure the investors move, the more intensive the
common basis . . . credit considerations should become.
The framework for credit analysis of subordinated tranches of unsecured
consumer loans and credit cards in most generic terms include the following
aspects:
• Quality of initial collateral pool and originator’s historical pool performance;
• Originator and servicer capabilities and credit standing;
• Structural enhancements (credit enhancement, liquidity facility, reserve
accounts, reserve account build-up triggers, early amortisation triggers, etc.);
• Legal enhancement (bankruptcy remoteness of the structure, ‘true’ sale
treatment); and
• Cash flow ‘waterfall’ and priorities of payment of principal and interest at
different subordination levels.
• The revolving (substitution) feature of these deals entails changes in the pool
composition within the respective eligibility criteria, hence the need to
ascertain:
• Tightness of the eligibility criteria.
• Consistency in the underwriting criteria of the originator and ability to
manage accounts in accordance with the changing economic and competitive
environment.
• Frequency of replenishment and account additions.
• Tightness of economic triggers.
• Ability to build up additional credit support (reserve accounts and excess
spread).
• Likelihood of the occurrence of trigger events.
. . . subordinated tranches, All of the above aspects are relevant to the analysis of both senior and
however, introduce unique subordinated ABS tranches. With regards to the analysis of the subordinated
analytical aspects tranches several additional considerations should be included:
• Excess spread stability or erosion, and, particularly, excess spread variability
and determining factors (yield, delinquency and loss management).
• Originator’s pool management capabilities - ability to manage yield,
delinquencies and losses.
• Potential levels of excess spread build-up in an additional reserve account
in accordance with the respective triggers.
• Applicability of the reserve accounts.
• Current, historical and expected losses compared to actual and potential
credit enhancement levels.

Refer to important disclosures at the end of this report. 56

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Likelihood of servicer bankruptcy/replacement and its effects on pool

performance.
• Effects of payment rates on the accumulation of losses and excess spread
variability.
• Break-even and sensitivity analysis, etc.
These aspects are more important to subordinated tranches investors than to the
most senior tranches investors given the lower loss cushion protection they have
and the higher likelihood of negative consequences of performance volatility. We
would discuss those particular aspects in more detail below.

n Excess Spread Stability

Excess spread is the first line of The main source of loss protection for an investor in the lower tranches of the
defence against losses ABS capital structure is the excess spread. That is, the difference between the
revenue from the asset pool (pool yield, which depends on interest rate charged on
the underlying accounts, as well as other charges – late payment penalty, annual
fees, interchange, etc.) and the expenses under the securitisation (ABS coupon,
servicing fees, losses, etc.). Hence, excess spread is a function mainly of the pool
yield and pool losses, which both depend on the ability of the originator/servicer
to manage the asset pool under different economic scenarios, see below.
A key aspect of the analysis of excess spread is its variability. We are concerned
with excess spread variability for a number of reasons:
• It determines the level of the available cushion against loss increases;
• It determines the extent, to which a reserve account can be built up in case of
need; and
• It determines the likelihood of the occurrence of an early amortisation event,
i.e. hitting the economic trigger associated with specified level of excess
spread typically on a three-month rolling average basis.
Volatility in excess spread can A high volatility of excess spread or its rapid drop may prompt early amortisation
prove negative without sufficient time to build up additional reserves against future loss increases
and depletion of the available protection to the subordinated tranches.
Furthermore, highly volatile excess spread, and especially a steep rapid drop in the
levels of excess spread would not allow sufficient build-up of the reserve account,
i.e. would limit the available additional cushion against loss increases.

n Originator’s Underwriting and Servicing

As mentioned above the unsecured consumer loan and credit card deals are
usually structured with a revolving period, during which the principal collections
from the existing pool are used to purchase new receivables from the original
accounts or new accounts altogether. In the case of a master trust, the originator
can add new accounts and the associated receivables to the pool on an ongoing
basis, which could alter the pool composition and performance over the term of
the transaction. Any performance variability would first of all affect the
subordinated tranches.
Servicing is crucial to the Originator’s ability to manage the pool is crucial and depends on a number of
subordinated tranches’ factors:
performance • Type and purpose of the unsecured loan or credit card, e.g. general credit card
or retail credit card, instalment consumer loans or revolving unsecured loan
facility. For example, one would expect a lower yield and higher first bucket
delinquencies for retail cards as compared to general credit cards.
• Market competition has effects on the way loans are originated and the rates
charged, e.g. teaser rates, lower interest rates, annual fees, etc.
• Ability to enforce risk-based pricing, thus reflecting the riskiness of the pool.

Refer to important disclosures at the end of this report. 57

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Balance between ‘convenience users’ (borrowers who pay their balances in

full every payment date) and ‘credit users’ (real users of the credit facility and
payers of interest rate charges) – the effect on pool yield is obvious.
• Ability to retain good quality accounts, especially among the ‘credit users’.

n Likelihood of Servicer Bankruptcy/Replacement

Though remote, the ability to Servicer bankruptcy is not only a trigger event in revolving structures but also
replace a servicer should be could lead to a significant deterioration of the collateral pool performance at least
considered on a temporary basis, e.g. an increase in delinquencies and losses, which affects
the lower tranches in the ABS structure first.
Overall, the ability to replace a servicer of a credit card pool in the US, for
example, is fairly high without much delay, while that may not be the case in
Europe. This likelihood, though, should be evaluated against the current
creditworthiness of the servicer (usually also the originator) – in Europe all the
credit card issuers for example are investment grade banks.
In both cases above, originator’s ability to manage the pool and a servicer
bankruptcy would effect the ability to build-up an additional reserve account
among other things.

n Applicability of the Reserve Accounts

A key issue for subordinated investors to understand is the applicability of the
reserve account, i.e. whether it is available to the senior investors or is available
only to the subordinated investor, be it the mezzanine or the equity one. It is fairly
obvious that a reserve account ‘reserved’ for the sub investor reduces the amount
of losses it can experience. On the other hand, such an account may be providing
a false support for the senior investor.

n Losses and Credit Enhancement

Investors need to evaluate their In order to determine the level of protection provided by the credit enhancement,
loss expectations against investors should compare the current, expected and historical portfolio losses to
available enhancement the available credit enhancement at their position in the capital structure.
In this regard, we would like to stress several points:
• When sizing the credit enhancement the rating agencies generally assume
zero or negative excess spread.
• In case of amortising assets, the seasoning of the pool with reference to the
relevant static loss pool curve becomes crucial in determining the expected
loss developments. That should also be valid for revolving pools which allow
addition of new receivables but not of new accounts.
• Pool losses should be adjusted for pool growth. Rapid pool growth could
mask loss increases because newly added accounts have initially low losses
and delinquencies. In case of an early amortisation, however, no new
additions are allowed and then the losses of the underlying pool should
increase and test the sufficiency of the available credit enhancement.

n Effects of Payment Rates

One of the key variables of the pool performance is the payment rate. Payment
rate has potentially different effects on investors:
• On the one hand, it determines how fast the pool pays down, i.e. how quickly
investors are out of their positions.
• On the other, it determines how quickly the pool quality deteriorates, as the
better borrowers tend to pay first.
The combination of the two determines the level of losses that the pool could
accumulate – the lower the payment rate, the higher the accumulated losses, the
higher the investor exposure.

Refer to important disclosures at the end of this report. 58

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Waterfall Priorities
ABS capital structure determines the prioritisation of cash flows among different
tranches. The senior tranche (usually rated AAA) is the one, which usually
receives its payments of interest and principal first – that assures that any cash
collections first service the senior tranche. As for the subordinated tranches, one
should understand how the interest and principal cash payments are distributed,
for example:
• Whether interest is paid pari passu or sequentially for the senior and
mezzanine tranche;
• Whether principal is paid pro rata or sequentially for the senior and
subordinated tranches;
• Whether interest payments for the junior tranches can be subordinated to
principal payments for the senior tranches; and
• Whether there is some kind of a trigger, which changes the priority of interest
or principal distribution during the life of the transaction.
As a rule of thumb, with the risk of stating the obvious, it is best for a
subordinated tranche to receive its interest on a pari passu basis and its principal
on a pro rata basis.

n Likelihood of an Occurrence of a Trigger Event

The trigger events (early amortisation events) in unsecured consumer loan and
credit card ABS are associated with two types of negative events – one effecting
the originator and/or servicer and their ability to perform their obligations, and the
others associated with the quality of the asset pool.
The potential negative impact for the occurrence of an early amortisation event is
more likely to be greater at subordinated tranche level than on senior tranche level.
Hence, it is more important for the subordinated tranche investor to evaluate the
likelihood of such an event taking place. Their occurrence will lead to changes in
the average life of the ABS notes. Depending on the nature of the event, it could
have the consequences of increasing volatility in pool performance and rapid
reduction in excess spread (reduced cushion against losses and inability to build
additional reserves).

n Specific Features of Master Trust Structures

• When evaluating the credit quality of the different subordinated tranches from
series issued by a master trust, several additional aspects come into
consideration.
Utilisation of Excess Spread
As discussed in the appendix all master trusts are not the same. They differ in the
ways cash flows are allocated and distributed among the numerous series of notes
outstanding. In particular, investors should be aware of the ability to allocate
excess spread from series with a surplus to others that face shortfalls in their
required cash flows.
In that respect a proper question to ask is how the performance of the outstanding
Investors should evaluate series of the trust affect the performance of the particular series held by the
structural mechanics that could investor? The different series in a trust are exposed to the same level of yield and
result in the redirection of same level of losses and servicing fee expenses, but to different levels of coupon
excess spread payments, i.e., interest expenses and other expenses (swaps, etc.); as a result the
servicing shortfall/excess may differ among series. In some cases, the shortfalls
(when they are very large) may be shared by all series, in other cases they may be
limited to the respective series facing a shortfall. Yet in other cases, excess spread
from one series may be re-directed to series facing shortfalls.

Refer to important disclosures at the end of this report. 59

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Spread traps can be unique to a In addition, the series may have certain specific excess spread capture
series, reducing the amount mechanisms, which require accumulation of excess spreads in a designated reserve
available to other series account for the benefit of that series or its most junior tranche under certain
circumstances. Consequently, the excess spread from that series available to other
series in the master trust would be reduced. In most cases of master trust
structures any residual excess spread is released to the originator. One exception
is the aircraft master trust structure, where in some cases the excess spread is used
to ‘turbo’ the principal repayment of the senior bonds outstanding.
• The location of principal from non-amortising to amortising series, or from
non-accumulating to accumulating series.
Utilisation of Principal Collections
Another key point in a master trust structure is the allocation of principal to repay
the series, or put differently, the type of amortisation. The principal can be:
• Accumulated in a special account to ensure a bullet payment at expected
maturity of the notes (the so-called accumulation structure meant to establish
a soft bullet), for example: the CARDS Master Trust; or
• Passed-through to investors up to a specified amount (controlled or regulated
amortisation structure), for example: Arran One MT, or fully (rapid
amortisation structure, usually a result of the occurrence of a trigger event in
all master trusts).
In this regard, it is the level of principal payment rate that determines the speed of
accumulation of the principal to achieve the soft bullet structure or the speed of
amortisation in case of rapid amortisation. In case of controlled amortisation the
required principal payment rate necessary to meet the required controlled principal
payment amount is usually much lower than the actual pool principal payment
rate.
Principal can be re-directed to In a typical credit card master trust structure the revolving period is followed by
support series in their an amortisation or an accumulation period. In a mortgage master trust structure the
accumulation phase revolving period may be interrupted in order to accumulate collected principal and
prepayments, and resume upon payment of the soft bullet. Under normal
circumstances, in a master trust there are series in a revolving period and in an
accumulation/amortisation period. Hence, principal from the revolving series may
be re-directed to support series in their amortisation or accumulation periods.
Such redistribution mechanism will not work if all series are accumulating or
amortising simultaneously.
With regards to the amortisation or accumulation period, a corollary question is
related to the speed of amortisation or accumulation. The payment rate becomes
important as it determines the likelihood of extension affecting in a domino
fashion the senior and the junior tranches.

4.2 Break-even and Sensitivity Analysis

Many of the issues we discussed above can be put into perspective through cash
flow simulations of any given deal. We believe that such simulations are
particularly important in the analysis of subordinated tranches. They involve the
application of stress scenarios to one key variable of the asset pool in order to
determine the effect on cumulative losses for the pool, hence potential losses for
investors at a different levels of the ABS capital structure.
We use in-house developed credit card ABS default model, to stress certain key
portfolio statistics to determine the break-even loss rate that a hypothetical credit
card ABS can withstand before the most junior class loses a dollar of either
interest or principal. We illustrate our approach through an example.
Break-even and sensitivity We assume a sample capital structure of 10% subordination, hence 90% AAA
analysis are based on a specific rated Class A, 6% A rated Class B, and 4% BBB rated Class C. Initially, we
example assume a base case of 8% payment rate, 6% charge-off rate, no reserve account,
and no trapping of excess spread. First, we alternatively stress the payment rate
and the charge-offs rate to determine loss accumulation and compare it to the

Refer to important disclosures at the end of this report. 60

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

available credit enhancement level (excluding excess spread and reserve account).
Second, we evaluating the role of the reserve account and excess spread trapping
on the cumulative loss levels.

n Effect of Payment Rate on Losses

As noted earlier, the payment rate is the percentage of outstanding balances paid
(collected) on average monthly. Our first stress test aims to evaluate the impact of
fluctuations in the payment rate on principal amortization and cumulative losses.
To this end we stressed our base case payment rate of 8% by 50% and 100%.

Chart 26: Payment Rate Analysis

Principal Outstanding Cumulative Losses

100% 10%
8%
80% 12% 8%
16%
60% 6%

40% 4%

20% 2%

0% 0%
0 5 10 15 20 25 30

Source: Merrill Lynch

Our example reveals that payment rates are inversely related with cumulative
losses and length of amortization period. Intuitively as the payment rate increases
the time required to make the necessary principal payments decreases and,
therefore, investor’s exposure decreases as well. Notably, with less outstanding
principal exposed to time, cumulative losses decrease. We are left to draw the
conclusion that any increase in payment rates is positive for our investor.
A decrease in payment rates Investors in subordinated tranches need to evaluate the factors that could cause the
accelerates the accumulation of payment rates in portfolios to fluctuate, and in particular slow down. These range
losses from the seasonal (Christmas time, summer holidays), to the fiscal (taxes) and
macroeconomic (unemployment, interest rates). For example, in any period of
economic slowdown, the subordinated investor must evaluate the extent to which
the changing conditions could slow payment rates. Will consumers feeling the
pinch slow down repayment, or will they defensively seek to clear debt burdens
cutting corners elsewhere? Answering these questions requires an examination of
the portfolio’s aggregates and understanding consumer behaviour in the respective
country. The seasoning of the accounts, the average balances outstanding, the
typical utilization of the credit line or credit card, the average interest rate are all
factors bearing upon payment rate volatility.

n Effect of Monthly Charge-Offs on Losses

In this stress test, we evaluate the extent to which changes in charge-off rates
increase cumulative losses. Intuitively we know that cumulative losses will rise.
The question, however, is the extent to which charge-offs can fluctuate without
compromising the performance of a subordinated piece. To this end, we
increased our base case of 6% by 100% and 200% respectively.

Refer to important disclosures at the end of this report. 61

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 27: Charge-Off Analysis

Oustanding Principal Cumulative Losses

100% 20%

6% 12% 18%
80% 16%

60% 12%

40% 8%

20% 4%

0% 0%
0 5 10 15 20 25 30

Source: Merrill Lynch

In all three cases losses wipe out Class C principal in its entirety. In our base case,
losses reach Class B principal very near to the transaction’s maturity. Increasing
our base case by 50% produces a large impact on losses. The first dollar of loss
occurs substantially earlier in the transaction (approximately the ninth month).
Cumulative losses for the B piece increase to nearly six dollars of principal,
essentially eating through the entire piece. Further increasing the stress to double
the base case causes losses to eat through both the Class B and C pieces. Class A
losses amount to over four dollars of principal.
Cumulative loss rates lag the What is notable in this nightmare scenario is that cumulative losses never increase
change in charge-off rates by the same magnitude as the change in charge-off rates. This is due to the
constant amortisation of principal. In a real world example, similar results would
hold true accumulation period of a transaction. By extension this suggests that
increases in charge-off rates are most damaging to a transaction when they occur
during the revolving period, leaving the full principal amount exposed to the
increase in loss.
For subordinated tranche investors, the question is whether an increase in charge-
off rates is a temporary fluctuation or an indication of erosion of the underlying
portfolio’s credit quality. In this regard the investor needs to turn to the
transaction’s performance monitoring for guidance. The investor needs to not
only evaluate changes in charge-off rates, but the evolution in the trust’s arrears
profile. A steady increase of amounts in all arrears bands suggests erosion of
portfolio credit quality. No change in the arrears profile suggests a blip.
The investor can look further into the matter by examining the same sort of detail
considered for changes in payment rates. However in this case the question is not
what will cause borrowers to slow down repayment, but cease it all together? As
with all forms of default, it is a matter of willingness and ability to pay. Though
ability to repay (liquidity) is not captured as a credit card variable, a proxy exists
for willingness in the distribution of the portfolio’s credit scores.
It is worth, however, emphasizing one more time that this simulation assumes zero
excess spread and no reserve account. In a more realistic scenario, the monthly
excess spread would dampen the accumulation of losses, by absorbing monthly
loss fluctuations out of the reserve account. We discuss this issue in the following
section.

Refer to important disclosures at the end of this report. 62

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Reserve Accounts Effects

Reserve accounts delay the first Our final example seeks to quantify the support that exists beneath a transaction’s
and reduce the cumulative loss rated tranches, namely reserve accounts and excess spread. Hitherto our examples
to investors have assumed that neither a reserve account nor the trapping of excess spread is in
place. For analytical simplicity we will assume the inclusion of a fully funded
reserve account sized at 5% of the portfolio’s notional amount.
For the first time in our stress testing our Class C notes emerge with their principal
virtually intact. Unlike our base case (where cumulative losses are just over 5%),
cumulative losses with the 5% reserve account total 20 bps. Clearly, the amount
of credit enhancement beneath the “C” piece is key to evaluation of subordinated
tranche performance.
Typically structures that include “C” pieces provide for the provision for charge-
offs out excess spread and for an excess spread funded reserve account. As
pointed out earlier the level of excess spread in the transactions typically dictates
the required funding level of the reserve account. This generally ranges between
0% to the equivalent of 4% of the portfolio’s notional value.
Table 16: Sample S&P With charge-offs met out of excess spread funded reserve account, the majority of
Reserve Account capital losses borne by our Class B and Class C notes in our example would not
3-month Excess Required Reserve have occurred. Further, had excess spread fallen below zero on a three-month
Spread Amount rolling average basis the transaction would have gone into early amortisation.
> 4.54% 0.00%
4.0% to 4.5% 1.50% The risk for investors, however, is a sudden drop in excess spread that prevents
3.5% to 4.0% 2.00% full funding of the reserve account and triggers an early amortisation of the
3.0% to 3.5% 3.00% transaction. In that respect, comfort with the excess spread volatility of a
< 3.0% 4.00% particular transaction and the adequate management of the servicer of the factors
Source: S&P
that affect excess spread, as discussed above, become crucial.

Chart 28: Reserve Account Analysis

Principal Outstanding Controlled Accumulation

100% 10%
0% Reserve Account
80% 8%
5% Reserve Account
60% 6%

40% 4%

20% 2%

0% 0%
0 5 10 15 20 25 30

Source: Merrill Lynch

An important question for investors is the cause and duration of any changes in
Subordinated investors should
excess spread. A one-off drop in excess spread due to accounting adjustment is no
focus on the cause and duration cause for concern, while gradual yet rapid decline would create problems. For
of excess spread declines example, trust excess spread on MBNA International Bank’s master trust dropped
from 6.65% in November 2000 to 1.41% the following month. The dramatic
decline is no cause for alarm - it is the result of an exceptional item resulting from
MBNA’s implementation of the FFEIC’s common charge-off policy. The
implementation expedited the charge-off of long term arrears as a one-off item.

Refer to important disclosures at the end of this report. 63

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Indeed trust excess spread returned to 6.5% during January 2001. Of greater
concern would have been a gradual erosion in excess spread from 6.6% to 1.4%
over a period of a few months, which would suggest an erosion of portfolio credit
quality, diminishing the credit enhancement available beneath the subordinated
pieces.

n In Conclusion
ABS capital structure allows investors to assume different risks and achieve
different awards by taking a view on the performance of a specific asset pool and
its originator/ servicer. The lower the investor moves down the capital structure,
the more intensive the credit analysis becomes and a more sophisticated cash flow
review is required. The a priori understanding of the collateral pool and its
expected performance, and the subsequent close monitoring of its actual
performance are more time consuming. However, the rewards, as evidenced by
the yield pick-up when moving down the credit levels, would compensate
investors for that.

Refer to important disclosures at the end of this report. 64

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

5. Key Consideration for Master

Trust Structures

Refer to important disclosures at the end of this report. 65

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Following its introduction in the mid-1990s, the Master Trust quickly became
the typical structure for credit card securitisations. Since then its
applicability has expanded to embrace other asset classes: unsecured
consumer loans, corporate loan obligations, residential mortgages, etc. As a
result, inevitable evolution in the structure has taken place. Though many
investors are comfortable with the basic operation of the Master Trust
structure, we believe that a detailed understanding is necessary when
focussing on the subordinated tranches of series issued by a master trust.

n Basic Structure

Master Trust Structures rely on Generally, a credit card ABS is issued by a trust as a series of investor certificates
and a seller certificate (also known as the seller’s interest). Certificates give
a series of building blocks
investors undivided beneficial interests in a pool of assets. This element of shared
ownership differentiates them from the Notes (which are typically issued by ABS)
are debt obligations. The use of certificates is intentional as, unlike a discrete
trust, Master Trusts can be used to issue multiple series of notes. The same pool
of receivables generated by the accounts specifically assigned to the trust secures
each series of notes issued. Further each series of notes has a pari passu claim on
the pool’s yield, less liability for trust’s costs and delinquencies.

n Seller Certificate
The seller certificate’s primary purpose is to absorb the daily fluctuations in the
amount of receivables in the trust. Fluctuations in the amount of receivables are
due to many factors including seasonal effects, returns of merchandise, and the
reversal of fraudulent charges. As the amount of receivables outstanding declines
(increases) the principal amount of the seller certificate also declines (increases).
The minimum amount of the seller certificates for most Master Trusts is generally
around 4-7% of the principal receivables transferred to the trust.

n Investor Certificates
Investor certificates are the Each series of investor certificates may have several classes. The senior, Class A,
notes sold to the market is rated AAA while the junior classes, the B and Class C, are generally rated A
and BBB, respectively. Prior to the introduction of BBB pieces, issuers used cash
collateral accounts (“CCA”) or collateral investment amounts (“CIA”). Though all
three methods work to provide the same credit enhancement, they differ in both
cash flow mechanics and marketability. For the purposes of our discussion, we
will only focus on the Class C structure which has grown to be the European norm
over the past year through transactions by Barclays, Royal Bank of Scotland, and
MBNA.
Part of a Master Trust’s appeal for issuers is that it can issue multiple series,
thereby spreading the cost across various transactions. Series issued by the same
trust can have different characteristics (e.g., coupons, principal amounts, and
maturities) despite representing a claim on a common pool of assets. For investors
this raises questions as the manner and extent to which cash flows can be allocated
amongst the series.

n Source and Uses of Cash

Principal receipts and finance There are four - two major and two minor - sources of cash for monthly
charges are the main sources of collections. The first is the payment of principal. The speed at which this occurs is
cash called the payment rate, which is the percentage of outstanding principal repaid
on average each month. The second major source of cash is finance charge
income – the interest on the outstanding receivable balances. The two minor
sources of cash are any fees payable on the cards and interchange.
Interchange is a payment of fees generated from the usage of credit card networks.
Payable to issuers by the networks, interchange is not always included in the
definition of income. The sum of interest charges, fees and interchange are the
portfolio’s income. Dividing income by the total nominal size of the portfolio
gives its yield.

Refer to important disclosures at the end of this report. 66

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Similarly, there are four uses for the cash received. They are ranked in order of
priority for disbursement. Foremost is payment for the servicing of the portfolio.
Typically the portfolio’s seller remains its servicer. Next in priority is payment of
the fees for the maintenance of the trust. This covers all expenses from corporate
services (accountants, lawyers, etc.) through to trustee fees. The first two uses are
administrative and logistical charges. The final two uses represent the actual
performance of the transaction, namely the payment of interest and principal.
Their treatment is detailed at length below. Any funds left over after meeting all
costs is called Excess Spread. Its treatment is also detailed below.
It is important to note that it is convention to express all sources and uses of cash
as a percentage of the notional value of the portfolio.
n Redemption Profiles
Master Trusts serve to extend The first role of cash flow allocation in a Master Trust is to extend the term the of
the lifetime of the receivables to (mostly) monthly credit receivables to match the term of the notes. Credit Card
the term of the notes ABS do not contain a static pool of assets. Instead, with each monthly collection
by the servicer, existing receivables are individually discharged either in whole or
in part depending upon the principal collected. Rather than pass these principal
receipts directly on to certificate holders, the Trustee uses principal receipts to
purchase additional receivables, thereby maintaining the required level of
receivables in the trust. “Revolving” asset pools in this manner can create Credit
Card ABS of varying maturities.
Credit card ABS typically redeem as “soft bullets” in a single repayment of
principal. Sinking fund redemptions, however, are possible. A series that has a
bullet repayment of principal utilises a controlled accumulation feature, while a
series that has multiple principal payments uses a controlled amortisation
feature.
Soft Bullets use controlled A controlled accumulation structure traps cash received from monthly collections
accumulation until it has enough collected to repay the principal in full. The cash accumulates
in a principal funding account (“PFA”), where it is invested in highly rated short-
term or money market instruments. This results in negative carry as cash returns
are typically less than the interest rate on the certificates. The time period required
to accumulate the required amount of cash depends upon the payment rate of the
underlying collateral.
Chart 29: Sample Soft Bullet Redemption
Principal Balance
Ex pected Legal
Maturity Maturity
100%
Rev olv ing Period Accumulation
Period

0%
1 11 21 31 41 51
Time 61 71 81 91 101 111

Source: Merrill Lynch

Sinking funds rely on With a controlled amortisation structure, equal instalments of principal are
controlled amortisation passed-through to investors over a specified period of time (e.g., 6, 12 or 24
months). This is an attractive feature for issuers since a shorter accumulation
period reduces the amount of negative carry experienced by the trust and
ultimately the seller/servicer. Common in the early days of credit card
securitisation, sinking fund redemptions are now rare.

Refer to important disclosures at the end of this report. 67

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

“Soft” redemption schedules Key to credit card redemption profiles is the concept of soft maturity schedules.
reflect extension risk All credit card ABS carry an expected maturity date as well as a legal final
maturity date. The expected maturity date is the planned date of the bullet
redemption or the final payment under the sinking fund. The trust’s ability to
meet the payment is a function of the payment rate of the underlying collateral. If
payment rates severely under perform from expectations, there is a risk that that
the trust will have insufficient cash to make the payment. To mitigate this risk
there is a time buffer between the expected and legal final maturity to give the
trust an opportunity to accumulate the required cash. Indeed, a transaction’s rating
typically address its ability to repay principal in its entirety by the legal final
maturity not the expected maturity date. Further, missing an expected maturity is
not an act of default, though would understandably be viewed very negatively by
the market. The risk that that a transaction will not be able to meet its expected
maturity date is called extension risk.

n Cash Flow Allocations

The cash flow derived from the underlying receivables and charged-off amounts
are allocated between the seller certificates and each series of investor certificates.
The allocation process occurs in stages. At the highest level, the principal balance
of the seller certificate and, assuming more than one series has been issued, the
aggregate principal balance of the investors certificates are used to allocate cash
flow and charged-off amounts. Unless the cash flow allocated to the seller
certificate is being used as credit enhancement, or is needed to maintain the seller
interest at a specified level, it is released to the holder of the seller certificate.
There are three common The next level of allocation occurs among each series of investor certificates.
methods for allocating cash There are three primary methods to allocate cash flow and charged off amounts.
flow across series These methods are used to distinguish three different types of Master Trusts.
Under the first and second method, the amount allocated to each series depends
upon its size relative to the total investor interest. The trust allocates monthly
principal receipts among series based upon a floating percentage during the
revolving period and a fixed/floating percentage during any other period (e.g., the
accumulation or amortisation period). The trust uses the same formula to allocate
finance charge collections (the interest on the underlying collateral) and charge
offs amounts. Once allocated to a series, principal collections can be used to
purchase new receivables, fund the PFA, or passed-through to investors. Finance
charge collections (interest on the underlying receivables) are used to cover
coupon payments, charge offs, and servicing fees.
Methods 1 and 2 differ in their Up to this point the first method and the second method have been the same. They
treatment of excess spread differ in their treatment of excess spread, which represents any excess in finance
charge receivable collections. Under the first method (Type 1 Master Trusts),
excess spread is related to either the credit enhancement provider or the seller.
The holders of the senior class of any series do not have any rights to the excess
spread.
In the second method of allocation (Type II Master Trusts), before excess spread
is released to the seller, excess spread for one series is made available to any other
series that cannot cover its expenses. Varying coupons of each series issued by
the same master trust is the primary reason for different excess spread among
series. By allowing the sharing of excess spread, early amortisation risk is
reduced for series with relatively high expenses. If the situation continues to
deteriorate, however, delaying an early amortisation may expose investors to
increased credit risk.
Method 3 allocates excess In the third method of allocation (Type III Master Trusts), principal collections are
spread in proportion to allocated and used in the same manner as the other two methods. The allocation
expenses of finance charge receivable collections, however, is quite different. Under the
third method, allocation among series depends upon each series’ expenses (i.e.,
coupon, charged off amounts and servicing fees) relative to the combined
expenses of all series. By allocating finance charge receivables in this manner,
series that have more expenses receive more dollars of finance charge collections.

Refer to important disclosures at the end of this report. 68

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

As such, series with relative low expenses are subsidising the series with relatively
high expenses. The higher expenses are usually the result of relatively high
coupons. Although this allocation method can reduce the credit risk and early
amortisation risk for series with relatively high expenses, it increases those same
risks for series with relatively low expenses.

Table 17: Sample Allocation of Excess of Spread

Method 1 Method 2 Method 3
Series 1 Series 2 Series 1 Series 2 Series 1 Series 2
Yield 18% 18% 15% 15% 20% 20%
Expenses 14% 15% 14% 16% 10% 15%
Difference 4% 3% 1% -1% 10% 5%
Allocated Excess Spread 4% 3% 0% 1% 6% 9%
Source: Merrill Lynch

The sharing of principal A final note on the allocation of principal collections, most Master Trusts allow
collections is also possible for the sharing of principal collections. The sharing of principal collections allows
a series that is currently in its accumulation or amortisation period to utilise the
principal collections allocated to other series that would normally be used to
purchase additional receivables from the seller. Recall principal collections
allocated to a series are used to purchase additional receivables from the seller
when such series is in its revolving period or are allocated principal collections
that exceed either their controlled accumulation amounts or controlled
amortisation amounts. By sharing principal collections, the receiving series can
shorten its accumulation or controlled amortisation period.
Once the cash flow has been allocated to each series, cash flow is further allocated
to each class of certificates within a series. Generally, principal payments are
made on a sequential basis, with principal payments made to the holders of the
Class A then to the holders of the Class B and then finally to the holders of the
Class C. Generally, the interest payments on the Class B is pari passu with the
interest payments on the Class A and the interest payments on the Class C is
subordinated to the interest and principal payments on the Class A and B.

n Credit Protection
As previously mentioned the credit enhancement for most credit cards ABS
consists of the subordination of cash flows and, in certain circumstances, reserve
cash accounts. The subordinated cash flows represent excess spread and junior
classes.

n Subordination and Reserve Accounts

Excess spread typically Generally, excess spread is the amount of finance charge collections remaining
represents the first layer of after paying for the current period’s interest payments on Class A and Class B
credit protection interest and servicing fee payments and covering the current period’s charged off
receivables. Excess spread is also used to cover unpaid amounts of interest
payments, servicing fees and charged off receivables. In certain structures, excess
spread from one series can be shared other series issued by the trust. As discussed
above, the method used to allocate finance charge collections determines at what
point excess spread will be shared. The amount of excess spread that is not used
by investors can either be used to fund a spread account, pay fees to providers of
credit enhancement or be released to the seller/servicer.
The junior “tranche” in a The junior classes include a Class B and a class C or a collateral investment
transaction can be structured in amount (CIA). A Class C is generally structured for institutional investors, while
different ways a CIA is structured for traditional bank providers of credit enhancement. As
noted, we will be focussing only on the investment merits of the Class C. Cash, if
available, is usually deposited into a cash collateral account (CCA) which supports
the Class A and Class B and a spread (reserve) account, which supports the Class
C. The CCA is usually fully funded at closing. The spread account may be fully
funded, partially funded or unfunded at the time of issuance. If the spread account

Refer to important disclosures at the end of this report. 69

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

is partially or unfunded at the time of issuance and additional deposits are

required, such amounts are funded by excess spread. Additional amounts are
usually required, if the excess spread is below certain predetermined levels.
The amount of credit enhancement depends upon a variety of factors, including
the capability of the seller/servicer to manage its credit card business, collateral
performance (e.g., defaults, delinquencies, portfolio yield, and payment rate) and
the transaction’s structure.
Table 18: Sample S&P Reserve As previously mentioned, the reserve account supporting the Class C can either be
Accounts fully-, partially, or unfunded at the time of issuance. If it is partially or unfunded
3-month Excess Required Reserve the structure will provide a mechanism to trap cash flow in the spread account
Spread Amount until it reaches a specified level. The ability to fully fund the spread account will
> 4.54% 0.00% depend upon how quickly the portfolio’s credit quality deteriorates. Though
4.0% to 4.5% 1.50% excess spread is expected to be able to fund the reserve account during a short
3.5% to 4.0% 2.00% period of time, it is susceptible to the risk of a sudden, sharp erosion in portfolio
3.0% to 3.5% 3.00%
yield. As such there is a risk that the structure might not be able to fund the
< 3.0% 4.00%
reserve account to the required levels, thereby weakening the transaction’s
Source: S&P expected credit enhancement.
Early Pay-out Events
Credit card ABS benefit from structural protections that require the early
repayment of principal if certain adverse events occur. These events are classified
as either series related or trust related. A series related pay out event would result
in the early repayment of only the affected series issued by the master trust.
Both series and trust related Series related events include the following:
events could prompt the early • The average excess spread for any three month period is less than zero.
amortisation of principal
• The principal balance of any class of certificates is greater than zero after the
expected final payment date.
• Failure by the seller to add receivables when required.
• Failure by the seller/servicer to make required payments.
• Failure of the seller/servicer to cure any breach of its representations or
warranties.
• Failure by the servicer to replace any interest rate swap or interest rate cap
provider counterparty that does meet the minimum credit rating requirements
(in some cases).
• Trust related events include:
• Insolvency of the seller/servicer.
• The trust is deemed to be an investment company.

n Seller/Servicer Support
Despite the legal separation between seller/servicer’s and the transaction, most
have a strong interest to have their transactions perform successfully. Reasons
range from reputational risk to the desire not to compromise future access to the
ABS capital markets. This results in an unwritten form of credit enhancement
comes in the guise of “moral” support from the transactions’s seller/servicer. This
consists of providing “support” to credit their credit card ABS, which was not
required under the legal documentation. Possible methods include over-
collateralising investor certificates or substituting assets of a superior credit
quality into the pool.
Though not an explicit form of Though this is yet to publicly occur in European Credit Card ABS market, a US
enhancement, seller “support” example is First Union who subordinated its right to receive a servicing fee to
is an important protection ABS holders. For example, Mercantile discounted principal receivables which
increased the trust’s portfolio yield. Similarly, Banc One (First Chicago) added
better performing seasoned accounts to its trust.

Refer to important disclosures at the end of this report. 70

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Providing moral support is not without some cost to seller/servicers. By

supporting deals after the fact, bank issuers of credit card ABS risk losing their
accounting and regulatory off balance sheet treatment. Although investors
generally prefer deals that do not need support, most investors welcome such
support as its maintains the rating of the ABS. Investors should be aware that at
the time of “support” these deals generally trade at a concession to the overall
market. Further, though in many cases “support” can be expected, it should never
be counted on. This point should be made even more firmly given the recently
published second draft of the proposed BIS capital adequacy guidelines. In
revolving structures with early amortisation features, for example, triggering early
amortisation leaves banks exposed to the need to fund newly generated receivables
on-balance sheet (since collections are used to rapidly repay the outstanding
securitisation bonds). This could result in liquidity pressures. In addition, there is
a potential for reputational damage due to an early amortisation triggered by poor
pool performance raising the stakes for the originating bank to find ways to
implicitly support the transaction. As a result, the Committee seeks to address
these risks by introducing a minimum capital requirement of 10% (or higher, up to
20%) of the notional amount of the off-balance sheet securitised pool of such
securitisations.

n Master Trust Evolution

Following its introduction in the mid-1990s, the Master Trust quickly became the
typical structure for credit card securitisations. Since then its applicability has
expanded to embrace other asset classes: unsecured consumer loans, corporate
loan obligations, residential mortgages, etc. As a result, inevitable evolution in the
structure has taken place. Launched in September 2000, Citibank Credit Card
Issuance Trust (CCCIT) represents the latest evolution. The structure is notable
for its ability (subject to required credit enhancement levels) to issue the different
tranches at individual points in time, rather than all at once. The program debuted
in September 2000 with the placement of a $800 mln C piece. Over the remainder
of 2000, the trust intermittently issued $4.5 bln in Class A and $500.0 mln in Class
B notes on the back of pre-placed Class C notes.
The structure operates by re-packaging certificates purchased from Citigroup’s
Citibank Credit Card Master Trust. These certificates are used as the collateral to
back the issuance of notes from CCCIT, for which it has a shelf registration. As
noted, the CCCIT is able to issue tranches at its discretion, provided that the
following credit enhancement levels are available. They are 12.25% below the
AAA rated Class A notes, 7.0% below the A rated Class B notes, and the
availability of a reserve fund below the BBB rated Class C notes. As with other
structures, the required funding level of the reserve account is a function of the
amount of available excess spread.
We believe the structure provides a positive model for European transactions for
the following reasons.
Timing Flexibility – Structured to be MTN-like in its ability to issue on demand,
the structure allows the issuer to discriminate between changes in appetite along
the credit curve and respond appropriately. This enables an issuer to respond
tactically to favourable demand conditions as well as meeting on-going funding
requirements. Further, it reduces the pressure on issuers to place subordinated
pieces when attempting to fund during difficult market periods.
ERISA Eligibility – Eligible investments for US pension plans is governed by the
Employment Retirement Investment Savings Act (ERISA). Understandably,
securities enjoy a liquidity benefit from ERISA eligibility. ERISA provides
different treatments for equity and debt. Due to their shared ownership properties,
ERISA regulations treats certificates as equity not debt. While investment grade
notes may be purchased by funds, pass-through certificates typically must pass the
“100 holders rule”. Under the 100-holders rule securities must be held by at least
100 investors to be an eligible investment for pension funds. Though the rule
applies to all certificates issued from trust regardless of credit rating, it has

Refer to important disclosures at the end of this report. 71

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

historically been hardest for subordinated tranches to pass the test given their
relatively small sizes. By issuing all tranches within a series as notes, CCCIT by-
passes this issue. In past other issuers have used a second trust to package
certificates into notes, but primarily at the BBB level.
Application to MBS – One of the most important innovations in the securitisation
of mortgages is the application of the master trust structure, typically associated
with credit card ABS, to residential MBS. Pioneered in the UK, it quickly gained
favor with the largest mortgage originators and securitisers, leading to the
establishment of large securitisation programs with regular multi-currency and
multi-tranche issuance.

Refer to important disclosures at the end of this report. 72

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

6. Commercial Real Estate and

Securitisation

Refer to important disclosures at the end of this report. 73

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Securitisation techniques are often used to package disparate exposures

together, allowing for diversification. However, the range of commercial real
estate securitisations available is even wider than for traditional MBS or
ABS: from single-loan or single-borrower transactions to multiple-loan
portfolio transactions. The investment landscape is complex too. Such
transaction diversity requires different analytical approaches and investor
considerations. Nevertheless, one can always begin with an assessment of
debt-service coverage, tenant quality, lease types, and property valuation.

6.1 Investment Landscape

The European investment market for commercial property is dominated by real
The largest commercial
estate in six countries: the UK, Germany, France, the Netherlands, Sweden and
property market remains the Italy. Of these, the UK investment property market accounts for roughly one-
UK, but market forces will fourth of total investment property. Several different aspects determine the
continue to increase supply and landscape of the European property investment market:
liquidity on the Continent
n Owner-Occupied vs. Tenant-Occupied
Greater focus on corporate profitability on the part of UK corporates has been
partly responsible for the relatively large supply of UK investment property, as a
result of the transition over the last twenty years from owner-occupiers to tenant-
occupiers. This shift is only beginning on the Continent, and France, Italy and
Spain, among others, still consist of a preponderance of owner-occupied
properties. Over the next decade, we expect increased supply and liquidity in
Continental markets, as the Continent experiences its own shift of occupiers out of
the role of property owners, particularly governments or ex-government
corporates, and into the role of lessee.

n Property-Type Mix
The UK and Continent also differ by types of property available in the real estate
investment market. The UK market remains the most transparent market, and
long lease structures have facilitated historic property lending in the retail and
office sectors. Office properties represent just under 40% of UK commercial
property stock, and retail properties a further 40%. Industrial property accounts
for much of the remaining 20% of stock. In contrast, Continental commercial
property markets consist of 50 to 70% office properties, less than 10% retail
properties (excluding Ireland), and less than 5% industrial properties. The balance
of Continental properties consists of multi-family/housing association residential
real estate.

n Effects of Consumer and Corporate Cycles

European commercial property markets exhibit unique cycles because of varying
degrees of correlation with the corporate credit and consumer credit cycles. The
Commercial property is neither correlation reflects the characteristics of the underlying tenant mix. In particular:
directly correlated to the • Office property rental values are correlated to the fortunes of the service
corporate cycle, nor to the industries.
consumer cycle
• Industrial properties are correlated to manufacturing and distribution.
• Retail and multi-family properties are correlated to the consumer economy.
• Hotel properties are correlated to tourism and business and consumer
confidence.

Refer to important disclosures at the end of this report. 74

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Supply and Demand

The degree of supply in a property market further influences the amplitude of the
market cycle. In general, prime retail properties with limited competition (due to
Limited competition through
restrictive planning regimes or limited space) have exhibited lower volatility than
restrictive planning regimes office properties. Due to historically buoyant consumer spending and low price
results in European retail rental inflation, retail properties are less subject to speculative development than office
growth that flattens, rather properties, which have a greater bias towards oversupply and the related property
than falls, in a downturn market cycle.
Again, whereas the office property cycle exhibits positive rental growth at the
peak and negative rental growth at the trough, the retail property cycle shows
instead moderately positive rental growth at the peak, and flat rental growth at the
trough.

n Illustrating the Cycles in Europe

European national markets follow different cycles not only on a national level, but
also on a property type level. This in itself is a diversification opportunity for a
sophisticated investor to explore.

Chart 30: European Office Property Markets, Chart 31: European Office Property Markets, Chart 32: European Office and Retail Property
Position at December 1993 Position at December 1997 Markets, Position at September 2002
Madrid, London Amsterdam, Frankfurt, Milan, Paris Madrid, London Amsterdam, Frankfurt, Milan, Paris
Munich Munich
Milan Milan
Berlin Berlin
Amsterdam, Amsterdam,
Dublin Dublin
Dublin Dublin
Rental Rents Rental Rents Rental Rents
growth falling Berlin, Berlin,
growth falling Madrid,
growth falling
Madrid,
slowing Berlin slowing London slowing London
Frankfurt Frankfur

Rental Rents Munich Rental Rents Rental Rents

growth bottoming Paris growth bottoming Paris
growth bottoming Frankfurt
accelerating out accelerating out accelerating out
Amsterdam
Retail Retail
Madrid
Office Office
Paris
Milan

London Dublin

Source: Jones Lang Lasalle Source: Jones Lang Lasalle Source: Jones Lang Lasalle

Refer to important disclosures at the end of this report. 75

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

6.2 CMBS: Securitised Property Portfolios

Unlike other types of securitised collateral, European CMBS are not homogeneous
asset pools that are automatically suited to statistical analysis. In fact, CMBS in
Europe come in a variety of shapes and sizes, and often different property types
and geographic locales are packaged together to enhance diversification benefits
for investors.

n Major Transaction Types

Some CMBS portfolios may be In creating European property securitisations, there are three major types of
suited to a statistical analysis, transactions employed, each of which require a somewhat different analysis.
others require more Chart 33 succinctly lays them out and illustrates each with actual transactions
fundamental property executed in Europe in recent years. With each type of transaction, a different risk
is contemplated, and each transaction type may include one or more property
assessment types:
• Portfolio transactions generally refer to multi-borrower, multi-property
diversified loan pools with more than 50 to 100 individual loans.
• Multi-borrower diversified loan portfolios (UK balance sheet, German and
Italian transactions to date) best employ a similar analysis to that of US
conduit transactions – primarily an actuarial analysis.
• Property transactions generally refer to multi-borrower loan pools with less
than 20 loans with a potentially wide or narrow range of properties. These
transactions are often classified as “conduit” transactions.
European conduits, despite their moniker, in many ways do not resemble the
European conduits do not
US-style CMBS conduits, which are more akin to the portfolio-type
resemble US conduits, but are transactions above. European conduits warrant instead a large-loan, or
more akin to a large-loan fundamental, analysis. These large-loan transactions may be further broken
portfolio down into two sub-groups, based upon whom the cash flows of the loan and,
hence, the notes rely:
• - Those containing properties with a diversified tenant base.
• - Those with a few tenants.
The former group requires predominately a property analysis, whereas the
latter requires both a property and a tenant analysis.
Multi-borrower pools with between 20 and 50 loans require a hybrid
approach, both an examination of the large and/or riskier loans, as well as
scenario analysis to determine portfolio impact of loan delinquencies and
defaults.
• Tenant transactions generally refer to sale-leaseback, single-borrower
single-property loans and single-borrower multi-property loans with long
tenancies.

Where a single tenant is a long- • Single-property transactions resemble traditional securitisations even less, and
are really more of an investment in a particular property, particular tenant or
term occupier, the property
group of tenants and precise property market (or business district). Although
itself becomes secondary to the risk is tranched, the normal benefits of property diversification seen in most
credit quality of tenant securitisations are not present. These transactions do, however, vary
considerably if the property in question is an office building, retail
establishment, industrial warehouse or hotel property, each of which has a
different cycle, expenses and tenant mix, as noted in the chapter on The Four
Cornerstones of Commercial Property.
• Finally, due to an emergent trend of European corporations and governments
(particularly on the Continent) shifting from owner-occupants to tenants, sale-
leaseback CMBS have become increasingly common. Sale-leasebacks are
effectively secured bonds of the property tenant – again, very different from
other CMBS transactions.

Refer to important disclosures at the end of this report. 76

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 33: Types of CMBS Exposures

•Global Commercial One (Synthetic)
Diversified •Bamburgh
Loan Portfolios •Duke
•Monument
(>50 loans)
•Dolerite
•SCIP
•GECO (Synthetic)
•Wuertthyp EU (Synthetic)
Portfolio •Dutch Dream (Synthetic)
•Europa (Synthetic)

•Nymphenburg (Synthetic)
Other Loan •Eurohypo 2001-1 (Synthetic)
Portfolios •Real Value One (Synthetic)
•Paternoster
(<50>20 loans)
•Acres
•Heritage Mortgage Securities

CMBS Property
•ELOC 1 to 8, and 11
Conduits & Large •Windermere
Loan Portfolios
(<20 loans)

•ELOC 9 and 10
Sale-Leaseback •IMSER
•Telereal

Tenant •Canary Wharf

Single •Broadgate
•Trafford Centre
Property •La Defense
Single-Borrower •ELOC 12

•Pan-European Industrial Properties

Multi •France Industrial Properties
Property •HOTELoC
•CIT

Source: Merrill Lynch

n Transaction Motivation
Not only do the risk profiles differ among the three exposure types, but the
motivation behind the transactions and, often, the ongoing servicer commitment
also vary. We differentiate, by originator, among:
Commercial Banks
Diversified multi-borrower portfolios generally represent part of the commercial
property book of a financial institution seeking to transfer risk from its balance
sheet to the capital markets.
Developers
Developers often employ large-loan “property exposure” CMBS in order to
achieve lower financing costs than in the traditional bank lending market.
Corporates and Governments
Corporates and governments often become the sole and long-term tenants of
properties within a “tenant-type” CMBS transaction. In these cases, such
insititutions have sought to divest of non-core assets, restructure the balance sheet,
or have desired a source of long-term financing.

Table 19: Typical Motivation Behind CMBS Transactions

Exposure Type Institution Purpose of Transaction
Portfolio Bank Capital management
Property Developer Lower cost of funds
Tenant Corporate Divest of non-core assets, long-term financing
Source: Merrill Lynch

n Common Aspects of the Analysis of Different Transaction Types

Despite the difference in required analytical framework mentioned above, an
investor in any commercial property transaction must first take a number of
fundamental credit views.

Refer to important disclosures at the end of this report. 77

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Location is the Most Important Determinant of Property

Performance
Location is the number one factor determining current and long-term viability of
commercial real estate. Location implies both the country, as well as the local
market area and/or city where the properties are situated. In European CMBS, the
geographic location not only impacts the relative position within the commercial
property cycle, but also the legal and regulatory environments and valuation rigor
to which the mortgage loan and property are subject.
Location, Location, Location is so important because the success of any property location depends on
LOCATION the economy of the area, whose employment, planning and demographic trends
impact the property’s ability to attract stable tenants on a regular basis.
A portfolio geographically diversified across the major European commercial
property markets is key to a European CMBS investor’s ability to deliver
consistent returns and avoid the peaks and valleys of credit deterioration
associated with any single market.

n Debt-Service Coverage
The second most important indicator of commercial property performance is cash
DSCR is generally more flow. The first year net cash flow figure can be derived for any property
investment and represents the EBITDA of the property (or Net Operating Income,
important than LTV
“NOI”, in real estate parlance) less capital expenses (e.g. tenant improvements).
This figure is then compared with the required debt-service under the loan to
arrive at debt-service coverage (“DSCR”) and interest coverage (“ICR”) measures.
Of note, Moody’s DSCR often differs from the published underwriter’s. In
Components of rental income
calculating the DSCR, Moody’s assumes a standard annuity amortisation,
differ by jurisdiction and regardless of the terms of the loan, as well as uses its own forecast of future
impact coverage ratios and interest and capitalisation rates in order to arrive at the debt-service figures.
valuation
It is important to inquire as to the components of NOI and rental income, as they
are not standard across Europe (see the discussion of Lease Terms in the chapter
on Legal Aspects of European Commercial Property).

n Loan-to-Value Ratio
In addition to interest and debt service coverage, the other key credit metric is the
loan-to-value ratio (“LTV"). The LTV measures the property value relative to the
debt financing. Though an important factor in determining recovery values upon a
loan default, the LTV is subject to varying degrees of error depending on the:
• Quality of the appraisal in the related country.
• Elapse of time since the most recent or initial appraisal.
• Relative illiquidity of the commercial real estate market.
• Presence (or lack) of comparable transaction data.
Furthermore, “exit” or “balloon” LTVs do not estimate future increases, or more
importantly, decreases in property value. Although some properties are re-
appraised annually, in general, the LTV is not a mark-to-market measure. Exit
LTV is often not marked-to- LTV therefore indicates only that principal amortisation is occurring, and loans
are being re-paid. Though important credit indicators, these are largely captured
market already in DSCR and, as time progresses, speak little about the saleability of a
property upon the default of a loan (the primary purpose of the LTV ratio being an
estimate of recovery value upon default). Investors must therefore take their own
view as to the likelihood of a significant decline in realisable value.

Refer to important disclosures at the end of this report. 78

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n More About Valuation

Valuation of the properties within a CMBS portfolio is particularly important in
portfolios consisting of smaller properties, which may be sold upon enforcement,
if necessary. Investors should not, however, place too much emphasis on property
valuation for large, unique or illiquid properties. For these properties, e.g. trophy
office assets or custom fit industrial properties, the liquidity of the asset is very
low, and alternative uses may be quite limited. In such cases, the credit quality of
the tenant and the lease terms to which they are contractually bound should
provide the basis of investors’ cash flow and credit analysis.
Regardless, for multi-property portfolios with smaller, more liquid individual
property exposures, consistent valuation practices are central to the reliability of
the LTV measure as an indicator of recovery value upon loan default.
A number of general issues should be noted when considering LTV measures for
European CMBS:
• In general, loans with greater than 80% LTV (or DSCR below 1.1) are
considered higher risk. Sale-leaseback transactions are the exception to this
rule (DSCR’s as low at 1.0 are acceptable, due to the low credit ascribed to
the property value in question).
Disposition proceeds may differ • The value of the collateral will likely decrease during the foreclosure period
significantly from an out-of- due to property deterioration and declining rental income. Building permits
date market valuation may expire, capital improvements will likely be deferred, and additional funds
to bring the property up to market standards may not be available.
• Disposition proceeds may differ significantly from a dated market valuation.
• If a property is left vacant, due to the default of a single occupier tenant, the
possession value and liquidity of the property will diminish significantly, as
investment purchasers of the property may be limited to speculative investors.
• Differing views on value are typically captured in the capitalisation rate used
to discount the future rental and other cashflows.
• Fully amortising loans are considerably more common in Europe than in the
US. Loans that fully amortise over the term avert the need to refinance the
property to re-pay the loan, and hence reduce the risk associated with volatile
property markets. Nevertheless, such loans may be longer term, increase
pressure on debt service ratios during the life of the loan, and introduce the
risk of obsolescence and the need for capital improvements.
Amortising loans are more • As there are fixed costs to foreclose and liquidate a property, which vary by
common in European property jurisdiction, smaller loans may experience proportionately higher losses.
than in the US In addition to these points, a view must always be taken as to the type and quality
of property, outlook for such properties, and the industries to which they are
exposed over the near and long term.

n Tenant Concerns in Single Borrower Transactions

In a single asset transaction, the property is exposed to two major risks,
concentrated tenant default and tenant non-renewal. The first depends on the
credit quality of the tenant and its position within the industry in which it
competes. The second often depends on market rents and the presence (or lack) of
new, competitive properties built in relative proximity to the existing building. A
diverse and granular tenant base may address the former concern, such that any
single tenant default impacts the property’s vacancy in only a small way. The
second risk may be mitigated by long-term leases, the bulk of which should
terminate after the loan maturity.
Tenant, Tenant, TENANT Rating agencies will consider both the duration and timing of the lease roll-off,
which are ideally long (most after loan maturity) and wide-ranging, respectively.
Assumptions will also be made with respect to re-leasing periods. All else being
equal, a variety of tenant businesses and a high credit quality tenant or tenants will
diminish defaults and increase the likelihood of lease renewals.

Refer to important disclosures at the end of this report. 79

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Sale-leasebacks are Secured Corporate Debt

Sale-leasebacks are a special case of a single borrower. During the lease term the
rental cashflows from a single occupying entity flow through to the securitisation
investor. In European transactions to date, the underlying property has often been
highly customised for the tenant (such as a utility), and alternative tenants may be
hard to find. In this case, a European sale-leaseback becomes analogous to a US
credit tenant lease, where the rental income is the sole source of repayment for the
issued securities, and the rating volatility of the CMBS is highly correlated to the
credit quality of the tenant. To some degree, structural enhancements can mitigate
the risk of tenant default, e.g. credit enhancement and liquidity facilities sized to
cover a void or liquidation period. Liquidity facilities should be placed at the
issuer, rather than the loan, level as loan level facilities may have the effect of
putting junior tranche interest payments ahead of senior liabilities in the payment
waterfall.
In addition, for sale-leasebacks involving regulated entities, the regulator may be
able to interfere with the ability of the creditor to evict the tenant or acquire and
dispose of the property. The regulator may also be able to modify (i.e. reduce) the
rent without regard to the property’s cost base or market rents.

n Other Risks to Keep in Mind

In addition to the areas explored above, a number of other areas should be
contemplated in assessing the investment merits of a given CMBS transaction,
including:
• Construction Risk. Construction of improvements can have a deleterious
effect upon issued existing securities. Cash reserves or other forms of
enhancement may be sized to accommodate this, but unless this risk is
transferred to the tenant, the risk remains that the developer spends above and
beyond intention.
Synthetic risk transfer may • Form of risk transfer. CMBS risk transfer may be executed in cash or
often reduce the credit exposure synthetic format, each requiring careful analysis as to the nature and the
from that of a typical cash timing of the loss to which the noteholder (or swap provider) is exposed.
Often, in CMBS, synthetic transactions expose the investor to a more limited
transaction
number of credit events than the typical cash transaction. This is because
accrued interest, costs of foreclosure and recovery may be limited or excluded
from the definition of loss.

Refer to important disclosures at the end of this report. 80

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

6.3 Legal Aspects of European Commercial

Property
Understanding of the legal framework of one European commercial property
market does not transfer directly to that of another, except perhaps in terms
of the questions that an investor must ask. The answers will differ according
to the structural differences in lease terms, tenant rights, planning
regulations, valuation methodologies, and asset security and enforcement
procedures. These also influence the timing and nature of commercial
property cycles between countries and cities and consequently impact the
appropriate timing of investments. Understanding these features allows
investors to properly compare the plethora of European CMBS transactions,
their underlying properties and to make adequate investment decisions.
In analysing CMBS, experience in one European country’s property market does
not transfer directly to that of another. The most fundamental tenet, that of fully
understanding location desirability and competition, remains applicable regardless
of the country in which a property resides. However, the next several major
concerns of a credit investor may be influenced by the statutory environment in
which the property resides, including:
• Security of the rental stream;
• Projected growth of rental flows (DSCR); and
• Maintenance or improvement of property value (LTV).
Commercial implications of different legal systems must therefore be borne in
mind in order to ensure that rental income flows, loan security and the structural
protections of similar CMBS investments in different jurisdictions are comparable.

n Lease Terms Vary

Lease terms range from Commercial leases across Europe vary as to the balance of property rights and
costs allocated between the tenant and landlord. Lease terms and tenant rights
landlord-friendly in the UK and also vary between office or industrial tenants, and retail tenants, the latter of which
Ireland to tenant-focused in the may have greater rights in some jurisdictions. All leases other than UK and Irish
southern European countries leases are generally indexed to inflation, in the form of a proportion of the cost of
living index in the respective country. In any case, the basic factors determining
the strength of a lease include:
• Term;
• Rental rate;
• Rent review;
• Landlord obligations and tenant rights; and
• Tenant quality.
Lease terms range from considerably landlord-friendly in the UK and Ireland, to
more tenant-focused as one moves south across the Continent. The landlord-
friendliness of a selection of European lease terms are arranged from top to
bottom, generally most to least favourable.

n Customary and Statutory Recovery Procedures Vary

Unlike the US, there are very few non-performing commercial property specialists
in Europe. Those that do exist usually have an element of US ownership or
management. As such, like in the residential market, commercial mortgage
originators are vertically integrated, managing all aspects of the property loans.
Originating lenders typically As most bank originators felt the pain of the last commercial property recession,
service both performing and many are well equipped to handle (or avoid) problem loans. In fact, as vacancies
trend upwards in a number of jurisdictions, LTVs on new financings have
non-performing loans declined, as lenders required additional equity in the properties.

Refer to important disclosures at the end of this report. 81

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Regardless, in most CMBS transactions, it is crucial that an investor understands

the likely recovery value of the property upon loan default. Three key elements
determine this value:
• Procedure to obtain the property;
• Time from property possession until sale; and
• Costs associated with the property sale.
The right to obtain the property lies in the insolvency framework of the relevant
jurisdiction. The extent and duration of the possession process varies by legal
jurisdiction, ranging from a lengthy court-driven process in Italy (as much as
several years in length), to creditor friendly foreclosure environments in the UK
and the Netherlands (Table 20).
Depending on the length of the foreclosure process, potential losses relating to
accrued interest could be substantial. In jurisdictions where long foreclosure and
recovery periods are possible, the maturity date on underlying loans should be
several years earlier than those on the issued securities.

Table 20: Foreclosure Periods Vary by Country

Country Foreclosure Process
UK Legal repossession process commences, with court order required, and lasts 12 months, on
average. Eviction and sale in open market. No government mandated auction system.
Ireland Period of beginning of arrears status to forced sale no longer than 18 months. Presentation
of civil bill, notice of trial/court hearing, order for repossession, execution order,
The UK and the Netherlands repossession by county sheriff.
exhibit the shortest foreclosure Germany 1 to 2-year recovery period.
periods; Italy, the longest France Foreclosure, 50% more than 1 year, 33% more than 2 years. Creditor must have court
decision to seize property.
Italy Lengthy court process with a period of as long as 8 years from time of default to recovery.
Netherlands Foreclosure process usually within 3 months.
Spain Regulated, new law reduces number of required auctions from 3 to 1 to speed up process.
However, foreclosure or auction is rare, normal course of action renegotiation of rate or
term of debt. Typically 3 years to foreclose.
Source: Merrill Lynch

The costs of selling the possessed property depend on loan size and cost of carry,
as well as legal fees, commissions, improvement expenses, and management fees.
As an example, stamp tax payable upon property transfer in the event of a
possession, differs by jurisdiction and ranges from 3% of the property value in
Sweden to 10% in Italy (Table 21).

n Valuation Frequency and Methodologies Vary

The key factor to determining the quality of valuation is market liquidity.
Liquidity is imperative when determining comparison values and appropriate
capitalisation rates for properties similar to those within a selected portfolio.
Liquidity may be assessed by the frequency of commercial properties changing
hands within a given jurisdiction. This ranges broadly across Europe, from
London and Paris, with the most liquid commercial property markets, to Italy and
Scandinavia, with the least liquid markets. Although less liquid than Paris and
London, the key German markets, including Berlin and Frankfurt, exhibit a fair
degree of transaction evidence, as does Amsterdam and, increasingly, Spain.
Professional organisations Some European countries offer greater impartiality (and frequency) in the
provide for consistent valuation appraisal process than others (Table 21). Most Continental appraisers use initial
methodology in the UK, yields and recent prices of similar properties to value a property. The UK, Ireland,
the Netherlands, and Germany have national professional organisations of which a
Ireland, the Netherlands and valuer must be a member. On the other end of the spectrum lies Italy, where the
Germany appraisal process is far more subjective and national guidelines are not present.
Spain and France fall somewhere in the middle.

Refer to important disclosures at the end of this report. 82

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

In Spain, although valuation companies are independent, they lack a professional

association to ensure uniform practices. In the absence of a professional
association, valuers may market themselves based upon relatively aggressive
valuations.

Table 21: Appraisal Varies by Country

Country Responsibility for Appraisal
UK Valued by surveyors who are members of the Royal Institute of Chartered Surveyors
Ireland Independent valuers from Irish Assocation of Valuers
Germany Federal Office for Supervision of Credit Sector verifies appraiser’s expertise
France Appraisal non-standard, varies by region
Italy Surveyors may be in-house, pre-selected, or selected by borrower
Netherlands Dutch Association of Estate Agents (NVM) and lenders determine foreclosure value
Spain Several large independent valuation companies
Sweden Independent valuation companies, discounted cash flow driven
Source: Merrill Lynch

These discrepancies in valuation methodologies are of particular import when

analysing Continental bank portfolio transactions, often containing loans from
more than one European locale. For continental transactions where the originator
of the commercial loan is a developer or a conduit, internationally recognised
agents and their approved appraisers are used almost exclusively. These property
appraisers typically use valuation techniques established by the UK Royal Institute
of Chartered Surveyors, without regard for the location of the property.
Unlike office and retail n Appraisal for Industrial Properties is Cost and Tenant-Based
properties, industrial values are
Apart from London and Geneva, across Europe, industrial rents are relatively
driven most by building costs
uniform, with prime rents averaging Euro 70 per square meter. Space for
and lease quality industrial properties is not limited, again apart from London. As such, unlike
retail and office properties, rather than being demand driven, industrial valuations
are more heavily based upon the cost of the building, the length of the lease and
any alternative value. For custom fit properties, the credit quality of the tenant
and strength of the lease are of utmost importance.

n Country comparison
The table below summarized concisely several structural elements (including lease
terms, tenant and landlord rights, enforcement of security, and planning
permissions) of the commercial mortgage markets in France; Germany; Italy; the
Netherlands; Spain; Sweden; the UK and Ireland (see Table 22). Other aspects,
which need to be examined include: standard lease terms, tenant rights, subletting
and assignment, planning permissions, appraisal and valuation, security and
enforcement. This is a subject to a much more detailed report.

Refer to important disclosures at the end of this report. 83

 Table 22: Commercial Property in Europe
Repair and Service Charges Obligations of:
Full Landlord Tenant
Valuation Typical Freq. of Statutory Basis of Indexa- Stamp Tax Recovery Recovery insures, but Local VAT Right to
and Standard Lease Rent Right to Rent tion of for Property From Excl. recovers Property payable on Assign/ Termin.
Re-valuation Lease Length Review Renewal Review Rent Transfer Tenant Structural from tenant Landlord Tenant Taxes Rent Sub-Let Early
UK Net, OMV, Yes 15 years 5 years Yes Open None 4% Yes No Yes Minimal Structural Tenant 17.50% Yes/Yes ** Only at
Independent, (upward market and internal break
Annually only) rents repairs clause
Ireland Net, OMV, Yes 20-25 5 years Yes (in Open None Up to 6% Yes No Yes Minimal All Repairs Tenant - 21% Yes/Yes Only at
Independent, years (upward most market 20-25% of **) break
at least once (break at only) cases) rents annual rent clause
every 3 years year 10
and 15)
US Gross or Yes 5-10 years None (step No (in COLI*** or Annual Varies by Yes, in Yes Yes Depends Depends Depends 0% Usually Only at
Refer to important disclosures at the end of this report.

Net, OMV, (three rent usual) practice, 2-5% fixed state/ county many (unless on whether on whether on whether prohibited break
Independent, types) usually for percentage cases triple net) gross or gross or gross or ** clause
Annually at least 1 increase net lease net lease net lease
term)
Germany Net, “Long- Yes 5-10 years Rare No COLI Annual 3.50% No Yes Yes Structural Normal Landlord, 16% No/Yes ** No
term” value, repairs maint. but (where
Infreq Costs reimbursed parties opt
by tenant to tax)
Netherlands Net, Yes 5-10 years 5 or 10 No (except Cost of Annual 6% No Yes No Structural Minor Real estate 19% Yes/Yes ** No
Independent, years (if at for retail Living repairs internal tax: 2-3.5% (where
Annually all) property) Index repairs (owner parties opt
(market 55%, to tax)
review*) tenant
45%)
Italy Gross, Yes 6+6 years 6 years Yes 75% of Annual 10% No Yes Yes Structural Normal Tenant 20% Usually Only at
Historical COLI (individual) repairs maint. (some prohibited break
cost/ Costs landlords clause

ABS/MBS/CMBS/CDO 101 – 16 July 2003

Independent, are exempt)
Infreq
Sweden DCF, Yes 3-5 years 3-5 years Yes (in 75-100% of Annual 3% No Not usual No Structural Internal Landlord 25% (where NA NA
Historical (end of most COLI repair and repairs and (1% of landlord
cost/ lease) cases) (above normal unusual assessed opts to tax)
Independent initial maint. wear value)
lease*)
Spain Gross, OMV, No 3, 5 or 10 3-5 years No COLI Annual 6% No Yes Not Usual Structural Normal Landlord 16% Usually Only at
Historical years (long and internal maint. prohibited break
cost/ leases repairs Costs clause
Independent, only)
Infreq
France Gross, Yes 9 years 3 years Yes INSEE 3 yearly, 4.8% (slight No Yes Yes Structural Normal Negotiable 19.60% Usually At break
Historical (tenant Cost of or regional repairs maint. restricted clause
cost/ break every Constructio annual* variation) (land tax*) Costs (tenant
Independent, 3 years) n Index only)
Annually
84

Source: Jones Lang Lasalle, DTZ Research, Merrill Lynch * By agreement ** Subject to landlord approval *** Cost of Living Index
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

7. Basics of Synthetic
Collateralised Debt Obligation

Refer to important disclosures at the end of this report. 85

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

7.1 Synthetic CLOs Explained

A synthetic CLO is effectively a combination of a short position in a credit
default swap (allowing the bank to transfer credit risk) and a long position in
a portfolio of highly rated bonds (which generates the cash necessary to repay
the synthetic bond). These structures can be fully or partially funded. Key
features to consider include the flow of funds mechanism within the capital
structure, and the determination and allocation of credit loss. Synthetic
CLOs can be de-linked from the bank sponsor through a number of
mechanisms, which assure the synthetic bonds’ credit independence.

n Why Issue Synthetic CLOs?

Synthetic CLOs allow banks to meet their primary objective of managing credit
exposure while overcoming many of the limitations of true sale and CLN
structures. Some view them as a refined version of pure CLN structures (limiting
or eliminating the credit linkage to the sponsor bank), while others consider them
a logical step in the development of credit derivatives structures (from single name
credit default swaps through to portfolio credit default swaps to public partially
funded credit default swap structures).
Synthetic CLOs can be In particular, synthetic CLOs can be structured to eliminate or limit any investor
structured to be de-linked from exposure to the credit worthiness of the bank, allowing for credit ratings that
the sponsor, even though the exceed that of the sponsor. Synthetic CLO structures emerged, not surprisingly,
during the credit bearish period in autumn 1998 when investors’ were beginning to
bank remains the lender of penalize CLOs with any credit linkage to the bank. Yet like CLNs, however, the
record bank remains the lender of record and can maintain its confidential business
relationship with its client that comes with continued servicing of the loan. Non-
disclosure of obligor identity is particularly useful in jurisdictions with stringent
banking secrecy regulations.

Table 23: Cash versus Synthetic Securitisation

Cash Synthetic
Credit Line Capacity Management Yes Yes
Regulatory Capital Management Yes Yes
Economic Capital Management Depends on First Loss Retention Depends on First Loss Retention
Funding Yes No
Off Balance Sheet Treatment Yes No
Transfer of Ownership Necessary Not Necessary
Ease of Execution/Administration Medium High
Flexibility Medium High
Source: Merrill Lynch

A broader range of asset types One of the other key advantages of synthetic CLOs – and one that is arguably the
can be used as reference primary driver behind growth of European bank synthetic CLOs – is that such
collateral structures allow for on balance sheet credit hedging of an asset pool that may
otherwise be unsuitable for securitisation, or funding off-balance sheet. This
would include unfunded assets (guarantees, derivative positions, etc), loans with
restrictions on assignment or loans from jurisdictions with legal or other obstacles
on transferability. Indeed, many synthetic CLOs are referenced to a portfolio of
multi-jurisdictional loans.
Transaction execution is As a matter of course there is much less transaction documentation involved in
simpler European synthetic CLO deals compared to the true sale variety – it is enough to
mention the absence of an asset sale and transfer agreement. Synthetic CLOs
allow banks to avoid the onerous process of combing through each underlying
loan document and unraveling assignment clauses or other restrictions in order to
sell the asset. As a result, transaction execution is that much simpler.

Refer to important disclosures at the end of this report. 86

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n The Cost-Benefit Economics of Synthetic CLOs

Chart 34: Example of Regulatory Capital Relief

E x a m p le o f fu n d in g c osts
F u lly F u n d ed (pa ) assu m in g 5% L IB O R P artially F u n d ed

S u p er S en io r C red it
S e n io r N o tes D efa u lt S w a p
(9 0 ) (8 7 )
L+25 10 bps

29 bps 18 bps S en io r N o tes (3 )

L +27

M ezza n in e N o te s M e zza n in e N o tes

(4 ) (4 )
L+60 L +60

J u n io r N o tes (3 ) L + 1 5 0 J u n io r N o tes (3 ) L + 1 5 0

R eta in ed E q u ity (3 ) R eta in ed E q u ity (3 )

F u lly F u n d ed / C on v e n tio n a l P ar tia lly F u n d ed

P re-D ea l R e g ulato ry C ap ita l = 8 .0 0 % = 8 .0 0%
P o st-D ea l R e gu latory C ap ital = 3 .0 0 % = 4 .3 9%
R e g u la to ry C ap ita l R e lief = 5 .0 0 % = 3 .6 1%
C o st P e r U n it of C ap ital R elie f = 0 .1 7 b ps 0 .2 0 b ps

Assumes (1) dollar-for-dollar capital charge on retained equity, (2) 20% risk weight counterparty for super senior swap
and (3) assets that collateralize the notes pay LIBOR – 15 bps.
Source: Merrill Lynch

Benefit of regulatory capital The economics of issuing a synthetic CLO is basically a function of the cost of
relief . . . buying protection versus the benefit from regulatory equity release. Banks
typically get full capital relief in synthetic transactions where the collateral
comprises zero risk-weighted securities, such as government bonds. Pfandbriefe
normally attract a 10% risk weighting, unless the securities are issued by one of
the bank’s mortgage subsidiaries, in which case we understand the bank achieves
full capital relief on a consolidated basis. The super senior swap, if underwritten
by an OECD bank, carries a 20% risk weighting.
. . . versus the cost of To be sure, borrowers would face lower all-in costs using a synthetic structure
buying protection compared to a conventional structure, quite simply because the risk is
underwritten using swaps as opposed to bonds (the cost of issuing bonds equals
the risk free rate plus a credit spread). In our example above, we have used 10 bps
as the cost of swapping the super senior risk (this being the most often quoted
figure) and, together with other assumptions, have arrived at a lower cost for a
partially funded structure compared to a fully funded synthetic transaction. But
really, the difference in cost between fully funded and partially funded structures
would depend on the extent of leverage in the latter, the yield accruing to the
collateral pool relative to the costs of the funded liabilities (that is, the interest
deficiency in the capital structure) as well as the cost of the super senior swap.

n Partially and Fully Funded Structures

Synthetic CLOs can be either partially or fully funded structures – this is a
somewhat confusing terminology considering most sponsor banks do not actually
realize any funding. A fully funded structure is where the risk in the entire
portfolio of reference assets is hedged through the issuance of bonds. This
structure is used less often – only three transactions in the European synthetic
marketplace have been fully funded.

Refer to important disclosures at the end of this report. 87

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

A partially funded structure is where CLO investors underwrite the riskiest

(bottom layer) portion of the reference assets, with the top layer of the asset
portfolio (so-called the ‘super senior’ tranche) hedged through a credit default
swap contracted with a separate counterparty. Noteholders in a partially funded
structure are effectively in the first loss position (subordinate to the ‘super senior’
credit default swap). Of course, this first loss position is further credit tranched
into notes of different seniority and ratings. Compared to fully funded synthetic or
conventional CLOs, the most senior tranche of a partially funded synthetic CLO is
much smaller in size.

Chart 35: A Typical Partially Funded Synthetic CLO

‘Super Senior’ Credit
Default Swap
Swap
‘Super Senior’
Counterparty
Unfunded Originator’s
Portion Reference
Loan
Portfolio
Pay-out of realised loss
Sale of
under ‘Credit Event’
Collateral
Asset Backed
Funded Portion
Credit Default Swap Bankruptcy Remote
Interest
Agreement Issuing SPV Noteholders
Commercial Bank &
Reference Loans Collateral
Swap Premiums Principal

Interest Payments due

Principal Payments due
to Noteholders
to Noteholders

Investors

Source: Merrill Lynch

In a partially funded structure, investors are typically long the first 6% - 15% of
credit exposure in the reference portfolio. We understand that full regulatory
capital relief on an asset pool may only be granted when the credit hedge covers
the entire asset pool – this is primarily why a credit default swap is taken out for
the (near riskless) senior most portion of a reference asset portfolio. The extent of
capital release may vary between jurisdictions depending on the domestic
interpretation of guidelines regarding capital treatment of credit derivatives.

n Key Features of Synthetic CLOs

The Credit Risk Transfer Mechanism
Reference portfolio risk is Most synthetic CLOs use credit default swaps to effect credit risk transfer. In the
normally transferred through HypoVereins Bank’s Geldilux transactions, however, the SPV issues limited
credit default swaps recourse guarantees to the bank, which protect against any credit loss in the
reference portfolio. And in Deutsche’s Blue Stripe deal, a ‘financial contract’ is
drawn out allowing the bank to short the credit risk in its reference portfolio.
There is little material variation in all cases, however. The sponsor bank makes
regular premium payments to the counterparties underwriting the credit risk in its
reference asset portfolio (i.e., the SPV and the super senior swap counterparty).
What Are Credit Default Swaps?
Credit default swaps are derivative instruments used to hedge credit exposure on a
reference asset, be it a single credit, a basket of credits or a portfolio of assets. In
a plain vanilla credit default swap, the buyer of protection pays regular premiums
(the swap quote) to the seller of protection in return for the seller agreeing to cover
certain ‘Credit Events’, as defined in the swap documentation. Payments to the
protection buyer, contingent on the occurrence of losses following a ‘Credit
Event’, may be physically settled (being the norm for single name or basket
swaps) or cash settled (being the norm for portfolio swaps).

Refer to important disclosures at the end of this report. 88

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 36: Credit Default Swap Cash Flows

Reference Credit

Occurrence of
PayOut under‘Credit Event’ ‘Credit Event’

Protection Buyer No PayOut if Credit Protection Seller

Event does not occur

Credit Default Swap Premiums

Source: Merrill Lynch

Credit Loss Allocation

A key element in a synthetic CLO is the definition, determination and allocation
of losses, against which the sponsor bank is seeking protection. Losses in this
instance are contingent upon the occurrence of a ‘credit event’, as defined in the
credit default swap documentation. Such ‘credit events’ would normally
constitute some or all of the International Swaps and Derivative Association
(ISDA) definitions of credit events, to include a failure to pay, bankruptcy,
insolvency related reorganization of the underlying entity, repudiation or
moratorium and obligation accelerations. Failure to pay and bankruptcy are the
‘standard’ credit events normally documented in a synthetic CLO. The
restructuring of an underlying loan, repudiation, moratorium, obligation
acceleration or a cross default with another obligation of the borrower may or may
not constitute credit events in synthetic CLOs.
Losses in the reference portfolio are normally covered by reducing the collateral
Coverage of losses in the amounts held by the SPV. Depending on the particular deal structure, the SPV
reference portfolio can:
• Sell the collateral to the extent required, remitting proceeds to the bank as a
buyer of credit protection.
• Deliver the collateral in kind to the bank to cover any losses.
• Under a repo agreement, the bank repurchases the collateral less the value of
any losses experienced in the portfolio.
The remaining collateral proceeds are sequentially allocated to redeem the notes at
maturity in order of the notes’ seniority, and hence the credit losses first affect the
junior note holders.
Under credit default swap or equivalent agreements, non-payment by an
underlying borrower would normally have to exceed a minimum threshold before
investors become liable. The amounts payable by the SPV for underlying
reference credit losses in synthetic CLOs may be computed by the calculation
agent in a number of ways:
• Cash settlement of the underlying defaulted loan. Such settlement can take
place a defined number of days following default (ISDA definitions call for a
60-day period) based on market bids, or can be based on actual (realized)
recovery values. If there are insufficient market bids for the defaulted assets,
it is not uncommon for independent auditors or appraisers to value the asset
for the purposes of loss determination.
Determination of credit losses • A fixed percentage of loan face value initially (i.e., a “digital” pay-out),
following which any excess recoveries or losses within a specified time frame
(12 months for example) may be allocated to investors. We understand that
regulators do not usually give full capital relief for structures with digital pay-
out provisions. (Digital payout structures are rare in synthetic CLOs, an
example being Triangle Funding II).

Refer to important disclosures at the end of this report. 89

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Realized losses usually include any costs associated with the recovery or sale
of the defaulted reference claim, but may or may not include accrued interest.
The timing of loss allocation may also vary. Losses may be allocated against
the collateral at the time of default, upon the crystallization of recoveries or
only when the notes are redeemed (i.e., at the termination of the deal).
n Collateralization of the Notes and Credit Event Settlements
The separate collateral pool is In a synthetic CLO, the issuing SPV either purchases the collateral on the market
used to provide note debt or from the sponsor bank. Purchases of collateral from the sponsor bank are
service and any loss protection normally transacted through a repo agreement where the originating bank is
obligated to repurchase the collateral at some future date that would coincide with
to the sponsor the bond redemption.
Assets used to collateralize synthetic CLOs are usually highly rated (typically
triple-A), low risk weighted debt instruments, such as – but not limited to –
government securities. A number of German bank synthetic CLOs have been
collateralized by triple-A rated public sector Pfandbriefe. Where the Pfandbriefe
is issued by the bank or one of its subsidiaries, the sponsor bank benefits
ultimately from the issue proceeds and thus funding, in addition to risk transfer.
n Flow of Funds
The mechanics of synthetic structure cash flow can be used to describe the
generation and payment of interest and principal to investors.
Sources of cash flow to service Interest on the issued notes in synthetic CLOs is met from:
note interest . . . • Coupons due on the purchased collateral.
• The premium paid by the bank to the SPV for credit protection.
• Any net repo interest payable by the sponsor bank should the structure
incorporate a repo agreement between the Issuer and the bank.
Synthetic CLO transactions are normally structured such that the premium
payments, coupled with any repo interest, covers the residual negative carry (or
interest deficiency) that would inevitably exist given lower coupons accruing to
the collateral versus interest due under the capital structure.
. . . and principal payments Principal repayment on the issued notes is normally met from the sale of the
collateral, either in the market or through a repurchase agreement with the bank. In
the absence of a repo agreement, note redemption would be met by redemption
and/or sale of the collateral into the market (examples of this include Scala 1,
CitiStar 1 and the Blue Stripe transactions). As described earlier, the sale,
redemption or repurchase of the collateral will be in effect net of any amounts
needed to cover eligible losses accruing to the reference pool.
In a synthetic CLO referenced against a multi-currency asset pool, the sponsor
bank usually assumes all foreign exchange risk (in cash structures, an adequately
rated swap counterparty always assumes this risk). As an illustration, losses on
foreign currency denominated reference assets can be based on the lower of
market exchange rate or the rate determined at the outset (or replenishment date).
Examples of such deals are Natix and Globe-2000.
n Early Amortization and Acceleration Events
Broadly similar to cash flow based CLOs, early amortization events in synthetic
CLOs would include economic triggers, such as accumulated reference portfolio
loss or default thresholds. Should such triggers be breached, reference pool
replenishments (or substitutions) are ceased and the notes begin to pay down.
Note redemption in this instance usually mirrors the amortization of the reference
pool. That is, the amount of collateral liquidated or repurchased to pay
noteholders reflects the redemption profile of the reference pool, taking into
account any losses. Under an early amortization event, therefore, synthetic CLO
notes will in effect become pass throughs. This resembles what happens under
similar events in conventional cash flow CLOs.
Refer to important disclosures at the end of this report. 90
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

But unlike cash flow CLOs, certain ‘acceleration’ events in synthetic CLOs – such
as seller insolvency or material non-compliance of warranties – triggers the
immediate redemption of the notes using available collateral.

n De-linking the Ratings of the Synthetic CLO from Sponsor Bank

Here we discuss how de-linkage of the synthetic CLO rating from the rating of the
sponsor bank is achieved, looking specifically at the generation of cash flows
necessary to service the notes and the features that mitigate risks associated with
these cash flows.
Higher Reliance on the Purchased Collateral
As a first step, de-linkage can be achieved for the notes provided the purchased
collateral is itself de-linked from the creditworthiness of the sponsor. Government
securities clearly are, and so are the Pfandbriefe, at least given the current
approach to their credit analysis and ratings. Of course, a sound legal ownership
or charge over the collateral is a necessary pre-requisite before de-linkage can be
achieved.
Market risks to both the value and liquidity of the purchased collateral represent
credit risks to the transaction and are usually mitigated through:
• Margin call provisions, where the sponsor bank would be required to post
additional assets in order to maintain the required collateral value should this
Protection against shortfalls in
value fall below a certain threshold. Debt collateral backing synthetic CLOs
the value of the collateral over is normally marked-to-market on a daily basis.
the term of the deal
• A hedging agreement (essentially, put option) allowing the SPV to sell the
collateral at purchase price plus any accrued interest at the redemption date.
This option is underwritten either by the bank or by another adequately rated
counterparty.
If the transaction’s structure does not sufficiently hedge mark-to-market risk
inherent in the collateral, there may be some measure of overcollateralisation to
further protect note holders should the bank or counterparty fail to perform under
its obligations as described above. The extent of overcollateralisation would
incorporate rating agencies’ assessment of the potential decline in market value of
the collateral during the time it takes the trustee to liquidate such collateral and de-
lever the transaction sequentially.
Should the bank default and there is no ready market for the collateral (when, say,
the collateral consists of non-government bonds), the transaction may allow an in-
kind delivery of the collateral to investors in lieu of cash redemption.
It should be noted that synthetic CLOs could attain triple-A ratings even if the
collateral is rated lower than triple-A (examples include KBC’s Cygnus Finance
where AA1/AA+/AA- rated Belgian OLOs are used). This is limited, however, to
cases where the joint probability of collateral and bank default is deemed
sufficiently remote to be consistent with a triple-A rating of the synthetic CLO.
Further provisos would include a repurchase or similar agreement in place, and
that the collateral bears no default correlation with the bank (government
securities, for example).
Lesser Reliance on the Sponsor Bank
The sponsor’s bank credit comes into play in its ability to meet the residual
interest payments (or, alternatively, the credit protection premium) under the
capital structure.
Mitigating sponsor bank risk Sponsor bank risk can be adequately mitigated by requiring the bank to fund its
obligations one or more payment dates ahead of schedule. So if the bank defaults
(a trigger event), payments already made allow for continued interest servicing of
the notes while the trustee sells the collateral to repay noteholders. Interest
deficiency in case of bank insolvency can also be adequately mitigated by
insurance taken out with a suitably rated third party (HypoVereinsbank Geldilux

Refer to important disclosures at the end of this report. 91

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

deals being examples of this). Another method of hedging sponsor bank risk vis-
à-vis credit protection premiums is by sufficiently overcollateralising the notes
such that interest coupons accruing to the issuing SPV’s asset side matches
payments due under its liabilities.
In a number of synthetic CLOs, investor exposure to the sponsor bank may extend
to the bank being able to perform under a repurchase or hedging agreement.
Additionally, certain transactions are structured such that payments made by the
bank include a measure of ‘excess spread’ which the notes are dependent on for
the purposes of credit enhancement.
De-linking the notes in transactions where there is greater reliance on the bank is
normally supported by adequate rating triggers built into the transaction. In
particular:
• If the sponsor bank is downgraded, the risk to its obligations may be mitigated
by, for example, the full pre-funding of such payments due to the SPV, or
other similar measures (joint-and-several or performance guarantees, or a
letter of credit, from adequately rated third parties) that would protect
Downgrade language investors. Failing to do so would normally be captured by an early
amortization trigger.
• A downgrade of the sponsor bank as repo/option provider is normally
countered by finding an adequately rated replacement counterparty, or by
cash collateralization of the obligations.
From a credit perspective, the bank’s role in such synthetic transactions is,
therefore, not dissimilar to the role of a swap counterparty in a more conventional
ABS structure1, that is, its ratings need not be constrained by the ratings assigned
to the notes provided adequate protection is in place. However, where the bank’s
role extends to being a hedging counterparty (or any other similar additional
obligations), we would expect the bank to be rated in the highest short term rating
category in order to support triple-A rated notes in the capital structure.

1
See for example Standard & Poor’s report titled ‘New Structured Finance
Interest Rate And Currency Swap Criteria Broadens Allowable Counterparties’,
January 1999.
Refer to important disclosures at the end of this report. 92
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 37: A Recap: Main Variations in Synthetic CLO Structures

Types of Types of
Possible protection protection Possible timing
Types of sponsor bank Types of Credit Types of loss
against against sponsor of loss
collateral roles Events settlement
collateral risks risks allocation

• Government bonds • Payer of credit • Repurchase agreement • Credit protection • Failure to pay * • Cash settlement, • At loss settlement
• Public sector protection premiums with sponsor / third party premiums paid in • Bankruptcy * based usually on • At deal termination
Pfandbriefe • Agreement to • Margin calls advance • Restructuring of reference market bids a
• Other triple-A repurchase collateral • Hedging agreement • Overcollateralisation asset defined number of
credits at deal termination (e.g., put option) • Full matching of • Repudiation or moratorium days after default,
• Bonds linked to • Hedge counterparty • Overcollateralisation collateral and note • Obligation acceleration recovery values
sponsor (e.g. put option • Matching collateral and profiles following workout or
provider) against risk note maturities • Interest deficiency independent
to value of collateral insurance valuation
• Requirement to • Downgrade language • Fixed percentage of
maintain collateral loan face value **
mark-to-market value
• Other payment
obligations, like the
provision of ‘synthetic’
excess spread, etc.
* Credit Events that are standard in synthetic CLOs. ** Very uncommon type of loss settlement
Source: Merrill Lynch

Refer to important disclosures at the end of this report. 93

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Finding Value in European CLOs

The first step in bank CLO asset selection is to differentiate between linked
and de-linked CLOs. We prefer de-linked structures. Through de-linked
CLOs investors can gain exposure to the European corporate economy
through bonds structured to be more credit resilient – yet typically cheaper –
than plain vanilla bonds. We view diversification as an important
consideration when comparing CLOs to other structured finance
instruments. The synthetic structures add an additional layer of investment
considerations for the European CLO investors. We view the lack of spread
tiering within the synthetic CLO marketplace as an opportunity for investors
to exploit synthetic CLOs with inherently stronger credit profiles, thus
enhancing their risk-return exposures.

n The Question of Credit Linkage

As we already discussed, bank CLOs can be structured as linked or de-linked from
the credit of the sponsor bank. A third variation, credit linkage that is contingent
on specified events, the so-called contingent perfection structures, are unlikely to
be seen on the European market. It should be noted that, each deal is different and
even the de-linked structures may exhibit different levels of linkage to the sponsor
bank. There are two key points to note:
Sponsor Linked CLOs Offer Credit Exposure to Both the Underlying
Portfolio and to the Unsecured Risk of the Sponsor Bank
While tranching of the capital structure allows investors to select their desired
seniority vis-à-vis the allocation of portfolio credit loss, all investors in credit-
linked CLOs with ratings equal to that of the bank are exposed to the
creditworthiness of the bank, and thus any rating volatility that goes with it. The
bulk of downgrades in the global CLO market have occurred in the credit-linked
sector (examples include ROSE2 and York Funding).
Many Sponsor Linked CLOs Have High Credit Ratings Only Because of the
High Supporting Ratings of the Sponsor Bank
Investors should understand how the ratings are achieved, noting that sponsor
linked CLO ratings are capped at those of the bank. This is particularly important
in synthetic structures, where certain notes in the capital structure may be directly
linked to the sponsor bank. Examples of such structures include the
mezzanine/junior notes in SPV-less structures, or transactions where the notes are
collateralized by bonds linked to the sponsor. A case in point: the mezzanine
tranche (Class B) of the GeldiLux 1999-2 transaction was recently downgraded
from Aa2/AA-/AA to Aa3/A+/AA- given a downgrade of the sponsor bank. The
mezzanine and junior notes in this transaction are collateralized by cash deposits
held at the bank and unsecured bonds issued by the bank under its MTN
programme, thus underlining its credit linkage to the bank.
Sponsor-linked CLOs should On a relative value basis, sponsor-linked CLOs should therefore trade at a
trade cheaper than de-linked reasonable discount to the de-linked market, though to us such tiering is not
paper immediately observable in today’s marketplace. This is particularly apparent
among the subordinated CLO market, where sponsor-linked paper (from a number
of synthetic deals, for instance) typically price similar to comparable de-linked
CLOs.
We continue our analysis by focusing on the relative value of de-linked CLOs.

n Comparison with Spread Products

European de-linked CLOs offer the credit investor an opportunity to gain a
diversified exposure to the European corporate economy. Many CLOs also give
investors exposure to a sector of the European corporate community that may
otherwise be inaccessible – that is, credits originated to entities that rely wholly on
bank financing (small-to-medium sized companies, etc.).

Refer to important disclosures at the end of this report. 94

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Credit resilience compared to Credit enhancement and structural support in de-linked cash or synthetic CLOs
plain vanilla corporate bonds underpin their ability to endure greater underlying credit deterioration compared to
unsecured corporate or bank bonds. The degree of credit resilience will vary with
bond seniority and ratings. Our point is that, given a rating, the CLO product is
structured to be able to hold out against more credit stress compared to an
unsecured bond and is therefore less at risk of being downgraded compared to an
unsecured credit of the same rating (ceteris paribus). We believe this strength of
CLOs makes it an excellent defensive instrument for the credit investor,
particularly in more credit volatile periods.

n Revisiting Synthetic Structures: Synthetic Versus Cash Flow CLOs

Let’s recap: In piecing together investors’ exposure in synthetic structures, we
note that the credit profile of a synthetic CLO is a function of:
• The reference assets (whose losses are allocated to noteholders).
• The collateral (liquidation or redemption of which pays-down the notes).
• The sponsor bank as credit protection buyer (payments from which cover
residual interest servicing on the notes).
Investor considerations for The schematic below captures the credit considerations for synthetic structures
synthetic CLOs differ somewhat compared to the more traditional cash based structures. The credit quality of the
from analysis used in reference assets is, of course, an investment consideration in any structure.
Beyond that, investor considerations for synthetic CLOs differ somewhat from
conventional structures, largely analysis used in cash structures. Certain aspects of cash flow as well as servicer
given the synthetic exposure to risks inherent in cash based structures are principally absent in synthetic
the reference pool but cash structures. On the other hand, investors in synthetic CLOs may be potentially
exposure to the collateral in the exposed to additional layers of risk associated with the collateral backing the notes
former and the sponsor bank’s role in the transaction.
Below, we describe some of these key features that differentiate synthetic CLOs
from cash based structures. These differences stem largely from the fact that
investors in synthetic CLOs take synthetic exposure to the underlying reference
pool, and cash exposure to the collateral backing the notes.

Refer to important disclosures at the end of this report. 95

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Chart 38: Key Risk Considerations in Bank CLOs – Conventional Versus Synthetic
Originator / Risks
Reference Pool Sponsor Bank Structural Market
Servicer Associated with
Credit Quality Credit Risks Considerations Considerations
Considerations Collateral

• Historical • Underwriting standards • Ability to perform as swap • Bond profile (pass

• Credit enhancement as a
performance • Competence in counterparty (if applicable) through vs bullet, call
cushion for potential losses
• Credit quality of administrating collections options, etc)
(including excess spread)
obligors and remitting payments • Potential for headline
• Eligibility / substitution criteria
• Diversification • Relationship with and/or event risks
• Priority of payments under
• Type of reference borrowers • Secondary liquidity
different scenarios
assets • Effectiveness of
• Early amortisation triggers
• Portfolio payment procedures in dealing with
• Protection against any risks of
rate and profile arrears CASH
commingling, set-off, interest
• Margin on • Write-off policies CLO
rate mismatch, forex, etc)
reference loans • Potential for breach of
• Role of supporting parties
duties and obligations
(swap counterparty, etc.) and
(eligibility criteria,
mitigants against associated
warranties, etc.)
risks
• Availability of back up
• Legal integrity (‘true sale’)
servicer
• Retention of first loss
• Ability to call the notes, if
applicable

• Historical • Underwriting standards • Ability to pay interest • Credit quality, and • Credit enhancement as a • Bond profile (pass
performance • Relationship with deficiency (credit scope for rating cushion for potential losses through vs bullet, call
• Credit quality of borrowers protection premiums) volatility • Eligibility / substitution criteria options, etc)
obligors • Effectiveness of • Ability to perform other • Reduction in market • Priority of payments under • Potential for headline
• Diversification procedures in dealing with roles in transaction (for value different scenarios and/ or event risks
• Type of reference arrears eg., a repo or hedge • Liquidity or • Early amortisation and • Secondary liquidity
assets counterparty) marketability acceleration triggers • Extent of senior note
• Potential linkage to the • Definition of Credit Events leverage in partially
sponsor bank • Loss determination and timing funded structures
of settlement, allocation • Reliance on trustee in
• Protection against credit being able to protect SYNTHETIC
deterioration of bank noteholder interests CLO
• Protection against
deterioration in value or
saleability of collateral
• Role of supporting parties (like
a hedge c/party, if different to
sponsor) and mitigants
against associated risks
• Legal integrity (‘perfection of
security interest in collateral,
use of SPV, etc’)
• Retention of first loss

Text in italics denote considerations that are unique to the respective structure type. Source: Merrill Lynch

n Some Differences Between Synthetic and Conventional CLOs

We have identified several key investment differences between a typical synthetic
CLO and a conventional, true sale CLO:
Bullet Redemption, Liquid Collateral
The availability of readily saleable collateral in synthetic CLOs allows for bullet
redemption of the notes in most cases, which in itself is investor positive. A
powerful feature here is that full note redemption – assuming a default neutral
scenario – will be immediately realizable under certain acceleration trigger events
(sponsor bankruptcy for instance). By contrast, in cash based transactions the
length of an early amortization payout window will be determined by underlying
pool payment rates.
Of course, full note redemption in this instance will depend on the value and
saleability of the collateral in the open market and/or on the ability of the put or
repo counterparty to meet its obligations.
Potential Exposure to Collateral Credit Risks
The credit quality of the collateral determines the credit quality of the synthetic
CLO notes.
From an investment perspective, the collateral presents less credit concerns when
it is comprised of selected government securities. But a number of transactions
are backed against non-government securities including most often Pfandbriefe.
Refer to important disclosures at the end of this report. 96
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

The triple-A ratings on these non-zero risk weighted securities allow for the senior
notes of synthetic CLOs to achieve similarly high ratings. Yet non-government
collateral is likely to be more credit volatile than true ‘risk free’ paper. For
example, it is far from certain whether Pfandbriefe ratings can withstand a
significant deterioration of the respective bank’s rating, or indeed a change in the
legal or regulatory framework supporting these instruments2. A downgrade of the
collateral may result in a downgrade of the notes.
Less Scope for Servicing and Structure Related Risks
Adequate servicing of the asset pool (administering collections, remitting
payments, etc.) in order to meet debt servicing under the capital structure is an
important consideration in cash flow CLOs. The servicing aspect of a CLO is less
of an issue in a synthetic structure:
• In case of the outright insolvency of the seller/servicer, there is no need to
transfer the servicing function. Using the collateral pool, as described above,
investors can be taken out of the transaction immediately. Hence, the risks
inherent in cash based structures that a proficient substitute servicer cannot be
found are of no relevance to synthetic CLOs.
• Any breach of servicer duties or warranties can be easily ‘reversed’ as no sale
/transfer of assets would have taken place. Having synthetic – as opposed to
cash – exposure to the reference pool also means that any breach of structural
criteria (such as eligibility conditions) can be quickly ‘reversed’. A reversal
in this instance amounts simply to canceling the default protection covering
those particular assets.
• As there is no cash flowing from the reference pool, risks associated with cash
transfer, commingling or set-off inherent in true sale CLOs are not relevant
for synthetic structures.
Synthetic CLOs also have less scope for other structural cash flow related risks
such as basis or currency mismatches. In multi-currency denominated deals, for
instance, foreign exchange risk can be assumed by the protection buyer. In cash
structures, any collections denominated in currencies other than the issuing
currency will need to be hedged, of course, exposing investors to counterparty
risk.
Potential Exposure to Sponsor Bank Risk
The exposure to the sponsor bank depends on the roles it performs in a synthetic
CLO structure. Such exposure could include and be mitigated in the following
ways:
• The risks to the payment obligations of the sponsor bank as credit protection
buyer can be fully mitigated in synthetic CLOs, as noted earlier in the report,
through – for example – an arrangement to deposit the payment well before
the payment date.
• In addition to being the buyer of credit protection, the sponsor bank may also
be a repo or hedging counterparty (vis-à-vis collateral mark-to-market risks)
in the transaction. The risks associated with the sponsor bank being a repo or
hedging counterparty based on its short-term rating can be mitigated through
appropriate downgrade language. While not a constraint to rating the
synthetic CLO notes, any deterioration in the bank’s credit quality will require
remedial action, failing which the collateral is liquidated and the notes paid
down.

2
In its report titled ‘German Pfandbriefe, Moody’s Analytical Approach’ (June
1996), the rating agency stated that “the probability of default for Pfandbriefe can
not be isolated from the creditworthiness of the issuing entity”.
Refer to important disclosures at the end of this report. 97
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Clearly, investors are directly exposed to the sponsor bank’s credit in any
synthetic structure where the collateral pool consists of bonds credit-linked to
the bank. Where the collateral comprises highly rated non-government
collateral that is ‘related’ – though not technically ‘linked’ – to the sponsor
bank (to include, in our opinion, Pfandbriefe), the ability to pay down the
notes under seller credit deterioration or insolvency may not be entirely
assured either. Such collateral is likely to be sensitive to the performance
and/or credit profile of the bank (a hedge provided by a suitably rated third
party should mitigate this risk). In a number of synthetic CLOs, note holders
may ultimately have to take physical delivery of the collateral in lieu of
redemption if a market sale or repurchase cannot be effected3 (examples of
such deals include Globe 2000 and Cast 2000).
In conclusion, many synthetic CLOs are structured in such a way that there is
multiple reliance on the sponsor bank. That is, there may be many ‘layers’ of
bank risk in the transaction – for instance, credit dependence given the nature of
the collateral, reliance on the bank as a hedging and/or repo counterparty, etc. The
more such layers, the greater the ultimate linkage to the bank, in our view, even if
such risks are mitigated at each level. True de-linked cash flow CLOs do not
typically have exposure to the sponsor bank to the extent that many synthetic
transactions may do.
Loss Determination and Settlement
Synthetic CLOs have more complex settlement and valuation issues related to the
credit events in comparison to the cash based structures, where loss crystallization
and impact on the deal’s liabilities are relatively straightforward:
• On the one hand, credit events and payout amounts in synthetic CLOs are pre-
defined, whereas in cash or ‘true sale’ based structures any form of
underlying non-payment or default adversely affects investors. This feature
of synthetic CLOs is clearly investor positive.
• But on the other hand, credit events that are too broadly defined may trigger
premature and more severe loss for noteholders. In transactions where the
credit events are not clearly described, investors would have to rely on a third
party for the interpretation and validation of a covered event.
The credit events, therefore, need to be carefully examined. In our view, synthetic
CLOs with broad credit event definitions and complex workout procedures prove
a challenge to rating agency analysis – default probabilities and loss severity
assumptions in such deals are far from being an exact science. Structures where
the loss payout is fixed from the outset mitigate risk of loss volatility, but we
realize that only very few deals are structured this way.
Greater Reliance on the Trustee
There is a greater reliance on trustees in being able to protect the interests of
noteholders in synthetic CLOs. Among other duties, the trustee in synthetic CLOs
will be required to be vigilant vis-à-vis:
• Monitoring and selling the collateral as required.
• Verifying the determination and allocation of losses.
But there is one important exception – under a seller bankruptcy (as mentioned),
the trustee’s performance in a conventional structure becomes crucial. Continued
deal servicing post seller insolvency, by contrast, is of no relevance in synthetic
transactions.

3
In our analysis, we do not take into consideration the potential benefits of being
delivered collateral in lieu of payment. Such benefits include the replacement of
notes by lower risk weighted paper.
Refer to important disclosures at the end of this report. 98
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Leveraging the Triple-A Tranche in a Partially Funded Structure

In a partially funded structure, senior noteholders are in a more leveraged position
compared to senior noteholders in a similar fully funded or cash flow CLO. Put
differently, senior noteholders could face greater relative losses (as a % of
principal) compared to senior investors in fully funded or conventional structures
should defaults reach a level where all subordination is wiped out. To be sure, the
likelihood of such high defaults is extremely remote being as it is consistent with
triple-A ratings.
As for the subordinated tranches, noteholders in partially funded structures are no
more aggressively long the underlying credit exposure from either a ‘first dollar of
loss’ or ‘expected loss’ position compared to their counterparts in a conventional
or fully funded structure.
Regulatory Call Option
One last point we would make on investing in synthetic CLOs is the potential for
issuers to exercise regulatory call options built into many such transactions. As
we pointed out earlier in this report, these options typically relate to proposed
changes under the new BIS Accord and are normally exercisable after 2002.
Given that funding the option (or refinancing the deal) will not be a deterrent to
calling a synthetic CLO transaction, we would expect issuers to call their
transactions at the earliest opportunity should the new regulatory capital regime
adversely impact on the economic benefits of keeping the synthetic securitisation
in place.
These call options may therefore limit the opportunities for synthetic CLO
investors to benefit from any spread rally post BIS. On the other hand, we note
that callable synthetic CLO paper usually price to scheduled expected maturity
rather than the call date, which in turn makes certain transactions look cheap
relative to their call tenor. But identifying precisely which deals would be called
is a difficult exercise currently given the ambiguity of what the final BIS proposals
may look like.

7.2 Synthetic Static and Dynamic CDOs -

Structural Variations
We address some of the key structural variations in synthetic CDOs under
several rubrics: static and dynamic, cash and synthetic, liquidity, etc. The
delineation, though, is not as clear-cut now as it was in the past. We suggest
that, now more than ever, investors should go beyond the classifications and
focus on the key issues and search for the right analytical questions to
address.

n Static and/or Dynamic CDO

The key differential in synthetic CDOs, as with their cash counterparts, is whether
they are static or dynamic, including both those, which allow for substitution and
the ones, which are managed. We doubt that the investor debate as to which one
is superior will ever be satisfactorily resolved, given the pros and cons of both.
The direction the market seems to be taking, though, is to find ways to mitigate the
downside of each of the two structures.
Hence, the key questions an investor should ask is not whether the pool of
referenced obligations is static or dynamic, but rather:
1. Who selects the referenced obligations? and
2. Who has the right to change one or more of them, when and how? We will
add more details to the answer of the second question later on.
All kinds of variations in the answers are possible:

Refer to important disclosures at the end of this report. 99

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Referenced pool may be selected by CDO equity investor, CDO arranger or

the CDO manager. In all cases, there is some ground for moral hazard due to
the possibly diverging interests of these parties.
The case where CDO equity investor or CDO manager selects referenced
obligations is the classic case of moral hazard demonstrated in many cash CDO
deals and motivated by reward. The mitigation to this risk or, in other words, the
ways to re-align the interest of the equity and debt holders are discussed below.
• Either CDO investor or manager may be allowed to replace referenced names
only under certain conditions, and within strict pre-determined
eligibility/selection criteria. In fact, in some cases the role of the manager
may be assumed by the CDO arranger or CDS counterparty (which saves on
management fees), but may result in the selection of the cheapest to include
names. In fact the ability to make changes in the portfolio can vary from
broad to increasingly restricted, the latter being more widely used recently.
• CDO manager may or may not have trading discretion or such discretion may
be limited to a certain percentage of the pool (trading bucket) or may be
restricted by certain conditions for trading (e.g., selling only reference credits,
which may become impaired according to manager’s judgment).
Investors should consider this range of options to be inversely proportional to their
confidence in the manager, and related to the quality of the referenced pool and
their investment objectives. A wider discretion in trading is probably more
appropriate for high yield portfolios (although they are not yet in the CDS domain)
and more experienced managers.
• The ability of the manager to change the portfolio and to what degree also
depends on the investment strategy permitted. In this regard, the range can
also be wide not only in terms of market sectors as mentioned above, but also
in terms of investment strategies allowed (long and/or short trades, naked
short CDS, basis positioning – adding and unwinding basis positions, etc.).
In any individual case, it is a balance between flexibility granted to the manager
and complexity of the deal associated with the given level of flexibility, and
whether the results of the manager’s use of all flexibility he is given will
materialize in rewards sufficient to compensate for that complexity.
Manager’s ability to trade may be limited to ‘credit improvement trades’ (linked
to spread tightening beyond certain predetermined level, premium ratios and other
tests) or ‘credit deterioration trades’ (risk of a credit event occurring as assessed
by the manager).
Furthermore, trading may be done either by termination or offsetting exposures.
In fact, different transactions can be differentiated by the level of prescription of
the use of termination and/or offsetting trades. For example, the deal structure
may allow the manager to choose one of the above actions. Or, alternatively, it
may direct manager to:
• Pursue a termination in case of an improved credit, thus generating cash into
the structure, and
• Pursue an off-setting trade in case of a deteriorated credit, which does not
require any cash outflow from the structure, contrary to the termination of a
deteriorated credit.

n Cash and/or Synthetic Pool

Initially, when synthetic securitisation was launched, the question cash or
synthetic referred to the form of transfer of the credit risk of the securitisation
pool.

Refer to important disclosures at the end of this report. 100

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

More recently, the question cash or synthetic is equally applicable to the way the
exposures in the CDO portfolio are created through:
• CDS: selling protection on given names (CDS); buying protection (which is
the equivalent of cash bond shorting) is also possible in some cases.
• Cash: buying exposure to given names in the form of bond purchases (short-
selling of bonds has not been allowed, to our knowledge), or
• Hybrid: a combination of the two above above.
To these two broad categories of assets, one can also add total rate of return
swaps, bi-variate risk positions, etc., which we believe will be of limited use in
future deals, as have been in past ones.
In this regard, we believe that the key question is not so much how the reference
pool is created (we will address this issue later), but rather
1. What is its credit quality? and
2. What level of diversification can be achieved and maintained?
Given that CDS exist primarily on investment grade corporates, in order to
improve the diversification of a CDS pool, cash bonds are a welcome addition. In
fact, cash exposures can take the form of bonds, ABS, convertibles (stripped), etc.
We note that particularly ABS, in its broadest meaning (that is, all asset classes),
can add to the portfolio exposures to credit sectors not easily available through
CDS or other cash alternatives. Such sectors include consumer loan portfolios as
in the traditional ABS/RMBS, corporate portfolios on names, where CDS are
scarce (high yield) or non-existent (mezzanine loans, SME bank loans, CMBS,
etc.). Such a portfolio should benefit from broader defensive diversification and
higher rating stability, judging by recent studies of rating transitions of different
fixed income instruments performed by the rating agencies. These studies also
emphasize the difference in transition ratios and in spread volatility for ABS and
corporates under the same economic conditions of duress experienced in the last
few years – this can be viewed as evidence of lower correlations between the two
sectors.
In addition, a hybrid managed CDO allows CDO manager to explore possible
relative value opportunities that may emerge between CDS and cash assets to the
benefit of investors in the respective CDO or achieve better returns through more
efficient leverage through CDS. Overall, a hybrid structure should allow for the
optimization of asset allocation given CDO manager’s skills and sector
knowledge.
One should not forget that regardless of how a given deal is structured it is, after
all, mainly about credit. A portfolio of credits – cash bonds or reference entities –
require similar analysis in terms of credit quality, diversity, single credit exposure,
total spread, default and recovery expectations, etc. However, the way one arrives
at them may require somewhat different analysis.
n One/Multiple CDS – One/Multiple CDS Counterparties
In its early days synthetic securitisation usually involved the transfer of the risk of
the entire portfolio through a single basket credit-default swap to one
counterparty, often the SPV. A possible variation, in the case of partially funded
synthetic CDOs for example, would be two separate risk transfers: one CDS
directly with the super-senior protection provider and one with the SPV.
While these structures are applicable to static and dynamic arbitrage synthetic
CDOs, the choice depends on the level of management anticipated in a given
transaction.
For static deal one CDS counterparty is the rule, as could be with lightly managed
portfolios, where the SPV can trade with or through the CDS counterparty. As the
level of management and manager discretion increases, the number of
counterparties to the SPV can increase into multiple counterparties.

Refer to important disclosures at the end of this report. 101

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Hence, the simplicity and ease of execution of the static balance sheet structures
are replaced with higher documentary complexity of actively managed CDS
structures to allow for an active and more efficient portfolio management. While
the need for competitive pricing for trading and settlement of CDS requires
several CDS counterparties, an excessive number of counterparties used may
inadvertently create the additional problems of documentation complexity and
management of counterparty risk. Hedging and offsetting trades with two
counterparties creates additional difficulties in determining the credit risk of the
position.
So the question is not so much how many CDS there are in the deal, but rather:
1. How many swap counterparties- one or multiple - are used? and
2. How is counterparty risk managed?
3. Is it necessary to have multiple conterparties, i.e. is the portfolio expected to
be actively managed?
Given that a CDS is an agreement between a protection seller and a protection
buyer and the related payment flows going in both directions, the buyer and the
seller have credit exposure to each other, that is – counterparty risk. A CDS
counterparty default has implications for:
• Regular payments of protection premium by the protection buyer. A mitigant
to the protection buyer counterparty risk is the requirement to make protection
payments in advance, i.e. at the beginning of the protection period.
• Event-related, one-time protection payment due from the protection seller.
Such risk can be addressed through a requirement of a minimum rating trigger
of the protection seller. A trigger breach requires a substitution of the
protection seller with a higher rated one, a guarantee of the protection seller’s
obligation by an appropriately rated third party, or posting of collateral by the
protection seller (the level of collateral depends on the level of deviation from
the minimum required rating).
• CDS MTM valuation in case of termination and funding of the termination
payment (repayment at par) through a liquidity line or through a senior
position in the deal’s cash flow allocation. MTM should not be eroding the
subordination levels in the deal.

n P/L SPV vs. Pass-Through SPV

In traditional securitisation structures the SPV is set up as a bankruptcy remote
entity, so that it has no obligations other than those to ABS noteholders. All
potential liabilities it may have in a given transaction should be satisfied from the
cash flows generated by the assets supporting the securitisation and all cash flows,
profits and losses should be passed-through on to the ABS noteholders and
equityholders.
In case of a synthetic securitisation and particularly managed synthetic CDO profit
or loss can arise at the SPV level due to unwinding of a CDS or an offsetting
trade.
The question is here is not whether a loss can arise, but rather
1. How does loss arise? and
2. How is the loss covered?
In case of a loss arising from offsetting trades, it can be covered by the premium
flows and hence absorbed by the excess margin. A loss due to termination must
be funded in cash, but can be amortized over time through the interest flows or be
funded by the liquidity facility.

Refer to important disclosures at the end of this report. 102

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Excess Margin vs. Excess Spread

In both cash and synthetic securitisations the excess cash flows arise from the
excess of the cash flows generated by the assets over the cash flows needed to
service the liabilities of the structure. In both cases key elements of the investor
analysis, particularly by investors at the lower steps of the capital structure, are:
• At what level are the excess cash flows likely to be during the life of the
transaction?
• Which are the key variables, which can affect excess spread negatively?
• How are excess cash flow used in the structure?
In synthetic CDOs the excess cash flows are a function of, among other things, the
credit quality of the pool, trading strategies, counterparty risk, CDS shorting,
MTM variations, etc. Some of these aspects are different from the traditional cash
securitisations and should be carefully examined in conjunction with the other
elements of the synthetic CDO structure discussed elsewhere in this report.
The use of excess cash flow in the structure can fall into one of three categories:
• Never used – released to equity holders and manager.
• Always used – retention in a reserve account for future use or used to purchase
additional collateral, thus creating additional credit protection in the form of
overcollateralisation.
• Used under certain circumstances, such as increase of portfolio losses beyond
certain level or breach of OC or other tests.
In synthetic managed CDO the level of excess margin may be used as a key
trigger as to whether shorting can be executed. Shorting CDS may lead to spread
erosion in the pool and to additional counterparty risk, both of which could
translate into liquidity shortfalls.

n Single Currency vs. Multiple Currencies

Most securitisations usually have assets and liabilities in the same currency, or if
there is a mismatch between the two it is easy to address on an SPV level. In cash
CDOs, the mismatch can occur on part of the assets, given CDO funding in, say,
Euros and CDO pool in a combination of Euro and USD denominated bonds,
whose relative share of the pool can change. The currencies are usually known
from the outset and the respective hedging mechanisms can be established. In an
CDS pool, the potential currency mismatches and the currencies involved may not
be known in advance, given that they can spontaneously arise in case of a delivery
of a foreign currency denominated bond should a credit event occur and is settled.
So the question is not how currency risk is hedged, but rather
1. What currency mismatches can potentially arise? and
2. How are investors protected against them?
As mentioned above currency risk can arise from the delivery, in case of a credit
event, of a bond, whose currency is not predetermined and could be different from
the currency of the CDO’s assets and liabilities. We stress that FX risk only
applies to recoveries, so it may play a small role. The simplest solution is to opt
for a cash settlement. Alternatively, it is through the reliance on the liquidity
facility that the currency risk, translated into the risk of cash shortfall due to
conversion risk, can be addressed.

n Liquidity
In typical cash CDO structure, given the pass-through nature of the cash flows
emanating from the asset pool and directed to servicing the CDO liabilities,
liquidity is usually needed to cover temporary cash flow shortfalls usually due to
timing mismatches. In a synthetic CDO structure and particularly managed

Refer to important disclosures at the end of this report. 103

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

synthetic CDOs, particular needs for liquidity may or may not arise. So instead of
an assumption that there is a liquidity facility in the deal, the questions instead
should be:
1. Is there a need for a liquidity facility in a given synthetic CDO structure? and
2. How is the liquidity facility sized and procured?
Particular need for liquidity in a synthetic CDO are created by:
• Shorting CDS,
• Trading losses that are crystallized by termination,
• Credit events requiring cash settlement or bond delivery (bond must be
purchased).
Furthermore, liquidity may play an active part in the execution of portfolio
management strategy and specifically asset allocation changes.
Certain trading strategies, for example uncovered shorts, the liquidity needs may
be substantial and can be satisfied only through a dedicated liquidity facility. The
size and cost of such facility are of key importance, as high liquidity cost may
reduce the benefits of some savings achieved on the liability side of the synthetic
CDO through the low cost super-senior tranche hedging.
Another aspect of liquidity management is the shortfall of Classes B and C interest
payments. A potential solution is introducing pikeable interest for these classes in
the form of the capitalization of the missed interest for two consecutive periods for
the single-A rated tranche and indefinitely for the BBB-rated tranche.

7.3 Investor Protection and Strategies

One of the key concerns derived from the experience of the cash CDO market is
that of the perennial conflict between the interests of debt and equity investors.
We explore potential measures implemented in synthetic CDO structures to re-
align their interests. We also suggest a simple decision-making tree for investors
in finding their way through the world of synthetic CDOs.

n Re-Alignment of Debt and Equity Investors’ Interests

The potential or real conflict between the interests of debt investors and of CDO
manager/equityholder as well as the mechanism of manager compensation have
been often used to explain partially the underperformance of arbitrage cash flow
CBOs. Static and managed synthetic CDOs are not immune to similar
shortcomings; in fact, they can be magnified given the higher leverage in a
managed synthetic CDO due to predominantly investment grade credits.
In some cases of a static synthetic CDO, the equity tranche may be written down
with the difference of the notional amount and settlement amount as a credit event
is crystallized. The-write down depends on the level of recoveries and reduces the
CDO liabilities (lowest tranches pay highest coupon), thus reducing the CDO
liabilities by potentially larger degree than CDO income and boosting excess
spread flowing out of the structure, if not captured under specific CDO features.
On the other hand, in a static synthetic structure, the equity investor is more
concerned with defaults rather than rating migration, while the latter, especially in
its more acute form, is the primary preoccupation of the mezzanine investors.
In a managed synthetic structure a CDS manager may be have the incentive
through the remuneration structure (being an equity investor and receiving a
performance fee) to provide exposure on riskier names of similar rating, that is,
buy CDS paying higher premium for the same perceived risk in order to boost
excess margin. This is an investment strategy similar to deep-discount buying in
managed high yield CBOs. This may have a negative longer-term effect
particularly on the mezzanine tranches of synthetic CDOs.

Refer to important disclosures at the end of this report. 104

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Some mechanisms embedded in the structures to re-align investors and managers

interests are:
• Limitations on trading – the extreme points of the range are: full discretionary
trading to investor approved trades only. Within this range fall such measures
as: limitations of the trading bucket; limitations to trading based on credit
quality of portfolio, limiting trading to impaired credits only, for materializing
gains or substitution trades (selling lower credit quality and buying higher
credit quality credits), etc.
• Definition of ‘credit impaired’ assets – definition can be changed in case
losses undermine portion of the subordinated tranches or an O/C test is
breached to incorporate rating downgrade or spread movement concept.
• Tranches retained – managers may buy both at the equity and debt level of
the capital structure of the respective synthetic CDO. In some cases managers
have purchases a piece of every single tranche of the capital structure, while
in others, they have purchased pieces of the equity and mezzanine tranches.
• Equity IRR – reasonable level of promised IRR on the equity tranche entails a
more conservative CDO management.
• Controlling position – in a typical CDO only the controlling class – senior
and super senior holders, have the right to remove a CDO manager and
approve a replacement. In some cases, such rights are granted to the
mezzanine tranche investors, giving them additional rights of control and
protection and putting them in par with the senior investors in some respects.
Granting of such rights may be linked to a threshold of losses due to credit
events or to a notional amount of reference entities that have experienced
credit events.
• Utilization of excess margin – as addressed above.
• Management fee – key issues here are not different from the ones in the cash
CDO market. It has become a norm to split the fee into two components: one
flat fee up in the waterfall and a second one based on performance at the
bottom of the waterfall. Variations of the latter are possible, and the key issue
is how manager’s motivation or course of action may be altered by the fees
structure.

Refer to important disclosures at the end of this report. 105

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Refer to important disclosures at the end of this report. 106

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

8. MBS Investor Consideration

Refer to important disclosures at the end of this report. 107

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

The purpose of analyzing mortgage-backed securities is to define their credit

and non-credit-related characteristics, and to consider how they fit with
investment objectives in the context of the current and expected status of an
investor’s portfolio. Mortgage-backed securities require a detailed analysis
of:
• The underlying mortgage pool, its servicing and expected performance.
• The financial and legal structure of the deal.
• The relationship between the performance of the underlying mortgage
pool and changes in the performance of the MBS and each of its tranches.
Furthermore, the broad range of mortgage products, mortgage securities
structures, regional housing price differences, obligor behavioral differences,
and foreclosure proceedings, necessitate a thorough understanding of the
differences between mortgage markets.

8.1. Analytical Framework

Investors interested in mortgage-backed securities need to determine the
likelihood of full repayment of the bonds as well as the likely timing of the
repayment cash flows. Investors should also be concerned with changes to any
aspect of the deal that could have an effect on the bonds’ pricing initially or on
secondary market pricing. These factors include the bond’s rating and the potential
for rating changes during its life, the performance of the underlying collateral and
the performance of the servicer, among other things.

n Credit Rating Agencies

The credit rating agencies – such as Standard & Poor’s, Moody’s Investors
Service, Fitch IBCA – are an important source of information for investors in
mortgage-backed securities. These agencies publish credit reports, which aim to
determine the likelihood of the timely payment of interest and the timely or
ultimate payment of principal (reflected in the credit rating) on the rated tranches
of the mortgage-backed securities. But first an agency must assess the credit
quality of the underlying pool of mortgages and, on that basis, assess the necessary
level of credit enhancement relevant to the assigned rating level.
To do this, the agencies evaluate the credit characteristics of the collateral pool,
the servicing capabilities of the servicer, the underwriting criteria of the
originating entity, the degree to which the cash flows generated by the collateral
pool match the cash flows promised to investors under the mortgage-backed
securities, credit events, collateral and counterparty risk in synthetic risk transfers,
and the soundness of the legal and financial structure of the deal. The rating
agencies subject the mortgage pool to a variety of stress scenarios, the severity of
which is directly related to the assigned final rating on the bonds.
The agencies’ credit rating of the bonds provides an indication of the
creditworthiness of those bonds. However, investors should also look at a number
of other features to better understand the current and future performance of their
investments. Such analysis should include the following three factors:
• The potential for rating changes affecting the bonds, associated with:
− The status and ratings of the parties involved in the MBS transaction, for
example, the originator, the servicer, the swap counterparty.
− The collateral backing the MBS transaction (the assumptions of the expected
performance of the pool) and any change in those assumptions over time.
• The liquidity characteristics of the MBS, associated with:
− The size of the initial issuance.
− The syndicate – its breadth and commitment as a market maker.

Refer to important disclosures at the end of this report. 108

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• The prepayment characteristics of the MBS, such as:

− The prepayment speed assumed for pricing purposes.
− The effects of changes in prepayment speed on a bond’s duration, weighted
average maturity (WAM) and convexity.
The remainder of this chapter is devoted to the specific aspects and features of
mortgage-backed securities that investors should analyze grouped according to
their relevance to the collateral pool, the legal and financial structures of
mortgage-backed securities, origination, servicing and monitoring of the pool, and
prepayment analysis.

Key Characteristics of the Collateral Pool

n Mortgage Balance Diversification
We start by considering the number and average size of the mortgages in a
collateral pool and the distribution by mortgage balance. All things being equal, a
greater number of mortgages in the pool correlates to higher diversity and lower
default levels. The larger the mortgage loan with a similar maturity profile, the
larger the debt service payments and the more potential stress created, given a
change in borrower circumstances or a rise in interest rates.

n Loan-to-Value Ratios
Loan-to-value (LTV) averages and distribution within a pool are also important.
The LTV ratio is considered the determinant of the borrower’s willingness to pay,
reflecting the portion of the borrower’s equity in the property. LTVs are, for many
reasons, not directly comparable across geographic borders; therefore the LTV
profile must be examined against benchmark performing pools in each jurisdiction
to determine whether pool LTVs are more or less aggressive.
A key element of the LTV ratio is the determination of property value. In the UK
and Ireland, for example, property value is based on appraised values by
professional surveyors. In France, some properties are valued simply on the basis
of market values of properties in the surrounding area. In Germany, lenders are
required by law to value property based on ‘sustainable’ rent. In Sweden, policies
differ from lender to lender, but generally use tax valuations, which are performed
every six years but are indexed annually, based on recent sales. In the Netherlands,
the concept of loan-to-value is instead sized to loan-to-foreclosure value, resulting
in a value between 80%-90% of the market value (not dissimilar to the practice of
German lenders).

n Income Ratios
Price-to-income or net income-to-mortgage payment determines the borrower’s
ability to pay debt service out of current income. Price-to-income ratios are less
commonly used by non-US mortgage lenders, where debt service as a proportion
of monthly income is a more powerful determinant of the approved loan amount.

n Interest Rate Composition

The weighted average interest rate on the mortgage pool when the deal closes
must be considered, along with the expectation of change, as determined by the
type of mortgages in the pool – whether fixed rate or variable rate mortgages, or
fixed rate mortgages with periodic resets – and the rate at which loans are repaid.
Higher interest rate loans are generally first to be prepaid, thereby reducing excess
spread. The remaining time to the first reset should also be noted, as refinancing is
more likely to occur at a reset or ‘change of condition’ date, when little or no
penalties are attached to the repayment.

Refer to important disclosures at the end of this report. 109

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Characteristics of Underlying Property

The type and quality of the underlying property are key elements of the collateral
pool. Detached single-family or two-family property, semi-detached property, a
block of flats, a single flat, investment property, residential property, multi-
purpose property (such as part commercial, part residential, which is common in
Italy) – each type of property has different characteristics from the point of view
of the borrower’s motivation to service the mortgage debt and from the
perspective of the property liquidation and recovery value in case of borrower
default.

n Mortgage Products
The type of mortgage products – such as annuity mortgages and life insurance
mortgages – and the effects of the mortgage type on debt servicing and debt
recoveries also factor into an investor’s consideration. Redraw and equity-release
loans, for example, are available in a number of markets, including Australia, the
UK and Ireland, and increase the leverage of a given borrower. In the Netherlands,
savings and investment-linked loans are the norm, while in Italy mortgage loans
may be linked to family businesses.

n Tax Implications
Tax implications relating to mortgage debt must be considered. In several
jurisdictions, tax advantages drive the structure of the mortgage product offered,
as well as the manner in which debt is repaid. Tax regimes have a heavy influence
on payment behavior and loan structure. For example, in the Netherlands,
mortgages are used as a tax-planning instrument and are linked to savings
accounts, investment portfolios or life insurance contracts where the related
investment returns are generally not taxed. Hence interest-only loans are very
common, and loans outstanding do not decrease, as rapidly over time as in the US,
despite being similar in that it is largely a fixed rate market. Another example
occurs in Australia where mortgages receive a tax deduction on principal repaid
but not on interest, thus providing an incentive for early repayment.

n Pool Seasoning
The seasoning, or ageing, of a mortgage is associated with the building up of the
borrower’s equity in the acquired property, which has a strong influence on the
borrower’s motivation to continue servicing the mortgage or to abandon it. A
pool’s ageing is also associated with a loss curve; mortgage loss curves tend to be
front-loaded, in that losses occur early in the life of the mortgage pool and
decrease and stabilise later in its life.

n Obligor Profile
The characteristics of the obligors in the pool – whether they are sub-prime or
non-conforming, salaried or self-employed, private or social housing – are
important; some characteristics are more negative, others are more positive. For
sub-prime and non-conforming mortgage pools, although not an established sector
outside the UK and Australia, performance varies dramatically from that of prime
mortgage pools. Generally, these differences include a higher expectation of
defaults, and prepayments are typically faster and are driven by a change in
borrower credit quality, due to a change in circumstances rather than interest rates,
which are the dominant factor for prepayment in prime mortgage pools. In contrast
to pools with riskier obligors, some pools carry a large proportion of high-quality
obligors, such as civil-servant mortgage loans. These borrowers are considered
higher quality because they are less likely to become unemployed.

n Nature of the Security

The nature of the mortgage security has a bearing on performance. For instance, is
it a first, second or third lien mortgage? In Germany, some loans are upper LTV

Refer to important disclosures at the end of this report. 110

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

(exceeding 60% LTV), but are secured by first-ranking mortgages. Is the mortgage
for a tenant-owner or a single family? In Sweden, tenant-owner loans are secured
not by a lien on the property, but by a perpetual right to occupy the property. For
these loans it is important to scrutinize the financial condition of the cooperative
and the credit quality of the borrower. In Italy, the distinction between fondiario
and ipotecario is critical in that fondiario loans usually achieve beneficial
treatment during the enforcement process.

n Geographic Concentration
The geographic concentration of mortgage loans is related to the demographic and
economic characteristics of the region. For instance, Italian mortgages for
properties in the south and the islands have historically exhibited higher default
rates than those for northern and central Italy. Hong Kong Island housing has
traditionally been more resilient to price downturns than Kowloon and the New
Territories. In Portugal and Sweden, the population is heavily concentrated in one
or two large cities, so mortgage pools may be subject to housing market dynamics
that are less well diversified than in other markets.

n House Price Movements

House price movements by region are another consideration, but there is a
disparity in the availability of house price indices across Europe. Belgium,
Sweden, the UK, the Netherlands and Ireland all have national and regional house
price indices, while Italy has none. Germany has some pricing information, but no
official national statistics.

n Recovery Value
Recovery value and the time it takes to recover a property should not be
overlooked. Property values could be reduced by market value declines,
foreclosure costs and carrying costs from delinquency to foreclosure. Recovery
times vary widely across markets; in the Netherlands, foreclosure to recovery
proceedings can take as little as three months, while in Italy the process can take
up to 10 years.

n ‘Set-Off’ Risk
‘Set-off’ is the risk associated with the borrower’s ability to offset mortgage debt
against a deposit held with the bank originator of the mortgage. In Germany and
the Netherlands, set-off is allowed between savings or insurance contracts and
mortgage debt in the event that the originating bank defaults. Structural steps
taken to reduce or eliminate this risk should be examined. That said, set-off is only
an issue for cash transactions. Synthetic transfer (which is typical in Germany)
obviates the need for any additional legal provisions.

n Substitution
Substitution is a structural provision, common in master trust and revolving
transactions, allowing for the addition of new mortgages to the original pool.
Eligibility criteria should be examined carefully to minimize the potential for pool
deterioration over time.

n Insurance Policies
Insurance policies may be present in a number of forms. Private mortgage
guarantee insurance is used almost exclusively in Australia, on both a pool and
individual loan basis. Private mortgage guarantee insurance is also used, although
not as broadly, in France, the UK, Ireland and Hong Kong. State mortgage
insurance is available in Belgium and the Netherlands. In jurisdictions such as
Germany, Spain and Portugal, lenders may require life insurance policies in
conjunction with certain mortgages, for example, to assure balloon payments on
interest-only mortgages.

Refer to important disclosures at the end of this report. 111

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Static Pool Data

Static and/or dynamic pool default information should be provided by mortgage
issuers but might also be available on a national basis for comparison purposes.

n Consumer Protection
Consumer protection laws vary by jurisdiction. In Sweden, over indebtedness
legislation excludes mortgage payments, in that it leaves the borrower responsible
for the debt even if otherwise over indebted; in France, similar legislation includes
mortgage debt, as well as other consumer debt.

Key Features of the Financial Structure of the

Securitisation Pool
Investors must not only be aware of the nature and peculiarities of the underlying
collateral, but also:
• Credit enhancement and the method of its provision. Via subordination,
insurance or reserve funds. When credit enhancement is established through
subordination of several tranches with significantly different maturities, the
pay down of the senior tranches and pool performance within, or better than,
initial expectations are prerequisites for subordinated tranches upgrade or
spread tightening.
• Cash flow mechanics of the structure. Or the priority of cash flow
distribution – this includes interest and principal payments on the different
tranches, trustee, servicer and issuer expenses, swap payments, and missed
payments of interest and principal.
• For synthetic transactions. Definitions of credit events must be examined
and compared with default recovery assumptions; realized loss definitions
should be examined to size collateral and counterparty risk.
• Liquidity support. This includes liquidity line or servicer advances for
liquidity support for payments of interest and/or principal for borrowers in
arrears
• Swaps. These address potential cash flow mismatches between the weighted
average coupon (WAC) of the underlying pool and WAC on the MBS.
Actions to be taken should be clearly identified in case of swap counterparty
default or downgrade and/or swap termination.
• Application of excess spread, if any. The excess spread is the difference
between WAC of underlying mortgage pool and the sum of WAC of MBS,
servicing fee and other trust expenses, and serves as a first level of protection
against current and potential future losses on the mortgage pool.

n Outline of the Legal Structure of the Transaction

• To securitise a pool of receivables, a number of steps must be taken to ensure
that the legal and economic interest in the pool is removed from the originator
and transferred to an unaffiliated ‘bankruptcy-remote’ issuer. In particular,
investors should examine the following:
• Legal transfer and perfection of security interest. Two broad categories of
legal transfer exist: those that have been established by securitisation laws
(for example, Italy, Spain, France, and Portugal), and those in common-law
jurisdictions (for example, Ireland, the UK, Australia and Hong Kong). In
common-law jurisdictions, a true sale is not effected, but an equitable
assignment is used instead. A true sale (retitling) can only be effected with
borrower notification.
• Transfer of mortgages securing the loans. True transfer of the mortgage
security usually only takes place on the occurrence of certain events, such as
Refer to important disclosures at the end of this report. 112
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

the downgrading of the originator below a certain level (for example, A or

A-). Collateral remains with the seller until such an occurrence – borrowers
may or may not have been notified about transfer.
• Representations and warranties by the seller/servicer. Investors should
ensure that these are provided by an appropriately capitalized entity.
• Bankruptcy remoteness of the issuing entity.
• Data protection trustee. The trustee has custody over certain data lists in
order to identify borrowers and enforce the loans and the collateral.

Key Features of Originating and Servicing the

Collateral Pool
Although the pool of receivables is segregated from the originator in any
securitisation, there typically remains a link to the quality of the originating entity
due to the continued requirement for servicing the pool receivables as well as the
potential impact on the MBS of negative news related to the originator. As such, a
number of issues that directly or indirectly related to the originator services should
be monitored:
• Credit quality of the originator. This should be examined in the context of
the likelihood of requiring a replacement of the servicing function.
• Underwriting criteria and guidelines. The conservativeness and the
consistency of the underwriting criteria inevitably affect the performance of
the mortgage pool. Mortgages originated through a bank branch network, and
based on a long-term relationship with a borrower, including the provision of
other financial services to that same client, tend to perform better than
mortgages originated by a broker and sold to a mortgage consolidator for the
purposes of an MBS deal. The trend of underwriting standards should also be
monitored. In some jurisdictions, such as Ireland, average debt-to-income
ratios and LTVs for approved loans have gradually risen over the past five
years. Absolute levels must also be considered, however; in the case of
Ireland, despite rising LTVs, these remain lower than their UK counterparts.
• Payment methods. Such as direct debit, where borrowers have a current
account with the originator, ensuring timely mortgage debt service.
• Availability of a credit scoring system. For faster and standardized
assessment of borrowers’ credit quality – for instance, credit scoring is widely
used in Holland, selectively used in the UK, and rarely used in Germany, Italy
or Spain.
• Availability of information regarding borrowers’ debt burdens and
credit performance. Through a centralized, nationwide, data-sharing system
among the credit institutions – detailed nationwide credit information is most
widely available in Scandinavia (where a large electronic infrastructure
provides timely information) and least available in southern Europe (Italy,
Portugal and Spain).
• Property value assessment. The availability of established practices and
mechanisms for current determination, and subsequent update, of property
value information. In Germany, for example, the established valuation
practice sets value between 10%-15% below the market value.

Key Features of MBS Bonds

In addition to a thorough understanding of the credit quality of the collateral and
qualifications of the servicer, investors should be aware of the implications of the
structuring of these cash flows into an MBS bond, and the resultant maturity and
risk profile of the investment under evaluation.

Refer to important disclosures at the end of this report. 113

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Prepayments and their effect on floating rate MBS (mainly weighted average
life, or WAL) and fixed rate MBS (duration and convexity).
• Prepayment assumptions used in credit enhancement modeling and MBS
pricing.
• Clear understanding of the differences between expected maturity and legal
maturity for pass-through structures, for example, pass-through structures are
rated by legal, not expected, maturity.
• Investment characteristics of specific MBS tranches – for example, senior,
subordinated, interest only (IO) payments or principal only (PO) payments; in
the case of IO, this includes the effects of prepayments, resets on the fixed
rate loans, excess spread, and WANM (weighted average net mortgage rate).
• Syndicate composition and commitment to secondary market making
associated with the expected liquidity of the bonds.
• Finding an appropriate benchmark for pricing disparate asset classes, all
falling under the label of MBS, e.g. performing versus non-performing loans,
originator experience, geographic distribution and loan seasoning. As default
curves vary by jurisdiction (e.g. defaults peak, on average, two years from
origination for Italian mortgages, and four years from origination for UK
mortgages), similar seasoning for different jurisdictions may have different
credit implications.

n Monitoring Ongoing Transactions

To ensure that an investment broadly reflects the characteristics under which the
initial commitment was made, monitoring the pool’s performance is crucial. Prior
to investment, investors should examine the following:
• Mechanisms for receiving timely information regarding pool performance
(trustee reports, rating agencies reports, reporting on Bloomberg or a specified
website).
• Sufficiency of the information received for evaluation of pool performance –
reporting criteria, clarity about calculation of different performance
indicators, comparability of reported information among similar deals in the
respective country and across countries.

n Explanation of Selected Key Review Points

The following is a closer examination of a few of the points mentioned in the
above sections (for a summary of the key aspects of the analysis of the asset and
liability side of MBS, see figure 2.1 at the end of this chapter):

n House Price Sustainability

Borrower inclination to remain current on a mortgage is a direct result of the
perceived equity captured in the property securing the loan. If the value of a home
is rising, borrower equity is growing and if the borrower were to default on the
mortgage, the equity invested would be lost. On the other hand, if housing prices
are deteriorating and equity is diminished (as in eastern Germany), eliminated, or
even negative (as in Hong Kong), there is less incentive to remain current on the
mortgage. In these circumstances, the loan amount on which servicing is required
may be in fact larger than the value of the property itself.
As residential mortgage market cycles differ widely depending on jurisdiction, to
identify the likelihood of a sharp reversal in prices it is important to monitor the
development of both structural and cyclical factors driving house prices. In short,
the greater the cyclical element behind the level of prices in a given locale, the
greater the likelihood that an economic slowdown would precipitate a reduction in
prices. A number of trends may be analyzed to determine the degree of structural
influence, among them:

Refer to important disclosures at the end of this report. 114

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• Household growth and size of household

• Home ownership trends
• Availability of a secondary market for homes (upgrade mentality)
• Population growth
• Migration to (or from) country and/or major cities
• Divorce and separation rates
• Age and background of individuals when purchasing a first mortgage
To the extent that the first six factors on this list are increasing, and the last is
decreasing, such structural support in the markets under review may support home
prices even in the face of moderately rising interest rates, unemployment, and
increased house construction (supply).

n Analyzing MBS Prepayments

An important aspect of MBS deals and underlying pools for investors to
understand is that of prepayments. Mortgage borrowers have the right, and are
often induced by a number of circumstances (such as market interest rates falling
well below the borrower’s initial mortgage rate, increased income and desire to
move to a bigger house, or change of job location and the need to move house) to
prepay their mortgage loans voluntarily. Alternatively, mortgage borrowers may
face another set of conditions (such as increasing mortgage payments due to
interest rate hikes, loss of income due to unemployment or disability) that may
force them to default on their mortgages and lead to involuntary mortgage pay
down as a result of a foreclosure on the mortgaged property.
Both voluntary and involuntary pay downs of outstanding mortgages have the
effect of changing the composition of a given mortgage pool backing an MBS,
thus changing the cash flows generated by the pool and ultimately affecting the
MBS amortization schedule. This is why investors must understand the effect
prepayments have on MBS, and form adequate expectations as to when they will
be repaid.
An analysis of the effect of prepayments on fixed and floating rate MBS, is
summarized as follows:
• Prepayments may affect the weighted average yield generated by the
mortgage pool, compared with the weighted average coupon as paid under the
MBS. The difference between the two is the excess spread (after deduction of
losses and servicing fees), which serves investors as the first protection
against losses – reduction in excess spread diminishes the cushion protecting
investors against losses. Reduction in excess spread due to prepayments
occurs as mortgage borrowers tend to prepay or default sooner on higher
interest rate loans – thus reducing the weighted average yield generated by the
mortgage pool – in a respectively falling or rising interest rate environment.
• Prepayments shorten the WAL of MBS, as investors receive principal
repayments earlier than expected.
• Receiving principal repayments earlier than anticipated may have little effect
for floating rate MBS investors, as the reinvestment risk of the repaid
principal is relatively small – the floating rate coupon on an MBS should
closely trace the reinvestment rate, save for a spread over a common index.
• The effect on fixed rate MBS could, however, be significant and can be
expressed in:
- Reinvestment risk. MBS prepay mainly when interest rates are declining and
low, hence an MBS investor may have to reinvest prepaid principal at a rate
below that of the MBS coupon.

Refer to important disclosures at the end of this report. 115

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

- Convexity risk. Prepayments lead to decreasing convexity for MBS. In a

declining interest rate environment, the value of a portfolio with decreased
convexity increases by a smaller percentage than that of a portfolio with
larger convexity.
Hence the need for investors in fixed rate MBS to understand the prepayment
dynamics of a pool backing the respective MBS, the prepayments of which give
rise to the prepayments under the MBS itself.

n Analyzing Interest-Only (IO) MBS

Issuers launch IOs with a view to receiving the excess spread on the mortgage
pool upfront, instead of receiving it on a monthly or quarterly basis. The excess
spread is the difference between the revenue from the mortgage pool WAC on the
mortgages and the expenses under the MBS (servicing), WAC on MBS, and
losses. In an IO, the issuer can receive excess spread at the beginning of the deal
by packaging the stream of excess spread payments into interest-only security. IOs
can also be created by stripping the interest component of an MBS and creating
two complementary securities – one entitled to the principal payments only (PO)
and another entitled to the interest payments only.
Here we discuss only the IOs created on the basis of excess spread. By doing this,
the issuer in effect passes on to investors:
• Risk associated with the generation of excess spread – changes to WAC of the
underlying mortgage pool due to voluntary and involuntary prepayments
(because of borrower default and subsequent repossession and liquidation of
the underlying property), as well as losses.
• Market interest rate risks (repayments on mortgages are heavily influenced by
a changing interest rate environment, and differences in indexes used for
pricing the underlying mortgages and the MBS).
IOs are mainly structured as notional amount securities, which promise to pay
investors specified interest on the outstanding notional amount of securities. The
outstanding notional amount is determined based on a specified formula.
IOs are usually rated securities. However, investors should take little comfort in
the assigned rating, as they address the likelihood of investors receiving a
promised coupon rate on the outstanding notional balance, but not the return of the
principal invested nor the timing of the cash flow receipts, nor the internal rate of
return.

n Banded Swaps and Yield Maintenance Agreements (YMA)

The mortgage pools underlying some MBS deals typically contain at least some
fixed rate mortgage loans. With the MBS notes structured as floating rate
obligations, the issuer must hedge the interest rate risk.
The hedging instrument is a swap between the issuer (the SPV) and a
counterparty, which absorbs the risk of mismatch between the interest received by
the issuer on the underlying mortgages and the coupon paid to the noteholders. For
noteholders this means that the issuer is able to meet its payments irrespective of
the interest paid on the underlying mortgage loans.
The main difference between the familiar plain vanilla interest rate swap and an
interest rate swap embedded in a MBS transaction lies in the notional amount on
which periodical payments are based.
• Under a simple plain vanilla interest rate swap, periodic payments between
the two swap parties are based on a notional amount, which is determined at
the outset and remains constant through time.
• Under an interest rate swap embedded in an MBS transaction, the notional
amount of the swap should ideally be equal to the outstanding collateral. The
difficulty is that this amount varies with mortgage pre- and repayments and
these variations cannot be predicted exactly.
Refer to important disclosures at the end of this report. 116
 ABS/MBS/CMBS/CDO 101 – 16 July 2003

• MBS transactions outstanding as at the end of 2002 deal with the

unpredictability of the notional amount of the swap using two techniques:
periodical swap payments can either vary without restrictions with the actual
outstanding amount of collateral (a yield maintenance agreement), or the
variations can be restricted, as is the case under a banded swap. In a banded
swap, the issuer is fully hedged against interest rate risk, as long as the
outstanding principal remains within a certain corridor, based upon maximum
and minimum conditional prepayment rate (CPR) assumptions. If the
outstanding principal moves out of that corridor, the issuer is exposed to
interest rate risk on the difference between the actual collateral amount and
the upper/lower limit of the corridor. This leads to extra revenues or costs for
the issuer, thus increasing or diminishing excess spread.

n Guarantor-Dependent Ratings
In a number of jurisdictions, insurance is provided by either state or private
insurers to encourage mortgage lending. In Australia, for example, it is a
widespread practice for the mortgage loans in securitised pools to be insured on an
individual basis and on a pool basis. Private insurers also play a role in the Irish,
French and UK mortgage markets. State guaranteed mortgages are often present in
mortgage pools in the Netherlands and Belgium. These additional parties within
the mortgage chain emphasizes the need for a clear understanding of the role of
mortgage insurers in the credit performance, credit rating, and rating volatility for
such MBS.
Australian mortgage pools are generally insured by pool policies (up to a defined
limit) or individual loan policies. As such, any credit-related aspects of their
performance defer to the performance of the respective insurance companies under
the insurance contracts. Changes in the process of mortgage origination, mortgage
product features, and the consequences of increased competition have specific
credit effects on mortgage performance, the effects of which are dampened or
absorbed by the respective mortgage insurers. In this regard, it is important to
follow market changes that could affect the motivation of mortgage insurers to
honor claims or increase their ability to reject claims.
Investors’ exposure to the mortgage insurers is to the extent claims are submitted
and are expected to be honored by the insurance companies. As the credit ratings
of MBS depend on the insurance coverage of the underlying mortgage pools, MBS
ratings may be correlated with the ratings of the insurance providers for a specific
MBS pool. The rating of the insurance providers reflects their ability to pay claims
under the granted insurance policies. Hence changes in the ratings of the claim-
paying ability of insurers could lead to changes in the credit quality of the insured
mortgage pools and the related ratings of the respective MBS.

Refer to important disclosures at the end of this report. 117

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Analytical Framework for Evaluating Mortgage Lenders

Criteria Considerations Aim
Company Story Background To develop an opinion on the company’s
Previous incarnations character and its positioning in the market.
Commenced lending The aim is to understand the key drivers that
differentiate a given lender from the rest.
Target sector
Company philosophy
Market positioning
Brands
Funding To determine a lender’s flexibility in
Warehouse lines financing and how this impacts its origination
Capacity targets and funding exercises.
Advance rates
Whole loan sales to date
Securitisations to date
Ownership To identify the driving forces behind the
Status company, its objective, and the potential for
Major shareholders/‘parent’ support in challenging times.
Parent’s rating
Guarantees/comfort letters
Historical experience
Corporate governance To understand the company’s management
Management structure structure and decision-making process, and
Management experience how it and management’s profitability targets
affect the company’s risk appetite.
Policy-setting procedure for
Risk targets
Credit committees
Profitability targets
Outlook To establish where the lender sees the
Market strategy industry at present and in the future going,
Competition and where it aims to position itself therein.
Market direction
Projected volumes
Origination Target mortgage sector To determine the company’s full range of
Products originated products, and how that corresponds to the
Unique products company’s stated target sector and risk
appetite.
Exclusions
Brands
Cross-sale products
Distribution To identify the company’s distribution model,
Sources its capacity, and viability based on the
Financial intermediary network company’s outlook for the market.
Procurement fees
Origination cost
Packaging procedures
Use of the third-party services
Process capture
Underwriting To determine how the company deals with
Process capture specific factors associated with its target
Number of underwriters market in originating mortgages; in particular
emphasis is placed on the basis for credit
Risk banding decisions.
Credit scoring
Explicit exclusions
Use of the third-party services
Unique features

Refer to important disclosures at the end of this report. 118

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Analytical Framework for Evaluating Mortgage Lenders

Operations Servicing To evaluate the strength of a lender’s
Model servicing platform and its ability to extract
Out-source company value from a mortgage portfolio throughout a
range of economic scenarios.
In-source services
Total book serviced
Costs
Systems
Management information systems
Developments
Unique products
Arrears management To confirm a lender’s ability to effectively
Payment method follow a viable arrears-management process
Structure to manage the risk on the overall book, and
minimize losses.
Work-out accounting
Recovery procedures
Shortfall procedures
Use of third-party services
Unique features
Performance To draw insight from the historical
First-payment defaults performance of the company’s portfolio and
Total 30+ arrears how it compares to the overall market and
the company’s risk appetite.
Total 90+ arrears
Possessions
Sales, and time to sell
Losses
Prepayment To understand the relationships between
Historical rates prepayment rates and various
Drivers macroeconomic factors and how these
impact portfolio performance.
Securitisation Structure To identify any unique features, strengths,
and weaknesses in MBS structures
employed.
Transactions to date To document the evolving capital structure
and execution levels across all of a lender’s
transactions, and compare these to the
broader market.
Source: Merrill Lynch

8.2. Calculus for MBS

As the supply of European MBS increases across the board (meaning countries,
different type of originators, different types of mortgage products, varying bond
structures, etc.) the opportunities for comparisons abound. And such comparisons
are becoming a must from a relative value perspective within the MBS sector.
The quantitative comparisons, though, cannot be done directly given the different
methodologies and conventions that exist in the different markets within Europe.
Hence, the need for adjustments. Below we discuss simple, back-of-the-envelope
adjustments that would help improve the comparability of the different
quantitative characteristics of the mortgage bonds in Europe.
One of the key features of MBS collateral pool is the weighted average LTV
(loan-to-value ratio) of the mortgages at deal closing. The LTV is a key feature of
a mortgage loan or a mortgage pool given its role of main indicator of the default
probability of the mortgagee. That is, the higher the LTV the lower the equity
stake of the mortgagee in the property and the more likely s/he is to hand back the
keys in case of a financial trouble.

Refer to important disclosures at the end of this report. 119

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

So, comparing the LTV of the mortgage pools of different MBS, all other things
being equal, should give investors a good indication of their respective default
probabilities. If only the LTVs were calculated the same way in all MBS!
Unfortunately, they are not. And in order to use them in a meaningful way in
comparing different MBS, investors should make proper adjustments.

n Recalculating MBS LTV

But let’s start from the beginning, the most generic loan-to-value ratio, i.e. the
amount of the mortgage loan divided by the value of the mortgage property. So the
first set of questions is:
• What does the amount of the mortgage loan include? (ratio numerator); and
• How is the value of the mortgaged property determined? (ratio denominator).
When buying a property, the mortgagee faces a number of expenses – tax, broker,
legal, appraisal, environmental evaluation, insurance, etc. fees and expenses.
These expenses can be quite substantial varying from approximately 5% in the UK
to 14%-23% in Belgium. And they can be financed differently: out-of-pocket
expenses for the mortgagee in Germany and Belgium or fully included in the
mortgage loan in the Netherlands, or a combination thereof in France. In the
extreme, high out of pocket expenses associated with the purchase of a property
increase the borrower’s equity in that property even though they are not reflected
in the LTV calculation. On the other extreme are high expenses fully funded by
the mortgage loan – they not only increase the LTV, but also decrease
mortgagee’s commitment to the property – s/he has made no expenses, no up-front
financial commitment to the property in question. Somewhat off the focus of our
discussion, it is worth mentioning that the high up front out-of-pocket expenses of
the mortgagee act as a deterrent to housing turnover, i.e. moving house at the spur
of the moment.
Determination of the property value evokes a series of other questions:
• Who does it? – the lender himself, branch manager, internal evaluation
department, certified specialist; or an outside party, qualified or certified
appraisers.
• How is it done? – using current market value, market comparisons, adjusted
market value (normalized for long-term trends) or liquidation value (fire-sale
assumptions).
Needless to say, the most conservative valuation would be the one performed by a
certified outside appraiser using the liquidation value approach.
Now, let’s summarize: an MBS from the Netherlands has an LTV of 80%, so does
the MBS from Germany. Depending on how the numerator and denominator of
the two LTV ratios are calculated they may not be really the same. In the
Netherlands the loan calculation may include all up-front mortgagee expenses
(assume 10% of the loan) and the property is valued at liquidation value (15%
below current market value), while in Germany the mortgage does not cover any
up-front mortgagee expenses and the property is valued at liquidation value (15%
below current market value).
Therefore, in order to make the two LTVs comparable, one should adjust the
Dutch LTV by decreasing the amount of the loan by 10% and decreasing the value
of the property by 20%. Hence 80/100 = 80% would be after adjustments (80 –
8)/(100 – 15) = 85%. For Germany, the LTV would be 80/(100 – 15) = 94%. So if
the same LTV calculation method is used the MBS in Germany would be riskier
with LTV of 94% compared to the LTV in the Netherlands with 85% LTV. Such
back-of-the-envelop calculations can be applied to the pool LTV of two MBS
from different countries, as well.

Refer to important disclosures at the end of this report. 120

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

n Adjusting MBS LTV for Seasoning

The adjustments, however, do not stop here. The next logical questions to ask are:
• When was the LTV calculated?
• What is the mortgage (or mortgage pool) seasoning?
• Have there been any adjustments made to the LTV reflecting the mortgage
seasoning – the repayment of principal or the changes in the property value?
As before these questions reflect the differences in the practices of different
countries and mortgage lenders, and directly affect the comparability of the
mortgage pools. Let’s look at the extremes again: an MBS from England where
the LTV is determined at the time of the mortgage origination and no further
adjustments are made, and an MBS from Sweden, where the LTV is re-calculated
regularly to reflect the value of the house and the mortgage loan amortization.
And again let’s assume that the mortgage from both countries have an LTV of
80% reported in the second year of their lives. An LTV of 80% in Sweden is the
actual current mortgage LTV. For England, LTV of 80% must be adjusted to
reflect:
• One year of scheduled loan amortization – we assume that this is 20 year level
principal payment mortgage, so 80/20 = 4;
• Prepayment of principal – we assume a CPR of 20%, i.e. 16; and
• Housing price index change – we assume 20% index appreciation year-on-
year.
• Therefore, the LTV of the mortgage loan in England would be (80 – 4 –
16)/(100 + 20) = 50%. If we perform this adjustment on the basis of a
mortgage pool weighted average LTV for two MBS reported with 80% LTV,
it is clear that the MBS from England would be less risky than the MBS from
Sweden, all other conditions being equal. But the analysis does not stop here.

n Reflecting the Housing Market Cycle

We move on to determine the stage of the housing market development and more
precisely, its point in the housing market cycle. The stage in the housing market
cycle is indicative of various factors affecting the mortgage loans and MBS
mortgage pools: solidity of underwriting, competitive pressures, housing price
inflation or deflation, product innovation, interest rate environment. All of the
above affect current and more importantly future mortgagee behavior and
mortgage pool performance.
So, with regards to future MBS performance expectation, adjustment should be
made to reflect the respective mortgage pool LTV as an indicator of the
probability of default and the housing market cycle as an indicator of the
conditions under which that probability will be played out.
Furthermore, it is necessary to take a view on the status of the respective housing
market cycle in conjunction with the respective timing of the mortgage pool cash
flows matched against the required cash flows under the bond in its average life.

Refer to important disclosures at the end of this report. 121

 ABS/MBS/CMBS/CDO 101 – 16 July 2003

Analyst Certification
We, Alexander Batchvarov, Jenna Collins and William Davies, hereby certify that
the views each of us has expressed in this research report accurately reflect each of
our respective personal views about the subject securities and issuers. We also
certify that no part of our respective compensation was, is, or will be, directly or
indirectly, related to the specific recommendations or view expressed in this
research report.

Copyright 2003 Merrill Lynch, Pierce, Fenner & Smith Incorporated (MLPF&S). All rights reserved. Any unauthorized use or disclosure is prohibited. This report has been
prepared and issued by MLPF&S and/or one of its affiliates and has been approved for publication in the United Kingdom by Merrill Lynch, Pierce, Fenner & Smith Limited,
which is regulated by the FSA; has been considered and distributed in Australia by Merrill Lynch Equities (Australia) Limited (ACN 006 276 795), a licensed securities dealer
under the Australian Corporations Law; is distributed in Hong Kong by Merrill Lynch (Asia Pacific) Ltd, which is regulated by the Hong Kong SFC; and is distributed in
Singapore by Merrill Lynch International Bank Ltd (Merchant Bank) and Merrill Lynch (Singapore) Pte Ltd, which are regulated by the Monetary Authority of Singapore. The
information herein was obtained from various sources; we do not guarantee its accuracy or completeness. Additional information available.
Neither the information nor any opinion expressed constitutes an offer, or an invitation to make an offer, to buy or sell any securities or any options, futures or other
derivatives related to such securities ("related investments"). MLPF&S, its affiliates, directors, officers, employees and employee benefit programs may have a long or short
position in any securities of this issuer(s) or in related investments. MLPF&S or its affiliates may from time to time perform investment banking or other services for, or solicit
investment banking or other business from, any entity mentioned in this report.
This research report is prepared for general circulation and is circulated for general information only. It does not have regard to the specific investment objectives,
financial situation and the particular needs of any specific person who may receive this report. Investors should seek financial advice regarding the appropriateness of
investing in any securities or investment strategies discussed or recommended in this report and should understand that statements regarding future prospects may not be
realized. Investors should note that income from such securities, if any, may fluctuate and that each security’s price or value may rise or fall. Accordingly, investors may
receive back less than originally invested. Past performance is not necessarily a guide to future performance.
The bonds of the company are traded over-the-counter. Retail sales and/or distribution of this report may be made only in states where these securities are exempt from
registration or have been qualified for sale. MLPF&S usually makes a market in the bonds of this company.
Foreign currency rates of exchange may adversely affect the value, price or income of any security or related investment mentioned in this report. In addition, investors in
securities such as ADRs, whose values are influenced by the currency of the underlying security, effectively assume currency risk.

Refer to important disclosures at the end of this report. 122