European Credit Derivatives Strategy

London
22 November 2004

Revisiting Credit Maturity

Curves

• Now having a larger and higher quality data sample across high
grade and high yield 5- and 10-year CDS spreads, we recalibrate
our regression based curve model. We adjust single name curves
for various assumptions about recovery rates

• Distressed and non-distressed credits have very different curve

shapes. We split the regressions into these two categories

• The shape of the credit curve changes through the business cycle. Jakob Due*
We develop a Relative Value model for single name curves, based +44-207 325 7043
on a 3-month rolling data sample [email protected]

Introduction Lee McGint y

+44-207 325 5482
One of the big themes during 2004 was taking views on credit maturity [email protected]
curves. This was in part spurred by the launch of 10-year DJ TRAC-X
Europe in 2003 when index trading boosted liquidity in single name CDS at Mike Harris
+44-207 777 1025
the longer maturity. We have recommended and traded a large number of
[email protected]
curve steepeners and flatteners and forwards both at the index and single
name level throughout the year. Given our strategists’ prediction of a
continuation of the low spread and low volatility environment in 2005 –
making it difficult to generate alpha - we believe that credit curves again
will be a dominating market theme next year.

Chart 1: DJ iTraxx Europe and Crossover, 10y-5y CDS spreads, bp

35
30 DJ iTraxx Europe
DJ iTraxx Crossover
25
20
15
10
5
0
2-Jan 2-Mar 2-May 2-Jul 2-Sep 2-Nov

Source: JPMorgan

The certifying analyst(s) is indicated by an asterisk (*). See last page of the
report for analyst certification and important legal and regulatory disclosures. http://mm.jpmorgan.com
 Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

JPMorgan’s Credit Maturity Curve Model

Credit curves were relatively volatile during 2004, both at the aggregate (index) level
and at the single name CDS level. This provided for numerous attractive trading
opportunities, but also lots of questions such as ‘what drives the slope of a credit
curve?’ or ‘what is the normal slope?’. In The Curve of DJ TRAC-X Europe, 12
January 2004, we introduced JPMorgan’s regression based framework for analysing
credit maturity curves. The main finding of this model is that the slope of a credit
curve primarily is driven by 5 year CDS spreads (which in itself is a proxy for credit
quality), but in a non-linear way. We have always said that this model will require
further refinement as we have access to more data points. We have started to move
through the cycle and therefore, we are now able to announce our first refinements to
the model. The model shows the average relationship between curve slope and spread
level through the business cycle. This summer most individual curves, and certainly
the market on aggregate, steepened. Our existing model tells us that all single name
curves currently trade steep compared to the model’s prediction. This information is,
however, less useful than when curves were closer to the cyclical averages.

Recalibrating the Model

We have recently revisited this subject and have made a couple of changes to the
existing framework. First of all, we have extended the model to also include the
credits within the DJ iTraxx Crossover index, so that we now have a combined high
grade / crossover model. As such, the regression is now based on two year history of
125 DJ iTraxx Europe names and 30 DJ iTraxx Crossover names.

Credit Curves and Recovery Rate Assumptions

The second extension of the model is to adjust the model for different assumptions
about recovery rates. As demonstrated by Rock Bottom Spread analysis in The Curve
of DJ TRAC-X Europe, 12 January 2004, credit maturity curves are impacted by the
assumed level of recovery rates. More specifically, relatively low recovery rates
imply steeper curves with inversion at wider spread levels (because they are pricing
in lower probabilities of default for the same spread). Adjusting single name curves
for such differences therefore better enables us to compare single name curve
steepness across credits, and allows us to incorporate more names into our model.

In our original model from January we excluded insurance companies from the
model as their recovery rates differ from the standard 40% used for corporates. Runs
on Recovery Swaps (see Trading Recovery Rates – Digital Default Swaps and
Recovery Swaps, 19 May 2004 for details) from JPMorgan’s financial traders show
tradable recovery rates between 36% and 50%. We can, however, take the credits
whose assumed recovery rates differ from the standard 40% and transform them into
40% recovery rate equivalent spreads by the use of some simple relationships.

The probability of default is roughly equal to the CDS spread divided by the loss
given default (par less the recovery rate). If we fix the default probability and let the
CDS spread vary, we get:

CDS X CDS y 1 − Rx
(1) : pdef = = Þ CDS x = ⋅ CDS y
1 − Rx 1 − Ry 1 − Ry

2
Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

In Table 1, we show examples of names within DJ iTraxx Europe and Crossover

where the recovery rate assumption is different from 40% together with actual and
40% equivalent CDS spreads.

Table 1: From traded spreads to ‘40% equivalent’ spreads

Assumed Recovery Rates, 5- and 10-year traded CDS spreads, and ‘40% equivalent spreads’, bp

Recovery Rate Traded Spread 40% equivalent spread Multiplication

Company Assumption 5y CDS 10y CDS 5y CDS 10y CDS Factor
RWE AG 50% 19 34 23 41 1.20
ALLIANZ AG 39% 23 31 22 31 0.98
ASSICURAZIONI GENERALI SPA 37% 17 24 16 23 0.94
AVIVA PLC 37% 18 24 17 22 0.94
AXA SA 43% 25 34 26 35 1.04
HANNOVER RE 48% 27 39 31 45 1.14
MUNICH RE 50% 26 38 31 45 1.19
SWISS RE CO 30% 18 26 15 22 0.86
ZURICH INS CO 30% 34 48 29 41 0.86
ALSTOM 30% 375 404 321 346 0.86
BRITISH AIRW PLC 30% 210 247 180 212 0.86
HAVAS 30% 140 173 120 148 0.86
SAS 30% 600 609 514 522 0.86

Source: JPMorgan

If we look at SAS, for example, we can see that with an assumed recovery rate of
30%, the traded 5- and 10-year spreads are 600bp and 609bp respectively. When we
convert these into 40% equivalent spreads we get 514bp and 522bp. This makes
sense intuitively as we have decreased the risk on credit by increasing the recovery
rate. Going through this process and adjusting for differences in recovery rates allows
for a more like for like comparison of credit maturity curve steepness. In Chart 2 we
show how levels and curve slopes are affected by the conversion.

Chart 2: Effect of converting spreads into 40% recovery rate equivalents

10y-5y versus 5y CDS spreads, bp
40
BA 30%
35 HAVAS 30%
BA 40%
30 ALSTOM 30%
25 HAVAS 40%
ALSTOM 40%
20

15
SAS 30%
10

5 SAS 40%

0
0 100 200 300 400 500 600 700

Source: JPMorgan

3
 Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

Functional Forms
With this converted and extended data set in place, we can now run the regressions in
order to obtain the coefficients for the model. A third change we have made from the
original model is to change the functional forms of the regressions. A logarithmic
function provides the best fit to the relatively low spread datapoints whereas a cubic
function provides the best fit to the high spread points. Where these two curves meet,
we join them together (Chart 3).

Chart 3: Regression based curve model; a combination of two functional forms

Log function

Cubic function

Source: JPMorgan

Distressed and Non-Distressed Curves

With the extra liquidity that a 10 year crossover index has provided, we have been
able to observe how different distressed and non-distressed credit curves behave
(even when trading at the same spread level) during a sell-off. We have therefore
split the regressions into two categories. Doing so reveals how distressed credits tend
to flatten and invert at much tighter levels than non-distressed credits. For example,
during the summer 2002 we saw a number of telecoms (such as DT, FT and KPN)
widening out significantly with their curves going into inversion at the 250-300bp
level. In contrast, many high yield (but non-distressed) credits trade at wide spread
levels but with normal upward sloping 5s10s curves.

The reason for these behavioural differences lie in the structure of the companies’
balance sheets. High grade companies tend to have relatively more short term
funding compared to high yield credits. When high grade credits suddenly go into
distress, for example on the back of liquidity or reputational concerns, curves will
quickly invert. As the credit migrates from high grade to high yield and in the process
deals with the event, the balance sheet structure changes. Access and reliance on
short term funding gets reduced and relatively more of the credit’s debt is being
moved toward the longer end of the curve. Once a company has gone through this
process, and trades as a high yield credit, it needs to get really distressed before
inverting (for example, SAS inverted at 750bp). This updated distressed and non-
distressed model is shown in Chart 4.

4
Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

Chart 4: Extended distressed and non-distressed curve model

Based on regression of 155 single name HG and HY 5- and 10-year CDS curves, 2 year data history

40

Non distress
20

-20

-40 Distress

-60

-80
0 100 200 300 400 500 600 700

Source: JPMorgan

While this model is based on a relatively long history of data, our research shows that
the shape of the credit curve in fact is dynamic and moves around as the business
cycle turns. In addition to this long term model, we have also created a relative value
model based on just the latest 3 months of data. This model allows us, first of all, to
track how the shape changes through time. For example, in 2002 when we were in an
economic downturn and default rates were high, we had many high grade credits
going into distress and the aggregate model showed relatively flat curves and early
inversion. At current, the economy is in much better shape, and the few high grade
names which have widened significantly this year have done so on grounds of
speculation of leveraged buyouts (M&S and Sainsbury for example) rather than
liquidity or default concerns. When calibrating the model with the short term 3-
month data sample we see relatively steeper curves and later inversion (Chart 5).

Chart 5: Relative Value Curve Model vs. current single name curves
Based on regression of 155 single name HG and HY 5- and 10-year CDS curves, 3-month history
10y-5y
60
Rhodia Colt
50 Invensys
Corus
M&S
40 Fiat Cablecom

30 DCX
Seat

20 Alstom

10 Publicis
SAS
5y
0
0 50 100 150 200 250 300 350 400 450 500 550 600 650 700
-10

-20

-30

Source: JPMorgan

5
 Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

Secondly, this model allows us to answer the fundamental question of which curves
are relatively steep and which relatively flat today? For example, is a credit trading at
40bp and 58bp in the 5- and 10-year points relatively steeper or flatter than another
credit trading at 100bp and 120bp? Well, if we ask the model, the low spread credit is
trading in line with its peers (same spread level) whereas the higher spread credit is
trading 5bp flatter than its peers at 100bp. The latter is in fact the flattest curve when
we have adjusted for the level of spreads.

We do not wish to create the impression that such steep or flat curves inherently are
mispriced. This curve model does not obviate the need for fundamental credit
analysis: there may well be legitimate reasons for the relative steepness or flatness of
an individual curve. Some companies may systematically trade steep - telecoms have
been a good example of this phenomenon this year. We think that they do so for two
reasons. Firstly, because telecoms have relatively more long dated debt on their
balance sheets. This increases default risk at the longer end of the curve, hence
pushing up 10 year CDS spreads relative to 5 year spreads. Secondly, because there
seems to be general concern in the market that the telecom sector is at risk of
releveraging and push even more debt toward the longer maturities. This could occur
on the back of an increase in acquisitions or M&A. While we therefore think it
possible that telecoms will continue trading steeper than our model’s prediction, a
view on the fair level of steepness involves taking a view on the releveraging story as
well.

Going forward, we will be recalibrating this relative value model on a monthly basis
by using a 3-month rolling data sample.

6
Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

7
Jakob Due European Credit Derivatives Strategy
+44 20 7325 7043 London
[email protected] 22 November 2004

Analyst Certification
The research analyst who is primarily responsible for this research and whose name is listed first on the front cover certifies
(or in a case where multiple analysts are primarily responsible for this research, the analyst named first in each group on the
front cover or named within the document individually certifies, with respect to each security or issuer that the analyst
covered in this research) that: (1) all of the views expressed in this research accurately reflect his or her personal views
about any and all of the subject securities or issuers; and (2) no part of any of the research analyst's compensation was, is, or
will be directly or indirectly related to the specific recommendations or views expressed by the research analyst in this
research.

[Important disclosures]

JPMorgan uses the following rating system: Overweight (we expect the bond to outperform the average total return of the bonds in the analyst’s (or
analyst’s team’s) relevant sector), Neutral (we expect the bond to perform the same as the average total return of the bonds in the analyst’s (or analyst’s
team’s) relevant sector), and Underweight (we expect the bond to underperform the average total return of the bonds in the analyst’s (or analyst’s team’s)
relevant sector).

Analysts’ Compensation: The research analysts responsible for the preparation of this report receive compensation based upon various factors, including
the quality and accuracy of research, client feedback, competitive factors and overall firm revenues. The firm’s overall revenues include revenues from its
investment banking and fixed income business units.
Principal Trading: JPMorgan and/or its affiliates normally make a market and trade as principal in fixed income securities discussed in this report.
Legal Entities: JPMorgan is the marketing name for JPMorgan Chase & Co. and its subsidiaries and affiliates worldwide. J.P. Morgan Securities Inc. is a
member of NYSE and SIPC. JPMorgan Chase Bank, N.A. is a member of FDIC and is authorized and regulated in the UK by the Financial Services
Authority. J.P. Morgan Futures Inc., is a member of the NFA. J.P. Morgan Securities Ltd. (JPMSL) is a member of the London Stock Exchange and is
authorized and regulated by the Financial Services Authority. J.P. Morgan Equities Limited is a member of the Johannesburg Securities Exchange and is
regulated by the FSB. J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated by the Hong Kong Monetary Authority. JPMorgan
Chase Bank, Singapore branch is regulated by the Monetary Authority of Singapore. J.P. Morgan Securities Asia Private Limited is regulated by the MAS
and the Financial Services Agency in Japan. J.P. Morgan Australia Limited (ABN 52 002 888 011/AFS Licence No: 238188) (JPMSAL) is a licensed
securities dealer
General: Information has been obtained from sources believed to be reliable but JPMorgan does not warrant its completeness or accuracy except with
respect to any disclosures relative to JPMSI and/or its affiliates and the analyst’s involvement with the issuer. Opinions and estimates constitute our
judgment as at the date of this material and are subject to change without notice. Past performance is not indicative of future results. The investments and
strategies discussed here may not be suitable for all investors; if you have any doubts you should consult your investment advisor. The investments
discussed may fluctuate in price or value. Changes in rates of exchange may have an adverse effect on the value of investments. This material is not
intended as an offer or solicitation for the purchase or sale of any financial instrument. JPMorgan and/or its affiliates and employees may act as placement
agent, advisor or lender with respect to securities or issuers referenced in this report.. Clients should contact analysts at and execute transactions through a
JPMorgan entity in their home jurisdiction unless governing law permits otherwise. This report should not be distributed to others or replicated in any
form without prior consent of JPMorgan.
U.K. and European Economic Area (EEA): Investment research issued by JPMSL has been prepared in accordance with JPMSL’s Policies for
Managing Conflicts of Interest in Connection with Investment Research. This report has been issued in the U.K. only to persons of a kind described in
Article 19 (5), 38, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2001 (all such persons being referred to as
"relevant persons"). This document must not be acted on or relied on by persons who are not relevant persons. Any investment or investment activity to
which this document relates is only available to relevant persons and will be engaged in only with relevant persons. In other EEA countries, the report has
been issued to persons regarded as professional investors (or equivalent) in their home jurisdiction.
Australia: This material is issued and distributed by JPMSAL in Australia to “wholesale clients” only. JPMSAL does not issue or distribute this material
to “retail clients.” The recipient of this material must not distribute it to any third party or outside Australia without the prior written consent of JPMSAL.
For the purposes of this paragraph the terms “wholesale client” and “retail client” have the meanings given to them in section 761G of the Corporations
Act 2001. Korea: This report may have been edited or contributed to from time to time by affiliates of J.P. Morgan Securities (Far East) Ltd, Seoul branch.

Revised November 12, 2004

Copyright 2004 JPMorgan Chase & Co. All rights reserved. Additional information available upon request.