Keep Your Eyes On The Size: The Weekly Publication of High Yield Strategy February 13, 2004 Vol. 2, No. 7

TM
The Weekly Publication of High Yield Strategy February 13, 2004 Vol. 2, No. 7
Contact us at: BIG PICTURE…

LeverageWorld@FridsonVision.com
Learn more at:

Keep Your Eyes on the Size
www.FridsonVision.com The turbulence of the past three weeks has intensified high yield investors’
focus on the risk side of the risk/reward tradeoff. As the Merrill Lynch High
Yield Master II Index's level plunged from 104.803 on January 26, 2004 to
102.303 on February 6, then rebounded to 103.026 on February 12, portfolio
managers shed any remaining illusions that prices would replicate last year’s
nearly uninterrupted ascent.1 (July was the only month of 2003 in which the
index’s weighted-average dollar price declined. By contrast, the last time the
high yield market posted a double-digit return – in 1997 – the index suffered
four month-over-month price drops.)
Deciding to put a greater emphasis on risk is the easy part. The real challenge
consists of deciding which risk to emphasize. In general, both rising interest
rates and increasing credit risk represent stumbling blocks to good performance
for high yield managers. The difficulty is that the standard strategies for
Martin@FridsonVision.com
defending against them are diametrically opposed.
The Historical Evidence

CONTENTS
Exhibit 1 shows returns by rating category for all years available in which
ten-year Treasury yields rose and Moody’s issuer-based default rate declined.2
BIG PICTURE
In each case, the “safer” Double-Bs performed worse. The simple explanation
Keep Your Eyes on the Size.......... 1 of this paradoxical result is that the Double-Bs had greater interest rate
Senior Versus Subordinated sensitivity by virtue of their lower coupons, lower yields, and longer maturities.
Valuation ...................................... 5 Using modified duration -to-worst as shorthand for interest rate sensitivity, we
SECURITY SELECTION find that as usual, the Master II’s Double-B subindex, at 4.76, is currently more
rate-sensitive than its Single-B counterpart, at 3.89. Based on this evidence,
Rich/Cheap: investors who are worried primarily about the possibility of a rise in government
Cadmus Communications / bond yields should underweight Double-Bs relative to Single-Bs.
Alliance Atlantis
Exhibit 1: Total Return by Rating*
Communications ........................... 9
Years of Rising Interest Rates, Declining Default Rates
Focus Issues
Year BB (%) B (%) Difference (%)
↑ Chesapeake Corp, Extended 1994 -2.16 -0.19 -1.97
Stay America, Hollywood 1996 7.98 13.68 -5.69
Entertainment, Rogers Wireless
2003 19.46 25.83 -6.37
Calculations subject to rounding.
↓ Frontier Oil .............................. 13 Source: Merrill Lynch & Co.
SECTOR ALLOCATION
1
These price swings may not connote extraordinary volatility to some readers, but note that the
Industry Value Tracker ............... 16 January 26-February 6 drop translated into an annualized total return of -41.36%. The subsequent
Credit Ratings Value Tracker...... 17 rise represented a 66.50% annualized return.
2
We use returns on the rating subindexes of the cash-pay-only Merrill Lynch High Yield Master
MARKET TIMING Index because no rating breakdown is available on the cash-pay-plus-non-cash-pay Master II prior to
High Yield Sector Value 1996.
Tracker........................................ 19
ODD LOTS...............................20
LEGAL NOTICES ...................21
Copyright 2004 by FridsonVision LLC. All Rights Reserved.

See last page for usage restrictions and other legal terms.
BIG PICTURE…
The indicated strategy is exactly the opposite if the main concern is the
possibility of a rise in default rates. Exhibit 2 shows that in six out of seven
years in which default rates rose while interest rates declined, the less-default-
prone Double-Bs outperformed Single-Bs.3 In summary, high yield investors
should overweight Single-Bs if they fear an interest rate surge and overweight
Double-Bs if they fear a default rate increase.4
Exhibit 2: Total Return by Rating*

Years of Rising Default Rates, Declining Interest Rates
Year BB (%) B (%) Difference (%)
1989 12.24 2.87 9.37
1991 22.38 38.43 -16.05
1995 22.50 18.17 4.34
1997 12.96 12.65 0.31
1998 6.58 2.79 3.79
2000 2.17 -6.24 8.41
2001 11.06 2.59 8.46
Calculations subject to rounding.
Source: Merrill Lynch & Co.
A counter-argument currently being advanced focuses on today’s

exceptionally tight spread between Double-Bs and Single-Bs. At +175 basis
points, the differential is much tighter than its 1980-2003 mean of +239. Note,
however, that at the beginning of 1994, the spread was even tighter, at +161.
The implication was that Single-Bs were rich, but they nevertheless
outperformed Double-Bs by nearly 200 basis points for the year. The
conclusion stands that portfolio managers who perceive the greatest risk on the
interest rate front should avoid the mistake of “playing it safe” by concentrating
in Double-Bs.
Current Assessment
On the face of it, the possibility of a rise in credit risk should not dominate
high yield portfolio managers’ current thinking. Moody’s currently projects a
year-ahead default rate of 3.4%, which would represent a decline from 2003’s
5.19%. The agency’s thinking is supported by the likelihood that U.S. economic
growth will accelerate this year. Currently, the 2004 consensus forecast for
Gross Domestic Product (GDP) growth of 61 economists surveyed by
Distributed by FridsonVision LLC. Bloomberg is 4.6%. The most pessimistic prognosticator in the group sees GDP
Martin Fridson, CEO.
Greg Braylovskiy, Analyst.
matching last year’s 3.1%. Since 1987, GDP growth has measured 3.1% or
greater 11 times and in nine of those years, the default rate has been lower than
The material contained in this publication is
protected by the copyright laws of the United the 2003 rate. Moreover, Exhibit 3 shows a clear improvement in the quality
States of America and by international treaty.
Any unauthorized use, reproduction or
mix of high yield new issues in 2001-2002. Given the statistical link between
distribution is punishable by civil and criminal quality mix in a given year and the default rate three years later,5 the default rate
penalty. See last page for usage restrictions and
other legal terms. should not escalate materially as early as this year.
3
In 1991, the sole contrary case, Single-Bs returned 38.43%. That is a level almost certain not to be
approached in 2004, given the Single-B index’s beginning-of-year, weighted average price of
105.265.
4
There were just two years in which both interest rates and the default rate rose. In 1990, Double-Bs
handily outperformed Single-Bs, while in 1999, Single-Bs beat Doubled-Bs by a small margin.
5
See Martin S. Fridson, M. Christopher Garman, and Sheng Wu, “Real Interest Rates and the
Default Rate on High Yield Bonds,” Journal of Fixed Income (September 1997), pp. 27-34.
Leverage World –February 13, 2004 Copyright 2004 by FridsonVision LLC. All Rights Reserved. 2
BIG PICTURE…
Exhibit 3: Percent of New Issuance Rated B- or Lower
Senior-Equivalent Basis
35
30
25
20
Percent
15
10
0
1983 1987 1991 1995 1999 2003
Source: Merrill Lynch & Co.
Fears of a ramp-up in interest rates, on the other hand, have a plausible basis.
Leverage World does not get involved in rate-guessing, but it is not overstating
the case to say that investors are generally asking when, rather than if, the
Federal Reserve will raise the fed funds rate from 1.0%, a 43-year low, over the
next 12 months. The average forecast of ten-year Treasury rates in the fourth
quarter of 2004, among 60 economists surveyed by Bloomberg, is 4.88%, up
from 4.07% currently.
An Anomalous Result
The greater apparent danger of a rise in interest rates, vis-à-vis a rise in default
rates, argues for favoring Single-Bs over Double-Bs. Very recent experience
raises at least a small red flag, however. During the January 26-February 6 high
yield market downturn, Single-Bs underperformed Double-Bs, -2.01% to
-1.13%. Triple-Cs fared even worse, at -3.25%, despite having shorter duration
than either of the other rating categories. The selloff was precipitated by a
change in nuance in the Federal Reserve’s policy statement that caused investors
to accelerate their estimated timetables for a hike in short-term rates. On the
face of it, a rise in ten-year Treasury rates, from 4.09% on January 26 to an
interim peak of 4.20% on January 29, resulted in a superior return for the more
interest-rate-sensitive (Double-B) sector.
Exhibit 4 sheds light on the anomaly. The table divides the Lehman Brothers
U.S. Corporate High Yield Index into size deciles for the period January 26-
February 6, 2004. Strikingly, total return declines with each step down in issue
size. By contrast, there is no noticeable correlation between issue size and
quality. (The “Average Quality” column is based on numerical equivalents of
the ratings, as explained in the accompanying key.) Apparently, the connection
between ratings and returns is a fallout of the dominant tendency in the period,
which is size-related. The big declines by large, liquid issues are consistent with
reported selling by aggressive investors such as hedge funds.
BIG PICTURE…
Exhibit 4: Average Quality and Total Return by Amount Outstanding
Deciles, January 26 to February 6, 2004
Average Amount Average Average Total
Decile # Outstanding (000's) Quality Return (%)
1 888,601 5.74 -2.86%
2 487,311 5.74 -2.70%
3 384,813 5.48 -2.00%
4 322,517 5.45 -1.80%
5 278,697 5.28 -1.57%
6 242,602 5.46 -1.42%
7 204,914 5.98 -1.37%
8 196,266 5.43 -1.13%
9 165,979 6.31 -0.91%
10 150,000 5.42 -0.76%
Source: Lehman Brothers.
Quality Key
Rating Point Rating
Baa3 1
Ba1 2
Ba2 3
Ba3 4
B1 5
B2 6
B3 7
Caa1 8
Caa2, NR 9
Caa3 10
Ca 11
C 12
Source: FridsonVision LLC.
Conclusion
During the brief downturn of late January to early February, the highest-rated
bonds performed best, despite having the greatest interest rate sensitivity. The
episode, however, represented only a limited test of the historically supported
proposition that Double-Bs can be expected to underperform Single-Bs in an
environment of rising interest rates. The ten-year Treasury yield ended the
period at 4.09%, almost unchanged from 4.07% at the beginning, despite
climbing as high as 4.20% in the interim. We expect that if Treasury yields rise
substantially over the full course of 2004, the comparatively high interest rate
sensitivity of the Double-Bs will prove a decided disadvantage.
Over the very near term, however, the size effect is likely to remain dominant.
For any portfolio managers hoping to capitalize on hedge-fund-generated price
swings lasting just a week or two, the focus for the time being should be amount
outstanding, rather than ratings. Over the full course of 2004, however,
overweighting Single-Bs should prove beneficial unless Treasury yields fall
materially in 2004. If interest rates hold steady, the Single-Bs should prevail on
the basis of yield, without giving up a lot of ground in comparative credit losses.
If rates rise, the record indicates, Double-Bs should underperform because of
their longer duration.
LW
BIG PICTURE…
Senior versus Subordinated Valuation

Swapping within the capital structure of a single company is a tantalizing
proposition because it neutralizes credit risk as a factor in the trade’s success.
One obvious form of intercapital arbitrage worth trying is to capitalize on
discrepancies in the pricing of senior and subordinated debt.
Comparing the values of the two issues of a single issuer would seem
achievable through a straightforward analysis. Newcomers to the market expect
to learn from the veterans a reliable rule-of-thumb for the correct spread.
Disappointed to find that no one makes an across-the-board statement about how
much extra yield the subordinated issue should carry, the novices undertake to
study the spread between selected companies’ senior and subordinated issues.
They hope to document historical averages, enabling them to declare a given
senior bond cheap (rich) versus the subordinated bond if the spread is narrower
(wider) than average. Data gaps and statistical noise tend to thwart such efforts,
however.
A more sophisticated approach is to create a quantitative model of the senior-

versus-subordinated spread. Surely, one would think, the market rationally
rewards subordinated bondholders for the incremental risk of lower recoveries in
the event of bankruptcy. It should therefore be easy to determine when the
market has temporarily turned irrational by overstating or understating the
disadvantage of subordination.
In theory, the risk premium for subordination (lower expected recovery in

event of bankruptcy) should increase as the probability of default increases. At
the extreme, a bondholder should expect to get almost no premium for owning
the subordinated bond of a Triple-A issuer, because there is so little chance that
one’s standing in liquidation will ever matter. On the other hand, if an issuer
has already missed the coupon and the grace period is running out, there is a
huge value in being senior rather than subordinated.
Alas, the empirical evidence stubbornly refuses to conform to the models

rattling around in investors’ brains. To test the very reasonable hypothesis that
spread is a function of default risk, we analyzed the senior-versus-subordinated
yield differentials on a sample of 60 company-matched pairs of nondistressed1
senior and subordinated issues, priced as of February 11, 2004. Our explanatory
variable was a proxy for the market’s assessment of the issuer’s default risk,
namely, the option-adjusted spread (OAS) of the senior issue. The simple
regression generated the following output:
Basis-Point Difference in Spread = 201 – 0.510 OAS Senior
t-statistic: -2.49
P-Value: 0.005
R2: 18.1%
1
Throughout this report on yield differentials between senior and subordinated issues, we exclude
distressed issues. Their pricing is dominated by dollar-price considerations, resulting in
astronomical basis-point spreads that would introduce unmanageable statistical noise.
BIG PICTURE…
The good news is that default risk has a meaningful correlation with the
senior-versus-subordinated spread, with statistical significance at the 95%
confidence level. The bad news is that the sign is negative. Contrary to
hypothesis, the senior-versus-subordinated spread decreases as default risk rises.
(The P-Value indicates that there is no material likelihood that the coefficient is
equal to zero.)
We suspect that the explanation of this anomalous result is yield-chasing on

the riskier issues. For example, current-income-maximizers who observe a 50-
basis-point spread between a Double-B company’s senior and subordinated
issues may accept a similarly slim yield pickup on the subordinated issue of a
Single-B company. Dazed by their hunger for yield, they may ignore the
substantially greater likelihood, in the case of the Single-B issuer, that priority in
bankruptcy will matter one day. The historically low risk-adjusted returns of
Triple-C bonds support the hypothesis that yield-chasing sometimes trumps
fundamental analysis.2
An Alternative Approach
If the market does not reflect one’s model of the senior-versus-subordinated

spread, it seems pointless to us to set up trades on the premise that it will in the
future. We believe the best way to identify senior-versus- subordinated
opportunities is to value the two issues independently, rather than analyzing the
spread between the two.
We tackle the problem by running the 60 pairs in our sample through the
Focus Issues Model,3 a four-variable, multiple regression analysis of the option-
adjusted spread of an individual high yield bond. For each bond in the sample,
the output consists of model-estimated spread and an actual spread. A large
excess of actual spread over estimated spread implies that the bond is
undervalued, while a large excess of estimated over actual implies
overvaluation. We then compare the actual-versus-estimated spread of each
senior issue with the actual-versus-estimated spread of its subordinated
counterpart. The greater is the difference between the two, the more out of line
the valuations are with each other.
Exhibit 1 lists the five most compelling cases of the senior issue representing
the better relative value and the five most compelling cases of the subordinated
issue representing the better value. In the case of American Greetings, for
example, the large gap in coupon between the 6.10s of 2028 and the 11-3/4s of
20084 ordinarily ought to result in a spread differential of 447 – 256 = 191 basis
points. (In fact, the average spread during September 2003 was 154 basis
points, versus a paltry 31 basis points currently.) With the 6.10s trading 67 basis
points wider than their model-estimated spread and the 11-3/4s trading 93 basis
points tighter than their model-estimated spread, however, the actual spread
differential (pickup) is -161 basis points. (Calculations are subject to rounding.)
The conclusion, as explained in greater detail below, under “Featured Pair,” is
that the 6.10s are cheap relative to the 11-3/4s.
2
Over the period 1989-2003, the Triple-C subindex of the Merrill Lynch High Yield Master
produced a lower mean quarterly return than the overall index (2.11% versus 2.29%) with almost
double the standard deviation (7.28% versus 3.74%). Yield-chasing seems a likely explanation of
investors’ acceptance of less return for greater risk than they could earn by settling for a lower
current return.
3
See “Focus Issues Methodology” in Sample Research section of www.leverageworld.com.
4
Investors might expect the maturity difference between the two issues to affect the yield differential
between the two issues, but empirically, we find no effect on spread differential.
BIG PICTURE…
Note that we allow no more than one pair per category for any single
company. Currently, however, that CSC Holding is represented in both the
senior-more-attractive category (7-1/4s of 2008 versus 10-1/2s of 2016) and the
subordinated-more-attractive category (9-7/8s of 2013 versus 7-5/8s of 2018).
Future Reporting on Senior-versus-Subordinated Opportunities
Going forward, we plan to report on swap opportunities between senior and

subordinated issues in alternating weeks. Each report will consist of tables
analogous to those in Exhibit 1 and a more detailed writeup on a “Featured
Pair.” We believe this new recurring feature will provide Leverage World
readers with a regular stream of attractive relative-value trading ideas.
Exhibit 1: Senior versus Subordinated

Pairs with Senior Issue as a Better Value Spread February 11, 2004*
Issue Senior/Subordinated Estimated Actual Difference Pickup
American Greetings 6.1% 8/1/2028 Senior 256 323 67
American Greetings 11.75% 7/15/2008 Subordinated 447 354 -93 -161
MGM Mirage 6.95% 2/1/2005 Senior 253 292 39

MGM Mirage 9.75% 6/1/2007 Subordinated 348 253 -95 -134
CSC Holdings 7.25% 7/15/2008 Senior 274 325 51

CSC Holdings 10.5% 5/15/2016 Subordinated 384 304 -80 -131
Penn National Gaming 6.875% 12/15/2011 Senior 300 338 38

Penn National Gaming 11.125% 3/1/2008 Subordinated 444 358 -86 -124
Mandalay Resort Group 6.375% 12/15/2011 Senior 241 235 -6

Mandalay Resort Group 10.25% 8/1/2007 Subordinated 372 254 -118 -113
Pairs with Subordinated Issue as a Better Value Spread February 11, 2004*
Issue Senior/Subordinated Estimated Actual Difference Pickup
Rogers Wireless 9.625% 5/1/2011 Senior 333 258 -75
Rogers Wireless 8.8% 10/1/2007 Subordinated 305 517 212 287
Navistar International 9.375% 6/1/2006 Senior 409 280 -129

Navistar International 8% 2/1/2008 Subordinated 362 431 69 198
CSC Holdings 7.625% 7/15/2018 Senior 286 246 -40

CSC Holdings 9.875% 2/15/2013 Subordinated 363 496 133 174
Dura Automotive Systems 8.625% 4/15/2012 Senior 356 376 20

Dura Automotive Systems 9% 5/1/2009 Subordinated 368 538 170 149
Hovnanian Enterprises 9.125% 5/1/2009 Senior 320 239 -81

Hovnanian Enterprises 8.875% 4/1/2012 Subordinated 311 358 47 127
Note: In certain cases a comparatively high coupon will cause a senior issue to have a wider expected spread than its subordinated counterpart.
*Calculations subject to rounding.
BIG PICTURE…
Featured Pair: American Greetings’s 6.10s vs. 11-3/4s
If we were to judge solely by the model-generated numbers, the American

Greetings (NYSE: AM) 6.10% senior notes due August 1, 2028 would be 161
basis points too wide versus the AM 11-3/4% senior subordinated notes due July
15, 2008. As of February 11, 2004, the spread of the 6.10% notes was 31 basis
points narrower than the subordinated issue’s. Based on the Focus Issues
Model, a yield giveup of 191 would be predicted.
Exhibit 2 graphs the relationship between the spreads of this senior vs.
subordinated pair during the last six months. Note that for the majority of the
period 11-3/4s traded at a substantially wider spread than the 6.10s, which is in
line with the model’s estimates. However, in late December, the spread began
to narrow and actually turned negative for most of January. While the
subordinated issue has widened since the beginning of February, the senior issue
continues to represent a far superior value.
Exhibit 2: Option-Adjusted Spread for American Greetings notes
Source: Advantage Data.
LW
SECURITY SELECTION…
THE PRESENT 113-BASIS-POINT DIFFERENCE IN SPREADS BETWEEN

THE CADMUS COMMUNICATIONS 9-3/4% OF 2009 AND THE
ALLIANCE ATLANTIS COMMUNICATIONS 13% OF 2009
REPRESENTS THE MOST ATTRACTIVE SWAP POSSIBILITY AMONG
NONDISTRESSED, HIGH YIELD, MEDIA NON-CABLE ISSUES, BASED
ON OUR FINANCIAL VALUATION MODEL. CADMUS
COMMUNICATIONS HAS POSTED SOLID FINANCIAL RESULTS IN
THE LATEST QUARTER AND IS EXPERIENCING SUCCESS IN
GROWING ITS SPECIALTY PACKAGING BUSINESS. MEANWHILE
ALLIANCE ATLANTIS COMMUNICATIONS FACES EXTREME
UNCERTAINTY AFTER THE REORGANIZATION OF ITS
ENTERTAINMENT UNIT AND THE DEPARTURE OF THE BROADCAST
UNIT'S CHIEF.
Media Non-Cable: Cadmus Communications and Alliance Atlantis

Communications
Our Rich/Cheap methodology identifies on a weekly basis an industry pair

that offers potential as a long-short relative value trade.1 (Note that the analysis
does not preclude executing only one side of the proposed swap.) This week’s
analysis concentrates on the Media Non-Cable industry. Exhibit 1 presents two
potentially misvalued bonds with their Focus Issues model estimates and actual
spreads. If we were to judge solely by the model-generated numbers, the
Cadmus Communications (NASDAQ: CDMS) 9-3/4% senior subordinated
notes due June 1, 2009 would be 199 basis points too wide versus the Alliance
Atlantis Communications (NASDAQ: AACB) 13% senior subordinated notes
due December 15, 2009.
Exhibit 1: Focus Issues Model Estimates and Actual Spreads*

(Basis Points)
Potential
Relative
BUY Ratings Price Estimated Actual Value Pickup
Cadmus Communications 9.75% 6/1/2009 B2/B 105.625 448 492
Alliance Atlantis Comm. 13% 12/15/2009 B1/B 112.75 534 379
SELL Difference -86 113 199
* Spreads as of February 11, 2004.
Sources: Advantage Data, FridsonVision LLC.
As of February 11, 2004, the spread of the CDMS notes was 113 basis points
wider than the AACB issue’s. Based on differences in coupon, coverage ratio,
and EBIT (neither issue is rated below B-), a yield giveup of 86 would be
predicted. Exhibit 2 graphs the relationship between the spreads of this industry
pair since February 2002. Note that from February 2002 to October 2003,
CDMS’s issue traded at a tighter spread than AACB’s. Moreover, the two
issues traded at nearly identical spreads as recently as January 2004.
1
See “Rich/Cheap,” Leverage World (August 1, 2003), pp. 1-5.
Exhibit 2: Option-adjusted Spread for CDMS and AACB notes
Source: Advantage Data.
Fundamental Factors
AACB’s bond has experienced significant spread-tightening and price

appreciation since November 2002, with a large part of the rally concentrated in
the August 2003 to November 2003 period. That portion of the tightening
started on August 25, 2003, when the company reported financial results for
2004's first fiscal quarter. While the results were solid, they were only in line
with expectations and some weakness appeared in the performance of the
Entertainment unit. It is likely that investors were impressed by management’s
extremely optimistic guidance for the full year and focused exclusively on the
positives. However, the love affair ended abruptly when in November AACB
reported weaker-than-expected results for the second fiscal quarter of 2004.
Nevertheless, the bond was not punished severely, possibly because of the high
coupon it offers. Meanwhile, CDMS’s bond has stayed essentially unchanged
since June 2003, despite the company’s return to profitability after a difficult
fiscal 2003.
Two major fundamental factors suggest that the divergence in spreads will
disappear or decrease significantly:
· On December 10, 2003, AACB announced a reorganization of its

Entertainment group, which will amount to a nearly complete abandonment
of the company’s non-CSI production business. However, the
announcement did not include financial consequences of the decision.
Instead, the company said that it expected to provide detailed financial
estimates in early 2004. As of February 12, 2004, no such information has
been made public, adding to the uncertainty surrounding AACB’s near-term
results. Furthermore, just this week the company announced the departure
of Mark Rubinstein, head of the company’s Broadcast unit. According to
AACB’s spokeswoman, Rubinstein decided not to renew his three-year
contract after it expired. This is surely a negative sign, considering that the
Broadcast unit has been the strongest performing part of the company.
· Cadmus Communications has delivered improved results in its most recent

quarter, highlighted by positive earnings. The company also increased its
revenues, with strong growth in the Specialty Packaging business offsetting
continued weakness in the core Publisher Services business. Also, the
company successfully continued its international expansion.
Cadmus Communications
Description
CDMS, headquartered in Richmond, Virginia, is one of the world’s largest

printers of scientific and medical journals. The company gets its name from the
mythological figure, Cadmus, who founded Thebes and introduced the
Phoenician alphabet to the Greeks. For the full year ending in June 2003,
CDMS recorded sales of $447 million, while posting a net loss of $54 million.
Currently, the company has a single high yield issue in the market with $125
million outstanding.
Overview
CDMS was formed in 1984 by a merger of a printing company and a graphics

art firm. Since then, Cadmus has remade itself several times through
acquisitions and divestitures. The latest shift in strategy came in 1999, when the
company bought Mack Printing, a producer of academic journals, and sold off
its money losing units. Currently, the company publishes hundreds of titles for
more than 300 publishers. Cadmus has also concentrated on diversifying its
revenue stream by expanding its Specialty Packaging business, which caters to
the needs of business associations and special interest publishers. Furthermore,
the company has escalated its international presence, most visibly through a
joint venture with an Indian firm. After a tough fiscal 2003, the company has
recovered and taken advantage of the improved economy and successes in the
developing parts of its business.
Recent Developments
On January 29, 2003, Cadmus Communications reported solid financial

results for the second fiscal quarter of 2004. Revenues rose by 1% to $113.7
million from the preceding-year period, driven by the rapidly growing Specialty
Packaging business. Earnings per share were strong at $0.33, a marked
improvement over a $0.35 per share loss in the second fiscal quarter of 2003.
The company reported progress in its international expansion with new business
wins in India, Dominican Republic, and Costa Rica. Finally, the company
successfully refinanced its $100 million bank credit facility with new the
maturity set for January 2008.
Alliance Atlantis Communications
Description
AACB, headquartered in Toronto, Canada, is Canada’s largest movie

production and distribution company. The company produced such hits as CSI:
Crime Scene Investigation and Bowling for Columbine.
For the full year ending in March 2003, AACB recorded sales of $606 million,
while posting a net loss of $13 million. Currently, the company has a single
high yield issue with $300 million outstanding.
Overview
AACB was formed when Alliance Communications acquired its biggest

competitor, Atlantis Communications, in 1998. The combined company has
multiple interests in the Canadian media world, including specialty cable
networks, a chain of movie theaters, and a television production business. The
company restructured in 2002 by consolidating its TV and motion picture
production divisions into one unit called Entertainment Group. While the move
was intended to reduce costs and turn the business around, it has clearly not
worked out with nearly 50% of the unit’s workforce being cut in December
2003. While the exact impact is not yet known, the latest reorganization will
almost certainly affect short-term results negatively. Finally, the recent
departure of the head of the Broadcast unit creates additional questions about the
company’s future.
Recent Developments
On November 24, 2003, Alliance Atlantis reported weak results for the second
fiscal quarter of 2004. Earnings per share were substantially below expectations
and negative. Even the strongest-performing Broadcast unit’s results were
merely in line with expectations. Meanwhile, the Entertainment unit was hard
hit with revenues declining by 42% from the preceding-year period. In the
announcement, the management largely blamed the shortfall in the
Entertainment results on the timing of deliveries. However, less than three
weeks later, AACB undertook a reorganization of the unit, including major
headcount reductions.
Media Non-Cable Industry Trends
According to index data provided by Lehman Brothers, the high yield Media
Non-Cable sector returned 21.50% in 2003. This is significantly less than the
total return of the broad Communications sector, at 40.23%. (The sector also
includes the Media Cable and Wireless industries.) In addition, the 21.50%
figure is materially below the overall U.S. Corporate High Yield Index’s 28.97%
2003 return. This relative underperformance partially reflects the Media Non-
Cable industry’s concentration in higher-priced issues, which underperformed
the index in 2003. The Media Non-Cable index started 2003 priced at 93.21, 10
points above the overall index. Furthermore, the industry’s return was hurt by
smaller than expected recovery in the advertising revenues.
LW
Focus Issues
Spread Wider than Estimated
by Financial Statement Data Spread February 11, 2004
Issuer Coupon Maturity Estimated Actual Difference
Boyds Collection 9.000% 05/15/2008 411 599 188
Calpine 8.750% 07/15/2013 311 571 260
Chesapeake Corp* 7.200% 03/15/2005 337 531 194
Extended Stay America* 9.150% 03/15/2008 379 592 213
Friendly Ice Cream 10.500% 12/01/2007 481 737 256
General Binding 9.375% 06/01/2008 435 651 216
Hollywood Entertainment* 9.625% 03/15/2011 351 549 198
KCS Energy 8.875% 01/15/2006 382 572 190
Nash Finch 8.500% 05/01/2008 376 639 263
Pantry 10.250% 10/15/2007 442 660 218
Philippine Long Distance 7.850% 03/06/2007 324 534 210
Pogo Producing 10.375% 02/15/2009 295 582 287
Rogers Wireless 8.800% 10/01/2007 305 517 212
Rogers Wireless* 9.375% 06/01/2008 325 531 206
Rural Cellular 9.750% 01/15/2010 502 748 246
Shopko Stores 9.000% 11/15/2004 379 646 267
Standard Commercial 8.875% 08/01/2005 381 648 267
Unisys 7.875% 04/01/2008 267 467 200
Spread Narrower than Estimated

by Financial Statement Data Spread February 11, 2004
Issuer Coupon Maturity Estimated Actual Difference
Amkor Technology 10.500% 05/01/2009 643 390 -253
Avista 9.750% 06/01/2008 392 191 -201
Frontier Oil* 11.750% 11/15/2009 518 335 -183
Manor Care 7.500% 06/15/2006 289 109 -180
Plains All American Pipeline 7.750% 10/15/2012 336 151 -185
Smurfit-Stone Container 11.500% 08/15/2006 461 193 -268
Addition since last update.
Model Update
The current version of the multiple regression formula is presented here:
Spread = 255.43 + 86.90a + 33.945b – 3.831c – 80.719d
Where:
255.43 is a constant
a = Dummy variable for CCC+ or lower rating (Yes = 1, No = 0)
b = Coupon, expressed without considering percentage sign, i.e., 7.5% = 7.5, not
0.075
c = Coverage, defined as EBITDA divided by interest expense
d = Earnings, defined as log of trailing-twelve-months EBIT in millions of
dollars
Regression Statistics:
Standard Error = 90.12 basis points

R2 = 54.2%
Adjusted R2 = 53.9%
Predictor t-statistic P-Value VIF

Constant 7.90 0.000
a 4.45 0.000 1.1
b 12.18 0.000 1.2
c -3.39 0.001 1.1
d -11.60 0.000 1.2
The analysis indicates that each explanatory variable is significant at the 99% confidence level or
greater. In no case is there greater than a 0.1% probability that variable’s coefficient is equal to 0,
which would signify that the variable has no explanatory power. Low values of variance inflation
factors (VIF) suggest that multicollinearity is not present in this model.
Focus Issues: Why Are They on the List?
The key to exploiting the Focus Issues list is fundamental analysis of factors
outside the historical financial statements. If, in the investor’s judgment, the
factors do not fully justify the disparity between the bond’s estimated and actual
yields, the investor should regard the bond as an opportunity to enhance relative
performance. The following comments provide the basic reason for each of this
week’s additions to and departures from the list.
Issuers with Issues That Entered the:

Yielding-More-than-Estimated List
Chesapeake Corp’s 7.20s widened without any apparent fundamental cause,

even though the High Yield Paper index outperformed the general index during
the last week.
Extended Stay America’s 9.15s joined the Focus Issues without any
apparent fundamental cause, despite the High Yield Lodging group performing
in line with the overall index.
Hollywood Entertainment’s 9-5/8s widened on uncertainty over the future

of its largest competitor, Blockbuster Entertainment.
Rogers Wireless’ 9-3/8s widened on fears that AT&T Wireless Services

might be looking to dispose of its 34% stake in the company.
Yielding-Less-than-Estimated List
Frontier Oil’s 11-3/4s tightened in conjunction with the High Yield Refining
group handily outperforming the high yield index during the past week.
Issuers with Issues That Exited the:

Yielding-More-than-Estimated List
Spread on Spread on
Issuer Coupon Maturity 2/4 2/11 Change
Dillard's 6.875% 06/01/2005 575 417 -158
Dillard's 7.150% 02/01/2007 511 468 -43
Dillard's 7.375% 06/01/2006 559 473 -86
Dillard’s issues gained one to two points in dollar price on news that the
company’s same-store sales rose by 2% in January.
Yielding-Less-than-Estimated List
Spread on Spread on
Issuer Coupon Maturity 2/4 2/11 Change
Pride International 10.000% 06/01/2009 160 325 165
Pride International’s 10s lost 5/8 of a point on news that Moody’s lowered
its liquidity rating to SGL-3 from SGL-2.
LW
SECTOR ALLOCATION…
Industry Value Tracker

The Industry Value Tracker ranks industries within the Lehman Brothers U.S.
Corporate High Yield Index according to the percentage of issues trading wider
(by even one basis point) than the spreads estimated for them by the Focus
Issues model. (For an explanation of this model, see “Focus Issues
Methodology” and “Performance of Focus Issues” in the Sample Research
section of www.LeverageWorld.com.) The underlying premise is that if an
industry has a pronounced concentration of undervalued (overvalued) issues, it
probably reflects a group effect whereby investors are shunning (flocking to) an
industry more energetically than the fundamentals warrant.
Portfolio managers should not attempt to fine-tune their portfolios to minor

differences in rankings. Instead, we recommend that they concentrate on
overweighting or underweighting major industries (those with large numbers of
outstanding issues) that appear near the top or bottom of the table.
Exhibit 1: Industry Value Tracker*
Total # of
Industry % Trading Wider than Estimated Issues
Consumer Cyclical Services 71.43% 21
Cheap Environmental 71.43% 14
Packaging 66.67% 18
Home Construction 61.54% 13
Independent Energy 60.71% 28
Retailers 57.38% 61
Media Cable 57.14% 21
Paper 53.49% 43
Technology 52.94% 17
Consumer Products 50.00% 16
Automotive 50.00% 22
Wireless 45.00% 20
Food/Beverage 43.75% 16
Aerospace/Defense 40.00% 10
Electric 40.00% 15
Oil Field Services 36.36% 11
Health Care 33.96% 53
Industrial Other 33.33% 15
Lodging 30.00% 20
Media Non-Cable 28.57% 28
Rich Chemicals 19.23% 26
Construction Machinery 18.18% 11
Gaming 15.38% 39
Diversified Manufacturing 0.00% 14
* Based on Spreads as of February 11, 2004.
Sources: Advantage Data, FridsonVision LLC, Lehman Brothers.
Update
Among the larger industries, Chemicals and Gaming currently stand out as
candidates for lightening up, with fewer than 25% of their issues trading wider
than their model-estimated spreads. Diversified Manufacturing has achieved the
dubious distinction of a zero percent ratio. On the undervalued side, Consumer
Cyclical Services is the largest industry that stands close to our preferred cutoff
of 75% of issues trading wider than estimated. Since last week, Media Non-
Cable’s ratio has moved from an extreme of 21.43% to 28.57%. As some of its
issues moved from rich to more reasonable valuations, the industry
underperformed the Lehman Brothers U.S. Corporate High Yield Index by a
margin of -0.26% to 0.18%.
LW
Credit Ratings Value Tracker

These rankings are not based on historical yield spreads, but instead on each
rating category’s degree of concentration in issues trading wider than the
spreads estimated by our Focus Issues model.1 (See accompanying conversion
table to determine a bond’s senior-equivalent rating, based on its nominal
rating.)
Exhibit 1: Credit Ratings Value Tracker*

Senior-Equivalent Rating % Trading Wider than Estimated Total # of Issues
BBB 27.27% 88
BB 41.64% 377
B 56.49% 154
CCC 56.25% 16
* Based on Spreads as of February 11, 2004.
Sources: Advantage Data, FridsonVision LLC.
Exhibit 2: Conversion Table

Senior-Equivalent Rating S&P Rating Seniority
BBB Senior
BBB- Senior
BBB
BB+ Subordinated
BB Subordinated
BB+ Senior
BB Senior
BB- Senior
BB
BB- Subordinated
B+ Subordinated
B Subordinated
B+ Senior
B Senior
B- Senior
B
B- Subordinated
CCC+ Subordinated
CCC Subordinated
CCC+ Senior
CCC Senior
CCC- Senior
CCC NR Senior
CCC- Subordinated
CC Subordinated
NR Subordinated
Source: FridsonVision LLC.
Update
Bonds of Single-B companies keep getting more attractive, relative to the

overvalued bonds of crossover names and Double-B issuers. For bonds of
Single-B companies (Exhibit 2, Conversion Table), the ratio of issues trading
wider than their model-estimated spreads rose to 56.49% from 54.19% last
1
See “Focus Issues Methodology” in the Sample Research section of www.LeverageWorld.com.
week, while the comparable ratios for Triple-B and Double-B companies
dropped from 28.41% to 27.27% and from 42.06% to 41.64%, respectively.
Alert readers will also note that the wider-than-estimated ratio for bonds of
Triple-C companies rose from a minority of 43.75% to a majority of 56.25%.
This does suggest some loosening up at the bottom, notwithstanding the very
small sample size involved. Note, though, that many CCC issues are
subordinated bonds of Single-B companies, and therefore in the B Senior-
Equivalent Rating Category in Exhibit 1. Companies rated CCC (or Moody’s
Caa equivalent) at the senior level are extremely default-prone. According to
Moody’s, issuers rated Caa to C at the senior level have a 25% probability of
defaulting within one year, based on 1970-2003 data.
LW
MARKET TIMING…
High Yield Sector Value Tracker

The High Yield Value Tracker is based on the Leverage World Model.1 This
econometric model explains 91% of the historical variance in the spread
between the Merrill Lynch High Yield Master II Index and ten-year Treasuries.
From indicators of default risk, market liquidity, and monetary conditions, the
model estimates the currently appropriate spread. The accompanying diagram
depicts the difference between the model's current estimate and the actual spread
observed in the market, expressed in standard deviations. (One standard
deviation equals 55 basis points.) Divergences of less than one standard
deviation are deemed immaterial
The High Yield Value Tracker's usefulness as a market-timing tool is

indicated by the following average annualized monthly returns, calculated over
the period 1986-2002:
Period Mean Return # of Observations

Undervaluation 19.63% 30
Overvaluation 4.26% 33
ALL PERIODS 8.45% 196
Actual minus Estimated Spread-versus-Treasuries*

Units: Standard Deviations
* Based on February 12, 2004 data.

Sources: Economagic.com, Investment Company Institute, Merrill Lynch & Co., Moody’s Investors
Service.
Update
The high yield sector remains rich relative to its prevailing risk. Over the past
week, the unfavorable gap between the index's actual and estimated spread-
versus-Treasuries increased by a nominal five basis points.
LW
1
For a description of the Leverage World model of the spread-versus-Treasuries, see “Record Short-
Run Overvaluation for High Yield” in the Sample Research section of www.LeverageWorld.com.
ODD LOTS…
Department of Corrections
Default Rates by Underwriter
In our report, “Default Rates by Underwriter” (February 6, 2004), we

incorrectly credited Jefferies as the underwriter of a $10 million floating rate
note priced at the same time as a $71 million fixed-coupon deal that the firm led
for Piccadilly Cafeterias on December 12, 2000. Both the floater and the fixed-
rate issue defaulted, but Jefferies’s default count should have included only the
former. With this adjustment, Jefferies’s actual default rate declines from
10.10% to 9.74%. The firm’s expected default rate is unchanged at 7.29%,
resulting in a gap between the actual and expected default rate of 2.46
percentage points. Jefferies’s ranking among underwriters does not change as a
result of this correction, nor do the investment implications of the report. We
nevertheless regret the error.
Confections of a Fashion Queen
It is our duty to correct errors that appear in Leverage World. We believe,

however, in going beyond the call of duty. To meet that standard, we also
correct errors that appear in other publications.
Accordingly, we take pride in drawing our readers’ attention to the Houston

Chronicle’s coverage of Fashion Week in New York. In addition to speculating
about whether Donald Trump had a comb-over and mentioning a wardrobe
malfunction on a Monique Lhuillier gown with a “(not-too) knitted bodice,” the
reporter commented on the Oscar de la Renta show:
If there was one complaint (other than the excessive use of

fur) about the show, it would be that the collection has an
almost over-the-top Marie-Antoinette-let-them-eat-cake
feeling.1
Insightful though this observation may be, Marie Antoinette never said, “Let
them eat cake.” Jean-Jacques Rousseau attributed the remark to a fictional
princess in his Confessions, published several years before the French queen
allegedly recommended a carb-heavy diet to the peasantry.
LW
1
Clifford Pugh, “Cameras Flash, Fur Flies Outside Shows,” www.chron.com (February 10, 2004).
LEGAL NOTICES…
The material contained in this publication is protected by the copyright laws of the United States of
America and by international treaty. Any unauthorized use, reproduction or distribution is
punishable by civil and criminal penalty.
The material contained in this publication is proprietary and confidential information of

FridsonVision LLC and may not be sold or resold under any circumstances. Except as provided
above or otherwise as expressly authorized by FridsonVision LLC pursuant to the terms of a written
license agreement, no part of the material contained in this publication may be reproduced,
duplicated, copied, disclosed, distributed, transcribed, adapted or transmitted in any form by any
means, electronic, mechanical, magnetic, optical, manual or otherwise.
Leverage WorldÔ and Brought to You in FridsonVisionÔ are trademarks of FridsonVision LLC.
THE MATERIAL CONTAINED IN THIS PUBLICATION IS FURNISHED “AS IS” WITHOUT

WARRANTY OF ANY KIND AND ALL WARRANTIES EXPRESSED OR IMPLIED, ARE
HEREBY EXCLUDED (INCLUDING, WITHOUT LIMITATION, ANY CONDITIONS OR
WARRANTIES RELATING TO MERCHANTABILITY OR FITNESS FOR A PARTICULAR
PURPOSE). FRIDSONVISION LLC SHALL NOT BE LIABLE FOR ANY TECHNICAL OR
EDITORIAL INACCURACIES OR OMISSIONS MADE HEREIN.
The material contained in this publication is subject to change without notice. FridsonVision LLC
assumes no obligation to keep customers informed of any inaccuracies, updates, or other changes or
modifications to any of the material contained in this publication.
FRIDSONVISION LLC SHALL NOT BE LIABLE FOR ANY SPECIAL, INCIDENTAL,

INDIRECT OR CONSEQUENTIAL, PUNITIVE OR EXEMPLARY DAMAGES OR FOR LOST
PROFIT OR REVENUE, IN CONNECTION WITH THE MATERIAL CONTAINED IN THIS
PUBLICATION, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH LOSS. IN NO EVENT
SHALL THE TOTAL LIABILITY OF FRIDSONVISION LLC TO ANY SUBSCRIBER FOR ALL
DAMAGES, LOSSES, AND CAUSES OF ACTION WHETHER IN CONTRACT, WARRANTY,
TORT (INCLUDING, BUT NOT LIMITED TO, NEGLIGENCE), STRICT LIABILITY, OR
OTHERWISE, EXCEED THE TOTAL AMOUNT PAID BY SUCH SUBSCRIBER TO
FRIDSONVISION LLC, IF ANY, FOR THIS PUBLICATION.
If you do not agree with these terms, please delete the publication.

Keep Your Eyes On The Size: The Weekly Publication of High Yield Strategy February 13, 2004 Vol. 2, No. 7

Uploaded by

Copyright:

Available Formats

Keep Your Eyes On The Size: The Weekly Publication of High Yield Strategy February 13, 2004 Vol. 2, No. 7

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Keep Your Eyes On The Size: The Weekly Publication of High Yield Strategy February 13, 2004 Vol. 2, No. 7

Uploaded by

Copyright:

Available Formats

TM

Contact us at: BIG PICTURE…

Learn more at:

The Historical Evidence

Copyright 2004 by FridsonVision LLC. All Rights Reserved.

Exhibit 2: Total Return by Rating*

A counter-argument currently being advanced focuses on today’s

Source: Merrill Lynch & Co.

Senior versus Subordinated Valuation

A more sophisticated approach is to create a quantitative model of the senior-

In theory, the risk premium for subordination (lower expected recovery in

Alas, the empirical evidence stubbornly refuses to conform to the models

Basis-Point Difference in Spread = 201 – 0.510 OAS Senior

We suspect that the explanation of this anomalous result is yield-chasing on

If the market does not reflect one’s model of the senior-versus-subordinated

Future Reporting on Senior-versus-Subordinated Opportunities

Going forward, we plan to report on swap opportunities between senior and

Exhibit 1: Senior versus Subordinated

MGM Mirage 6.95% 2/1/2005 Senior 253 292 39

CSC Holdings 7.25% 7/15/2008 Senior 274 325 51

Penn National Gaming 6.875% 12/15/2011 Senior 300 338 38

Mandalay Resort Group 6.375% 12/15/2011 Senior 241 235 -6

Navistar International 9.375% 6/1/2006 Senior 409 280 -129

CSC Holdings 7.625% 7/15/2018 Senior 286 246 -40

Dura Automotive Systems 8.625% 4/15/2012 Senior 356 376 20

Hovnanian Enterprises 9.125% 5/1/2009 Senior 320 239 -81

Featured Pair: American Greetings’s 6.10s vs. 11-3/4s

If we were to judge solely by the model-generated numbers, the American

Exhibit 2: Option-Adjusted Spread for American Greetings notes

Source: Advantage Data.

THE PRESENT 113-BASIS-POINT DIFFERENCE IN SPREADS BETWEEN

Media Non-Cable: Cadmus Communications and Alliance Atlantis

Our Rich/Cheap methodology identifies on a weekly basis an industry pair

Exhibit 1: Focus Issues Model Estimates and Actual Spreads*

Source: Advantage Data.

AACB’s bond has experienced significant spread-tightening and price

· On December 10, 2003, AACB announced a reorganization of its

· Cadmus Communications has delivered improved results in its most recent

CDMS, headquartered in Richmond, Virginia, is one of the world’s largest

CDMS was formed in 1984 by a merger of a printing company and a graphics

On January 29, 2003, Cadmus Communications reported solid financial

Alliance Atlantis Communications

AACB, headquartered in Toronto, Canada, is Canada’s largest movie

AACB was formed when Alliance Communications acquired its biggest

Media Non-Cable Industry Trends

Spread Narrower than Estimated

The current version of the multiple regression formula is presented here:

Spread = 255.43 + 86.90a + 33.945b – 3.831c – 80.719d

Standard Error = 90.12 basis points

Predictor t-statistic P-Value VIF

Focus Issues: Why Are They on the List?

Issuers with Issues That Entered the:

Chesapeake Corp’s 7.20s widened without any apparent fundamental cause,

Hollywood Entertainment’s 9-5/8s widened on uncertainty over the future

Rogers Wireless’ 9-3/8s widened on fears that AT&T Wireless Services

Issuers with Issues That Exited the:

Industry Value Tracker

Portfolio managers should not attempt to fine-tune their portfolios to minor

Credit Ratings Value Tracker