ABN-Amro - Mobile Networks - W

250 Bishopsgate
London EC2M 4AA, UK

Tel: + 44 20 7678 8000
Fax: + 44 20 7678 7353
Industry
Pan-European Telecoms October 2002
Pan-European Telecoms
3G Tsunami: the revolution begins Jamie Mariani
+44 20 7678 0243
jamie.mariani@uk.abnamro.com
Rodney Sherrington
+44 20 7678 1610
rodney.sherrington@uk.abnamro.com
Marketing Analyst
THE AMERICAS EUROPE AND AFRICA ASIA PACIFIC Simon Carrington
New York United Kingdom Hong Kong +44 20 7678 0395
ABN AMRO Inc. ABN AMRO Equities (UK) Ltd ABN AMRO Asia Ltd simon.carrington@uk.abnamro.com
Park Avenue Plaza 250 Bishopsgate 38th Floor, Cheung Kong Center
55 East 52nd Street, London EC2M 4AA, UK 2 Queen’s Road Central, Hong Kong
6th Floor Tel: +44 20 7678 8000 Tel: +852 2700 3000
New York, New York 10055 Fax: +44 20 7678 7353 Tlx: HX 74111
Tel: +1 212 409 1000 Fax: +852 2700 5000
Netherlands
Fax: +1 212 409 5224
ABN AMRO Bank N.V. Auckland
Boston Gustav Mahlerlaan 10 ABN AMRO Equities (NZ) Ltd
ABN AMRO Inc. 1082 PP Amsterdam +64 9 358 7500
+1 617 753 6900 The Netherlands Bangkok
Chicago Tel: +31 20 628 9393 ABN AMRO Asia Securities Public
ABN AMRO Inc. Fax: +31 20 628 7740 Company Limited
+1 877 711 1781 Brussels +66 2 679 5790
Houston ABN AMRO Bank N.V. Bombay
ABN AMRO Inc. +32 2 546 0460 ABN AMRO Asia Equities (India) Ltd
+1 832 681 7100 Bucharest +91 22 285 4363
Mexico ABN AMRO Bank (Romania) S.A. Jakarta
ABN AMRO Securities (Mexico) SA +40 1 2 020 400 PT ABN AMRO Asia Securities (Indonesia)
+52555 2577800 Cairo +62 21 515 1880
San Francisco ABN AMRO Delta Securities (Egypt) Kuala Lumpur

ABN AMRO Inc. +20 2 3317480 ABN AMRO Asia Equity Research
3G Tsunami: the revolution begins

+1 415 659 3086 Copenhagen (Malaysia) Sdn Bhd
Alfred Berg Bank A/S +60 3 2166 8213
Sao Paulo
Banco ABN AMRO Real +45 33 961000 Makati City
+55 11 31742600 Dublin ABN AMRO Asia Securities (Philippines) Inc.
ABN AMRO Stockbrokers (Ireland) Ltd +63 2 867 7800
+353 1 609 3700 Port Louis
Frankfurt SBM ABN AMRO Asia Securities Ltd
ABN AMRO Bank (Deutschland) AG +230 202 1473
+49 69 2690 00 Seoul
The 3G growth opportunity
ABN AMRO Asia Limited
Helsinki
Alfred Berg Finland Oy Ab +82 2 2131 6400
exists - in voice, not data,
+358 9 228 321 Shanghai services
Istanbul ABN AMRO Asia Ltd
ABN AMRO Bank NV +86 21 5049 6333
+90 212 2938802 Singapore Network capacity could grow
Johannesburg ABN AMRO Asia Securities
ABN AMRO Bank NV (Singapore) Pte Limited five-fold
+27 11 685 2000 +65 536 7688
Madrid Sydney
ABN AMRO Equities (Spain) S.A. SVB ABN AMRO Equities Australia Ltd
+61 2 8259 5000
58% price cuts may deliver
+34 91 423 6900/82
Milan Taipei 30% ARPU growth by 2006
ABN AMRO Bank N.V. ABN AMRO Bank NV
+39 02 722671 +886 2 25037888
Oslo Wireless stocks to rerate,

Alfred Berg Norge A/S
+47 22 005000 after initial period of
Paris
ABN AMRO Securities (France) S.A.
readjustment
+33 1 56 21 5000
Stockholm
Alfred Berg Fondkommission AB
mmO2 and Orange offer
+46 8 7235800
greatest long-term upside
Warsaw
ABN AMRO Securities (Polska) S.A. potential
+48 22 573 0500
October 2002
3G-tsnunami-cov 1 15/10/02, 12:03 pm

Industry
INDUSTRY DEVELOPMENTS > CORPORATE STRATEGY
Mobile Networks - W. Europe
3G Tsunami: The revolution begins 16 October 2002
Longer-term flat rate access plans and pre-to-post-paid migration

will drive structurally higher ARPUs and increased cash flow per sub. Jamie Mariani
jamie.mariani@uk.abnamro.com
In the interim price reductions will put pressure on margins. We +44 207 678 0243
maintain a neutral stance and highlight mmo2 (Buy from Add) and Rodney Sherrington
rodney.sherrington@uk.abnamro.com
Orange (Buy) as preferred plays. +44 207 678 1610
Summary of recommendations
New Recomm. Old Recomm. Price Target
Mmo2 Buy Add 60p
Orange Buy Buy €7.0
TIM Hold Hold €4.1

TEM Hold Hold €6.0
Vodafone Add Add 105p Telecom Networks
Source: ABN AMRO
W. Europe
Europe-Ds Telecom Services
605.74
The European cellular equity story has changed. While stale bulls
continue to cling to the 'blue sky' mobile data service vision, mobile voice
communication is the killer application. We believe wireless networks will
account for 30%-40% of originated traffic (from 10%-15% today),
structurally shifting ARPU by up to 30% by end 2006.
New entrant competition: We anticipate that H3G will adopt bucket

pricing plans throughout its operations in Europe (similar to VoiceStream in
the US). Aggressively priced bundles charged at a flat rate fee could result Sector Performance
in a structural shift in European cellular's share of GDP.
Aug-99 Aug-00 Aug-01
In Europe, 3G capacity enables voice substitution. We believe 3G 2435
will increase available capacity by up to 5x in some markets. As wireless

1935
voice prices decline towards a 50% premium to underlying fixed line prices,
1435
we expect a material displacement of volume from fixed to mobile.
935
Ahead of rising ARPU, lower prices will hit profitablity, due to

435
declining gross telephony margins. Ironically, operators with
disproportionate exposure to the corporate market could suffer most, as
they cannabalise higher-value customers.
(1M) (3M) (12M)

Prior to CYQ3 results we continue to prefer Orange (Buy) and mm02 Absolute -4.9% -6.6% -21.0%
(Buy from Add). We have reduced our price targets for TIM (€4.1 from €4.4) Source: Datastream
and TEM (€6 from €6.7).
Please refer to terms relating to the provision of this research at the end of the document.
ABN AMRO UK
250 Bishopsgate
London
EC2M 4AA
http://www.abnamroresearch.com
United Kingdom
In Brief…
What's Changed
Performance. European wireless stocks have outperformed strongly since we last
looked in detail at the subsector (European Cellular Bear Trap dated 9 July 2002).
Industry restructuring, strong Q202 results and scope for even better numbers at the
Q3s have underpinned this achievement. Our preference for beta has been partially
validated with strong relative performance from Orange (Buy), Vodafone (Add) and
mmo2 (Buy from Add). To the end of the Q302 results season, we expect continued
strong share price movements. Beyond November the key management challenge will
be to turn volume into value.
Chart 1 : Relative share price performance (09/07/02 to date)

35%
Top picks
25%
19%
15% 10%
6% 7%
4%
5%
-5% FTSE EURO DJ Stoxx 50 FTSE Global TEM TIM mmo2 Vodafone Orange
100 Telco Services
-15% -6%
-15%
-25% -20%
Percentage change
Source: ABN AMRO
Capacity and competition. The deployment of 3G networks throughout Europe will

increase capacity by up to 5x (with Italy most affected). Contrary to consensus
expectations (which focus on the long-promised data wave), we suspect that H3G
could introduce “bucket” pricing plans. It is our core belief that as cellular prices move
to a 50% premium to underlying fixed line prices, mass substitution could occur. We
believe that wireless networks will account for 30%-40% of originated traffic (from
10%-15% today), structurally shifting ARPUs by up to 30% by end 2006.
Unfortunately changing business model is unlikely to frictionless. We expect a
minimum 300-500bp contraction in margin.
Recommendations. We have upgraded mmo2 to Buy (from Add) and increased our
fair value estimate to 60p, reflecting recent consolidation in the German market and
the prosect of further restructuring. The Buy recommendation on Orange is
maintained, with a price target of €7. Price targets for the southern European
operators have been lowered to €4.1 and €6.0 for TIM and TEM respectively. We are
leaving our view on Vodafone unchanged (Add, 105p price target).
MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

2
E XE CUT IVE S UMMAR Y
A C T I O N P O I N T S
■ Ahead of CYQ302 results we prefer Orange and mmo2. This is framed

with reference to our core belief that in the event of (1) industry
restructuring, (2) strong Q3 (over Q2) ARPU expansion higher-risk stocks
will outperform.
■ During the medium term (into 2003), we would consider a change in

stance, reflecting our core belief that a change in business model (to
bucket pricing from tariff plans) could dramatically contract margins
(300-500bp). Operators with corporate exposure (eg Vodafone) look
particularly exposed. Longer term we continue to prefer the higher-beta
plays, recognising the potential of F2M substitution.
Issues to Consider
The mobile data services vision is increasingly bankrupt. The core problem
facing Europe’s wireless operators is simple, identifying opportunities large enough to
drive top-line growth today’s absolute level. In our view, data services should be
viewed as an incremental bonus. Disciplined capital investment should require focus
on revenue streams, where sufficient demand exists to validate investment.
Fixed to mobile substitution offers a compelling second leg of growth. In the

short term, pre-to-post-paid migration is the most important driver of the top line.
Our forecasts imply 4%-8% blended ARPU growth from this dynamic over the next
four years. During the medium term we expect wireless networks to grab 30%-40%
share of outgoing total traffic by year-end 2006, driven by price cuts of about 50%-
60%. This could structurally shift ARPU upwards by around 30% by year-end 2006.
Bucket pricing plans facilitate displacement and are superior. US operators

have maintained high ARPUs by not targeting lower value prepaid subs, and through
the introduction of bucket plans. The US model generates higher EBITDA per
subscriber. Similar to the impact VoiceStream had on the US, H3G could be the
catalyst in Europe for driving voice substitution through flat rate access.
3G provides capacity to drive wireline volume displacement. European 2G

capacity constraints have prohibited the introduction of bucket pricing plans. We do
not buy into the “3G is data” proposition (3G is capacity). We estimate that W-CDMA,
using a conservative 20% increase in spectral efficiency, results in a 1.7x-2.8x
increase in capacity. In densely populated areas, capacity could increase by up to 5x.
Switching between business models will cause friction. In our view (should
bucket pricing plans emerge), a margin contraction is inevitable. Dramatically falling
prices will drive structurally lower gross telephony margins. Ironically, operators with
exposure to the corporate market could be the worst affected. Notwithstanding this
uncertainty, longer-term investors should view fixed to mobile substitution as a
potentially unique opportunity. We cannot think of any other sector in the wider
economy that offers such easily identifiable growth.

3
Contents
I NDUST RY DY NA MI CS
Data services dream 7

The mobile data services vision appears increasingly bankrupt. Global SMS
revenue amounted to about US$13bn (approximately £8.7bn) in 2001,
contributing a mere 4% to total global mobile revenue of nearly US$300bn
(approx...
Introduction 7
MA CRO DY NA MI CS
Fixed to mobile substitution 21

In the short term, pre-to-post-paid migration is the most important driver of
revenue growth. We believe high prepaid call charges have stunted usage. Our
forecasts imply about 4% annual growth in blended ARPU over the n...
Introduction 21
I NVES T MENT VIEW
Bucket pricing plans 45

Despite intense competition, US operators have maintained high ARPU levels by
not targeting lower-value prepay subscribers, and through the introduction of
“bucket plans.” US operators have increased the number of free m...
Introduction 45
Bucket plans generate higher ARPU and usage 47
Average yield based on actual minutes used 51
SE CT OR DY NA MI CS
3G as the facilitator 59
In our view, European 2G capacity constraints have prohibited the introduction of
bucket pricing plans (and by association fixed-to-mobile substitution). While 3G
may facilitate the use of high-speed applications, the re...
Introduction 59
NEWS HIGH LIGH T
Fixed-line tariff rebalancing 71

Fixed-line tariff rebalancing is a significant topic that merits more space than we
have in this note (hence our intention to revisit it at a later date). However, we
must recognise that fixed line tariff rebalancing pro...
Background 71
RIS K A NA LYS IS
Economic impact 73
We see between 10% and 25% upside from our existing fair value estimates for
the European wireless community as a result of (1) our pre-to-post-paid
migration thesis and (2) our central case fixed-to-mobile substitution...
Timing 73

4
TAB LE O F CONT ENTS
A PPE NDIX
European 3G licensing conditions 81

In this section we have presented regulatory requirements for rollout by major
European market. License duration ranges between 12 and 20 years, with sharply
divergent coverage requirements by country. In Korea, we note...
Technical glossary 82
Erlang 82
TE AM
Acknowledgements 83

5
This page has been left intentionally blank

6
I N D U S T R Y D Y N A M I C S
Data services dream
The mobile data services vision appears increasingly bankrupt. Global SMS revenue
amounted to about US$13bn (approximately £8.7bn) in 2001, contributing a mere
4% to total global mobile revenue of nearly US$300bn (approximately £200bn). In
the E5 (France, Germany, Italy, Spain and the UK), we estimate mobile revenue of
€75bn was generated during 2001. The core problem facing Europe’s wireless
operators is simple, identifying opportunities large enough to drive top-line growth
from this absolute level. Contrary to some commentators’ expectations, we do not
believe that MMS will materially affect European operators’ top line during the next
24 months. Disciplined capital investment should require operators to focus on
identifiable revenue streams. In our view data should be viewed as an incremental
“bonus.”
Introduction
In this section of the note we present:
■ the size of the E5 voice and data market (€75bn and €6bn, respectively);
■ the absolute revenue required to generate a material upside surprise (€9bn);
■ the validity of operator-articulated targets for data as a percentage of revenue;
■ our view on the sustainability of the low volume/high prices of SMS; Data services look
increasingly like the ‘better
■ the problem of scale – why MMS will struggle to drive growth; mousetrap’ nobody wants
■ the potential cannibalisation of voice/SMS by MMS; and
■ a preliminary assessment of the margin/capex impact of take-up.

7
IN DUST RY D YNAMI CS
European revenue
In the E5, we estimate that mobile revenue of €75bn was generated during 2001, Wireless revenue of €75bn
was generated during 2001
accounting for about 1.1% of GDP. The problem facing Europe’s major operators is
in the E5
clear – identifying opportunities large enough to generate material growth from this
absolute number.
Table 1 : E5 total revenue (€m) calendarised

1999E 2000E 2001E
France 8,863 12,343 15,046
% Change 39% 22%
Germany 10,729 15,057 16,480
% Change 40% 9%
Italy 11,671 13,808 15,534
% Change 18% 12%
Spain 6,523 8,107 9,615
% Change 24% 19%
UK 11,975 16,384 18,582
% Change 37% 13%
Total E5 49,761 65,699 75,257 Scale creates growth

difficulties
% Change 32% 15%
Source: ABN AMRO estimate
Data revenue
Relative to the absolute size of the European cellular market, data pales by
comparison. We calculate that data revenue of €6.3bn was posted in 2001,
corresponding to a small 8.3% of the total top line, or nearly 10% of service
revenue.
Major operator revenue outlook

As the absolute size of operators’ revenue line grows further, future growth
becomes more difficult. We calculate that to deliver 5%-10% upside surprise at the
top line, Europe’s major operators would need to generate an additional €4.5bn-
9.0bn to December 2002/March 2003 (see following table), corresponding to the
total top line of the UK IT Hardware sector.
Table 2 : Revenue by major operator

Unit Revenue 5% Upside 10% upside
2002/03F
mmo2 £m 4,815 241 481
Orange €m 16,877 844 1,688
TIM €m 10,648 532 1,065
TEM €m 9,096 455 910
Vodafone £m 29,480 1,474 2,948 Revenue upside would need

Total Revenue €m 88,063 4,403 8,806 to register scale in line with
the top line of the UK IT
Source: ABN AMRO
hardware sector

8
Data as a percentage of revenue

Every major European operator has articulated a target “data as a percentage of
revenue” by year-end 2004/05 (see following table). We have asked the operator
community to clarify the methodology it has employed in deriving these targets (in
terms of GDP or disposable income). The absence of any meaningful response has
led us to openly question the validity and reliability of these targets.
Table 3: Data as a percentage of revenue

% of revenue from data (last
reported) Data as a % of target 2004/05
mmo2 Group 14.6% 24% Data targets guilty of

finger-in-the-air
Orange UK 13.9% 25%
methodology?
TIM Spa 8.5% 16%-20%
TEM Esp 14.6% 25%-30%
Vodafone Group 12.1% 20%
Orange France 8.6% 25%
Source: ABN AMRO
SMS: A lead indicator?

SMS is often cited as the positive lead indicator validating mobile data success. We
acknowledge that SMS has been a genuine smash hit for the European cellular
community. The vast proportion of existing data revenue is SMS-related (see
preceding table 3). Looking forward, there are few reasons why operators towards
the lower end of the usage range should not be able to close the gap with higher
SMS generating operators.
High SMS price hides low volume

However, before rushing to the conclusion that SMS validates the longer-term data SMS high price hides low
network volume
services story, it is worth putting this revenue stream in perspective. Incredibly
high tariff rates hide low network volume. We estimate that an average SMS is
priced up to 450 times higher than an average two-minute voice call, which in our
view is unsustainable (see following table). If in a 3G environment there is no
distinction between a voice or data bit, then tariffs per bit should converge,
suggesting SMS pricing pressure is inevitable.
Table 4 : SMS in perspective

SMS no of characters 160
Data capacity employed (kb/s) 1.28
Ave. price per SMS (€) 0.10
Implied price per kb (€) 0.08
2 minute voice call (length seconds) 120
Data capacity employed (kb/s) 9.6
Total capacity employed (Mb) 1.15
Ave. price for 2 minute voice call (€) 0.20
Implied price per kb (€) 0.0002
SMS per bit pricing multiple (x) 450 SMS cost/unit volume
Source: ABN AMRO priced at a 450x premium

9
Actual capacity used close to 0%

We estimate that SMS traffic is close to 0% of total traffic carried. Even in the SMS capacity used minimal
youth segment (the keenest users of SMS), we understand that SMS traffic
represents less than 0.5% of total traffic. We can demonstrate this with reference
to TIM, the leading Italian operator. To June 2002, TIM generated 17.8bn voice
minutes and 3.8bn messages (see following table).
Table 5 : SMS capacity in perspective

Jun-02
Voice minutes (m) 17,842
Average length (seconds) 60
Capacity employed per second (kb/s) 9.6
Capacity employed per minute (kb/s) 576.0
Total capacity employed (Mb) 10,277
SMS volume (m) 3,826
Ave. no of characters per SMS 160
Capacity employed per SMS (kb/s) 1.28
Capacity employed (Mb) 4.9
SMS as a % of total traffic 0.05% TIM SMS less than 1% of

total traffic
Source: ABN AMRO
We calculate that SMS as a percentage of total traffic accounts for about 0.05% of
usage. While Italy does not “lead the curve” in terms of SMS usage, our example
underlines an important point, namely, the tiny volume contribution of SMS.
Regulator interest in SMS

We also note that OFTEL, the UK regulator, is examining UK SMS termination rates, SMS termination rates may
come under increasing
following a complaint by an unnamed operator. At present, SMS termination rates
scrutiny
are unregulated, leaving operators free to set rates. The four UK operators
currently charge a termination rate of 3p per message. This is approximately 3x the
rate of fixed-line termination charges in the UK, and 30% of voice cellular
termination fees. A risk to forecasts is that the “cost price” of SMS becomes an
important determinant of retail prices as a result of regulatory intervention. We
also highlight the role OFTEL plays as an opinion leader in European regulation.
MMS: The silver bullet?

Multimedia Messaging Service (MMS) allows users to send and receive messages Data bulls continue to
believe in the ‘power’ of
with colour photos, voice, sound and text. Data bulls have suggested that the MMS
MMS
has immense potential (see following chart). According to Mobile Streams by 2008,
MMS volume globally is expected to be double that of SMS.

10
Chart 2 : Monthly mobile messaging volume at year-end, 2002-2008

1 00,00
90,000
80,000
70,000
60,000
50,000
40,000
30,000
20,000
1 0,00
0
Monthly totalglobal Monthly totalglobal

Monthly totalglobal
Source: Mobile streams
Our discussions with investors suggest to us that MMS is still perceived as the silver MMS silver bullet status
provides some hope ...
bullet for European cellular. Following the somewhat unexpected success of SMS,
the market appears to believe that MMS provides the next step on the long path to
build a credible mobile data equity story.
2002/03: An important year for messaging?

MMS was widely launched across Europe during 1H02. Historically, we (incorrectly) ... but ignores the
importance of handset
viewed the potential of MMS as the largest potential upside risk to our short-term
replacement and top-line
forecasts. Today we do not expect a material impact on group revenue for any of absolute scale
Europe’s major cellular operators this year (or next year) from MMS.
The problem of scale – Sha-mail lead indicator?

The J-Phone service Sha-mail is often cited as an example of the success of MMS,
although Vodafone (J-Phone’s owner) does not disclose the top-line contribution.
We understand the following.
■ Historically, J-Phone has charged between €0.12 and €0.33 per message (about Sha-mail is often cited as
example of MMS success
£0.21, at the top end).
■ At year-end (March 2002), J-Phone had 4 million camera phone customers

(about one-third of the subscriber base).
■ At the end of August 2002, J-Phone had 6m Sha-mail enabled handsets (at the
end of August 2001, J-Phone had approximately 1.8m Sha-mail handsets sold).
■ To August, J-Phone had a 12-month period-average number of Sha-mail

subscribers of 3.9m.
■ In early 2002 (at the time of the J-Phone presentation to analysts, 28 February
2002), Vodafone management indicated that the J-Phone picture messaging
service was being used about twice a week.
For simplicity, we have factored in a constant “two messages sent a week” scenario
to calculate the revenue generated from MMS.

11
Table 6 : Sha-mail example

J-Phone subscribers (June 2002) 12.189
J-Phone MMS subscribers (period average) 3.900
Avg MMS usage per week 2.0
Avg MMS usage pa 104
Avg price per MMS (£) 0.21
Avg MMS revenue per user (£) pa 22.1
Avg MMS revenue per user (£) per month 1.8
Total MMS revenue (£m) 86.4
Total J-Phone revenue (£bn) – end of March 2002 7.5
MMS revenue as a percentage of total sales 1.1% ... but MMS revenue is in
Source: ABN AMRO
the noise (1.1%)
We estimate that MMS, on a 12-month run rate, would generate approximately

£86m of revenue per annum using an average price of £0.21 per message. To put
this into context, it is important to recognise the absolute size of J-Phone’s top line
– around £7.5bn at the end of March 2002. In other words on a 12-month run rate,
Sha-mail’s contribution appears to be about 1% of J-Phone’s mobile revenue.
Despite the “success” of Sha-mail, J-Phone posted a 6% decline in ARPU yoy to
June 2002.
Networking effects – the importance of critical mass

A networking effect refers to the self-fulfilling cycle of penetration and usage We expect MMS to follow an
‘S’ curve of penetration
beyond a point of “critical mass” of adoption of a new technology. European cellular
telephony has followed the classic “S” curve of adoption. SMS is a good example of
a product that has also followed an “S” curve. MMS is likely to follow a similar
growth pattern, displaying slow penetration/revenue growth followed by a rapid
increase, as users become more familiar with the new technology.
Critical mass: 30% penetration of base

In our view, Europe’s operators must embed MMS handsets into 30% of subscriber In our view, 30%
penetration will be required
bases to achieve the critical mass required to “kick off” a networking effect of
to drive a networking effect
penetration and usage. This is a crucial point and has profound implications in
terms of timing of P&L impact.
Handset replacement cycles

We estimate that by year-end 2002, Western Europe will have 307m subscribers. Current Western European
handset replacement cycle
Of this base, 30% equates to 92m MMS-enabled subscribers. Our IT Hardware
is running at 34 months
team currently estimates the natural handset replacement cycle to be around 39
months on a global basis. This is skewed by Latin America (Latam), among other
developing continents (see following table). We understand that the Western
Europe replacement cycle is about 34 months (or slightly less than three years).

12
Table 7: Replacement cycle (months)

2000 2001 2002E 2003E
Europe (W & E) 24 31 35 35
- W Europe 24 30 34 34
- E Europe 34 39 48 40
Middle East & Africa 33 32 44 45
Asia/Pacific 34 37 46 47
North America 27 31 33 32
South America 30 37 50 48
Total 28 33 39 39
Source: ABN AMRO IT Hardware equity research team
A three-year handset replacement cycle suggests an annual rate of handset churn Even if 50% of all Western
European handsets sold in
of nearly 33%. Even if 50% of the replacement handsets sold during 2003 were
2003 were MMS, only 17%
MMS-enabled this corresponds to only 17% of Western Europe’s existing subscriber of Western Europe’s base
base owning an MMS-enabled handset by calendar year-end 2003, some distance would be MMS-enabled
below the critical mass level we estimate is required to drive usage/revenue.
Using the 34-month handset replacement cycle, Western Europe would need to MMS critical mass is
reached CY2005
wait until year-end 2004 to reach the point of critical mass in terms of MMS
penetration. This suggests that a material revenue boost will not occur until
CY2005. Bulls of MMS beware!
What about guidance of 18 months?

We have heard some operators talking handset replacement cycles of 18 months
today, ie every one and a half years. We can sanity-check the “reasonableness” of
this claim with reference to total global handset sales. Holding “new” European
sales and ROW handset sales constant, we can identify the implied replacement
handsets sold in Europe and the impact this has on total global handset sales. For
example, if we believe that 2002 and 2003 will both have an 18-month
replacement cycle, we can imply global handset sales of 474m and 559m
respectively (see following table), somewhat ahead of our IT Hardware team’s
forecasts.
Table 8 : Scenario analysis: 18-month handset replacement cycle

2002E 2003E
Suggested replacement cycle (months) 18 18

Replacement cycle implied replacement handset
153 187
sales
Add new sales 18 13
Implied total handset WE sales 171 200
ROW handset sales 303 359
Implied total handset sales 474 559 18-month replacement

cycle would lead to material
IT Hardware forecast 402 471
upgrades of global handset
Variance 72 88 sale forecasts
Source: ABN AMRO
Nevertheless we do not rule out a shortening of the handset replacement cycle in

Western Europe, should operators choose to aggresively subsidise growth.

13
Vodafone – a worked example

Using Vodafone group as a benchmark, we can quickly emphasise the importance Vodafone would need to
generate about £3bn of
of absolute scale. To the end of March 2003, we forecast Vodafone to generate
revenue to generate a 10%
£29.3bn of revenue (consolidated – see following table). To deliver a 10% upside upside revenue surprise.
surprise at the top line, Vodafone would need to generate another £2.9bn in
revenue.
MMS pricing structures vary on a case-by-case basis. The average price appears to
be around €0.40-0.60 per message (about 25p-40p; outliers such as Telenor in
Norway have recently dramatically reduced their price from as much as €1.33 per
message to €0.65).
Table 9 : Vodafone – the problem of scale (£)

Mar-03 Mar-04
Vodafone consolidated revenue 29,480 31,922
Vodafone average subscribers 91.7 95.8
Vodafone consolidated ARPU 321.5 333.4
Vodafone consolidated ARPU (Monthly) 26.8 27.8
10% upside surprise (revenue) 2,948 3,192
MMS Enalbed Handsets 11% 26%

Using an average price of 25p
MMS Subscribers (period end) 10.4 25.2
per message ...
MMS Subscribers (beginning period) 4.0 10.4
Ave. MMS Subscribers 7.2 17.8
MMS ARPU (Monthly) 34.2 15.0
Ave. price per MMS 0.25 0.25
Implied MMS volume per subscriber (Monthly) 136.9 59.9

... implied volume appears
Implied MMS volume per subscriber (Weekly) 31.6 13.8 challenging: 32 messages a
Implied total MMS volume 11,792 12,769 week could prove
demanding
Source: ABN AMRO
In our example we have factored in 11% and 26% of Vodafone’s subscriber base to
buy MMS-enabled handsets by the end of March 2003 and March 2004,
respectively, and an average price of £0.25 per message. Our basic analysis shows
that to deliver a 10% upside surprise to our existing revenue forecast, Vodafone
would need to encourage its initial MMS-enabled subscribers to pay for 31.6
messages a week to the end of March 2003 (or 13.8 messages per week to the end
of March 2004). We believe this level of volume is too challenging.
Best case: 1.5% impact

To fiscal 2004, we suspect that MMS provides at most 1.5% upside to our existing
forecasts. This view is based on three simple drivers:
■ factoring in a generous 26% of Vodafone’s base will buy (and use) MMS-enabled
handsets (implying 25.6m MMS subscribers at the end of March 2004 from
10.4m at the end of March 2003);
■ usage of 2x a week (resulting in 104 messages a year);.and
■ an average price of £0.25,

14
Pulling these drivers together suggests potential incremental revenue of nearly £2.2
per month per subscriber, corresponding to £26 per year. In total, this offers
£469m of additional revenue to the end of March 2004 (about 1.5% of consolidated
revenue). This realistic upside is in the noise (see following table).
Table 10 : Vodafone impact of MMS

Mar-04
Avg MMS subs (m) 18.0
Usage pa per sub 104.0
Total usage (m) 1,875.6
Price per MMS 0.25
MMS ARPU pa 26.0
MMS ARPU (monthly) 2.2
MMS revenue (£m) 468.9
Implied % of revenue 1.5% Realistic upside in the noise
Source: ABN AMRO
Voice cannibalisation
We also note the potential cannibalisation of existing SMS revenue by MMS, and the Evidence to date suggests
that messaging has
potential of MMS to substitute for voice usage. Arguably MMS can be viewed as the
cannibalised existing voice
classic “technology” push innovation, for which little or insufficient demand exists. revenue streams
Should customers not perceive they have a real "new” need for the technology,
they could either:
■ not buy the innovation; or
■ substitute MMS for existing cellular usage.
Arguably evidence to date suggests that data has been generated at the expense of
existing voice revenue streams. In the charts that follow we show that NTT
DoCoMo,1 the widely accepted leader in data service deployment, has not yet
registered an increase in ARPU, and data ARPU appear to have been generated
largely at the expense of voice ARPU. Similarly, Vodafone’s voice ARPU in its key
European territories has continued to decline at a faster rate than blended ARPU
(assuming constant spending patterns across time), suggesting data service
cannibalisation.
1
Our analysis of NTT DoCoMo ignores the impact of marginal investment costs for 2G. In Japan, a
high number of subscribers for available spectrum has led to higher capital costs (increased cost of
cell splitting), hence discouraging aggressive volume expansion. Nevertheless, data’s initial failure to
drive expanding ARPUs for DoCoMo provides a negative benchmark for European investors.
15
Chart 3 : NTT DoCoMo ARPU trend (yen) Chart 4 : Voice/Data ARPU trends CY Q2 02/Q2 01
14,000 CY Q2 02/Q2 01
4%
12,000
2%
10,000 0%
8,000 -2%
-4%
6,000
-6%
4,000
-8%
2,000
-10%
0 -12%
1997
1998
1999
2000
2001
Q1 02
Q2 02
Q3 02
Q4 02
Q1 03
Q2 03F
Q3 03F
Q4 03F
-14%
VOD - UK VOD - VOD - Italy DoCoMo
Germany
ARPU (12m, contract) VARPU DARPU
ARPU VARPU
Source: Company reports Source: Company reports
Notwithstanding this axiom, we do acknowledge that the higher pricing of MMS

relative to existing voice and SMS usage could reduce the risk of cannibalisation. DoCoMo has failed to drive
an ARPU upgrade from data
Table 11 : Relative price, voice, SMS and MMS (€)

Price Premium/Discount to voice
Avg price 2 minute voice call 0.20 0%
Avg price per SMS 0.10 -50%
Avg price per MMS 0.40 100%
Source: ABN AMRO
Margin impact of MMS

So far we have considered the revenue impact of MMS. It is also important to In our view, embedding
handsets requires a
consider the margin implications. As we have noted historically (see Floating at the
material investment in opex
edge, dated 14 January 2002) the process of embedding MMS-enabled handsets
into an operator’s subscriber base requires a substantial investment in opex
(handset subsidies, dealer commissions and marketing expenses).
Subscriber acquisition costs/subscriber retention costs (SAC/SRC) historically have SAC/SRC could increase
between 100-300bp as a
represented 7.0%-45% of an operator’s top line. The top end of this range has
percentage of sales
been provided by the least mature new entrants throughout Europe. Typically, for a
more mature operator, we see SAC/SRC running at 7.0%-20.0% of sales. A major
risk to margins is the cost of (1) acceleration in the handset replacement cycle (ie
volume) and/or (2) increasing handset subsidies (ie unit cost).
(1) Volume
Our views on the prospect of acceleration in the European handset replacement We suspect an increase in
churn is inevitable
cycle are well known. We have long argued that an increase in churn is inevitable,
if the European operator community is to achieve material data-related revenue
growth in the medium term. Our margin thesis has revolved around the concept of
technology migration (stimulating the migration of exiting 2G subscriber to next-
generation terminals).

16
We recognise that the speed of migration will vary on a country-by-country basis, Speed of migration will vary
on a country-by-country
reflecting:
basis
■ the level of market share concentration;
■ the balance sheet strength of the smaller players; and
■ the arrival (or absence) of new entrants.
We continue to believe that Spain and France will be the most benign markets in
terms of competition, with the UK confirming its reputation as the most competitive
market in Europe. Italy, currently a benign market, is a wild card that could see
much fiercer price-based competition looking forward.
In our view, the key metric on which investors should focus to assess our volume Churn becomes a key driver
of margin
thesis is gross additions. For the purpose of simplicity, our modelling of costs
includes “internal migrations” as external churn. While this is technically incorrect,
it does enable us to capture the volume driver of costs. Should the number of the
gross adds accelerate (all other factors held equal), operating costs will increase as
a percentage of sales. We highlight the structural shift in growth that Europe has
(and will) continue to undergo away from “new” growth, to the size and share of
the disconnection pool, driven by churn and SOGA (see following chart).
Chart 5 : German gross adds example
45 16% 35%
40 29%
30%
5.9
35
25% 25%
23% 24% 25%
30
25 19% 20%
14% 15%
20 15%
34.4
15
12.0 10%
3.5 15.2 15.7
10 14.2
2.1
5%
5 1.4 1.4 9.3
5.7 7.3
2.0 2.5 3.7 4.7 4.0
0 0%
1996 1997 1998 1999 2000 2001 2002 2003 2004
Net Adds (m) Disconnections (m) Churn %
Source: RegTP/Group3G
Our modelling incorporates full detail on the trend in gross adds across Europe. A Tracking gross adds crucial
copy of this model is available upon request. Investors requiring further detail of
our views on technology migration should refer to Vodafone – The cost of capital
challenge, dated 8 April 2002.
(2) Unit cost

We note recent comment from the Carphone Warehouse that by late January 2003, Handset retailers have
suggested that handsets
MMS-enabled handsets will be priced for the mass market. The Nokia 7650 is
will need to be retailed
currently priced at €450-500. In 2003, Carphone is seeking to sell camera-enabled below €100
handsets for €72-80 to contract customers. Similarly, first estimates of the unit
cost of the dual-mode 3G Nokia handset were €700-800 (about £500).

17
Our IT Hardware team has suggested an initial unit cost range of €600-700 per 3G Our IT Hardware team
handset. To put this in some context, existing unit costs for 2G handsets in Europe suggests that 3G handsets
range between €250 and €300. (Globally, costs range between €100 and €110, will cost €600-700
reflecting the low-end 2G handsets sold into markets such as China). While we
would expect this cost to fall as scale production commences, it could result in a
sharp short-term spike to subsidies.
Average subsidies at the last reported date have been running €24-350 per Post-paid subsidies
addition. At the (all-important) post-paid level, SACs have been €156-350 currently sit at €156-300
(excluding the Italian experience; see following table). Post-paid subsidies over the
past year have diverged between (1) falling subsidisation of lower-end handsets
and (2) accelerating subsidisation of higher-end handsets (eg the Ericsson T68i).
Table 12 : Major operator SAC trends FY2002/1H02

Vodafone mmO2 Orange
SAC as SAC as SAC as
Contract Prepaid Blended Contract Prepaid Blended Contract Prepaid Blended
(€) % of % of % of
SAC SAC SAC SAC SAC SAC SAC SAC SAC
Sales Sales Sales
UK 184 41 110 7.5% 286 84 156 13.1% 350 32 158 9.0%
Germany 156 24 81 6.6% 254 70 157 17.1% na na na na
Italy 35 35 35 3.0% na na na na na na na na
France na na na na na na na na 206 83 140 10.2%
Source: ABN AMRO
Using a unit cost of (say) €650 would result in a material spike in subsidies if the
operator’s objective is to retail the product below €300 (about £200). This could Average subsidies could
have a profound impact on margin. The implied subsidy would grow to about €350 materially increase to make
3G handset prices attractive
per handset. This is significantly ahead of existing average post-paid handset
to the end user
subsidies, although we note that it is more in line with existing high-end handset
subsidies. The actual margin impact would be driven by the number of gross adds
(volume) connecting to 3G. Our central case is for a 100-300bp reduction in
profitability (purely from an increase in SG&A as a percentage of sales).
A key risk to a positive outlook is that aggressive subsidisation does not result in
significant increases in ARPU (ie value destruction occurs). This risk is most
pronounced in markets such as the UK and Germany (in Europe) given the lower
level of industry concentration and the scope for the present equilibrium to be
destabilised by aggressive new entrants.
Capex impact of MMS

Investors should also be aware of the capex implications of successful MMS Successful deployment of
MMS could shift capex
deployment. GSM networks operate with eight time slots. A typical MMS can take
expectations upwards ...
up to three time slots, cannibalising existing time slots allocated to voice
communication.
While the operator community claim this “success-driven” capex is already built
into its models, we are sceptical given the lack of visibility on (1) MMS subscriber ... although capex would be
targets and (2) existing network capacity utilisation in major urban areas. Hence success-driven …
the successful deployment of MMS could lead to upward momentum in capex

guidance from the major operators, although we note that this spend would be
success-driven.

18
Strategic investor choices ... but we remain sceptical
Hence we encourage investors to sanity-check claims that MMS (or data) is the
silver bullet for the European wireless community. In our view, a fresh Investors should demand a
more disciplined approach
management approach is required. We suspect that operator business models (and,
to capital investment
for that matter, analyst forecasts) based on material contributions from mobile data
in the short term could be fundamentally flawed.
We are not writing off the medium- to longer-term prospects of mobile data. On the Mobile porn, sport and
gambling could still drive
contrary, we recognise the potential for data (as a longer-term driver) to surprise
‘niche’ usage
on the upside. In our view the risk look finely balanced. We can use a basic
portfolio analysis tool to consider the risks.
In our view, mobile data is a highly attractive market, but in an area of low
business strength (see chart below, top right quadrant). Bulls of mobile data could
categorise it as part of “developments,” recently developed services that may have
some future, but require greater investment to achieve that future. More bearish The biggest risk (already
commentators (like ourselves) could categorise data as part of “ego trips,” or mostly discounted in prices)
is that mobile data turns
services that have strong product champions among influential managers, but for out to be an ego trip
which there is little proven demand in the marketplace. The company, because of
the involvement of powerful managers, continues to pour resources into these new
services in the hope that they eventually come good. Hence the risks, today,
appear finely balanced between data following the “Death cycle,” down to becoming
a failure or moving across and then downwards via “Tomorrow’s breadwinners” into
“Today’s breadwinners.”
Figure 1 : The product portfolio
Business strength
High Low
High
Tomorrow’s Developments
breadwinners Sleepers
Ego trips
Market Life Death

attractiveness cycle cycle
Today’s Failures
breadwinners Yesterday’s
breadwinners
Low
Source: ABN AMRO/Drucker (1973)
Given this outlook, we believe that investors should demand a more disciplined Investors should demand a
more disciplined approach
approach to capital investment, until there is greater visibility that data is not a
to capital investment
classic “better mousetraps nobody wants” but actually offers scope to create value
where none existed before.

19

20
M A C R O D Y N A M I C S
Fixed to mobile substitution
In the short term, pre-to-post-paid migration is the most important driver of

revenue growth. We believe high prepaid call charges have stunted usage. Our
forecasts imply about 4% annual growth in blended ARPU over the next five years
from this dynamic. During the medium term, we suspect that consensus models of
revenue growth are fundamentally flawed, assuming 2G economics (lower volume,
low ARPU, low EBITDA per subscriber) will underpin European 3G. It is our core
belief that at a 50% premium to underlying fixed-line prices, mass displacement of
voice minutes and revenue will occur from fixed to mobile networks. In our view,
3G will be characterised by higher-volume, higher-ARPU, higher-EBITDA per
subscriber voice services, similar to the model that prevails in the US. Our central
case assumes mobile grabbing 30%-40% share of outgoing total traffic by year-end
2006, from 10%-15% today, driven by price cuts of about 50%-60%. This view has
a profound implication for European ARPU growth, which could structurally shift
upwards by about 30% by year-end 2006. The key is flawless execution to turn
volume into value.
Introduction
■ the ARPU impact of pre-to-post-paid migration;
■ US volume as a lead indicator for Europe; Fixed to mobile substitution

offers a compelling second
■ the defining characteristics of 3G over 2G economics; leg of growth
■ the relative price of mobile vs fixed;
■ the cross-elasticity of demand of mobile to fixed-line;
■ a sensitivity analysis on the scope for voice substitution by call type and the
impact on ARPU; and finally
■ the risk/opportunity presented by declining termination rates.

21
MA CRO DYNAMI CS
Background
The average European wireless subscribers still only use their handset for about 3.0 European wireless users
still only use 3 minutes a
outgoing minutes a day, or 4.5 minutes including incoming minutes of use (MOU).
day
Across Europe, mobile calls represent 10%-15% of total outgoing call volume. In
the UK there are seven times as many minutes carried on fixed-line networks than
on the UK’s wireless networks (see following chart).
Chart 6 : UK call volume by type (bn minutes)
300
250
200
150
100
50
0
1999 2000 2001
Fixed local calls Fixed Internet calls Fixed long distance calls
Mobile outgoing calls Fixed international calls Fixed to mobile calls
Source: ABN AMRO/OFTEL
At the top line, we believe that Europe’s wireless landscape will be characterised by
three major phases during the next four years, namely:
At the top line, the wireless
■ Phase one: prepaid to post-paid subscriber migration (2002-2006); landscape will be
characterised by three
■ Phase two: material fixed-to-mobile call volume substitution (2004-2006); and major phases
■ Phase three: dramatically declining termination rates (2002-2006).
Phase One: Pre-to-post-paid migration

It is our core belief that European cellular usage has been stunted by high prepaid Prepaid usage has been
stunted by high call charges
call charges. Including the handset subsidy offered on post-paid products, we
estimate prepaid prices are 40%-110% higher than “equivalent usage” post-paid
offers across Europe. In our view, lower call charges have played a key role in
driving materially higher volume in the US market.
Prepaid subscribers as a percentage of the total European base appear to have In Europe, prepaid
subscribers have peaked as
already peaked. Looking forward, we expect four key drivers to cajole Europe’s pre-
a percentage of total
paid subscribers into migrating: (1) handset subsidies, (2) dealer incentives, (3) subscribers
product exclusivity and (4) bucket pricing plans (value for money).
Phase One drives higher usage/blended ARPUs

We retain a firm belief that increasing pre-to-post-paid migration offers a Prepaid subscriber
migration offers a strong
compelling macro driver of volume and blended ARPU. The impact of migration can
macro driver of blended
be identified with reference to a “generic” model. ARPU

22
MA CRO DYNAMI CS
Our generic model is built on Country X. In Country X we have considered Operator

Y.
We have built a generic
model to show the
■ To capture the impact of pre-to-post-paid migration, we have moved operator importance of plan
migration
Y’s post-paid share of overall gross adds from 38% at x+0 (today), to 60% in
x+1 (today + one year).
■ Each new year’s post-paid subscriber connections are assumed to connect to a

lower-priced post-paid plan, reflecting incremental subscriber dilution (we have
factored in a 20% decline pa).
■ In terms of current period (x+0) post-and prepaid ARPU, we have used £40 and
£9, respectively.
In the following chart we show the evolution of subscribers by post-paid plan type We have grouped each
years connections by a plan
(post-paid subscribers joining high-value, Phase One, Phase Two and Phase Three
type, to model migration
plans).
Figure 2 : Generic subscriber evolution by plan type
10.0
9.0
8.0
7.0
Subscribers (m)
6.0
5.0
4.0
3.0
2.0
1.0
0.0
X+0 X+1 X+2 X+3 X+4 X+5 X+6 X+7 X+8 X+9 X+10 X+11
Time (Yrs)
High Value Contract Phase 1 Contract Phase 2 Contract Phase 3 Contract
Source: ABN AMRO
The chart shows the highest value post-paid subscribers (the bottom area) falling in Incrementally we expect
post-paid ARPU to fall
absolute numbers over time. Each year’s connections are assumed to connect to a
new type of plan (Phase One in x+1, Phase Two in x+2 and Phase Three in x+3).
In aggregate, total post-paid subscribers grow to nearly 60% of the total base from
31%. We have then factored in the different (declining) incremental post-paid
ARPUs for new connections (see following chart).

23
MA CRO DYNAMI CS
Figure 3 : Generic ARPU evolution by plan type
60.0 We have then modelled

ARPU by plan type, to
50.0 understand the impact on
blended ARPU ...
40.0
ARPU (GBP£)
30.0
20.0
10.0
0.0
X+0 X+1 X+2 X+3 X+4 X+5 X+6 X+7 X+8 X+9 X+10 X+11
Time (yrs)
High Value Contract Phase 1 Contract Phase 2 Contract
Phase 3 Contract Total Contract Total Blended
Source: ABN AMRO
Together these two charts provide an expected blended ARPU growth rate. Our Together these drivers
provide blended ARPU
analysis shows blended ARPU improving by between 4%-8% on an annual
growth of 4% on an annual
basis to year-end 2005 (see chart above). In the short term, this is the most basis
important driver of growth.
Given the sensitivity of the analysis to the rate of churn, the percentage of gross
adds connecting to post-paid tariff plans and post-paid ARPU decline, a full copy of
generic model is available to investors in the appendix of the note.
Call volume per pop

Our view on the potential for strong growth in wireless volume driven by lower US call volume, driven by
lower prices, lends
prices is lent some credence by existing US wireless/wireline volume. US telecom
credence to our view
volume is significantly higher than existing European volume. We attribute the gap
to the difference subscriber mix and the pricing structures offered (US local calls
are bundled with line rental). For example, in the US fixed local call volume per pop
is 3.5x higher than the UK fixed local volume (see table below).
Table 13 : Major market local call volume per pop pa

Local calls Volumes bn Population min/per pop min/pop/mon
UK 80 60 1,331 111
France 77 59 1,309 109
Germany 125 82 1,516 126
Italy 60 58 1,040 87
Spain 42 39 1,065 89
Netherlands 25 16 1,583 132
Finland 10 5 1,863 155
US 1,353 288 4,698 391
Japan 215 126 1,706 142
Source: ABN AMRO/IDC

24
MA CRO DYNAMI CS
Call volume per line

If we re-run our analysis using a per line metric, we can remove the distortion of
differences in penetration (see following table). On average we estimate US telco
volume is about 6x higher per line than the European average. For example, in the
US, fixed local call volume is 6.5x higher than the UK fixed local volumes per line,
US local call volume per line
or 6.9x the European average. Mobile traffic per line is about 3x European average
6x higher than in Europe
volume.
Table 14 : Major market wireline local call volume per line

No of Lines
Volume bn (m) Minutes/Line Min/Line/Mon
UK 80 84 951 79
France 77 73 1060 88
Germany 125 157 794 66
Italy 60 52 1,157 96
Spain 42 33 1,267 106
Netherlands 25 27 950 79
Finland 10 11 849 71
US 1,345 219 6,133 511
Source: ABN AMRO/Country regulators
In the following table we have summarised the multiple of US minutes per line by In aggregate call volume in
the US is between 2x and
call type relative to the European average by call type per line. The evidence makes
6x the European average
stark reading. For the wireless industry we do not believe that the difference in
usage is primarily attributable to differences in penetration.
Table 15 : US vs European telco volume (2000)

Euro ave. MOU per US ave. MOU per European average
line per line line per month multiple
Local - Euro average 84 511 6.1
National - Euro average 35 315 8.9
International - Euro average 5 11 2.1
Mobile - Euro average 107 284 2.7
Source: ABN AMRO
Our thesis is simple – in the event of subscribers moving from pre-to-post-paid Should pre-to-post paid
migration occur blended
plans (with large bundles), usage should increase. Evidence to date supports a
spend should increase
positive view. We highlight TEM’s Q202 results, which indicated that average usage
from migrated subscribers jumped more than 2x. Applying the US propensity to call
to each of Europe’s major markets results in a significant shift in European wireless
outgoing volume.

25
MA CRO DYNAMI CS
Table 16 : Mobile minutes 2001: applying a US multiplier

Annual mobile US multiplier (x) New annual mobile
volume 2001 (bn) volume (bn)
UK 69.4 2.7 187.3
France 65.6 2.7 177.1
Germany 74.1 2.7 200.1
Italy 49.1 2.7 132.6
Spain 27.6 2.7 74.6
European major market total 285.8 2.7 771.7
Source: ABN AMRO
Readers should recognise that we have not incorporated this volume increase into Material wireless growth
may not necessarily
our bottom-up models for each operator. We merely present the data as a further
terminally damage wireline
perspective. This is important from a wireline perspective, given the scope for the volume but ...
wireless subsector to gain (in terms of volume) from pre-to-post-paid subscriber
migration, but not entirely at the revenue expense of the wireline operators.
Phase Two: fixed to mobile substitution

While local access fees may be largely retained by wireline telephony operators, we ... As 3G is deployed across
Europe, we anticipate large
expect a large percentage of fixed-line voice calls to migrate to Europe’s cellular
F2M volume displacement
networks. This view has a profound implication for European ARPU growth and
fundamental valuation.
2G vs 3G economics
Increasingly, we suspect that consensus models of revenue growth are The assumption that 2G
mobile economics will
fundamentally flawed, assuming 2G economics will underpin European 3G. We
prevail in 3G could be
reject the view that blended ARPU will gently tick along at 3-5% growth pa, coupled fundamentally flawed
with margin progression and capex/sales falling to sub-10%. In our view, European
cellular models that follow this trend ignore the potential capacity for 3G to enable
the mass substitution of fixed-line voice minutes to the wireless network.

26
MA CRO DYNAMI CS
Chart 7 : 2G vs 3G economics
Growth via shift in

European 2G capacity 3G Economics
Economics
Low MOU/low volume Higher MOU/high volume

growth growth
Low ARPU/declining voice Higher ARPU/lower yield

ARPU
High level of market Lower level of market

concentration concentration
Shift in Capacity
2x-5x
High margins Lower margins
Lower EBITDA per Higher EBITDA per

subscriber subscriber
Capex: Capacity and Capex: coverage focused

maintenance
Low NOPAT/IC: good Higher NOPAT/Similar IC:

returns higher returns
Source: ABN AMRO
Elasticity of demand
Historically, a great deal of telco equity research has been dedicated to the
discussion of elasticity of demand, the percentage change in quantity demanded for
Elasticity of demand ranges
a percentage change in price. Analysis of elasticity of demand is obscured by
from negative 0.6 to
penetration growth. The authors of this study have seen estimates ranging from negative 0.9
negative 0.6 to negative 0.9, suggesting a high degree of elasticity at historical
price levels (implying volume growth somewhat compensates for price declines).
Cross-elasticity of demand
From a theoretical perspective, the key determinant of elasticity of demand is the
... but the key is cross
availability of substitutes, which is a key point missed by many industry experts. A elasticity of demand
discussion of elasticity of demand for one good in isolation is flawed.

27
MA CRO DYNAMI CS
In our view, the key factor to focus in on is the cross-elasticity of demand

between European wireless and European wireline (ie At what price will wireless be
preferred to fixed-line communication?). Cross-elasticity of demand refers to the
responsiveness of demand in one product to the change in price of another product.
Substitute goods have positive cross-elasticities. To better understand cross-
elasticity, we have focused our attention on the relative price of cellular to fixed
line.
Fixed vs mobile: A comparative analysis

The average price of a cellular minute of use vs a fixed-line minute of use has fallen We estimate wireless as a
multiple of wireline has
to nearly 3x today from about 6x in 1995 (excluding fixed-line access and
fallen to about 3x from 6x
connection fees; see following chart).
Chart 8 : Comparative price of mobile vs fixed

7.0
6.0
5.0
4.0
3.0
2.0
1.0
-
1996 1997 1998 1999 2000 2001
Euro average
Source: ABN AMRO estimates/IDC
France
In determining revenue, we can either include or exclude access and connection In France, the multiple has
fallen to nearly 3x
charges. Choice of methodology has a significant impact on the implied multiple of
cellular prices to fixed line. For example, in the French market, including access
fees, we calculate a multiple of around 2x vs a relative call price of nearer 3x
excluding access fees. In terms of the relative price of fixed voice calls to a mobile
voice calls, the French market offers a decent proxy for Europe in aggregate (see
following chart, including fixed-line access).

28
MA CRO DYNAMI CS
Chart 9 : France: mobile. price per minute relative to fixed line price per
minute (inc. fixed line access)
3.5
3.0
2.5
2.0
1.5
1.0
0.5
-
1999 2000 2001E
Local National International F2M Ave multiple
Source: ABN AMRO/IDC
The implied multiple of fixed to mobile calls and mobile international calls are close If cellular becomes the
primary point of local
to parity. Should the cellular handset become the primary point of local access for
access, we see significant
the customer, we would view both of these segments as having a very high scope for displacement
propensity to migrate towards mobile.
The revolution begins
Intuitively a 10% reduction in price, when cellular is priced at (say) a 6x premium

to equivalent wireline prices, is not a compelling reason to substitute usage.
However, it is our core belief that as cellular outgoing voice prices move to 1.5x At a 1.5x multiple we
average fixed call prices (ie a 50% premium), mass substitution of minutes and believe a mass substitution
of volume will occur
revenue could occur. Of course the actual premium customers will be willing to pay
will vary on a type-by-type basis (eg local, national etc).
Convenience
We believe that a 50% premium for the convenience of mobility will be accepted. In
our view, the overwhelming convenience and privacy of the mobile phone (relative We think customers will be
prepared to pay a 50%
to fixed) should not be underestimated as a driver of growth. We view this driver as
premium for the
further justification for increased migration from fixed to mobile usage. Remember convenience of mobility
a 50% difference at (say) 1p increases the spend to 1.5p, for the convenience of
mobility, so for a two-minute call 1p more is spent.
The UK – a proxy for Europe; elasticity in focus

Of Europe’s regulators, OFTEL provides the most detail on volume and revenue by OFTEL data enables to
assess the scope for
call type. To assess the potential impact on outgoing voice ARPU throughout
substitution by call type
Europe, we have used the UK as a proxy.
At 1Q02, UK mobile prices were around 3.6x the corresponding fixed line price (see
following table, average fixed line price per minute £0.04, average cellular price per
minute £0.15). To the end of Q102, the UK mobile market generated £1.8bn in
outgoing call revenue, corresponding to a monthly ARPU of £13.2. This includes
connection, access and call charge revenue.

29
MA CRO DYNAMI CS
To get the implied multiple of cellular to fixed line down to 1.5x (a 50% premium to
the underlying fixed line price), we calculate cellular prices must fall by about 60%
(1.5 x £0.04 = £0.06, which is a 50% premium to the fixed price, and a discount to
the current cellular price per minute of about 60%). Excluding any positive revenue
impact from increased volume, we calculate a revised ARPU of £5.6 per subscriber
(see following table).
Table 17 : UK mobile market ARPU (pre- and post-price decline)

Unit Q102
Total UK mobile minutes (all calls, exc termination) bn minutes 12330
Average UK mobile subscribers m 45.6
Minutes of use (MOU) minutes 90.2
Total UK outgoing mobile revenue £bn 1.8
Actual price per minute £ 0.15
% change in price % -58%
ARPU (pre-price change) £ 13.2 58% price reduction

reduces blended ARPU to
ARPU (post-price decline) £ 5.6
£5.6 from £13.2
Source: ABN AMRO
Central case outlook

Our central case is for large-scale substitution of call volume to occur between the
wireline and wireless networks. We believe the future will see most wireless
subscribers paying a flat rate fee for a large bundle of voice minutes. A high flat
rate entry point provides a marketing tool. Customers may never fully consume
their allocated minutes, but the large volume should create the impression of value
in their mind.
To assess the potential impact of fixed to mobile substitution on UK industry

monthly ARPU on a bottom-up basis, we have disaggregated UK call
volume/revenue by type, and taken a view on the scope for substitution. In the
following example we have used OFTEL data for Q102. Our substitution analysis
excludes any revenue for connection or access, which we assume remains
with the wireline operator.
OFTEL tells us that in Q102, total UK non-call revenue was about £1.1bn. Again we When assessing relative
prices, we must consider
have assumed none of this revenue migrates to the wireless network. However,
connection and access fees
methodologically we believe it would be incorrect to ignore the access fee for
comparison of fixed line prices with cellular prices. Hence at a theoretical level we
have to allocate this non-call cost to the fixed line, to get a feel for the price per
fixed-line minute the customer actually pays, when comparing the average fixed
line price per minute with the average cellular price per minute.
The reason we must include the access fee for price comparison is that the fee paid
by the customer is regular in nature, and is not a sunk cost for the customer.
Additionally, as the number of minutes originated on wireline networks diminishes,
the implied price per minute rises, and the value for money proposition
deteriorates. While we have ignored the subject of fixed-line disconnection, we do
recognise the risk displacement of originated minutes presents to the existing
number of fixed lines without DSL emerging.

30
MA CRO DYNAMI CS
Including access fees, we estimate that the implied multiple (of cellular price per
minute vs fixed price per minute) is about 1.5x (ie cellular at a 50% premium to
the underlying fixed line price). Excluding access fees from the relative price
comparison, the price multiple between the two platform grows from 1.5x to 2.6x.
We can calculate this from total fixed call revenue post-substitution excluding
access of £1.5bn, divided by the number of minutes equals an average price per
minute of £0.02, which is 2.6x less than the implied price per mobile minute of
£0.06 post price decline.
Including access fees, total fixed line quarterly revenue equalled £2.5bn in Q102,
divided by the number of minutes equals an average price per minute of £0.04.
This is 1.5x less than the implied price per mobile minute of £0.06 post price
decline.
Our central case for volume migration is based on the following assumptions:
■ 30% of all local call volumes substituting to wireless;
■ 60% of all F2M calls substituting to wireless; Our central case assumes
mobile grabbing around
■ 50% of all national calls substituting to wireless; 35% of total originated call
volume (from 13%)
■ 30% of all international calls substituting to wireless; and
■ 15% of “other” calls, including NTS and “free call numbers.”
This equates to 35% of UK call volume orignated by wireless networks from 13%.
Total call volume in Q102 equalled 97.3bn minutes, of which 13% were mobile-
originated. Our central case grabs an additional 22.2bn minutes, giving total
mobile minutes post price decline of 34.5bn minutes, or 35% of the total market
minutes for the quarter.
We are not suggesting that call volume at Q102 offers a good proxy for call volume
in the absence of fixed-to-mobile substitution looking forward. But given the
transparency of Q102 numbers, we have used the data to get a feel for the scope
for an ARPU shift.
Similarly we don’t want to become too bogged down by the question, What’s the
correct percentage for displacement by call type? The future is by definition
unknown, and hence the question is clearly open to debate at a bottom-up level.
In our view it is more important to focus on the bigger picture – What percentage
of total volume will migrate?, and What price reduction is required to drive this
displacement?
Factoring in a willingness to pay a 50% premium to the underlying fixed line price
in aggregate, we estimate conservatively outgoing voice ARPU growth of 18% This corresponds to
(outgoing voice ARPU £5.6 post price decline, from £13.2, adding incremental ARPU outgoing ARPU growth of
18% (using a price cut of
uplift from volume growth of £10.0, resulting in post price decline ARPU of £15.60,
58%).
which is 18% ahead of £13.2). Outgoing growth of 18% corresponds to 11%
change in overall ARPU. This would represent a significant upside surprise to
consensus expectations. We note comment from British Telecom that to date about
4% of existing volume pa has been cannibalised by mobile.

31
MA CRO
Table 18 : UK Market Outgoing Drivers
Local calls F2M calls National International Other calls Total Fixed Mobile Total All Calls
Q1 2002 OFTEL Statistics
calls calls calls
Starting Point
DYNAMI CS
Call volume bn mins 18.8 3.5 13.6 2.0 47.1 85.0 12.3 97.3
Call revenue (exc. access) £ bn 0.4 0.5 0.4 0.3 0.7 2.3 1.8 4.1
Implied revenue per minute 0.02 0.14 0.03 0.14 0.01 0.03 0.15 0.04
Impact of Subsitution
Mobile price per minute reduction -58%
MO BILE
New mobile price 0.06

Scope for local substitution 30% 60% 50% 30% 15% 26%
Average mobile subscribers 45.6 45.6 45.6 45.6 45.6 45.6 45.6
MOU substituted to mobile (bn mins) 5.6 2.1 6.8 0.6 7.1 22.2 22.2
Implied additional MOU per sub 41.3 15.2 49.9 4.3 51.7 162.4 162.4
NET WO RKS
Revenue lost to Fixed Line 0.1 0.3 0.2 0.1 0.1 0.8
Revenue substituted to mobile 0.3 0.1 0.4 0.04 0.4 1.4 1.4
Mobile Revenue lost due to price cut 1.0
Incremental ARPU uplift 2.5 0.9 3.1 0.3 3.2 10.0 10.0
End Game
-
Call volume bn mins 13.2 1.4 6.8 1.4 40.0 62.8 34.5 97.3
W.
Call revenue (exc. access) £ bn 0.3 0.2 0.2 0.2 0.6 1.5 2.1 3.6
E UROPE
Implied revenue per minute 0.02 0.14 0.03 0.14 0.01 0.02 0.06 0.04
Pre-price decline Post price decline
ARPU 13.2 5.6
Incremental ARPU uplift (from substitution) 10.0
New ARPU (post price decline inc. volume uplift) 15.6
% Change 18%
Difference
Total fixed line call volume 85.0 62.8 22.2
Total mobile call volume 12.3 34.5 22.2
16
97.3 97.3 -
O CT OBE R
Total fixed line call volume 87% 65%

Total mobile call volume 13% 35%
Difference
Total fixed line call revenue 2.3 1.5 0.8
200 2
Total mobile call revenue 1.8 2.1 0.3

Total telco call revenue 4.1 3.6 0.5
Total fixed line access revenue 1.1 1.1 -
Total telco revenue 5.1 4.7 0.5
Total fixed line call revenue (%) 56% 41%
Total mobile call revenue (%) 44% 59%
Total telco call revenue (%) 100% 100%
Total fixed line call revenue 44% 32%
Total mobile call revenue 35% 46%
Total fixed line access revenue 21% 23%
Total telco revenue 100% 100%
Source: ABN AMRO
32
MA CRO DYNAMI CS
UK local calls
Factoring 30% volume substitution at an average cellular price of £0.06 results in 30% local call volume
substitution equates to an
revenue for the quarter of £0.3bn substituting to mobile from fixed line. This
additional 41 MOU per
corresponds to 5.6bn minutes of fixed-line local voice calls substituting to the month
mobile network, or an additional 41 monthly MOU per wireless subscriber per
month (see following table).
Table 19 : UK local call market (quarter only £bn)

End-Q102
Fixed local call volume (bn mins) 18.8
Fixed local call revenue (exc. access) (£bn) 0.4
Implied revenue per minute 0.02
Scope for local substitution 30% 30% local call volume

substitution drives £0.3bn
Local MOU substituted to mobile (bn mins) 5.6
to mobile
Implied additional MOU per sub 41.3
Local revenue substituted to mobile 0.3
Using an average price per cellular minute of £0.06 implies a premium of 2.8x the
average price per fixed-line minute (fixed-line price per minute of £0.022). This
massively overstates the actual premium the wireless user is paying because it
ignores the wireline access and connection fee that the customer pays.
UK F2M calls
Using an average revenue per cellular minute of £0.06, and assuming the
displacement of 60% of fixed to mobile volume to mobile, we see about £0.1bn of
60% F2M substitution
fixed calls to mobile moving to mobile. This amounts to about 2.1bn of additional
corresponds to 15 MOU per
minutes, or 15 monthly MOU per wireless subscriber. It is important for readers to sub per month
recognise that as the wireless handset becomes the “accepted” point of access, we
believe that a significant networking effect could occur (in terms of mobile to
mobile communication). Hence we believe that our 60% migration assumption
could prove too conservative.
Table 20 : UK F2M call market (quarter only £bn)

End-Q102
F2M call volume (bn mins) 3.5
F2M call revenue (exc. access) (£bn) 0.5
F2M Implied revenue per minute 0.14
Scope for F2M substitution 60% 60% substitution results in

£0.1bn moving to mobile
MOU substituted to mobile (bn mins) 2.1
F2M revenue substituted to mobile 0.1

33
MA CRO DYNAMI CS
UK national calls
Our central case uses 50% substitution of national calls to the wireless network. We
view the national call market as a core opportunity for Europe’s wireless operators.
On this basis, we see 6.8bn minutes substituting to the UK’s wireless operators (50
monthly MOU per subscriber), or about £0.4bn in revenue terms.
Table 21 : UK national call market (quarter only £bn)

End-Q102
Fixed call volume (national calls) (bn mins) 13.6
Fixed call revenue (National calls) (£bn) 0.4
Scope for national substitution 50% 50% substitution of

national calls results in
£0.4bn moving to mobile
National revenue substituted to mobile 0.4
UK international calls
In our central case outlook, we see a less material opportunity from the
international call segment. We acknowledge that this could be overly conservative.
Our central case assumes 30% substitution of international call volume, or about
0.6bn minutes migrating to the UK wireless network operators (4 monthly MOU per
subscriber).
Table 22 : UK international call market (quarter only £bn)

End-Q102
Fixed call volume (international calls) (bn mins) 2.0
Fixed call revenue (International calls) (£bn) 0.3
Scope for international substitution 30% 30% substitution of

national calls results in
£35m moving to mobile
International revenue substituted to mobile 0.0
UK other calls
Other calls include Internet dial-up, pay phones, calls to premium services, calls
through the operator, calls to the speaking clock, calls to paging operators and free
phone numbers. While we view this market as “fair game” for the wireless operator
community, we have included only 15% of this traffic displacing to the wireless
operators, recognising that Internet traffic is a key driver. This corresponds to
about 7.1bn minutes (52 monthly MOU per subscriber).

34
MA CRO DYNAMI CS
Table 23 : UK other call market (quarter only £bn)

End-Q102
Fixed call volume (other calls) (bn mins) 47.1
Fixed call revenue (Other calls) (£bn) 0.7
Scope for 'other calls' substitution 15% 15% substitution of other

calls drives £0.4bn to
mobile
Other call' revenue substituted to mobile 0.4
UK market in aggregate
In aggregate, our central case estimates show mobile traffic growing to 35% of
total telco volume from 13%, or to 46% of total revenue (from 35%). This equates
to 162 monthly MOU per subscriber substituting from the UK wireline operators to
the UK wireless community. This corresponds to an additional 5.3 MOU per day
Together we see up to 18%
being substituted, or about 8.3 minutes a day in outgoing cellular usage in total growth in outgoing voice
(5.3 MOU plus the original 3.0 MOU). In revenue terms this equals £1.4bn of ARPU
additional revenue, or a monthly ARPU uplift of £10 per subscriber. This is 18%
growth on the “pre” price decline monthly ARPU scenario of £13.2 per
subscriber.
Table 24 : Total UK call market (quarter only £bn)

End-Q102
Total MOU substituted to mobile (bn mins) 22.2
Period average UK mobile subscribers (Q4/Q3) 45.6
Total MOU per sub substituted to mobile 162.4
Total revenue substituted to mobile (£bn) 1.4
Total ARPU uplift (£) 10.0 In total, substitution

Source: ABN AMRO estimates/OFTEL
increases ARPU by £10
We can also think of this substitution in terms of ARPU contribution by call type
(see following table). Our analysis clearly shows outgoing voice ARPU increasing
to £15.6 from £13.2 (+18%).
Table 25 : Central case: UK market ARPU by call type

Unit End Q102
Total ARPU (outgoing, exc. SMS) pre-decline £ 13.2

Total ARPU (outgoing, exc. SMS) post-decline (exc
£ 5.6
elasticity)
Add: local call ARPU £ 2.5
Add: fixed to mobile call ARPU £ 0.9
Add: national call ARPU £ 3.1
Add: international call ARPU £ 0.3
Add: “other call” ARPU £ 3.2
Total outgoing ARPU post decline £ 15.6
% growth in ARPU % 18% ARPU post migration is

18% higher ...

35
MA CRO DYNAMI CS
Industry impact
We can also think about the potential of fixed to mobile substitution in terms of the
total industry, recognising the importance of the 50% premium to the underlying
fixed-line price per minute. Our central case for mobile (holding all other factors
equal) suggests that UK mobile revenue grows to £2.1bn (for Q102 factoring in the
indicated substitution), and UK wireline revenue declines by 24% to £2.5bn from
£3.3bn. For the wireline operators this does not include any compensating benefit
from DSL. In aggregate, this results in total outgoing telecoms revenue in the UK
declining by about 9% (from about £5.1bn to £4.7bn).
Table 26 : UK fixed and mobile market post 58% cellular price decline
Fixed: post-price decline End-Q102
Fixed local call revenue (exc. access) (£bn) 0.3
F2M call revenue (exc. access) (£bn) 0.2
Fixed call revenue (national calls) (£bn) 0.2
Fixed call revenue (international calls) (£bn) 0.2
Fixed call revenue (Other calls) (£bn) 0.6
Total fixed call revenue (£bn) 1.5
Fixed access revenue (£bn) 1.1
Total fixed revenue (post-decline) £bn 2.5
% change -24.3% ... but migration (without

DSL growth) damages the
Total fixed minutes bn 62.8
wireline industry
Actual price per minute (£) 0.04
% change 3%
Mobile revenue £bn 2.1
% change 18%
Mobile volume bn 34.5
% change 180%
Actual price per minute (£) 0.06
Implied multiple 1.5
We can present this impact more clearly using the following pie chart.
Figure 4 : UK telecoms quarterly revenue (£bn): central case

Pre-Decline Post-Decline
Mobile revenue
(pre-decline) Mobile revenue
% total 35% (post-decline)
% total 46%
Wireline revenue Wireline revenue

(pre-decline) (post-decline)
% total 64% % total 54%
£5.1bn £4.7bn
Source: ABN AMRO

36
MA CRO DYNAMI CS
A word of warning
Intuitively, some readers of this study may ask why total revenue is shrinking. If Without DSL growth,
telecoms shrinks
UK wireline loses more than 22bn minutes to wireless, and UK wireless customers
are prepared to pay a 50% premium to the underlying fixed line price, then
(surely) the industry should grow (not contract). This is (of course) correct,
excluding the original UK wireless revenue (see following chart). Total fixed falls to
£2.5bn from £3.3bn, mobile adds £1.4bn, resulting in total revenue of £3.8bn
(from £3.3bn). But mobile also loses £1bn from its historical revenue stream.
Figure 5 : Q102 UK telco market revenue evolution (£bn)

3.5 3.3 0.8
3
2.5
2.5 2.2
1.4
£.bn
2 1.8 1
1.5
1 0.8
0.5
0
Total F.L..R Total F.L R Total F.L.R Mobile Mobile (Lost) Mobile Mobile Rev. Added Mobile Rev
(Pre-Decline) (Lost) (Post-Decline) (Pre-Decline) (Post Decline) (Post-Decline)
Source: ABN AMRO
Sensitivity analysis
In the following table we have presented a range of scenarios for outgoing voice We have also run a
sensitivity analysis on the
ARPU based on our analysis of the UK market at Q102. This enables investors to
price change required to
take their own view of the scope for traffic substitution, for a given price reduction drive migration
and identify potential outgoing voice ARPU. For example, assuming a (say) 55%
share of traffic and 58% price reduction suggests outgoing voice ARPU shifts to
£24.4 (+84%).
Table 27 : Absolute ARPU (£) scenario analysis

Price reduction
-78% -68% -58% -48% -38% -28% -18%
Volume share 15% 4.1 5.5 6.8 8.1 9.4 10.8 12.1
25% 8.5 9.9 11.2 12.5 13.8 15.2 16.5

For example, assuming a
35% 12.9 14.3 15.6 16.9 18.2 19.5 20.9
58% price cut results in
45% 17.3 18.6 20.0 21.3 22.6 23.9 25.3 55% of originated calls
migrating to mobile, moving
55% 21.7 23.0 24.4 25.7 27.0 28.3 29.7
ARPU to £24
65% 26.1 27.4 28.7 30.1 31.4 32.7 34.0
75% 30.5 31.8 33.1 34.5 35.8 37.1 38.4
Source: ABN AMRO

37
MA CRO DYNAMI CS
Table 28 : Percentage change ARPU: scenario analysis

Price reduction
0.2 -78% -68% -58% -48% -38% -28% -18%
Volume share 15% -69% -59% -49% -39% -29% -19% -9%
Using the same example,
25% -36% -26% -16% -6% 4% 14% 24%
we can observe 84%
35% -2% 8% 18% 28% 38% 48% 58% growth on the pre-decline
45% 31% 41% 51% 61% 71% 81% 91% ARPU
55% 64% 74% 84% 94% 104% 114% 124%
65% 97% 107% 117% 127% 137% 147% 157%
75% 130% 140% 150% 160% 170% 180% 190%
Source: ABN AMRO
What about the power of the network?

A key risk is presented by the management of this process. We recognise that
running volume into value will be a challenging task. To quote a famous Greek
proverb, “What’s small, dark and knocking at the door? The future.” People are
rightly cynical about forecasting, but forecasting is at the heart of the planning
process and competitive positioning. However, it is our core belief that the low case
scenario is unlikely to emerge, because it ignores the scope for a networking effect
on usage.
In our view, an analysis of existing call volume does not fully capture the potential
of the volume opportunity for the European wireless operator community. Why?
■ Mobiles are not generally shared and are most usually carried on the person, Should the wireless handset
become the primary access
which provides an opportunity for a networking effect of increasing usage.
point, large scale growth in
the size of the pie could
■ Our core thesis is that high prepaid charges have stunted European cellular
occur
usage.
Hence our earlier assessment of the percentage of fixed-line call volume that could
displace to mobile could fundamentally underestimate the overall size of the
wireline telephony opportunity. In other words, the total size of the voice
telephony/broadband data market could explode (see following chart).
Chart 10 : Hypothetical volume forecast

25,000
Thousands of Gigabytes per year
20,000
15,000
A classic bull market chart,
10,000 or maybe realistic upside?
5,000
-
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Total Fixed Narrowband Data Total Fixed Narrowband Voice Total Fixed Narrowband
Total Broadband Data Total Broadband Voice Total Broadband
Total Mobile
Source: ABN AMRO

38
MA CRO DYNAMI CS
F2M additive to pre-to-post paid migration

It is crucial to note that this fixed-to-mobile substitution growth is in addition to
underlying volume growth and the benefit we expect from pre-to-post-paid
migration (see above). In other words, there are two distinct drivers of ARPU:
■ Phase One: 4%-8% annual growth in blended ARPU from pre-to-post-paid

migration for the next four years’ and
■ Phase Two: an 18% structural shift in outgoing voice ARPU by 2006 (holding all
other factors equal 11% overall ARPU growth)
This results in up to 30% overall ARPU growth in total by year-end 2006.
Phase Three: Termination rate declines

Slowing subscriber growth at a time of sharply falling mobile termination prices We have to balance
improving outgoing voice
could further expose the cellular operators to a revenue crunch. Where regulated
ARPUs with potential for
reductions in mobile termination rates are exercised, they must be passed onto the termination rate cuts
consumer. Termination rates typically account for 15%-35% of operators’ revenue.
Termination rates in Europe remain substantially higher than is observable in other
major markets, including the US and Korea (see following chart).
Chart 11 : European termination rates (€)

35
30
25
20
15
European termination rates
10
remain at a signifiacnt
5 premium to other major
markets
0
US
Korea
Luxembourge
France
Denmark
Greece
Germany
Norway
Sweden
UK
Ireland
Austria
Netherlands
Portugal
Spain
Italy
Finland
Switzerland
Belgium
Source: ABN AMRO/country regulators
In Regulation – Friend or Foe? dated 16 August 2002 we ran a detailed sensitivity Lower quality (primarily)
new entrant businesses are
of various operators’ revenue to reductions in termination rates. Simply put, all
most exposed
operators have material exposure – but new entrant operators have
disproportionately higher exposure. Ironically, as many regulators aim to promote
competition, they cannot cut rates too aggressively. We can assess the impact of
cuts to termination rates with reference to the UK.
A UK example
In the UK, OFTEL has suggested that RPI negative 12% should be applied to all UK The UK (again) offers an
excellent benchmark for
operators’ termination fees. Historically, UK operators have been subject to RPI
understanding the
negative 9% for fixed line calls to wireless only. The extension of the mandate will economic impact of cuts
encompass about 48% of total incoming volume from 21% under the previous
regulation (see following tables).

39
MA CRO DYNAMI CS
We can work through the impact of RPI-12% using OFTEL data. First, we know Using OFTEL data we can
work through the proposed
from OFTEL that mobile outgoing calls amounted to 12.3bn minutes in Q102 (for
RPI-12% cut
the quarter, see following table).
Table 29 : UK mobile volume (m mins) Q102

(m mins) %
UK calls 11,931 97%
Outgoing international 166 1%
Roaming 233 2%
All calls (outgoing) 12,330 100%
Source: OFTEL
Of these minutes we know the percentage split between fixed calls, and on-and off- OFTEL tells us the split
between fixed, on- and off-
net mobile minutes (see following table).
net mobile minutes
Table 30 : UK mobile volume call split % (excluding SMS)

FY Dec 01 Jan - Mar 2002
UK fixed calls 49% 6,042
On-Net mobile 26% 3,206
Off-Net mobile 13% 1,603
Outgoing international 1% 123
Roaming abroad 2% 247
Other 9% 1,110
100% 12,330
Source: ABN AMRO
We know that total interconnection volume amounted to 6.1bn minutes. Using the We know that total
interconnection minutes
data above we can imply about 1.3bn minutes attributable from the fixed line (see
amounted to 6.1bn minutes
following table). (Q102), implying 1.3bn
minutes attributable from
the fixed line
Table 31 : UK interconnection volume
Jan - Mar 2002 % split
Portion from on network mobiles 3,206 52.4%
Portion from off network mobiles 1,603 26.2%
Total Portion from other mobiles 4,809 78.6%
Portion from fixed line 1,309 21.4%
Interconnection total 6,118 100.0%
Source: ABN AMRO
We also know that quarterly interconnect revenue of £672m was generated during Using an inflation rate of
2%, and RPI-12% suggests
1Q02, and that £361m was attributable to off-net mobile. Hence factoring in a
a 2.4% reduction in
2.0% RPI rate (ie a 10% price cut), suggests a 2.4% reduction in total quarterly quarterly revenue, which is
revenue (24% of revenue attributable to interconnect, multiplied by the percentage in the noise
subject to regulation [100%], multiplied by the percentage fall in regulated

interconnection [10%]). Hence double-digit declines in termination rate (all other
factors held equal) results in minor negative revenue growth for the industry.

40
MA CRO DYNAMI CS
Table 32 : UK interconnection calculations

Volume Revenue ppm
On-net mobiles 3,206 0 0.00
Off-net mobiles 1,603 361 0.22
Mobile-to-mobile 4,809 361 0.07
Fixed-line 1,309 311 0.24
Total 6,118 672 0.11
Source: ABN AMRO
Given much of the doom and gloom surrounding the issue of termination rates, it is We put 60% probability on
the RPI-12% ruling being
important not to lose sight of the minor impact at the top line of RPI-12%. We put
confirmed
a 60% probability on the RPI-12% cut being endorsed by the UK’s competition
commission. If this were confirmed, we would view it as an upside surprise to the
market’s current expectation.
We recognise that in a worst-case scenario, the competition commission could

introduce a draconian one-off price cut. Using a 20% price cut (for the purpose of We put a 40% probability
scenario analysis) we would expect revenue to decline between 4% and 5%, and on a draconian price cut ...
EBITDA by 8%-10%. Either way, we do not expect any announcements before

Q103.
Increasingly we think investors should question this analysis that excludes volume … but remind investors of
the potential for increasing
uplift. Focusing on the impact to the profitability of a reduction in termination rates
volume to at least partially
without any compensating volume increase is appealing (due to simplicity), but it offset losses
ignores the possibility that growth in volume could occur, potentially compensating
operators for the loss in price.
Impediment to substitution
We recognise that the one of biggest impediments to potential substitution is the Ironically, cuts in
termination rates could r-
high rate of mobile termination fees throughout Europe. We expect termination
focus management on
rates will continue to decline at a double-digit rate for the foreseeable future. fixed-to-mobile substitution
Ironically, cuts in termination rates could refocus management on the fixed-to-
mobile substitution opportunity. Hence we believe falling termination rates are key
to enabling growth in outgoing voice ARPU.
The US experience
While there is a weight of evidence showing that cuts to terminations rates do
materially affect short-term profitability in markets with relatively high levels of The US model suggests
termination (see Portugal as an example), we believe there could be a silver lining there could be a silver
to the cloud. In our view, the US cellular market provides a benchmark that could lining to the cloud ...
provide an indicator of Europe’s future. US blended ARPUs are more than 50%
higher than UK ARPUs (see following chart).

41
MA CRO DYNAMI CS
Chart 12 : UK vs US – blended ARPU and penetration
90% 40.0
78% 81%
80% 35.0
67%
70%
30.0
Penetration (%)
60%
50% 25.0
ARPU (£)
50% 45%
40% 39% 20.0
40%
31%
25% 15.0
30% 23%
20%
20% 14% 10.0
10% 5.0
0% 0.0
1997 1998 1999 2000 2001 Q1 02
Year
Penetration US Penetration UK US Monthly ARPU UK Industry ARPU
Source: ABN AMRO & company reports
We estimate that approximately 60% of all US traffic is outgoing. New entrant US incoming volumes are
materially higher than in
players, such as VoiceStream, generally have a greater reliance on incoming
Europe, as a result of lower
volumes (about55% of traffic is outgoing). On a per minute basis, US termination prices
rates average about €0.005. The key driver of US incoming ARPUs is higher
volumes (see following table). Our thesis is that European incoming ARPUs
(and usage) have been stunted by excessive charges.
Table 33 : US wireless trends

1999 2000 2001
Call volume (bn) 221.5 372.6 561.2
% Change 68% 51%
Call revenue (bn) 40.6 49.4 64.0
Implied revenue per minute 0.18 0.13 0.11
% Change -28% -14%
Mobile subscribers 86.0 109.5 129.2
Ave. lines 77.6 97.8 119.3
MOU per line 285.4 381.1 470.3
Outgoing MOU (ABN AMRO estimate) 171.2 228.7 282.2
Incoming MOU (ABN AMRO estimate) 114.1 152.5 188.1
Source: ABN AMRO/CTIA
The problem Europe’s operators face is to manage the decline in rates without
The key challenge in Europe
significantly damaging returns. The danger is that in the event of relatively small
is to manage declining
cuts to termination rates, demand fails to materialise. This is akin to the firm that termination rates without
gets caught in the middle of the road (between excessively high termination rates damaging returns
and rates similar to the fixed line operator community) – getting run over!
While declining termination rates will create a “tough” top-line growth environment, Using the US as a
benchmark suggests upside
given the US experience, it is plausible to argue that a significant decline could be
to bearish market
proportionately offset by an increase in volume. perception

42
MA CRO DYNAMI CS
Top down: share of GDP

While getting into the detail is useful, it is important not to forget the implied share We have sanity-checked our
bottom-up work to share of
of GDP that we are suggesting is attainable for Europe’s wireless operators.
GDP
According to our estimates, wireless’s share of GDP ranges between 0.7% and
1.3% in Europe’s major markets.
Wireline spend ranges between 1.2x and 2.3x wireless spend as a percentage of Wireline spend ranges
between 1.2x and 2.3x
GDP. Hence in aggregate, telecoms spend accounts for 2.2%- 3.3% of GDP. In the
wireless spend
UK, telco spend accounted for around 2.4% at year-end 2001 (see following table).
Table 34 : UK telco spend

(£) 1998 1999 2000 2001
Wireline revenue (£) 13.4 14.2 14.9 15.0
Wireless revenue (£) 3.9 5.3 6.9 8.2
Nominal GDP (£) 859 902 950 988
Telecom spend as %GDP 2.0% 2.2% 2.3% 2.4%
Wireline spend as %GDP 1.6% 1.6% 1.6% 1.5%
Wireless spend as % GDP 0.5% 0.6% 0.7% 0.8%
Source: ABN AMRO
This can be compared with the higher share of GDP that is observable in the US. US wireless share of GDP is
particularly impressive
Wireless share of GDP is particularly impressive given the materially lower rate of
given the lower rate of
penetration (c.50% vs Europe c.81%). penetration
Table 35 : US telco spend as a % of GDP

1998 1999 2000 2001
Wireless 0.4% 0.5% 0.6% 0.8%
LD 1.1% 1.1% 1.0% 0.9%
Local 1.3% 1.3% 1.3% 1.3%
Total 2.8% 2.9% 3.0% 3.0%
Source: ABN AMRO/FCC

43
MA CRO DYNAMI CS

44
I N V E S T M E N T V I E W
Bucket pricing plans
Despite intense competition, US operators have maintained high ARPU levels by not
targeting lower-value prepay subscribers, and through the introduction of “bucket
plans.” US operators have increased the number of free minutes of service, rather
than reduce the monthly access fee. In Europe, “tariff plans” have resulted in
significant ARPU dilution, and higher margins but lower pre-capex cash flow per
subscriber. In our view, the bucket pricing plan offers superior long-term returns to
Europe’s tariff plan model. As a result, we believe that Europe will move towards
flat rate access fees. We highlight that success can breed failure as the historical
success model becomes the major obstacle to a firm’s adoption of the new reality.
The introduction of the “your plan” flat rate fee by Orange in the UK provides early-
stage validation of our view. In addition we flag that an analysis of VoiceStream,
the aggressive US new entrant, provides a benchmark for understanding the
potential impact of Hutchison in Europe, and the domino effect we forecast
throughout the European wireless landscape.
Introduction
■ the difference in subscriber mix by major market;
■ US vs European ARPUs; Flat rate access plans

enable fixed-to-mobile
■ bucket plan volume; substitution
■ capacity/capex as a driver of pricing; and
■ US bucket plan yields vs Europe’s major markets.

45
IN VESTM ENT VI EW
The US experience
US operators, while reporting falling historic ARPU, have consistently reported US operators have
consistently reported
higher and more stable ARPUs than their European counterparts. We attribute this
higher and more stable
to two key drivers, post-paid focus and bucket plans. ARPUs than their European
peers
Post-paid focus
US operators have not aggressively targeted lower-value prepay subscribers. This
is easily identifiable from a comparative analysis of subscriber mix by geography
(see following chart). In the US, this product is unattractive because prepaid
customers would also have to pay for incoming calls.
Chart 13 : Worldwide subscriber bases
89%
15% 79%
75%
69%
28%
60%
35%
33%
53%
46%
48%
85%
72%
65% 67% 97% 91%
52%
9%
3%
Italy UK Spain Germany France Japan US
Total 50 48 29 58 37 62 130
Subs (m)
Pre-Pay Post-Pay
Source: ABN AMRO and Deutsche Telekom
Flat rate fees/increasing volume

The US cellular business model has increased the number of free minutes of service The US model is volume
included in access plans (bucket plans), rather than reducing the cost of access based
plans. By comparison, European operators have traditionally structured mobile

plans with cheaper access fees, but with much fewer free minutes included and
charging per minute for calls made (tariff plans). The US model is based on
increasing volume offsetting declining prices to maintain ARPU.
US vs UK blended ARPUs
Blended ARPUs in the US are materially higher than in Europe. To December 2001, US monthly industry ARPU
over 50% higher than the
in the US monthly ARPU stood at £29.7, compared with £19.6 in the UK, ie over
UK industry average
50% higher (see following chart). At 2Q02, Vodafone was the highest ARPU
operator in the UK (£23 per month) vs monthly ARPU at Nextel [NXTL-US$9.20,
N/R] of £48 and AT&T Wireless [AWE-US$4.94, N/R] of £43 per month. This is a
direct function of the higher-volume/lower-margin business model that has been
adopted (in marked contrast to the established European model).

46
IN VESTM ENT VI EW
Chart 14 : UK vs US - Blended ARPU and penetration
90% 40.0
78% 81%
80% 35.0
67%
70%
30.0
Penetration (%)
60%
50% 25.0
ARPU (£)
50% 45%
40% 39% 20.0
40%
31%
25% 15.0
30% 23%
20%
20% 14% 10.0
10% 5.0
0% 0.0
1997 1998 1999 2000 2001 Q1 02
Year
Penetration US Penetration UK US Monthly ARPU UK Industry ARPU
Source: ABN AMRO & Company reports
Our time series analysis clearly shows that the US market increased penetration by Differences in ARPU
between the US and the UK
over 30% between 1997 and 2Q02, without material dilution to ARPU. This can be
are not primarily
contrasted to the UK’s performance. The difference in ARPU between the two attributable to penetration
markets is not entirely attributable to differences in penetration. This can be quickly
demonstrated by looking at US monthly ARPU at year-end 2000, relative to UK
monthly ARPU at year-end 1999, with similar levels of penetration. Hence our
thesis is that average European cellular MOUs/ARPUs have been stunted by tariff
plan pricing. Looking forward, we believe the introduction of higher flat rate access
fees, with a higher level of “free” minutes (bucket pricing plans) should have a
profound impact on ARPU.
Bucket plans generate higher ARPU and usage

As noted in our recent study VoiceStream – The American Dream dated 2 August Evidence that bucket plans
significantly increase
2002, the structure and pricing of bucket plans encourages mobile subscribers to
mobile usage
use their wireless phone for all their voice calls. The average US subscriber uses
three times as many minutes per year as a UK subscriber (see following chart).
VoiceStream subscribers (who are more geared toward bucket plans) use 1.5 times
the US average. UK industry volume has remained static during the past nine
quarters. In our view, the introduction of bucket-style pricing plans should
dramatically shift UK (and European) cellular usage.

47
IN VESTM ENT VI EW
Chart 15 : Monthly MOU (incoming and outgoing)
700
600
500
400
Minutes
300
200
100
0
Q1 00A Q2 00A Q3 00A Q4 00A Q1 01A Q2 01A Q3 01A Q4 01A Q1 02A
Period
VoiceStream US Industry Average UK Industry Average
Source: ABN AMRO, company reports and OFTEL data
Following AT&T Wireless, VoiceStream introduced more aggressive bucket pricing Two main types of bucket
plans
plans into the US market, creating a domino effect. US national operators offer two
main types of bucket plans, which generally provide more minutes than most
subscribers can consume. This is a marketing ploy used particularly well by
VoiceStream to make the subscribers feel they are receiving greater value for
money.
■ National one-rate plans eliminate long distance charges. These plans consist
of a monthly fee that includes an allocation of “anytime minutes,” off-peak
minutes and one standard overage rate. It is difficult to compare plans across
operators because some, like VoiceStream, charge extra for off-network
roaming, while others include it in the price.
■ Regional plans offer users free minutes in their region (multi-state, not
national) but typically charge for roaming out of one’s region and for long
distance minutes. This is the plan that has given VoiceStream the most publicity
– ie 3,000 “anytime” minutes for $59.99 – but accounts for less than 20% of
new customers.
Recent pricing plans announced by the big six operators are very similar as each VoiceStream the most
fights for market share. VoiceStream has been the most aggressive US operator, aggressive US operator
setting industry-low price points for both national one-rate and regional plans.
The evidence available from the US market model to date suggests that new 80% of new subscribers are
choosing national plans
subscribers are choosing national instead of regional plans. Customers are using
these national plans to replace their landline phones for long distance calls, but still
using wireline for untimed local calls.

48
IN VESTM ENT VI EW
US operator national rate plans

When discussing bucket pricing plans we are often asked, “What distinguishes the Scale of the number of free
US model from what we already have in Europe?” The answer is simple – the scale minutes and high flat rate
of the buckets and the high flat rate entry fee differentiates the US. Europe is entry prices differentiate
the US model
charecterised by low flat rate entry fees and relatively high overage rates. To
provide investors with a feel for the magnitude of the typical US offering and
minimum price entry point, we have taken a sample of various plans offered at the
end of July 2002.
Table 36 : National one-rate plans – roaming included

Yield/ Anytime Off-peak Weeknight
Entry point minute Monthly min. Overage minutes minutes
Included in any
AT&T Wireless $0.13 $59.99 450 $0.35 NA time minutes
Cingular $0.10 $29.99 300 $0.40 1,000 Included in off-peak
Sprint PCS* $0.17 $34.99 200 $0.40 3,300 Included in off-peak
Verizon Wireless $0.23 $35.00 150 $0.40 NA Included in off-peak

High usage minute Monthly min. Overage minutes minutes
Included in any
AT&T Wireless $0.10 $199.99 2,000 $0.25 NA time minutes
Sprint PCS* $0.07 $74.99 1,000 $0.40 6,500 Included in off-peak
Cingular $0.06 $199.99 3,500 $0.25 3,500 Included in off-peak
Verizon Wireless $0.10 $300.00 3,000 $0.20 NA Included in off-peak

* ABN AMRO does not cover this stock.
Source: ABN AMRO
Table 37 : National one-rate plans - roaming NOT included

Entry point minute Monthly min. Overage minutes minutes
VoiceStream $0.07 $39.99 600 $0.35 Unlimited None
Nextel* $0.25 $49.99 200 $0.35 2,000 Included in off-peak
Verizon Wireless $0.12 $35.00 300 $0.40 4,000 Included in off-peak

High usage minute Monthly min. Overage minutes minutes
VoiceStream $0.07 $99.99 1400 $0.30 Unlimited None
Nextel* $0.10 $99.99 1000 $0.35 3,000 Included in off-peak
Verizon Wireless $0.07 $200.00 3000 $0.20 4,000 Included in off-peak

Source: ABN AMRO
Most of the national operators have a national rate plan “anytime” yield per minute
of USD $0.07 (off-network roaming extra), following VoiceStream’s aggressive
pricing. Similarly, the established operators have increased their number of free
US national rate plans have
off-peak minutes (following VoiceStream), but none have matched VoiceStream’s an ‘anytime’ yield of
unlimited weekend minutes. As the US operators have experienced increasing US$0.07
capacity constraints (a function of rising usage and penetration), operators have
introduced a slight shift in off-peak times for new customers ie pushing back the
start time for off-peak to 9pm from 8pm, to manage network congestion.

49
IN VESTM ENT VI EW
US operator regional plans
Table 38 : Regional (Northeast neighbourhood) rate plans

Yield/ Anytime Off-peak Long
Entry point minute Monthly min. Overage minutes Roaming distance
VoiceStream $0.02 $59.99 3,000 $0.35 NA $0.49 $0.20
AT&T Wireless $0.30 $29.99 100 $0.40 2,000 $0.69 NA
Nextel* $0.10 $49.99 500 $0.35 NA NA $0.15
Sprint PCS* $0.14 $34.99 250 $0.40 3,250 NA $0.20
Verizon Wireless $0.10 $35.00 350 $0.45 4,000 $0.69 $0.20
Yield/ Anytime Off-peak Long

High usage minute Monthly min. Overage minutes Roaming distance
VoiceStream $0.02 $59.99 3,000 $0.35 NA $0.49 $0.20
AT&T Wireless $0.08 $199.99 2,500 $0.25 Unlimited $0.69 NA
Nextel* $0.10 $99.99 1,000 $0.35 NA NA $0.15
Sprint PCS* $0.09 $69.99 800 $0.40 6200 NA $0.20
Verizon Wireless $0.07 $100.00 1,500 $0.30 4000 $0.69 $0.20

Source: ABN AMRO
Many investors flag the low yield per minute ($0.02) generated by VoiceStream’s Points to consider about the
3,000 ‘anytime’ minute
controversial 3,000 “anytime” minute regional plan as being an unsustainable
regional plan
business model; however, the following points must be considered.
■ This plan was introduced in 3Q01 primarily as a marketing tool by VoiceStream Little impact on margins –
only 20% of customers
to reinforce its image as being the best value wireless provider in the US. The
choosing this plan
actual effect on VoiceStream’s network capacity and margins is limited as less
than 20% of new net adds choose this plan.
■ On average, only 1,000 (16.6 hours per month) of the 3,000 minutes (50 hours Only one-third of minutes
used, generating same
per month) are actually used, implying a real yield of $0.06 per minute, which is
actual yield as national call
almost in line with the competition and not dissimilar to the $0.07 yield plans
generated by national call plans.
We believe that national operators will continue to promote this offer as a

marketing tool while they have spare capacity. However, as penetration and usage
increase, operators will need to either increase prices and/or reduce the number of
bundled minutes. The speed of this process will be an individual consideration for
each operator based on spare network capacity available.
Spare network capacity in each market is a function of spectrum and the number VoiceStream is the smallest
national operator, but has a
of subscribers on the network. In the US, VoiceStream is by far the smallest
similar amount of spectrum
national operator in terms of subscribers, but has a similar amount of spectrum in in the top 50 markets
the top 50 markets as its competitors, and therefore has greater spare capacity to
carry additional minutes on its network (see following chart).

50
IN VESTM ENT VI EW
Chart 16 : MHZ per subscriber (average top 50 markets)
4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
Verizon Wireless Cingular AT&T Wireless Sprint PCS Nextel Voicestream
Wireless Communications
ABN AMRO does not cover Sprint PCS or Nextel Communications.

Source: ABN AMRO
Additionally, all of VoiceStream’s network is based on GSM technology, while some VoiceStream’s network
technology is more
operators such as AWE and Cingular are still partially using less capacity-efficient
capacity-efficient
TDMA. Cingular’s chief technology officer has stated that upgrading to GSM/GPRS
by 2004 will increase voice capacity on Cingular’s existing network by 120%, while
at the same time requiring 18% less spectrum.
Average yield based on actual minutes used

To calculate the average actual yield per minute being generated by the various VoiceStream the cheapest
US wireless provider per
plan options, we have divided average ARPU by the actual number of minutes used
minute
(rather than offered) for the past nine quarters. In the US market VoiceStream,
with its lower-priced bucket plans, is the cheapest US wireless provider per minute.
The effect of competition and bucket plans has led to the US market pricing per
minute (yield) being relatively cheaper.
Chart 17 : Yield per MOU
0.30
0.25
0.20
Yield (USD)
0.15
0.10
0.05
0.00
Q1 00A Q2 00A Q3 00A Q4 00A Q1 01A Q2 01A Q3 01A Q4 01A Q1 02A
Period
VoiceStream US Industry Average UK Industry Average
Source: ABN AMRO, company reports and OFTEL data

51
IN VESTM ENT VI EW
Linking yield to ARPU

It is important for investors to recognise that yield is the actual average revenue Yield is not a proxy for
per minute achieved and is not a proxy for ARPU. In fact (as we have already ARPU
observed), US ARPUs are substantially higher than in the UK. Hence lower yield in
the US does not result in lower ARPU. This is particularly pronounced at the
VoiceStream level. The key difference is volume.
US invested capital
The reason VoiceStream has been able to price so aggressively relative to its A more capital-light
network enables a lower
competition is its lower cumulative capex spend per covered pop (see following
unit operating cost
chart). This is a very important point. Intuitively, a more “capital-light” network proposition
requires a lower absolute level of net operating profit after tax to generate a cost of
capital return.
Chart 18 : Cumulative capex spend per covered population
120
100
80
US$ m
60
40
20
0
Voicestream T-Mobile Nextel Com. Orange Vodafone MMo2 AT&T Sprint PCS
Wireless
ABN AMRO does not cover Sprint PCS or Nextel Communications.

Source: ABN AMRO and company reports
Our analysis also highlights the lower cost proposition of T-Mobile and Orange in In terms of cumulative
capex per covered pop,
the UK relative to (say) both Vodafone UK and mmo2 UK. This is intuitive, given
there is no discernible
that the UK’s established operators (Vodafone and mmo2) have deployed their difference between the US
networks over 20 years, relative to about eight years for the new entrants. Hence and European model. The
key driver is length of
the new entrants have benefited from falling kit costs. The same relationship is
service
observable between AT&T Wireless (established operator) and VoiceStream (new
entrant) in the US.
US operators generate lower margins ...

It has often been pointed out to us that US operators generate lower EBITDA The US model appears to
result in lower EBITDA
margins than their leading European counter-parts. This is true, but percentage
margins
profitability only tells us half the story.
... but generate higher EBITDA per subscriber

Compared with Europe, more mature US operators generate significantly higher But higher EBITDA per
subscriber
absolute EBITDA per subscriber. US operators report lower margins because they
also generate much higher revenue/ARPUs from bucket plans. This makes it almost
meaningless to compare EBITDA margins from Europe with the US.

52
IN VESTM ENT VI EW
Chart 19 : EBITDA per subscriber
80.0
60.0
40.0
20.0
(Euro)
-
-20.0
-40.0
-60.0
-80.0
Vodafone - Germany
Orange - UK
MMo2
Vodafone - UK
Viag
Sprint PCS
Eplus
Cingular
AT&T Wireless
VoiceStream/Powertel
T-Mobile - UK
T-Mobile Germany
Verizon Wireless
ABN AMRO does not cover Sprint PCS.
Source: ABN AMRO & Company reports
Higher pre-capex cash flow per subscriber supports our thesis that the US model
could be superior to the existing European model. Unfortunately, it is difficult to Higher pre-capex cash flow
per sub supports our
compare absolute cash flow per subscriber between the US and Europe. This is a positive thesis
function of three key factors.
■ Differences in penetration. If the numerator is reasonably similar, differences

in the denominator distort comparison. As US penetration moves towards
maturity (c.75%), comparison would be more meaningful.
■ Differences in capex phase. US operators are completing network coverage

relative to more mature European network operators that are primarily investing
in capacity and maintenance. Coverage remains the most capital-intensive
phase of the cellular business model.
■ Technology migration. US operators such as AT&T Wireless and Cingular have

(and will continue) to be affected by the commitment of capital to migrate their
existing networks from TDMA to GSM.
Notwithstanding the difficulty of comparison, we suspect that the US model

(technology migration issues aside) is not proportionately more capital-intensive Cellular capex is primarily
coverage (not capacity)
than the European model. When coverage has been completed, capacity and
driven
maintenance is success-driven and represents a very small percentage of
incremental revenue generated. In other words, higher pre-capex cash flow per
subscriber should feed down to higher absolute cash flow per subscriber.
If our thesis is correct, we must address why US cellular assets trade at such a
material discount to their European peers. We attribute this to four easily
identifiable factors.
■ Uncertainty over incremental penetration growth remaining (driving

significant forecast downgrades to top-line growth)
■ A structurally more competitive market. We note nearly 50% of the US Weak US cellular valuations
do not affect our positive on
population has access to over six wireless operators (over 75% have access to
‘bucket pricing’ thesis
over five). The US Herfindahl index at year-end 2001 sat at 1700, suggesting
fierce competition, much lower than its European peers (eg Italy at 3729,
suggesting a high degree of pricing power).

53
IN VESTM ENT VI EW
■ Coverage and technology migration capex. As capital markets have closed

to wireless operators across the world, US operators have had to complete
network coverage and migrate technology platforms, which have damaged
returns.
■ Minimal spectrum on the horizon. US wireless operators have no clear path

to 3G, given the scarcity of spectrum (there is little new wireless spectrum set
to become available). Arguably this removes a second leg of growth (from mass
F2M substitution).
European outlook: H3G – the catalyst

VoiceStream’s role in the
In our view, reviewing VoiceStream and the US market provides a valuable insight US market provides a
into H3G’s strategic options as it enters the European market. In the absence of benchmark for
meaningful data revenue streams, we believe H3G will refocus on the fixed line understanding H3G in
Europe
voice opportunity via the provision of bucket pricing plans. We believe that Europe’s
operators will be forced to mimic H3G’s pricing plans, similar to the behaviour of
the established US operators in response to VoiceStream. This is a function of
available capacity. In our view, investors should think of H3G in terms of (1) spare
network capacity and (2) technology.
Spare network capacity. Our analysis of Mhz per subscriber in the US market High capacity (MHz per sub)
provides scope for
offers an insight into potential pricing in Europe. Similar to VoiceStream, we expect
aggressively priced flat rate
H3G to enjoy an abundance of capacity relative to its competitors. For example in access plans
the UK, investors can replace Verizon Wireless, Cingular and VoiceStream with
(say) mmo2, Vodafone and H3G (see chart below).
Chart 20 : UK MHz per subscriber

4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
Vodafone
One2one
Orange
BTCellnet
MHz per subscriber 2G MHz per subscriber
Source: ABN AMRO
Technology
H3G will enter Europe using W-CDMA, relative to its competition currently
employing GSM. Similar to VoiceStream, H3G should enjoy a more capacity- W-CDMA provides a more
capacity efficient
efficient technology. We very conservatively estimate capacity increases of 2x-3x
technology
(for a more detailed discussion of capacity we refer investors to the “3G as the
facilitator” section of this report).

54
IN VESTM ENT VI EW
European yield
As part of the preparation for this note, we analysed the yield observable in the UK, Other European markets
still have very high yield
France and Italy (see following chart). The markets with the highest level of
levels
concentration ie less competition have enjoyed the highest yield (France and Italy).
This is in line with our core belief that industry structure determines longer-term
sustainable returns.
Chart 21 : Key European market yields: the impact of competition
0.30
0.25
Yield (USD)
0.20
0.15
1999 Q100 Q200 Q300 Q400 Q101 Q201 Q301 Q401
Period
France Italy UK
Source: ABN AMRO
European yield outlook

The trend in yield per minute observable in the US provides an excellent lead
indicator of potential yields throughout Europe as a result of the entrance of The trend observable in the
US provides a lead indicator
Hutchison 3G. In the US, VoiceStream has been able to prosper with a lower yield of Europe’s potential future
as a result of its lower invested capital base. In Europe, we believe that H3G will be
able to play the same game.
Chart 22 : US % change in yield Q200A to Q102A
10%
5%
0%
-5%
-10%
-15%
-20%
Q2 00A Q3 00A Q4 00A Q1 01A Q2 01A Q3 01A Q4 01A Q1 02A
Voicestream % change US industry % change
Source: ABN AMRO

55
IN VESTM ENT VI EW
Over the past two years, US industry yields have fallen by a CAGR of 20%, relative In the US, industry yields
have fallen by a CAGR of
to VoiceStream posting an annual decline of about 15%. However, we note that US
20% over the past two
yields have stabilised in recent quarters, reflecting capacity constraints. years
Cumulative capex
We understand that HWL’s original business plan budgeted for about US$10bn of
cumulative capex to year-end 2005 for its European wireless operations. We
believe that approximately 55% of this spend will be dedicated to the UK operation H3G could enjoy a similar
lowest-cost operator
(US$5.5bn), 35% to Italy (US$3.5bn) and what remains to HWL’s more minor
position
operations (eg Denmark, Sweden, Ireland, Austria etc). More recently, HWL has
indicated that the operations in Italy and the UK will benefit from infrastructure
sharing, reducing capital expenditure by around 20%, implying new cumulative
capex of €4.4bn for the UK and €2.8bn for Italy.
Cumulative capex per covered pop

Using this data as a starting point we can pinpoint H3G’s position in the UK relative
to the existing competition.
■ Using Companies House data, we can identify cumulative capex for each of the
UK’s major operators to year-end 2001 (with the exception of Vodafone and T-
Mobile, for which we have used an estimate of 10% of their top lines for the
period to December 2001).
■ We then assume a maintenance level of capex at 6% of each operator’s forecast We have presented a
potential cumulative capex
top line per annum to year-end 2005.
per covered pop analysis for
the UK operators
■ Finally, we take a common view on the cost of deploying a 3G network in the UK
for the established operators, conservatively assuming 50% of existing 2G sites
can be re-used (see following table).
Table 39 : UK established operator 3G capex to year-end 2005

(€m) Cost
H3G capex 4,400
70% of capex “real estate build” 3,080
Existing operator (50% re-use of existing sites) 1,540
Other kit capex 1,320
Total existing operator 3G capex 2,860
As a % of total 65%
Source: ABN AMRO
While this analysis is “quick and dirty,” it enables us to identify the cumulative While this is ‘quick and
dirty,’ it enables us to
capex per covered pop by operator (see following chart). Reflecting its newer
better understand the
operator status, H3G enjoys a significantly lower cumulative capex per covered potential pricing
pop. environment

56
IN VESTM ENT VI EW
Chart 23 : Forecast cumulative capex per covered pop by UK operator
180
160
140
120
100
80
60
40
20
-
mmo2 Vodafone Orange T-Mobile H3 G
Source: ABN AMRO estimates
As demonstrated by the pricing of VoiceStream in the US market, an operator’s Over the longer term we
would expect a movement
returns are a function of its balance sheet invested capital (of which a large
towards convergence
proportion is capex); historical cost plays an important role in framing reported
returns. To 2005, we would expect H3G to have a “lighter” balance sheet than its
established peers. Over the longer term, as the UK’s established operators replace
the capital in their balance sheet, we would expect a movement towards
convergence.

57
IN VESTM ENT VI EW

58
S E C T O R D Y N A M I C S
3G as the facilitator
In our view, European 2G capacity constraints have prohibited the introduction of

bucket pricing plans (and by association fixed-to-mobile substitution). While 3G
may facilitate the use of high-speed applications, the real untold story of 3G is
capacity. 3G will enable Europe’s cellular operators to increase volume
aggressively. We estimate that employing W-CDMA with just one transceiver, 3G
will increase existing operator capacity between 1.7x and 2.8x, employing a
conservative 20% improvement in spectral efficiency. In densely populated areas,
capacity could increase by up to 5.0x employing only one transceiver. It is our core
belief that this will drive substantial displacement. In the medium term we do not
rule out the possibility that EDGE could be deployed in Europe to unlock existing 3G
spectrum allocation.
Introduction
In this section of this note we present:
■ existing operator capacity constraints;
■ modulation comparisons; 3G provides the capacity to

drive fixed-line
■ cell sites as a driver of capacity; displacement
■ W-CDMA vs GSM capacity analysis;
■ frequency allocations by operator;
■ the potential capacity uplift from W-CDMA and additional spectrum; and
■ the scope for EDGE as an interim solution.

59
SE CTO R DYNAM I CS
Existing operator capacity constraints

A method of assessing the capacity of existing operators is to look at the number of
cell sites per subscriber vs covered pops per cell site.
■ Covered POPS per base station. This calculation provides a rough proxy to We have presented a simple
relative capacity matrix for
compare the amount of capacity each operator has built into its footprint.
each major European
Operators with high covered POPs per base station can spend additional capex operator
on base stations to increase capacity, while operators with low covered POPs per
base station are closer to reaching maximum capacity.
■ Subscribers per base station. This calculation provides a rough proxy to the
current capacity each operator has available to service its existing subscriber
base.
Whilst our matrix does not gives us absolute capacity, in our view it provides a
good feel for relative capacity constraints across Europe (see chart below).
Figure 6 : Capacity matrix (1H02 2002)

Quadrant Two: Quadrant Three:
High capacity built into footprint Low capacity built into footprint
2,000 Low capacity per subscriber Low capacity per subscriber
1,800 TIM Spa

Vodafone (Germany)
Vodafone (UK) T-Mobile (UK)

1,600 T-Mobile (Germany)
Orange (France) mmo2 (UK)
1,400 Orange (UK)
TEM Esp
1H02 Sub/Cell
1,200
1,000
800 Bouygues
E-Plus
600
mmo2 (Germany)
400
Quadrant One: Quadrant Four:
High capacity built into footprint Low capacity built into footprint
200 High capacity per subscriber High capacity per subscriber
-
2,500 3,500 4,500 5,500 6,500 7,500 8,500 9,500
1H02 CPOP/Cell
Source: ABN AMRO
The graph can be interpreted by looking at the four quadrants.
■ Quadrant One. Operators in this quadrant have the most flexibility with regard Quadrant One. Excess
capacity enables operators
to medium-term capex spend because they have the greatest available capacity
to generate high returns as
per subscriber. They offer the best network access because they have already they add new subscribers
built a lot of capacity into their footprint and have low market share. As these
operators obtain greater subscribers and market share, they will move towards
Quadrant Two, and therefore report significantly higher-than-average returns on
new invested capital.
■ Quadrant Two. Operators in this quadrant are normally in highly penetrated Quadrant Two. Operators
are generating high returns
markets and are relatively capacity-constrained, requiring additional spectrum
but are relatively capacity-
or improved technology. They have already built a lot of capacity into their constrained
footprint (fewer covered POPs per cell site) but because of their high market
share have less available capacity for each subscriber (greater number of
subscribers per cell site). Subscribers will experience network congestion during
peak periods. TIM Spa fits into this category.

60
SE CTO R DYNAM I CS
■ Quadrant Three. Operators in this quadrant have not yet built a lot of capacity Quadrant Three. Operators
face imminent future capex
into their footprint and have low capacity per each subscriber. Subscribers will
spend to meet existing
experience network congestion during peak periods; however, by spending subscribers’ needs
additional capex on new cell sites, these operators can improve their network
capacity. This quadrant is an indicator of high future capex. Most of Europe’s
major operators fit in this quadrant.
■ Quadrant Four. Operators in this quadrant have not yet built a lot of capacity Quadrant Four. Operators
can delay capex if
into their footprint, but have high capacity for each subscriber. If these
experiencing slow
operators are not experiencing rapid growth, they can delay capex. Mmo2 subscriber growth
Germany fits in this category.
These quadrants show that Hutchison3G could be relatively well placed compared
with its European peers, admittedly recognising its lower intended population
coverage (around 80%). When it is established, H3G’s network/footprint should be We can clearly observe
able to handle many more subscribers without needing substantial additional capex. emerging capacity
constraints for Europe’s
It is also important to note that it is not all rosy for the remaining European
major operators
operators with regard to capex spend and, in particular, capacity when compared
with the H3G’s potential. We can clearly observe the relative capacity
constraints of Europe’s most established operators, such as TIM Spa,
Vodafone Germany, Vodafone UK and Orange UK. In our view this is major
incentive for these operators to accelerate the deployment of 3G.
Enabling volume growth
To have confidence in our fixed to mobile substitution thesis, investors must believe
in the prospect of an imminent material increase in available voice capacity. The
capacity of a cellular system is dependent upon several factors, including:
■ the signal (analog or digital);

We have considered four
■ the coding and modulation method used in air interface; potential drivers of
increased capacity
■ the total number of cells in the network system; and
■ the amount of frequency,
(1) The signal – analog - AMPS (1G) to digital

Signals are either digital or analog. Analog signal transmission is a form of
electronic transmission accomplished by adding signals of varying frequency or
amplitude to carrier waves of a given frequency of alternating electromagnetic
current. This compares with a digital signal, which consists of a discrete set of data
points that can take on only fixed values of one or zero. The ones and zeroes are
called bits, which comes from the phrase binary digits.
In all segments of communication, a migration to digital from analog has been

occurring for several reasons: (1) Increased capacity on digital networks is possible
because data compression allows multiple digital channels and more economic
benefits; (2) signal quality is better and is nearly independent of distance and
network topology; (3) there is better resistance to interference; and (4) frequency
is more easily shared with other signals and is more flexible. Depending on the
digital technology (the coding and modulation), digital cellular has three to
15 times greater capacity than analog service. Both GSM and W-CDMA are
digital.

61
SE CTO R DYNAM I CS
(2) Coding and modulation

Three major cellular digital technologies are currently in use: TDMA, CDMA and Three major cellular
technologies co-exist
GSM. A fourth technology used to a lesser extent is iDEN. Each has own their own
advantages and disadvantages. For a more detailed “bottom-up” discussion of each
of these technology platforms, we refer investors to the detailed descriptions within
VoiceStream – The American Dream.
Table 40 : Comparison of the various mobile systems

System Advantages Disadvantages
TDMA Capacity advantage of 3X over analog Average sound quality, more digitised, than normal human
voice
More operators on GSM and CDMA
Does not have the capacity of GSM or CDMA
CDMA Has a nationwide footprint in the US Newest digital technology. Less experience about what could
happen with regard to the potential technology glitches or in
Has the most significant capacity gains over analog systems, of
overcapacity situations.
approximately at least 6x. Therefore, cost per minute of service
and capital expenditures could fall to low levels Initial costs to build CDMA are significantly higher, expensive
handsets
Fewer equipment manufacturers for CDMA than for GSM.

Hence, less economies of scale and therefore higher
Not widely used in Asia or Europe
GSM Most widely used in the world Footprint still not comparable with CDMA or TDMA in the US
Proven technology Present system not entirely suited for wireless data
Cheaper handsets, cheaper equipment
Has capacity gains of 3-5x over analog systems
IDEN Only technology that can currently support 3-1 function of Limited use and support from equipment providers, handsets
dispatch, cellular and paging
Less spectrum that other players
Packet-switched based network should work well with wireless
data
Source: Company reports and ABN AMRO
Because radio spectrum is a limited resource shared by all users, a method must be
devised to divide up the bandwidth among as many users as possible. The method
used by GSM is a combination of TDMA and frequency division multiple access
(FDMA). The FDMA part involves the division by frequency of the (maximum) 25
MHz bandwidth into 124 carrier frequencies spaced 200 kHz apart. One or more
carrier frequencies are assigned to each base station.
The introduction of 3G should lead to a structural shift in spectral efficiency (see The introduction of 3G
should lead to a structural
following discussion, “The total number of cells”), enables scalability and provides
shift in capacity
additional spectrum (see frequency allocation discussion above). 3G provides the
additional capacity to deploy bucket pricing plans and drive mass fixed-to-mobile
substitution.
(3) The total number of cells

Network systems are generally designed with a sufficient number of cells, and radio
frequencies assigned to each cell, to serve the estimated number of subscribers
and network traffic in each service area. As the capacity requirements in each cell
increase, additional radio frequencies may be used to cover the same area.

62
SE CTO R DYNAM I CS
Coverage and capacity trade-off

There is clear trade-off between coverage, capacity and quality of service in the 3G
wireless business model. 3G operates at a higher frequency than GSM and as a
result requires a greater density of cell sites. GSM 900 normally has a maximum There is a clear trade-off
between coverage, capacity
cell radius if about 35km due to its frame structure. GSM 1800 has at least 8-10dB
and QoS in the W-CMDA
weaker link budget, resulting in a typical cell radius of about 3-5km. model
The authors of this study do not profess to have a doctrate in physics. We

acknowledge the work of Claes Beckman from the the Royal Institute of Technology
in Stockholm that has given us the scope to calculate the capacity benefit of W-
CDMA over GSM based on different coverage scenarios.
GSM capacity (static channel allocation)

If we look at a very simple GSM network with static channel allocation and 5Mhz of
spectrum, we find that:
■ 5MHz divided over 200KHz per radio channel gives us 25 carriers;
■ 12 of those are control channels;
■ the other 13 traffic channels can be allocated in a 1/3 re-use; and We have calculated the
capacity benefit offered by
■ in total we get five carriers per sector, W-CDMA over 2G
We can then derive total capacity:
■ maximum 5 carriers per sector; and
■ 5*8-1 = 39 voice channels, with 2% blocking equates to 28 Erlangs/sector
W-CDMA capacity
We understand that the soft capacity of W-CDMA gives flexibility between
coverage, capacity and quality. As a cell sites coverage area increases, capacity
drops dramatically. We can calculate the capacity of a simple W-CDMA network
with 5MHz of spectrum assuming a cell site radius of 3.5km and one transciever,
which implies 42 voice channels per sector, with 2% blocking giving 33
Erlangs/sector.
W-CDMA vs GSM 1800

In a large cell site (3-5km)
Therefore, in a network with a cell site radius of 3-5km, 5 MHz of spectrum and one W-CDMA offers a 20%
transceiver, we get only a marginal improvement in capacity of about 20% from improvement to capacity
employing W-CDMA from GSM. This immediately raises the question, Why bother
with W-CDMA? The key is that in smaller cells, W-CDMA capacity increases
substantially. For example, in a cell site of about 2.6km, W-CDMA could provide up
to 86 voice channels per sector and about 75 Erlang (ie an improvement of nearly
170%). And this is based on one transceiver, which is scalable. It has been pointed
out to us that GSM could deliver similar capacity using modern techniques such as In a smaller cell site W-
CDMA offers a 127%
AMR, frequency hopping and 1/1 re-use, but would require 11 radios per sector. improvement in capacity
Hence W-CDMA can offer material capacity improvements. The secret is the
wideband 5MHz radio channel. We flag that should European operators deploy more
than one transceiver per cell site, available capacity could explode (ie an operator
could double a cell’s existing capacity from about US$50,000 of incremental capex
on an additional transceiver).

63
SE CTO R DYNAM I CS
(4) The amount of frequency

Before Europe began the process of 3G licensing, total 2G spectrum available in
Europe was similar to the 170MHz allocated in the US. However, most European
countries have now allocated on average an additional 130MHz of spectrum for 3G
use, bringing the total amount of spectrum available to an average of 220-330MHz.
Chart 24 : Allocated unpaired spectrum per country - pre and post 3G auctions
350 Spectrum and capacity

issues for larger US
300 operators in each market
250
Spectrum (Mhz)
200
150
100
50
0
United States France Spain Germany Italy United Kingdom
2G Spectrum 3G Spectrum
Source: ABN AMRO
Existing frequency
A simplistic approach to assessing an operator’s capacity is to look at Mhz per
subscriber. This approach clearly favours operators with a smaller subscriber base
In our view, smaller
(eg VoiceStream in the US, T-Mobile in the UK). It is our view that the smaller operators have a greater
operators have greater ability to service their existing subscriber base and ability to dictate pricing
therefore dictate pricing. It also shows the magnitude of 3G spectrum waiting on-
line in Europe (total MHz per subscriber shows 2G and 3G specturm, vs. 2G only
MHz per subscriber). We can contrast Europe’s relative capacity depth to
constraints in the US.
Chart 25 : France MHz per subscriber Chart 26 : Germany MHz per subscriber
8.00 9.00
8.00
7.00
7.00
6.00 6.00
5.00
5.00
4.00
4.00 3.00
3.00 2.00
1.00
2.00
0.00
Mannesmann
E2
E-Plus
DeTe Mobil
1.00
0.00
Itineris SFR Bouyges
MHz per subscriber 2G MHz per subscriber MHz per subscriber 2G MHz per subscriber
Source: ABN AMRO estimates Source: ABN AMRO estimates

64
SE CTO R DYNAM I CS
Chart 27 : Italy MHz per subscriber Chart 28 : Spain MHz per subscriber
4.00 60.00
3.50
50.00
3.00
40.00
2.50
2.00 30.00
1.50
20.00
1.00
10.00
0.50
0.00 0.00
TIM OPI Wind Telefonica AirTel Amena
MHz per subscriber 2G MHz per subscriber MHz per subscriber 2G MHz per subscriber
Source: ABN AMRO estimates Source: ABN AMRO estimates
Chart 29 : UK MHz per subscriber Chart 30 : US MHz per subscriber
4.00
4.50
4.00 3.00
3.50
3.00
2.00
2.50
2.00
1.00
1.50
0.00
1.00
Wireless
Cingular
AT&T Wireless
Sprint PCS
Voicestream
Communications
Verizon Wireless
0.50
0.00
Nextel
Vodafone
One2one
Orange
BTCellnet
MHz per subscriber 2G MHz per subscriber
Source: ABN AMRO estimates ABN AMRO does not cover Sprint PCS or Nextel Communications.
Source: ABN AMRO estimates, average top 50 markets
On average, there are 6-8 competitors in each US market compared with individual The average European has
European countries of 3-4 for 2G and 4-6 including 3G licences. While the average 40-50MHz, US operators
have 10-30 MHz
European operator has 40-50MHz, US operators make do with 10-30MHz.
Frequency outlook
Clearly this leads to spectrum and capacity issues for the larger US operators in
each market. In Europe, the 1900-2100 spectrum secured is currently
When unlocked, 3G
“locked up” and is useable only over IMT-2000 compliant infrastructure. A spectrum should provide
term of the license awards prohibits the use of this spectrum over existing 2G significant additional
infrastructure. The International Telecommunications Union (ITU) has defined capacity
384kbps as the data rate limit required for a service to fulfil the IMT-2000 standard
in a pedestrian environment. The 384 kbps data rate corresponds to 48 kbps per
time slot, assuming an eight time slot terminal. The key point is that when
unlocked, this spectrum should provide significant additional capacity (see following
section on modulation for discussion of potential capacity increase).

65
SE CTO R DYNAM I CS
3G – 1.7x-2.8x increase in capacity

Having identified two key drivers of additional capacity (spectral efficiency and
We have considered the
additional spectrum) we can put them together to get a feel for the potential
potential capacity increase
increase in capacity from the introduction of W-CDMA across Europe. For example, using a 20% improvement
even in the event of W-CDMA (3G) only operating 20% ahead of existing GSM in spectral efficiency
infrastructure and one transceiver we calculate a capacity increase of between 1.7x
(France) and 2.8x (Italy) to existing capacity (see following tables).
Table 41 : France: potential capacity increase

2G Total Capacity New Capacity
spectrum paired increase spectrum increase
spectrum (1:1) (1.2:1) (1.2:1)
Orange 24 39 1.6 42.0 1.8
SFR 24 39 1.6 42.0 1.8
Bouyges 26 41 1.6 44.4 1.7
Sub-total 74.4 119.4 1.6 128.4 1.7

Source: ABN AMRO
Table 42 : Germany: potential capacity increase

spectrum (1:1) (1.2:1) (1.2:1)
T-Mobil 17.4 27.4 1.6 29.4 1.7
Vodafone D2 17.6 27.6 1.6 29.6 1.7
E-Plus 22.4 32.4 1.4 34.4 1.5
mmo2 (Viag) 22.4 32.4 1.4 34.4 1.5
Mobilcom 0.0 10.0 na 12.0 na
Quam 0.0 10.0 na 12.0 na
Sub-total 79.8 139.8 1.8 151.8 1.9

Source: ABN AMRO
Capacity is increased by uy
between 1.7x-2.8x
Table 43 : Italy: potential capacity increase

spectrum (1:1) (1.2:1) (1.2:1)
TIM 31.2 46.2 1.5 49.2 1.6
Vodafone-Omnitel 15.2 30.2 2.0 33.2 2.2
Wind 14.6 29.6 2.0 32.6 2.2
Andala 0.0 20.0 na 24.0 na
IPSE 0.0 20.0 na 24.0 na
Sub-total 61.0 146.0 2.4 163.0 2.7

Source: ABN AMRO

66
SE CTO R DYNAM I CS
Table 44 : Spain: potential capacity increase

spectrum (1:1) (1.2:1) (1.2:1)
Telefonica Moviles (Esp) 13.4 28.4 2.1 31.4 2.3
Vodafone Airtel 13.4 28.4 2.1 31.4 2.3
Amena 13.4 28.4 2.1 31.4 2.3
xfera 0.0 15.0 na 18.0 na
Sub-total 40.2 100.2 2.5 112.2 2.8

Source: ABN AMRO
Table 45 : UK: potential capacity increase

spectrum (1:1) (1.2:1) (1.2:1)
mmo2 (Cellnet) 5.8 15.8 2.7 17.8 3.1
Vodafone 5.8 20.8 3.6 23.8 4.1
T-Mobil 30.0 40.0 1.3 42.0 1.4
Orange 30.0 40.0 1.3 42.0 1.4
3 (Hutchison 3G) 0.0 15.0 na 18.0 na
Sub-total 71.5 131.5 1.8 143.5 2.0

Source: ABN AMRO
Of course, this dramatically understates the spectral efficiency of W-CDMA relative

to GSM, in the event of more dense cell site coverage in major urban areas. For
example, using the top end of expectations (75 Erlang), W-CDMA offers nearly
170% more capacity than existing 2G infrastructure (in other words, a ratio of
2.7:1.0) using only one transceiver.
Using a 2.7x we estimate that capacity is increased by between 2.6x and 5.0x. Capacity is increased
between 2.6x and 5.0x
Even at this conservative level this is a material increase in capacity. Looking
using more aggressive
forward, as operators add further transceivers to each cell site, an explosion in estimate of spectral
capacity could occur at a very low incremental capital expenditure (i. real estate efficiency
costs, and civil engineering costs have already been met). We also note that
eventually existing spectrum allocated to 2G will be re-farmed for use with W-
CDMA technology. This would drive a minimal 20% increase in the 2G spectrum to
which in the analysis above we apply a 1x multiple. Hence our capacity increase
estimates appear very conservative.
Capacity increase varies market by market

We also note the sharp difference observable between countries. Italy clearly
stands out as the market with the greatest increase in available capacity. This has
implications for the timing of the introduction of bucket pricing plans and the speed Italy stands out as having a
of fixed to mobile substitution. At the other end of the spectrum, France will see a particularly stark increase
in capacity
less significant increase in supply, suggesting a slower emergence of fixed-to-
mobile substitution.

67
SE CTO R DYNAM I CS
The scope for EDGE (2.5G) as an interim solution?

EDGE is the final network-upgrade-stage technology for GSM operators. The EDGE increases data
transmission rates and
objective of the new technology is simple – to increase data transmission rates and
frees up voice capacity
spectrum efficiency and to facilitate new applications and increased capacity for
mobile use.
EDGE uses a new modulation technique and improvements in radio protocol that
allows existing operators to use existing frequencies more efficiently (800, 900,
1800 and 1900 MHz). Software upgrades in the base station system enable use of
The core network does not
the new protocol facilitating packets over the radio interface in both the base require any adaptations
station and base station controller; new transceiver units in the base station enable
use of the new modulation technique. The core network does not require any
adaptations. EDGE increases theoretical end-user data rates up to 384Kbps (3
times speed of GPRS); however, realistic data rates range between 80Kbps and
150Kbps.
Compatibility of existing infrastructure

We note many commentators suggest that only GSM base stations deployed during
the past two years are capable of being upgraded to EDGE. This is in fact incorrect.
For example, the Ericsson RBS 2000 macro products (produced since 1995) can be
easily upgraded to EDGE.
EDGE: Capacity implications

EDGE also enables each transceiver to carry more voice and/or data traffic (see
following chart). In a typical TDMA system, like GSM, a single frequency is shared
EDGE enables operators to
by eight time slots. The capacity increase with EDGE means that a single
carry more voice
subscriber, or a number of subscribers using, (say) three time slots with GPRS,
could be handled with just one time slot. Consequently, capacity is freed up for
other uses (data or voice)., although we note coverage is lost.
Chart 31 : EDGE capacity benefits
Voice Voice Voice Voice Voice Data Data Data
Voice
Standard GSM Transceiver
Voice Voice Voice Voice Voice Free Free Data

TS TS
Voice
EDGE Transceiver
Freed up capacity!
Source: ABN AMRO/Nokia
Should scepticism grow further among investors and operators over the potential
Should scepticism grow
for revenue from 3G wireless services in the foreseeable future (2-3 years), some among investors over 3G,
operators may question if they need to incur the substantial costs of upgrading EDGE may be explored
their networks beyond EDGE to 3G. further

68
SE CTO R DYNAM I CS
US perspective
Operators see little need for higher data rates within the current suite of products
being developed. Supporting this view for US GSM operators (AWE, Cingular and Arguably US operators
provide a sanity check
VoiceStream) is that fact that timing of availability of sufficient additional spectrum
is very unclear. Cingular, the second-largest US wireless operator, has announced
that it had no plans after EDGE to roll out W-CDMA.
European options
European operators that have no such spectrum issues following the recent 3G
auctions have long been expected to bypass EDGE in their rush to deploy third-
generation WCDMA. However, there stands a greater chance now that European
operators also concerned with returns on additional capex may decide to delay
investing in 3G infrastructure. Obviously, this would require changes in
laws/regulations regarding scheduled build-out licence requirements and use of
new 3G spectrum. However, it is important that investors are aware of this issue
and its implications for free cash flow.
What about the regulators?

A fly in the ointment is the specified technology that each operator has nominated
for deploying additional spectrum. For example, in the UK, each of the existing
Regulatory changes would
operators specified UMTS as the technology platform they intended to deploy to
be required and ...
fulfil the IMT 2000 standard. To unlock the spectrum granted, each of the UK
operators would have to approach OFCOM (the UK telco regulator) to request a
change in their license conditions (ie to enable the use of the 3G spectrum, with
EDGE). Our conversations with the UK operators suggest that this is a possibility.
What about handsets?

To date, no EDGE handsets are available. This represents a major impediment to
EDGE deployment. In theory, the development of an EDGE handset should be
... new handsets would
easier than dual mode (3G and 2G) handset technology. Starhub, the No. 2 need to developed
Singapore operator, had indicated it is working with Nokia to develop an EDGE
handset for deployment in its market. We anticipate similar operator pressure from
the US.
Should more operators consider EDGE as a capacity-enhancing solution, the vendor

community would come under greater pressure to deliver an appropriate handset.
This would be similar to the process behind the development of the GAIT handset
(TDMA/GSM) in the US, when the vendor community was cajoled into developing
the terminal by AT&T Wireless and Cingular. While the widespread deployment of
EDGE throughout Europe remains a wildcard, it is important to be aware of its
implications.

69
SE CTO R DYNAM I CS

70
N E W S H I G H L I G H T
Fixed-line tariff rebalancing
Fixed-line tariff rebalancing is a significant topic that merits more space than we
have in this note (hence our intention to revisit it at a later date). However, we
must recognise that fixed line tariff rebalancing provides a potential third driver of
material fixed-to-mobile call substitution. Should fixed line players be forced to
rebalance their tariffs (ie increase the price of access and drop the price of calls),
we suspect the wireless handset could become the primary telephony access point
for many customers. Political pressure on regulators to maintain universal access
provides a barrier to rebalancing. Notwithstanding this issue, we suspect operators
could overcome the obstacle via the provision of a bundled wireless alternative.
Background
Long distance fixed-line voice calls in Europe typically do not reflect underlying
network cost structure (for example, a fixed-line long distance call may cost Fixed-line telephony prices
approximately €0.04, compared with as little as €0.02 for local traffic, without any do not reflect underlying
network cost structures
underlying cost difference). The difference in cost is an incumbent telephony
legacy, in which distance was a key driver of cost. The importance of distance as a
driver of cost has diminished (reflecting lower backhaul cost, driven by technology
innovation and the removal of physical operators connecting calls).
Arguably Europe’s wireline operator community are overcharging customers for

long distance and international call charges, and subsidising the provision of local Arguably wireline operators
are overcharging for long
access, fulfilling political pressure for universal access. We suspect there is an
distance and international
opportunity for Europe’s wireless operators to lobby country regulators to force calls, and subsidising local
incumbent operators to rebalance their tariffs to reflect underlying network cost access
structure.
Wireless opportunity
In our central case analysis of the fixed-to-mobile call market, we conservatively In the event of fixed-line
tariff rebalancing, our local
assumed the migration of only about 30% of all local call volume from the wireline
access substitution scenario
to the wireless network. We further assumed that no fixed-line access fees would (30%) could prove
be displaced to wireless. We suspect that in the event of tariff rebalancing, this conservative
could be overly conservative.
In our view, a core opportunity for Europe’s cellular operators pursuing fixed-to-
mobile substitution is provided by the local access/local call volume portion of voice
fixed-line calls and fixed to mobile calls. As we noted earlier in this research study,
as fixed-line voice call volume falls the implied price of each incremental unit of use
increases, thus encouraging savvy customers to re-assess their need for a wireline
connection to the home/office.

71
NE WS HIG HL IG HT
Easing of regulatory stance

Across Europe, the prices for cellular are dropping more quickly than fixed call
prices. Going forward we expect this to continue, with increasing prospect that the
fixed call prices may be increased. It is noteworthy that cellular prices are even
decreasing in countries like Germany and France, where mobile prices are not
regulated. This implies that cellular markets can be competitive in their own right.
Lobbying opportunity
Looking forward, this environment may help the cellular operators to lobby Reduced cellular tariffs and
wireline tariff re-balancing
regulators to rebalance local call charges to reflect the underlying network cost
could be powerful
structure. It is possible that if tariffs were rebalanced (to reflect underlying cost stimulants of growth
structure), local calls could move towards the price of long distance calls.
Dramatically reduced European cellular tariffs, coupled with rebalancing of fixed-
line voice tariffs, could be a powerful stimulant of the next phase of cellular growth.
US perspective
We note that in the US in 1993, President Clinton mandated that auctions be held
for frequencies within the PCS airwaves. The primary motive was to open up
competition in the cellular market, drive down the cost of wireless driving
individuals to use their cellular phone for local calls.
We suspect that more proactive lobbying of National Regulatory Authorities (NRAs)

by Europe’s cellular operator community could result in a more favourable
regulatory environment during the medium term. At a recent lecture given by Dave The future: one wireline
and three wireless
Edmonds (head of OFTEL) and Gregory Sidack (ex-deputy counsel of the FCC),
networks?
both commentators indicated a belief in one wireline network and up to three
wireless networks contesting telephony markets longer term. Recognition of this
dynamic is important if Europe’s wireline operators are to rip a more material chunk
of the local market away from the wireline incumbent.

72
R I S K A N A L Y S I S
Economic impact
We see between 10% and 25% upside from our existing fair value estimates for the
European wireless community as a result of (1) our pre-to-post-paid migration
thesis and (2) our central case fixed-to-mobile substitution belief. Our published
price targets do not factor in the upside from a structural shift in fixed-to-mobile
substitution. As evidence builds supporting our thesis we will re-visit our published
fair value estimates. Full bottom-up detail by major European wireless operator is
detailed in the sister publication, Wireless Model Builder: Edition 1. Current cellular
share prices bet against the emergence of mass fixed-to-mobile substitution. We
recognise that margins could contract in the event of a change in business model.
Investors will need to be nimble, timing will be crucial.
Timing
Two distinct positive drivers
As already noted we see two distinct positive drivers of ARPU looking forward:
of ARPU growth
■ Phase One (2002-2006): 4%-8% annual growth in blended ARPU from pre-
to-post-paid migration for the next four years; and
■ Phase Two (2004-2006): an 18% structural shift in outgoing voice ARPU by 1) Pre-to-post migration,
and
2006 (holding all other factors equal 11% overall ARPU growth)
2) F2M substitution
Holding incoming ARPU constant, this results in up to 30% overall ARPU growth in
by year-end 2006. This would be a material upside surprise to consensus
expectations.
Valuation: Published case

Reflecting our core views on the outlook for ARPU, we have re-set our fair value
estimates by operator. Given the profound impact of Phase Two (fixed-to-mobile
substitution), we have run our models excluding the structural shift in outgoing Our published price targets
voice ARPU. In the following table we detail out fair estimates by major operator. exclude the upside from a
structural shift in ARPU
Full bottom-up detail by major European wireless operator is detailed in the sister
publication accompanying this study, Wireless Model Builder: Edition 1.
Table 46 : ABN AMRO fair value

New price target Old price target New Old
recommendation recommendation
mmo2 0.60 0.52 Buy Add
Orange 7.0 7.5 Buy Buy
TIM 4.1 4.4 Hold Hold
TEM 6.0 6.7 Hold Hold
Vodafone 1.05 1.05 Add Add

Source: ABN AMRO

73
R IS K ANALY SIS
Market-by-market positioning
With the help of David Wright from our wireline telecoms team, we have developed
a schema to better understand the impact of the fixed-to-mobile substitution
opportunity/threat by major European market. This schema highlights two key
drivers of a market’s attractiveness:
■ its level of market share concentration (the Herfindahl index); and
■ the difference in price per minute between an average fixed and cellular call.
We have then disaggregated schema into four quadrants to enable easier

interpretation (see following chart). The location of an individual operator on the
schema determines the relative attractions of the asset from a wireline perspective.
Figure 7 : European market matrix: relalative resilience

7500
7000
Intermodal HH, outgoing voice minutes
6500
6000
5500
5000
4500
4000
3500
3000
2500
1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
mob/fix price factor
Spain Italy Portugal Germany Uk Holland France
Source: ABN AMRO
Quadrant One (top left): Bad long term, bad short term, is the worst possible
place for an incumbent wireline operator to occupy on the schema. Operators in
this quadrant have a high level of concentration (indicated by the monopoly style
HHI score on the y-axis) suggesting negative regulatory intervention and mobile
pricing has fallen below the crucial substitution level of 1.5x (see earlier discussion
on fixed to mobile for rationale).
Quadrant Two (bottom left): Good long term, bad short term, is marginally
better in terms of the level of concentration, suggesting adverse regulatory
intervention is unlikely. Against mobile enjoys large-scale substitution in markets in
this schema, as a result of the decline in relative price. Portugal Telecom sits in
this quadrant. As we noted in European Cellular Bear Trap, Portugal has been
charecterised by the emergence of fixed call volume substitution to mobile. Minutes
of domestic calls have fallen by 8% pa, and 10% of households use mobile only,
compared with the UK at an estimated 4%.

74
R IS K ANALY SIS
Quadrant Three (top right): Bad long term, good short term, is a sweet spot
for existing returns for incumbent wireline operators, as they enjoy high levels of
market concentration (ie low competition) and mobile is still priced at a high
premium to fixed, suggesting little imminent impact from fixed to mobile
substitution. Spain and the Netherlands fit into this category. Longer term, the
prospect of fixed-line tariff rebalancing (increasing competition, lowering the HHI)
and aggressive 3G-driven price cuts suggest that a downward and “left” trajectory
will be established.
Quadrant Four (bottom right): Good long term, good short term, arguably is
the sweet spot for both existing and future returns for incumbent wireline
operators, as they have relatively minimal regulatory risk, and continue to display a
high premium for wireless prices over wireline (delaying the impact of substitution).
BT and DT sit within this quadrant.
Fixed to mobile case

While local access fees may be largely retained by wireline telephony operators, we A structural shift in ARPU
are expecting a large percentage of fixed-line voice calls to migrate to Europe’s offers between 10%-25%
upside from our published
cellular networks. This view has a profound implication for European ARPU growth
valuations
and fundamental valuation. Including our central case outlook for outgoing voice
ARPU (based on our work in the UK market) suggests upside of between 10%
and 25% from our fair value calculations. As more evidence emerges during
the medium term, supporting our thesis, we will be looking to revisit our published
price targets in anticipation of crystalisation of upside.
... but look out for margins

We are bullish regarding the outlook for fixed-to-mobile substitution to provide the
“next leg of growth” for Europe’s wireless operators. Notwithstanding this view,
ahead of the improvement in returns, lower prices will hit profitability, due to
declining gross telephony margins.
In the following table we have presented a framework for identifying the risk to We have considered the
impact on margin
margins. In the left-hand column we have the potential volume increase from fixed
to mobile substitution; in the top row we detail the expected price decline per
minute. For example, using an initial price per minute decline of 30%, a volume
increase of 10% results in a margin decline of 13%. We can envisage a run-rate of
decline moving diagonally upwards towards the top right of the table. The exact
margin contraction will be determined by the volume increase that follows the price
cuts.

75
R IS K ANALY SIS
Table 47 : Generic example of margin impact (for given price and volume)
Price per minute fall
0% -30% -35% -40% -45% -50% -55% -58%
Volume increase 180% 22% 20% 18% 15% 13% 9% 7%
150% 19% 17% 15% 12% 9% 5% 2%
120% 15% 13% 11% 8% 4% 0% -4%
80% 9% 6% 3% 0% -5% -10% -14%
50% 2% -1% -5% -9% -15% -21% -26%
20% -8% -12% -17% -23% -29% -38% -43%
10% -13% -18% -23% -29% -36% -45% -51%

Source: ABN AMRO
Managing this transition could prove challenging. Ironically, operators with

exposure to the corporate market (eg Vodafone), could suffer most, as they
cannabilise higher value customers. This is the primary reason we have not
upgraded our recommendations for Europe’s wireless operators in aggregate.

76
Table 48: Generic adoption model
R IS K
Year end December Units X+0 X+1 X+2 X+3 X+4 X+5 X+6 X+7 X+8 X+9 X+10 X+11
Country X Population m 65.7 66.3 67.0 67.6 68.3 69.0 69.7 70.4 71.1 71.8 72.5 73.2
Penetration rate % 75.0 76.0 77.0 78.0 79.0 80.0 81.0 82.0 83.0 84.0 85.0 85.0
All subscribers (period end) m 49.2 50.4 51.6 52.8 54.0 55.2 56.4 57.7 59.0 60.3 61.6 62.3
ANALY SIS
Country X Net Adds per period m 1.14 1.16 1.17 1.19 1.21 1.23 1.25 1.27 1.29 1.31 1.33 0.62
Market Churn % per month % 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
Country X Churn per period m 11.5 11.8 12.1 12.4 12.7 13.0 13.2 13.5 13.9 14.2 14.5 14.8
MO BILE
Country X Gross Adds m 12.7 13.0 13.3 13.6 13.9 14.2 14.5 14.8 15.1 15.5 15.8 15.4
Operator Y Subscribers (period end) m 13.0 13.2 13.3 13.5 13.7 14.0 14.3 14.5 14.8 15.1 15.5 15.6
Net Adds per period m 0.05 0.11 0.16 0.20 0.23 0.25 0.27 0.28 0.30 0.31 0.32 0.14
NET WO RKS
-
Churn % per month % 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
W.
Churn per period m 3.1 3.1 3.2 3.2 3.2 3.3 3.4 3.4 3.5 3.6 3.6 3.7
Gross Adds m 3.17 3.24 3.32 3.39 3.47 3.55 3.62 3.70 3.78 3.87 3.95 3.85
Share of Gross Adds % 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% 25.0%
E UROPE
Share of Net Adds % 4.7% 9.8% 13.6% 16.5% 18.6% 20.2% 21.4% 22.3% 23.0% 23.5% 23.9% 23.2%
Overall Market Share % 26.5% 26.1% 25.8% 25.6% 25.5% 25.3% 25.2% 25.2% 25.1% 25.1% 25.1% 25.1%
Contract Subscribers (period end) m 4.17 5.11 5.88 6.50 7.02 7.46 7.85 8.19 8.49 8.77 9.04 9.18
Contract Net Adds per period m 0.27 0.95 0.76 0.63 0.52 0.44 0.38 0.34 0.31 0.28 0.26 0.14
16
Contract Churn % per month % 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
Contract Churn per period m 0.9 1.0 1.2 1.4 1.6 1.7 1.8 1.9 2.0 2.0 2.1 2.2
Contract Gross Adds m 1.2 1.9 2.0 2.0 2.1 2.1 2.2 2.2 2.3 2.3 2.4 2.3
O CT OBE R
Contract Share of overall net adds % 504% 834% 477% 318% 231% 178% 143% 120% 103% 92% 83% 100%
200 2
Contract Share of overall gross adds % 38% 60% 60% 60% 60% 60% 60% 60% 60% 60% 60% 60%
Contract % of Total Subs % 32% 39% 44% 48% 51% 53% 55% 56% 57% 58% 58% 59%
77
Table 49: Generic adoption model (cont’)
R IS K
Prepaid Subscribers (period end) m 8.87 8.04 7.44 7.01 6.72 6.52 6.41 6.35 6.34 6.37 6.42 6.42
Prepaid Net Adds per period m -0.22 -0.83 -0.60 -0.43 -0.30 -0.19 -0.12 -0.06 -0.01 0.03 0.05 0.00
Prepaid Churn % per month % 2.0 2.1 2.1 2.1 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
ANALY SIS
Prepaid Churn per period m 2.2 2.1 1.9 1.8 1.7 1.6 1.6 1.5 1.5 1.5 1.5 1.5
Prepaid Gross Adds m 1.97 1.30 1.33 1.36 1.39 1.42 1.45 1.48 1.51 1.55 1.58 1.54
Prepaid Share of overall net adds % -404% -734% -377% -218% -131% -78% -43% -20% -3% 8% 17% 0%
MO BILE
Prepaid Share of overall gross adds % 62% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40%
Prepaid % of Total Subs % 68% 61% 56% 52% 49% 47% 45% 44% 43% 42% 42% 41%
High Value Contract Subscribers (period end) m 4.17 3.75 3.25 2.77 2.42 2.16 1.97 1.83 1.73 1.66 1.62 1.23
NET WO RKS
-
High Value Net Adds per period m 0.27 -0.42 -0.50 -0.47 -0.35 -0.26 -0.19 -0.14 -0.10 -0.07 -0.04 -0.39
W.
High Value Churn % per month % 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
High Value Churn per period m 0.9 1.0 0.9 0.8 0.7 0.6 0.5 0.5 0.4 0.4 0.4 0.4
High Value Gross Adds m 1.2 0.6 0.4 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.0
E UROPE
High Value Share of overall contract net adds % 100% -44% -66% -76% -68% -59% -50% -41% -32% -24% -16% -273%
High Value Share of overall contract gross adds % 100% 30% 20% 15% 15% 15% 15% 15% 15% 15% 15% 0%
High Value % of Total Contract Subs 100% 73% 55% 43% 34% 29% 25% 22% 20% 19% 18% 13%
Phase 1 Contract Subscribers (period end) m 1.36 1.83 1.90 1.96 2.02 2.08 2.14 2.19 2.25 2.30 2.33
16
Phase 1 Net Adds per period m 1.36 0.47 0.07 0.06 0.06 0.06 0.06 0.05 0.05 0.05 0.03
Phase 1 Churn % per month % 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
Phase 1 Churn per period m 0.0 0.3 0.4 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.6
O CT OBE R
Phase 1 Gross Adds m 1.4 0.8 0.5 0.5 0.5 0.5 0.6 0.6 0.6 0.6 0.6
200 2
Phase 1 Share of overall contract net adds % 144% 61% 11% 12% 14% 15% 16% 18% 19% 20% 18%
Phase 1 Share of overall contract gross adds % 70% 40% 25% 25% 25% 25% 25% 25% 25% 25% 25%
Phase 1 % of Total Contract Subs % 27% 31% 29% 28% 27% 27% 26% 26% 26% 25% 25%
78
R IS K
Phase 2 Contract Subscribers (period end) m 0.80 1.11 1.37 1.57 1.74 1.88 1.99 2.10 2.18 2.24
Phase 2 Net Adds per period m 0.80 0.32 0.25 0.20 0.17 0.14 0.12 0.10 0.09 0.05
Phase 2 Churn % per month % 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
ANALY SIS
Phase 2 Churn per period m 0.0 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.5 0.5
Phase 2 Gross Adds m 0.8 0.5 0.5 0.5 0.5 0.6 0.6 0.6 0.6 0.6
Phase 2 Share of overall contract net adds % 104% 51% 49% 46% 43% 41% 38% 36% 34% 38%
MO BILE
Phase 2 Share of overall contract gross adds % 40% 25% 25% 25% 25% 25% 25% 25% 25% 25%
Phase 2 % of Total Contract Subs % 14% 17% 19% 21% 22% 23% 23% 24% 24% 24%
Phase 3 Contract Subscribers (period end) m 0.71 1.27 1.71 2.06 2.34 2.58 2.77 2.93 3.39
NET WO RKS
Phase 3 Net Adds per period m 0.71 0.56 0.44 0.35 0.28 0.23 0.19 0.16 0.45
-
Phase 3 Churn % per month % 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
W.
Phase 3 Churn per period m 0.0 0.2 0.3 0.4 0.5 0.6 0.6 0.7 0.7
Phase 3 Gross Adds m 0.7 0.7 0.7 0.8 0.8 0.8 0.8 0.8 1.2
E UROPE
Phase 3 Share of overall contract net adds % 114% 107% 99% 92% 84% 76% 69% 62% 318%
Phase 3 Share of overall contract gross adds % 35% 35% 35% 35% 35% 35% 35% 35% 50%
Phase 3 % of Total Contract Subs % 11% 18% 23% 26% 29% 30% 32% 32% 37%
Operator Y_Country X Subscriber Summary
16
High Value Period End Contract Subscribers m 4.2 3.7 3.2 2.8 2.4 2.2 2.0 1.8 1.7 1.7 1.6 1.2
Phase 1 Period End Contract Subscribers m - 1.4 1.8 1.9 2.0 2.0 2.1 2.1 2.2 2.2 2.3 2.3
Phase 2 Period End Contract Subscribers m - - 0.8 1.1 1.4 1.6 1.7 1.9 2.0 2.1 2.2 2.2
O CT OBE R
Phase 3 Period End Contract Subscribers m - - - 0.7 1.3 1.7 2.1 2.3 2.6 2.8 2.9 3.4
Total Period End Contract Subscribers m 4.2 5.1 5.9 6.5 7.0 7.5 7.8 8.2 8.5 8.8 9.0 9.2
200 2
Total Period End Prepaid Subscribers m 8.9 8.0 7.4 7.0 6.7 6.5 6.4 6.4 6.3 6.4 6.4 6.4
Total Period End Subscribers m 13.0 13.2 13.3 13.5 13.7 14.0 14.3 14.5 14.8 15.1 15.5 15.6
High Value Average Contract Subscribers m 4.0 4.0 3.5 3.0 2.6 2.3 2.1 1.9 1.8 1.7 1.6 1.4
Phase 1 Average Contract Subscribers m - 0.7 1.6 1.9 1.9 2.0 2.1 2.1 2.2 2.2 2.3 2.3
Phase 2 Average Contract Subscribers m - - 0.4 1.0 1.2 1.5 1.7 1.8 1.9 2.0 2.1 2.2
Phase 3 Average Contract Subscribers m - - - 0.4 1.0 1.5 1.9 2.2 2.5 2.7 2.9 3.2
Total Average Contract Subscribers m 4.0 4.6 5.5 6.2 6.8 7.2 7.7 8.0 8.3 8.6 8.9 9.1
Average Prepaid Subscribers m 9.0 8.5 7.7 7.2 6.9 6.6 6.5 6.4 6.3 6.4 6.4 6.4
Average Total Subscribers m 13.0 13.1 13.2 13.4 13.6 13.9 14.1 14.4 14.7 15.0 15.3 15.5
79
R IS K
KPI data & Assumptions
High Value Contract ARPU £ 40.0 41.2 42.4 43.7 45.0 46.4 47.8 49.2 50.7 52.2 53.8 55.4
% change PoP % 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0%
ANALY SIS
Phase 1 Contract ARPU £ 32.0 33.0 33.9 35.0 36.0 37.1 38.2 39.4 40.5 41.8 43.0
% change PoP % -20.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0%
Phase 2 Contract ARPU £ 25.6 26.4 27.2 28.0 28.8 29.7 30.6 31.5 32.4 33.4
MO BILE
% change PoP % -20.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0%
Phase 3 Contract ARPU £ 20.5 21.1 21.7 22.4 23.1 23.7 24.5 25.2 25.9
% change PoP % -20.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0%
Total Blended Contract ARPU (Calculated) £ 40.0 39.8 38.5 36.8 35.4 34.7 34.6 34.7 35.1 35.7 36.4 36.7
NET WO RKS
-
% change PoP % -0.4% -3.5% -4.4% -3.8% -1.8% -0.5% 0.5% 1.2% 1.6% 2.0% 0.7%
W.
Prepaid ARPU £ 9.0 9.2 9.4 9.6 9.7 9.9 10.1 10.3 10.5 10.8 11.0 11.2
% change PoP % 2.0% 2.0% 2.0% 2.0% 2.0% 2.0% 2.0% 2.0% 2.0% 2.0% 2.0%
Blended ARPU (subscriber revenue only) £ 18.6 20.0 21.4 22.1 22.5 22.9 23.4 23.9 24.5 25.1 25.8 26.1
E UROPE
% change PoP % 7.7% 7.0% 3.1% 1.6% 1.9% 2.1% 2.3% 2.4% 2.5% 2.6% 1.4%
Revenues:
Total Contract revenues £m 1,935 2,218 2,536 2,729 2,869 3,017 3,174 3,340 3,515 3,699 3,893 4,011
% change PoP % 14.6% 14.3% 7.6% 5.1% 5.2% 5.2% 5.2% 5.2% 5.2% 5.2% 3.0%
16
Prepaid revenues £m 970 932 870 828 802 789 786 791 803 820 841 862
% change PoP % -4.0% -6.7% -4.8% -3.1% -1.6% -0.4% 0.6% 1.5% 2.1% 2.6% 2.4%
Total service revenues 2,905 3,150 3,405 3,557 3,671 3,807 3,961 4,131 4,318 4,519 4,734 4,873
O CT OBE R
Source: ABN AMRO
200 2
80
A P P E N D I X
European 3G licensing conditions
In this section we have presented regulatory requirements for rollout by major

European market. License duration ranges between 12 and 20 years, with sharply
divergent coverage requirements by country. In Korea, we note that SK Telecom
must launch service by 2003 (with KT Freetel following in 2004 and LG Telecom by
2005). In Singapore, 80% coverage is required by 2004.
Table 52 : European 3G license summary by country

Country Duration (years) Requirements
UK 20 80% pop by 1.12.07
Germany 20 25% of pop by 31.12.03. 50% by 31.12.05
France 20 80% of pop by T1+8 years.
Italy 20 Regional capitals <30 months. Provincial capitals within next 30 months
Netherlands 15 60% of population by 2007
Switzerland 15 50% population coverage by end of 2004
Austria 20 25% by 31.12.03. 50% by 31.12.05

(Originally all cities with a population greater than 250,000 by August 2001 and total investment by
Spain 20
four operators of €16.29bn into 3g networks by 2010)
Government now saying that it is largely free of investment obligations and push back launch date to
sometime in 2003
Sweden 15 In licence application. Generally 99.98% population coverage by 31.12.03
20% of the national population at the end of the first year of the validity period of the license; 40% at
Portugal 15
the end of the third year; 60% at end of the fifth year
Finland 20 No coverage obligation
Norway
90% of the population of Greater Oslo, Bergen, Stavanger/Sandnes, Trondheim,
12 Fredrikstad/Sarpsborg, Porsgrunn/Skien, Drammen, Kristiansand, Tromsø, Tønsberg/ Åsgårdstrand,
Sandefjord, and Bodø within five years
Denmark 15 End-04 30% population coverage, end 08-80%
Belgium 20 Start service, Sep-02. 30% coverage after 3 years

A license- 53% of the national population (equivalent to the five major cities) by the end of 2005 and
Ireland 20
with the fulfilment of the minimum 80% population requirement by the end of 2007
Source: ABN AMRO/Cellular News.com

81
APP ENDI X
Technical glossary
Erlang
An Erlang is a unit of telecommunications traffic measurement. Strictly speaking,
an Erlang represents the continuous use of one voice path. In practice, it is used to
describe the total traffic volume of one hour. For example, if a group of users made
30 calls in one hour, and each call had an average call duration of 5 minutes, then
the number of Erlangs this represents is worked out as follows:
■ minutes of traffic in the hour = number of calls x duration;
■ minutes of traffic in the hour = 30 x 5 = 150;
■ in the hour = 150/60 = 2.5; which gives
■ a traffic figure of 2.5 Erlangs
Trunking
Erlang traffic measurements are made to help telecommunications network
designers understand traffic patterns within their voice networks. This is essential if
they are to establish the necessary trunk group sizes. The more efficient the
existing lines are used, the higher the trunking efficiency is of the network.
Erlang B
Several traffic models exist that share their name with the Erlang unit of traffic.
They are formulae that can be used to estimate the number of lines required in a
network, or to a central office (PSTN exchange lines).
Erlang B is the most commonly used traffic model, and is used to work out how
many lines are required if the traffic figure (in Erlangs) during the busiest hour and
the number of blocked calls are known. The model assumes that all unlocked calls
are immediately cleared.

82
T E A M
Acknowledgements
We would like to thank Nimeshh Patel, from ABN AMRO’s wireline telecom equity
research team, for his invaluable insights and resource in researching global telco
volume and revenue trends by market and being a generally good bloke.

83
Companies mentioned: 0013.HK, ORA.PA, VOD.L
© Copyright 2002 ABN AMRO Bank N.V and affiliated companies ("ABN AMRO"). All rights reserved.
ABN AMRO Bank NV and/or one of the companies in the ABN AMRO Group have specific interests with respect to
ENEL and TIM, as defined under article 69, paragraph 1 of Consob Resolution n.11971/1999, as subsequently
amended and supplemented, and Consob Notice n.DME/102955 or 20 April 2001. In particular, ABN AMRO Bank
NV and/or its holding company has issued financial instruments related to the securities of ENEL and TIM.
This material1 was prepared by the ABN AMRO affiliate named on the cover or inside cover page. It is provided for informational
purposes only and does not constitute an offer to sell or a solicitation to buy any security or other financial instrument. While
based on information believed to be reliable, no guarantee is given that it is accurate or complete. While we endeavour to
update on a reasonable basis the information and opinions contained herein, there may be regulatory, compliance or other
reasons that prevent us from doing so. The opinions, forecasts, assumptions, estimates, derived valuations, and target price(s)
contained in this material are as of the date indicated and are subject to change at any time without prior notice. The
investments referred to may not be suitable for the specific investment objectives, financial situation or individual needs of
recipients and should not be relied upon in substitution for the exercise of independent judgment. ABN AMRO may from time to
time act as market maker, where permissible under applicable laws, buy or sell as agent or principal securities, warrants,
futures, options, derivatives or other financial instruments referred to herein. ABN AMRO, officers, directors, employee benefit
programmes or employees, including persons involved in the preparation or issuance of this material may from time to time
have long or short positions in securities, warrants, futures, options, derivatives or other financial instruments referred to in
this material. ABN AMRO may at any time solicit or provide investment banking, commercial banking, credit, advisory or other
services to the issuer of any security referred to herein. Accordingly, information may be available to ABN AMRO, which is not
reflected in this material, and ABN AMRO may have acted upon or used the information prior to, or immediately following its
publication. ABN AMRO may also within the last three years have acted as manager or co-manager for a public offering of
securities of issuers referred to herein. The stated price of any securities mentioned herein is as of the date indicated and is not
a representation that any transaction can be effected at this price. Neither ABN AMRO or other person shall be liable for any
direct, indirect, special, incidental, consequential, punitive or exemplary damages, including lost profits arising in any way from
the information contained in this material. This material is for the use of intended recipients only and the contents may not be
reproduced, redistributed, or copied in whole or in part for any purpose without ABN AMRO's prior express consent. In any
jurisdiction in which distribution to private/retail customers would require registration or licensing of the distributor which the
distributor does not have currently, this document is intended solely for distribution to professional and institutional investors.
Should you require additional information please contact your local ABN AMRO account representative, unless governing laws
dictate otherwise.
Australia: This document is distributed in Australia by ABN AMRO Equities Australia Ltd (ABN 84 002 768 701), a participating
organisation of the Australian Stock Exchange Ltd. Australian investors should note that this document was prepared for
wholesale investors only.
Canada: The securities mentioned in this material are available only in accordance with applicable securities laws and may not
be eligible for sale in all jurisdictions. Persons in Canada requiring further information should contact ABN AMRO Incorporated.
Hong Kong: This document is being distributed in Hong Kong by ABN AMRO Asia Limited to persons whose business involves
the acquisition, disposal or holding of securities.
India: Shares traded on stock exchanges within the Republic of India may only be purchased by different categories of resident
Indian investors, Foreign Institutional Investors registered with The Securities and Exchange Board of India ("SEBI") or
individuals of Indian national origin resident outside India called Non Resident Indians ("NRIs") and overseas corporate bodies
("OCBs"), predominantly owned by such persons or Persons of Indian Origin (PIO). Any recipient of this document wanting
additional information or to effect any transaction in Indian securities or financial instrument mentioned herein must do so by
contacting a representative of ABN AMRO Asia Equities (India) limited.
Italy: Persons in Italy requiring further information should contact ABN AMRO Bank N.V. Milan Branch.
New Zealand: This document is distributed in New Zealand by ABN AMRO Equities New Zealand Limited, a member firm of the
New Zealand Stock Exchange.
Russia: The Russian securities market is associated with several substantial risks: legal, economic and political ; and, high
volatility. There is a relatively high measure of legal uncertainty concerning rights, duties and legal remedies in the Russian
Federation. Russian laws and regulations governing investments in securities markets may not be sufficiently developed or may
be subject to inconsistent or arbitrary interpretation or application. Russian securities are often not issued in physical form and
registration of ownership may not be subject to a centralized system. Registration of ownership of certain types of securities
may not be subject to standardized procedures and may even be effected on an ad hoc basis. The value of investments in
Russian securities may be affected by fluctuations in available currency rates and exchange control regulations.
Singapore: This document is distributed in Singapore by ABN AMRO Asia Securities (Singapore) Private Limited to clients who
fall within the description of persons in Regulation 44(i) to (iii) of the Securities Industry Regulations. Investors should note
that this material was prepared for professional investors only.
United Kingdom: This document has been issued in the United Kingdom by ABN AMRO Equities UK Limited, which is registered
in England No 2475694 and is regulated by the Financial Services Authority. The investment and services contained herein are
not available to private customers in the UK.
United States: Distribution of this document in the United States or to U.S. persons is intended to be solely to major
institutional investors as defined in Rule 15a-6 under the U.S. Securities Act of 1934. All U.S. persons that receive this
document by their acceptance thereof represent and agree that they are a major institutional investor and understand the risks
involved in executing transactions in securities. Any U.S. recipient of this document wanting additional information or to effect
any transaction in any security or financial instrument mentioned herein, must do so by contacting a registered representative
of ABN AMRO Incorporated, Park Avenue Plaza, 55 East 52nd Street, New York, N.Y. 10055, U.S., tel + 1212 409 1000, fax
+1212 409 5222.
1
Material means all research information contained in any form including but not limited to hard copy, electronic form,
presentations, e-mail, SMS or WAP.

84

ABN-Amro - Mobile Networks - W

Uploaded by

Document Information

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

ABN-Amro - Mobile Networks - W

Uploaded by

Copyright:

Available Formats

250 Bishopsgate

London EC2M 4AA, UK

San Francisco ABN AMRO Delta Securities (Egypt) Kuala Lumpur

3G Tsunami: the revolution begins

Oslo Wireless stocks to rerate,

3G-tsnunami-cov 1 15/10/02, 12:03 pm

Mobile Networks - W. Europe

3G Tsunami: The revolution begins 16 October 2002

Longer-term flat rate access plans and pre-to-post-paid migration

Mmo2 Buy Add 60p

Orange Buy Buy €7.0

TIM Hold Hold €4.1

New entrant competition: We anticipate that H3G will adopt bucket

In Europe, 3G capacity enables voice substitution. We believe 3G 2435

will increase available capacity by up to 5x in some markets. As wireless

Ahead of rising ARPU, lower prices will hit profitablity, due to

(1M) (3M) (12M)

and TEM (€6 from €6.7).

Chart 1 : Relative share price performance (09/07/02 to date)

Source: ABN AMRO

Capacity and competition. The deployment of 3G networks throughout Europe will

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

■ Ahead of CYQ302 results we prefer Orange and mmo2. This is framed

■ During the medium term (into 2003), we would consider a change in

Fixed to mobile substitution offers a compelling second leg of growth. In the

Bucket pricing plans facilitate displacement and are superior. US operators

3G provides capacity to drive wireline volume displacement. European 2G

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

Data services dream 7

Fixed to mobile substitution 21

I NVES T MENT VIEW

Bucket pricing plans 45

Bucket plans generate higher ARPU and usage 47

Average yield based on actual minutes used 51

NEWS HIGH LIGH T

Fixed-line tariff rebalancing 71

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

European 3G licensing conditions 81

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

Data services dream

■ the absolute revenue required to generate a material upside surprise (€9bn);

■ the validity of operator-articulated targets for data as a percentage of revenue;

■ the potential cannibalisation of voice/SMS by MMS; and

■ a preliminary assessment of the margin/capex impact of take-up.

MO BILE NET WO RKS - W. E UROPE 16 O CT OBE R 200 2

Table 1 : E5 total revenue (€m) calendarised

France 8,863 12,343 15,046

% Change 39% 22%

Germany 10,729 15,057 16,480

Italy 11,671 13,808 15,534

% Change 18% 12%

Spain 6,523 8,107 9,615

% Change 24% 19%

UK 11,975 16,384 18,582

% Change 37% 13%

Total E5 49,761 65,699 75,257 Scale creates growth

Source: ABN AMRO estimate

Major operator revenue outlook

Table 2 : Revenue by major operator

mmo2 £m 4,815 241 481

Orange €m 16,877 844 1,688

TIM €m 10,648 532 1,065

TEM €m 9,096 455 910