KPMG Landscape Report Template

STRATEGIC AND COMMERCIAL INTELLIGENCE

Project Star 2007 Results 18th June 2008

ADVISORY

This document was produced by KPMG LLP in 2008. It has DISCUSSION PURPOSES ONLY been placed here for reference.

The information contained herein is of a general nature and is not intended to address the circumstances of any particular individual or entity. Although we endeavour to provide accurate and timely information, there can be no guarantee that such information is accurate as of the date it is received or that it will continue to be accurate in the future. No one should act on such information without appropriate professional advice after a thorough examination of the particular situation. Project Star 2007 Basis of preparation

 This presentation of key findings (‘our report’) is given in accordance with our agreed written terms of engagement dated 18 October 2007 and our letter to OLAF dated 3 August 2005 detailing the scope of our review of the contraband and counterfeit segments of the tobacco market within the 27 EU Member States. We draw your attention to the limitations in scope set out therein.  Our work has been ongoing since 1 November 2005. Our fieldwork for 2007 results is now complete.  In preparing our report, we have used a range of sources. Details of our principal information sources are set out in the Appendix and we have satisfied ourselves, so far as possible, that the information presented in our report is consistent with other information which was made available to us in the course of our work in accordance with the terms of our engagement letter. We have not, however, sought to establish the reliability of the sources by reference to other evidence. The scope of our work was different from that for an audit and, consequently, no assurance is expressed.  Our report makes reference to ‘KPMG Analysis’; this indicates only that we have (where specified) undertaken certain analytical activities on the underlying data to arrive at the information presented.  Our report is provided solely for the benefit of the parties identified in our engagement letter and should not be copied, quoted or referred to in whole or in part without our prior written consent. We will not accept responsibility to any other party to whom our report may be shown or who may acquire a copy of our report.

©2008 KPMG LLP, the UK member firm of KPMG International, a Swiss cooperative. All rights reserved. KPMG and the KPMG logo are registered trademarks of KPMG International, a Swiss cooperative. 1 Project Star 2007 Glossary of terms

Cigarette industry Project Star Bootlegging Also called small-scale smuggling, bootlegging is the purchase of tobacco CAGR Compound Annual Growth Rate products in one country for illegal consumption or resale in another country C&C Counterfeit and Contraband without paying the applicable taxes or duties Consumption Actual total consumption of cigarettes in a market, including legal IMS and Cigarette Any factory-made product that contains tobacco and is intended to be burned or illicit products as well as those purchased overseas to be brought back and heated under ordinary conditions of use smoked in market Contraband (CB) Genuine product that has been bought in a low-tax country and illegally re-sold Consumption gap The difference between total consumption and legal domestic consumption (for financial profit) in a higher priced market. There are generally two types of contraband: bootlegging and large scale smuggling/organised crime Consumption index A corroboratory source for estimating the consumption gap in a market. The model compares an index of consumption based on historic smoking Counterfeit (CF) Cigarettes that are illegally manufactured and sold by a party other than the prevalence, average consumption and population against legal domestic sales original trademark owner to track changes over time Duty-free Purchases made outside the EU that have no state, local or provincial taxes, EU European Union federal import duties or any other type of taxation added to the cost of the item EU Flows Model The primary methodology for measuring consumption in a market. The model purchased. Subject to purchase volume restrictions details the volume of inflows and outflows of product for a given market by MPPC Most popular price category country of origin (the model only specifies flows to EU countries) NMA / TMA National Manufactures Association / Tobacco Manufacturers Association Inflows/Outflows Inflows of non-domestic product into a market/outflows of product from a market OTP Other Tobacco Products (RYO/MYO, cigarillos, portions, rolls and cigars; excluding smokeless tobacco and water-pipe tobacco) LDS Legal Domestic Sales are sales of genuine product with domestic tax stamps through legitimate, domestic channels RYO/MYO Roll-your-own/Make-your-own - loose tobacco for the purpose of hand rolling / loose tobacco for the purpose of tubing LDC Legal Domestic Consumption is defined as legal domestic sales net of outflows Smoking prevalence The percentage of smokers in the total adult population ND Non-Domestic product – product that was not originally intended for the Stick Preferred unit of measurement of tobacco product, either one cigarette or its market in which it is consumed equivalent for OTP ND(L) Non-Domestic (Legal) – product that is brought into the market legally by Tobacco taxes The sum of all types of taxes levied on tobacco products. There are two basic consumers, such as during a cross-border trip methods of tobacco taxation: Normal or specific taxes are based on a set amount of tax per unit (e.g. cigarette); these taxes are differentiated according OLAF Office Européen de Lutte Anti-Fraude (European Commission Anti-Fraud to the type of tobacco. Ad valorem taxes are assessed as a percentage mark up Office) on a determined value, usually the retail selling price or a wholesale price and PMI Philip Morris International includes any value added tax Market research PMI data sources CAPI Computer-aided personal interviewing EPS Empty Pack Survey CATI Computer-aided telephone interviewing GCTS Global Consumer Tracking Survey IMS In-Market Sales (the primary source of legal domestic sales volumes) Measurements Bn Billion Mn Million

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

 Appendices

The primary deliverable of PMI’s co-operation with the Project Star is an European Commission assessment of the level of counterfeit and  PMI is required to fund studies for Finland estimating the size and composition contraband cigarettes of the total tobacco market (including across the EU markets counterfeit and contraband tobacco products) for the EU Estonia  Phase 1 of the research was Sweden conducted in 2006 in the 25 EU Latvia Member States Denmark  2007 data includes Bulgaria and Luxembourg Lithuania Romania to reflect the accession of Netherlands these countries to the EU in January Ireland 2007 UK Poland Germany Belgium Czech Republic KPMG’s approach to meet this Slovakia challenge Austria France Hungary Slovenia Romania  Develop a methodology to measure Italy the size of the legal, contraband and counterfeit markets for tobacco Bulgaria products  Develop a programme plan for Spain management of delivery Greece Portugal  Carry out, with third party research providers as required, measurement of counterfeit and contraband across Key:  Country covered from 2006 (Phase 1) Cyprus all EU Member States Malta  Country added to coverage from 2007 (Phase 2)  Country not covered

Fieldwork commenced in Project Star 2007 results: research timetable April 2007 and was 2007 2008 completed in January 2008 when full year legal April May June July Aug Sept Oct Nov Dec Jan Feb March April May June July domestic sales data Methodology review became available

Data analysis and Empty Pack Surveys modelling was undertaken during Quarter 1 2008 Agency set up Preliminary results were then tested with both PMI Country Management Non-domestic legal research Teams and external experts PMI Management and Member State Customs meetings Full 2007 results have now been finalised and are Data analysis and modelling contained in this report

Draft results

Data review and external challenge May 2008 Full 2007 results

Member State Communication

10th June 2008 Member State Presentation

Methodology design steps The methodology has been designed and tested Methodology Information assessment Pilot and refine Implement through several steps and design according to five  underlying principles Review available internal  Develop preliminary approach  Test methodology in three  Roll out approach to remaining information in pilot markets to C&C measurement pilot markets (Finland, 25 EU markets for 2006  Assess quality of information  PMI and OLAF approval to Germany and Poland) during  Addition of Romania and 2006  Identify gaps in data test methodology in three- Bulgaria in 2007 market pilot process availability and coverage  Evaluate results and refine  Scope to add further methodology Accession States in future years where appropriate

Methodology design principles

Consistent Corroborated Fact-based Pragmatic Flexible

 Our approach must be  We will seek to  The area of C&C is  Perfect measurement of  We need to be flexible in able to be applied in as corroborate key sources characterised by limited the C&C trade is not our approach and standardised a manner and overall methodology information but possible. We need to thinking in order to as possible across results to limit excessive significant emotion. Our have a practical and identify situations where markets to ensure all reliance on individual approach and feasible approach that a rigid methodology Member States are sources conclusions need to be will deliver results that would fail to capture the treated equally and fairly data-driven and impartial are robust, credible and market reliably fit for purpose  This flexibility includes: − modifying and improving our approach through the pilot phase and beyond − customising our approach where necessary to cater for specific market differences

 Project overview and timing

 Methodology

− overview

− limitations of results

 EU cigarette market

 2007 Project Star results

 Appendices

We have developed a The methodology has been tested extensively and refined to ensure that it can deliver the most robust and defensible results possible methodology for  Our approach comprises four steps: initial information assessment, preliminary methodology design, pilot and refinement, and then implementation quantifying C&C incidence across the 27 EU markets  Our approach integrates multiple sources and custom-built analytical tools

The methodology is based The EU Flows Model is a dynamic, iterative model that is principally based on legal domestic sales and Empty Pack Survey results primarily on objective  Legal domestic sales are the starting point of the methodology, from which outflows of legal sales to other countries are then subtracted to determine evidence from legal legal domestic consumption domestic sales and Empty Pack Survey results  Empty Pack Survey results provide the most credible indication of the incidence of non-domestic and PMI counterfeit packs by country of origin

Consumption index Analysis of long-term GCTS consumption results provides a further source with which to triangulate our consumption gap estimates modelling and regression  Consumption index modelling based on GCTS data provides a useful indicator of market trends and helps to corroborate the scale of non-domestic modelling provide further consumption. GCTS data cannot be used to quantify total consumption due to exclusions of particular age cohorts and under-reporting of corroboration of the extent consumption, but is helpful in tracking broad developments over time of non-domestic consumption  Consumption regression modelling has been used to understand the relationship between consumption and changes in disposable income and cigarette prices. These relationships have in turn been used to estimate 2007 consumption − please note that as we have not yet assessed the impact on consumption from factors other than price and income, estimated consumption can only be considered a broad approximation  Analysis of external data sources and a focused interview programme also delivered good supporting evidence with which to corroborate results

Primary market research The key objective of the market research programme was to quantify genuine, legal non-domestic tobacco purchases in each market was used to quantify legal  ND(L) data for 2007 Project Star results is based upon over 11,300 full interviews and almost 135,000 gross respondents. This research was updated non-domestic cigarette in 14 Member States during 2007, while Bulgaria and Romania were also included within the scope of research for the first time purchases  Primary research was critical to deliver robust results as no other sources of sufficient detail and accuracy are available for legal cross border shopping In addition to the research programme, ND(L) data is adjusted to reflect inbound visitor inflows from higher cost markets

There are some specific Given the innate complexity of measuring C&C, some limitations to accurate quantification are to be expected limitations to the results  There are broadly two types of limitations, scope exclusions and source limitations, which are covered in more detail in this section that our methodology delivers − scope exclusions include areas which cannot or have not been accounted for in our approach, such as geographic, brand (non-PMI counterfeit), category exclusions (OTP) and legal domestic product flows out of the EU − source limitations cover potential errors inherent with any data sources such as sampling criteria, coverage issues and seasonality factors

In order to maximise the Triangulation of results from alternative sources identified a few markets where country-to-country flows required minor adjustment accuracy of results some  In nearly all instances, overall country results and flows from the EU Flows Model appear reasonable minor refinements were necessary at a country level  However, in a limited number of instances, specific adjustments were made to country-to-country flows on the basis of sound supporting evidence

Overview of Project Star methodology and key terms Our methodology is based on a dynamic, iterative model driven primarily from We have used a top-down actual legal domestic sales and empty pack survey results approach to estimate PMI  Legal domestic sales are the key starting point for our methodology and are derived in most instances from shipment data counterfeit and − legal domestic sales are defined as sales of genuine product with domestic fiscal contraband volumes Non-domestic stickers through legitimate, domestic trade channels starting from total Other C&C  Outflows are defined as purchases of domestic product which are taken out of the country of purchase and consumed elsewhere C&C consumption − EPS provides a consistent source across all 27 markets of non-domestic packs by country of origin from which we estimate total product outflow from a market to the other 26 countries

% − we have not quantified outflows to non-EU markets LDS PMI contraband − however, given the high prices of cigarettes in Europe relative to the rest of the world, LDS outflows to the rest of the world are not expected to be material ND(L)  Legal domestic consumption (LDC) is defined as legal domestic sales net of PMI counterfeit outflows  Non-domestic product is defined as product of non-domestic origin, i.e. does not have a domestic tax stamp or is duty free LDS outflow  ND(L) is defined as legitimate purchases of product in a non-domestic market for personal use on return to the home market − legal cross-border shopping: buying duty paid tobacco product in a neighbouring Total Non-domestic Illegal non-domestic country for buyer’s own consumption in amounts allowable under customs consumption consumption consumption regulations − legal tourist shopping: buying tobacco products in a non-neighbouring country for buyers own consumption in amounts allowable under customs regulations − legal duty free sales: buying tax free products in amounts that are allowed under travellers’ allowances  Contraband is defined as genuine product that has been bought in a country and illegally re-sold (for financial profit) in another market. There are generally two types of contraband: − bootlegging: the purchase of tobacco products in one country for consumption or resale in another country without paying the applicable taxes or duties − large scale smuggling/organised crime: occurs when tobacco products are sold without payment of taxes or duties, even in their country of origin − for the purpose of this investigation, we are not able to quantify accurately the split between smaller scale bootlegging and large scale smuggling. Consequently our contraband incidence may be larger than some external observations anticipate  Counterfeit product is defined as non-genuine product that is not produced by or with authorisation of the original trade mark owner

Methodology steps and key information sources We have sought to triangulate our findings against alternative sources 1. 2. 3. 4. 5. 6.

Measure legal domestic Adjust legal Add total non- Deduct legal non- Deduct PMI Deduct PMI wherever available sales in all markets domestic sales for domestic domestic counterfeit contraband product outflows consumption to purchases from volumes from total volume from adjusted legal total non-domestic illicit purchase remaining illicit domestic sales to volume to arrive at volume volume to derive estimated illicit purchase arrive at non- total consumption volume PMI C&C Calculation step

Directly measured from Directly measured from Legal domestic ND(L) market Measured from Directly calculated by shipment data or ‘on the ground’ empty consumption plus research programme Empty Pack Survey the KPMG equivalent packs by country of EU Flows Model results methodology origin in 27 country flow model Primary source

 Tax stamp receipts  PMI management  Consumption index  PMI Duty Free  Expert interview  Expert interview estimates modelling market estimates programme programme  Euromonitor  Border sales surveys  Consumption  External research  Seizure data  Seizure data  Federal statistics regression modelling  source Expert interview programme

Corroboratory

Preliminary results are subject to testing and review with local PMI management in each of the 27 EU markets

Our approach integrates Primary Inputs Data Modelling and Iteration Final Output multiple sources with custom-built analytical tools

The results have then been through extensive Empty pack iteration and testing to survey finalise KPMG EU Preliminary non- Flows Model domestic results Model refinements are informed by gap analysis, Legal external public research domestic sales and interviews with both Final results cigarette manufacturers/ distributors and independent market Non Preliminary experts domestic Brand share counterfeit and (legal) validation contraband Further corroboration is research results provided by the KPMG consumption index and econometric analysis

Corroboration of results

KPMG External Econometric consumption interviews modelling index

Empty Pack Surveys provide a highly objective Overview Empty Pack Surveys are a system of collecting discarded empty cigarette packs, the results of which are used to and robust view of the estimate the share of non-domestic and counterfeit packs in each of the markets population samples and,  Results are based on a large sample of packs collected in various cities throughout the countries, although the collection plan differs by country. Accuracy and credibility of results is driven by sound design of the sampling plan notwithstanding some scope constraints,  Results are not subject to subject to respondent behaviour and are therefore less prone to sampling errors than many other alternative methodologies represents the most  Evidence is based on collected packs: no discrepancies or scope for respondent confusion credible indication of the  Data reflects actual overall non-domestic share and provides good snapshot of brands consumed incidence of non-domestic and counterfeit packs Process Empty Pack Surveys measure shares of total consumption and avoids potential errors associated with estimating volumes  Once packs are collected, they are sorted by manufacturer and the number of packs with domestic versus non-domestic tax stamps are counted to determine the proportion of packs that did not originate from that jurisdiction (including duty-free variants) − in cases where tax stamps are not shown on a packet, health warning and packaging characteristics are used to define the source market  In markets where collection is handled centrally, packs are sent to the manufacturers for analysis to determine which are genuine and which are counterfeit. Only the manufacturers can determine this, based on inks, paper and other characteristics. Results of these analyses are not released to competitors  Empty Pack Surveys can also be used to extrapolate overall consumption in the market by projecting LDS (net of outflows) using the percentage of non-domestic cigarettes in the market as found through Empty Pack Surveys Coverage Empty Pack Surveys are designed to be fit for purpose and the coverage per market is tailored by the size of the market, the likelihood of high non-domestic incidence and PMI’s share of the legal market  Large surveys (10,000 packs collected; all cities with over 100,000 inhabitants covered / at least 20 cities): France, Germany, Italy, Poland, Spain, UK  Medium surveys (5,000 packs collected, all cities with over 100,000 inhabitants or top 10 cities by population): Belgium, Czech Republic, Finland, Greece, Lithuania, Netherlands, Portugal, Romania, Slovakia, Sweden − note that in Austria the Empty Pack Survey is carried out by the Tobacco Manufacturers Association based on a sample over 10,000 packs collected from 601 communes. In Hungary, the Empty Pack Survey is carried out by the GfK Hungária Market Research Institute and is based on a sample of nearly 20,000 packs  Small surveys (1,000 – 2,000 packs collected, top 4 – 5 cities covered): Bulgaria, Cyprus, Denmark, Estonia, Ireland, Latvia, Luxembourg, Malta, Slovenia In 2007, Empty Pack Surveys were conducted in 22 of the 27 EU Member States  Data from 2006 results was used in the following countries: Cyprus, Czech Republic, Denmark, Luxembourg and Malta

German consumption gap estimates by source During the pilot phase, the We believe that Empty Pack Survey provides the most reliable indicator of non-domestic consumption advantages of using 20 18 18  The standalone consumption survey does not deliver a credible Empty Pack Survey results 18 consumption result without a significant uplift for under-reporting (in the 16 2002 order of a further 15-20%) to monitor non-domestic consumption 14 gap − estimates of consumption gap (i.e. the difference between legal consumption became clear 12 sales and actual consumption) of 7bn sticks for Germany from 10 consumer research results alone was too low given external 7 The case of Germany 8 estimates of ND(L), C&C and results from border studies 6 11-13

provides a good example Volume (bn sticks)  4 Econometric analysis provides corroboration of a significant growth in the consumption gap over the last few years of the benefits of this 2 approach 0 − however, while this analysis can support the quantification of a gap, ConsumerGfK gap Econometric gap EPS+LDS gap it would be dangerous to use it as the primary source survey gap (2002-2005)  Given consumption limitations, we believe that Empty Pack Surveys provide the most credible indicator of the incidence of non-domestic and illicit packs. Benefits of Empty Pack Surveys include;

Sources: (1) GfK consumption survey 2006 − less potential for non-sampling error (e.g. results are not distorted by (2) Legal Domestic Sales – PMI In Market Sales, In Market Sales 2006 01; KPMG analysis behavioural error such as systematic under-reporting of (3) German EPS market study 2005 consumption) − evidence based on physical collected packs so there are no discrepancies or potential for respondent confusion − reflects actual consumption and provides good snapshot of brands consumed Implied EPS + LDS German consumption gap calculation  Therefore, we have used the Empty Pack Survey results to derive a consumption gap for our analysis, and have focused on this estimate rather than those obtained from consumer research EPS Non-domestic − Empty Pack Survey results for 2006 show that non-domestic Domestic implied (consumption gap) purchases represented 17% of overall German consumption. As a consumption result, the total consumption in 2006 is estimated at 103.5bn sticks, with a consumption gap of approximately 18bn sticks % of packs collected 82.7% + 17.3% = 100%

Sticks (bn) based on 85.6 + 17.9 = 103.5 net LDS

Germany historical Yellow Bag Survey results Empty Pack Surveys can The low level of variance in German results highlights the validity of using Empty Pack Surveys to monitor trends in cigarette consumption provide a method for 25.0% 22% 22% 23% 20% 19% 20%  Empty Pack Surveys based on the Yellow Bag approach are conducted in monitoring trends in 20.0% 18% 18% 17% 17% Germany on an ongoing basis using the country’s network of recycling centres individual markets 15% 15% 15.0%  Packs are collected monthly, with the results released on a quarterly basis In countries where  The emergence and strength of the trend for increased non-domestic 10.0% historical EPS data is consumption has been apparent in each survey at both national and regional levels 5.0% available, there is a strong There is a strong correlation between changes in the German results and other Share of total of Share consumption sources including legal domestic sales, PMI shipment data and the observations 0.0% correlation between of government bodies with respect to cross border flows results and observed  Underlying trends in terms of country of origin and brand of cigarette are market trends in domestic Q1 2005Q2 2005Q3 2005Q4 2005Q1 2006Q2 2006Q3 2006Q4 2006Q1 2007Q2 2007Q3 2007Q4 2007 consistent with expectations and corroborated by external sources

sales and legal non- Source: “Yellow Bag” survey, an Empty Pack Survey undertaken by the German Cigarette Industry Association (VDC), 2005 to 2007. Full PMI results were domestic flows available along with total non-domestic market size Empty Pack Surveys conducted in Hungary have identified and quantified the rise and subsequent fall in non-domestic incidence since 2004  The increase in non-domestic incidence to 2005 corresponded with an increase in excise taxes of 93.5% between 2002 and 2004

Hungary historical Empty Pack Survey results − the impact of this tax change was a price increase of 63%  In 2006, increased domestic sales and stricter border controls corresponded to a significant decline in non-domestic incidence 30.0% 27.5% 18 16 − new enforcement measures implemented by Hungarian Customs included 25.0% tightened border controls, vehicle confiscation powers, additional sniffer dogs

14 LDS - billion sticks 18.6% and increasing the administrative burden of importing cigarettes from Romania 20.0% 17.5% 12 10 EPS results from Q3 2007 show a continued decline in non-domestic incidence 15.0% 11.1% 8  This further reduction is in line with the understanding of Hungarian Customs and 10.0% 6 is further corroborated by legal domestic sales and PMI shipment data for the year 4 5.0% Share of total of Share consumption 2 0.0% 0 Q3 2004 Q3 2005 Q3 2006 Q3 2007

% non-domestic incidence Legal domestic sales

Source: GfK Empty Pack Surveys, 2004 to 2007

We have built a dynamic, Dynamic outflow and LDC calculation – EU Flows Model iterative model based primarily on objective evidence from legal Estimated total LDS PMI CF domestic sales and empty consumption Uplift using Apply EPS non- pack survey results EPS results domestic share

27 country model

LDC Non-domestic flows

Attribute inflows Subtract outflow as outflows from source country Outflows

The EU Flows Model is a dynamic, iterative model that is principally based on legal domestic sales and empty pack survey results  Legal domestic sales are the starting point of the model from which outflows of legal sales to other countries are then subtracted to determine legal domestic consumption in a market  EPS results provide a measurement of the share of non-domestic packs by country of origin in all markets - EPS results provide a consistent source across all 27 markets of non-domestic packs by country of origin from which we can calculate total product outflow from each market to the other 26 markets  We have then iterated the model to refine estimates for legal domestic consumption

Example of historic GCTS consumption index modelling for Hungary Historic GCTS data can provide a helpful indicator of market trends to Historic GCTS 1997-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market consumption index  The rate of change in the GCTS smoking prevalence and average smoking analysis was added to the volume figures provide a useful indication of observed changes in 25 methodology to provide a smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and further means of Consumption 20 has declined the problem of under-reporting of consumption corroborating changes in since 2005 while LDS have  The KPMG consumption index is calculated using the incidence of 15 consumption and net increased smoking and the average daily consumption (both from GCTS) and the population of each market (from Euromonitor GMID) outflows 10 17.8 − these figures are then indexed to the earliest year for which both

Volume (billion sticks) 5 GCTS and LDS data are available, in most cases 1997  If LDS trends differ from those shown by the consumption index, it

0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1997-2003 2003-2005 2005-2007  Comparing consumption and sales trends helps to understand market  LDS (2.8)% (14.1)% 9.3% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.0)% (2.8)% (0.3)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated

Example of consumption regression modelling for the Netherlands KPMG has investigated how changes in disposable income and cigarette Consumption regression 1997-2007(1)(2)(3)(a) prices affect cigarette consumption in a market modelling was added to  Historical data up until 2006 has been used to estimate the strength of the methodology during relationships between consumption and prices and income 30 − estimated incidence of smoking and average daily consumption are the year two research to Control period Forecast based on GCTS data. Population and disposable income are estimated provide a further means of 25 using Euromonitor GMID and Oxford Economic Forecasting data. corroborating changes in Cigarette prices are based on the weighted average pack price supplied 20 by PMI consumption and net 15 KPMG has used a linear regression modelling approach to understand the outflows relationship between consumption and changes in disposable income and 10 cigarette prices

Volume (billion sticks)  5 Linear regression modelling estimates the coefficients of a linear equation, involving one or more independent variables, that best predict the value of 0 the dependent variable 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 − in our analysis, consumption per capita was used as the dependent variable, with disposable income per capita and price per cigarette the independent variables CAGR (%) 1997-2001 2001-2006 2006-2007  The estimated coefficients for the independent variables, prices and income,  LDS (1.2)% (3.2)% 2.6% measure the strength and direction of their relationship with consumption

 Estimated consumption(b) 2.2% (3.8)% 2.0% − changes in prices can be either positively or negatively related to changes in consumption. However, the external research consensus is that there is a negative relationship between prices and consumption(4) Notes: (a) Relationship based on historical data from 1997 to 2006 i.e. increases in prices lead to lower consumption. KPMG regression (b) The estimated consumption line has been rebased so that it equals the 2006 results are consistent with this consensus Project Star estimate of consumption for the Netherlands (19.0 billion sticks) Sources: (1) KPMG EU Flows Model − changes in income can also be either positively or negatively related to (2) In Market Sales supplied by PMI changes in consumption (3) KPMG Consumption Regression Model (4) “Cigarette demand: a meta-analysis of elasticities”, Craig A. Gallet and John  It is important to note that factors other than price and income are likely to A. List, 25th June 2003 impact consumption levels in a market − examples of other factors likely to impact consumption levels include changes in smoking restrictions and changes in demographic profiles  Given that we have not yet assessed the impact on consumption from factors other than price and income, estimated consumption can only be considered a broad approximation The relationships between changes in consumption and changes in disposable income and prices have been used to estimate 2007 consumption in the market  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption

Primary market research  Our approach was to measure the number and volume of tobacco purchase occasions from a complete, nationally representative sample of males and females, aged 19 years and over who have travelled abroad in the past 12 months was conducted to quantify Approach − these results were then weighted and projected by age and gender to a national level to estimate the volume of legal non- legal purchases of non- domestic cigarettes brought back into each market by travellers returning from overseas domestic cigarette by a − during 2007 research was updated in 14 markets, while Bulgaria and Romania were included for the first time Member State’s

inhabitants  The sample was drawn from the most complete, nationally representative database available and was representative of both urban and rural areas, age and gender During 2007, research was − a fully random sample approach was used to ensure results were as ‘certifiable’ as possible and could be projected to the updated in 14 key markets total target population while Bulgaria and Sample  A target of 7,000 gross contacts (i.e. agreed to be interviewed and aged 19 years and over) or 500 net contacts (i.e. travelled Romania were included abroad and purchased tobacco products in the past year) was set within the study for the − these targets were considered sufficient to derive accurate volume estimates once projected to the national population first time and set based on past experience from the research agencies and findings from the pilot process − the target was achieved in all markets except the UK

 Computer Aided Telephone Interviewing (CATI) was the data collection method in each market except for Malta and Latvia which used pen and paper data collection Data collection  The interview script was consistent across all markets, translated into local language and back translated into English for quality control purposes

 Numerous validation tools were built into the script to enhance the accuracy of responses, for example Validation tools − respondents were asked to recall all trips abroad in the past year and purchase volumes and brands for each trip − for each trip, the purpose of visit was also recorded to ensure final results appear logical and within a reasonable range

 To ensure that we were recording legal personal purchases only, results were capped at an individual respondent level − total annual purchases were limited to a maximum of 1,000 packs per person as this was considered the absolute upper level for a heavy smoker who makes all of their purchases abroad Results capping − purchases to non-EU destinations were limited to a maximum of 10 packs per trip in line with duty free purchase restrictions − purchases were also capped for intra-EU purchases where limits are enforced

Gross versus net respondents by country (roll-out countries only) The project has involved 8,000 600 an extensive primary

research programme 7,000 500 ND(L) for Project Star Gross respondents 6,000 Net respondents results in 2007 is based 400 upon over 11,300 full 5,000 interviews and almost

130,000 gross respondents 4,000 300 carried out during the Net respondents

second halves of 2006 and Gross respondents 3,000 200 2007 2,000 ND(L) data is also adjusted 100 to reflect inbound visitor 1,000 inflows 0 0

Italy UK* Spain Estonia Finland France Ireland Latvia* Poland Malta* Romania Germany Belgium Lithuania Sweden Cyprus* Austria*Greece* HungarySlovakiaBulgaria Slovenia* Portugal* Denmark* Netherlands Luxembourg* Czech Republic Note: * indicates research not updated in 2007 The key objective of the market research programme is to quantify genuine, legal non-domestic tobacco purchases in each market  The 2007 market research programme incorporated an extensive interview programme across the 16 markets, using recognised market research specialists AC Nielsen and Synovate − research for the remaining 11 Member States was conducted during 2006 by either AC Nielsen or Synovate  ND(L) data for countries where research was not carried out during 2007 was updated in line with overall non-domestic trends for each country − in some examples further adjustments were made on the basis of additional corroborating data  In the EU 27 countries, ND(L) results are based on a total of 129,953 contacted respondents and 11,322 successful interviews with adults (age 19+) who had travelled abroad and purchased tobacco products in the preceding twelve months. In addition to the research programme, ND(L) data is adjusted to reflect inbound visitor inflows  Non-domestic product found in Empty Pack Surveys from high cost inbound tourist /visitor countries is likely to represent an incidental inflow and is therefore categorised as legal − flows attributable to inbound tourism and visitors can not be identified in the market research programme − adjustments to reflect inbound tourism/visitors total 3.7 billion sticks in 2007

Two parallel methodologies for calculating non-domestic Methodology comparison A combination of two brand share were used to ensure that the most reliable and brand share calculation Methodology one Methodology two realistic results were achieved methodologies, combined  While results at an overall market level were all highly robust Description Brand share of total non-domestic Non-domestic share of brand and credible, smaller sample sizes at a brand level have the with IMS / ND(L) analysis x x potential to introduce distortions at this lower level and applied with market  In order to maximise the accuracy of brand results, a dual Total consumption gap Domestic sales by brand understanding and methodology was used to estimate non-domestic brand- level results in each markets judgment in case of Key  Brand share of non-domestic in  Non-domestic share of a brand assumption the EPS is representative of the in the EPS is representative of  The results from both approaches were then compared to substantial variances, national picture the national picture both IMS and ND(L) brand results for corroboration ensures the most robust − any overstatement of − any overstatement of  In almost all markets and for the vast majority of brands the brand-level results domestic share of premium premium brands’ domestic results for the two approaches were highly consistent brands in EPS is not and non-domestic share is  In a few markets there were some brand-level reflected in their non- proportional discrepancies. In these markets, adjustments were made domestic shares based on the weight of evidence from both approaches and the IMS and ND(L) findings to determine the most credible Strengths  Brand totals tally to overall total  More robust for brands which non-domestic brand share of non-domestic are overweight in the EPS Brand share methodology two was used in a limited samples at a non-domestic  Can track flows by brand and number of instances to ensure that the results were both brand share level country as reliable and realistic as possible  Can calculate non-domestic  Methodology one is the most universally applicable and was volumes where no legal sales therefore used where both approaches were consistent. are present Where an adjustment from methodology one was required, a combination of both approaches or methodology two was applied as appropriate  For a few countries, methodology two was used as it Limitations  Some potential to overstate  Totals by brand will not appears to give more reliable and robust results. premium brands due to necessarily match total overall Methodology two was used in the following countries: concentration on cities in EPS − most effective as an − Denmark  Small IMS share / large EPS estimate of share of non- − France share discrepancies domestic for major brands − Ireland  Small IMS share / high EPS non- − Poland domestic level discrepancies − Spain − UK

Analysis of external data External public research Expert interviews sources has provided significant cross-validation  We have undertaken extensive research into external data sources in  In addition, we have undertaken structured interviews with industry each of the 27 EU markets specialists to canvas their opinions on C&C in each of the 27 EU of our research results  Research covered a wide variety of data sources, including: markets where possible  Our contacts were identified from multiple sources, including: Although the interview − third party information available within PMI − press articles − PMI recommendations programme delivered − retail trade and tobacco industry associations − OLAF recommendations good circumstantial − universities and other academic institutions − KPMG external search supporting evidence for − ministries of health and social affairs − other interviewee recommendations our findings, interviews − customs departments  We have interviewed specialists across a broad spectrum of areas and backgrounds, including: were not effective in − other government and policy-making institutions − governmental and policy-making organisations delivering consistent and − market research publications − academic research institutes accurate estimates for the − industry related journals and publications − trade and industry associations quantum of contraband − federal statistics − PMI management, both centrally and at a country level  We have reviewed, collated and used the information available to cross- and counterfeit check and test our research results  We devised a structured interview process for each interview category which underwent multiple iterations to ensure consistency and accuracy − we tested the reasonableness of our research results against a range of quantitative estimates obtained on the size and scale of of both questioning and capturing results C&C in each market

Conclusions

 Analysis of external research has been highly effective in: − improving our understanding of local market dynamics, trends and the nature of C&C in each country − facilitating our judgement on the potential limitations of our findings  However, external data is not sufficiently detailed on its own to obtain a credible estimate of the size and scale of C&C as: − basis for estimates is often unknown and may not be objective − data sources and estimates across countries lack consistency − data is often sparse and patchy  External expert interview programme has provided good soft corroboration of trends and issues − however, it has been less effective in delivering quantitative results

 Project overview and timing

 Methodology

− overview

− limitations of results

 EU cigarette market

 2007 Project Star results

 Appendices

We have designed a Scope limitations Source limitations methodology that is as robust and inclusive as we believe could practicably  There are specific scope exclusions which cannot  Limitations are, of necessity, present with any have been delivered be or have not been accounted for in our approach: primary information sources − geographic exclusions  This primarily affects EPS, LDS and ND(L) sources However, given the innate complexity of C&C, our − brand exclusions – non-PMI counterfeit  For example, limitations can arise from methodology does have − category exclusions – OTP − sampling criteria limitations − LDS product flows out of the EU − coverage issues − timing/seasonality factors − specific regional or demographic exclusions

Our methodology has Limitation Detail Impact Adjustment

certain specific scope Geographic  We have limited our geographic  Spanish results only cover mainland Spain and do not include the Canary Not adjusted limitations coverage coverage in some markets where Islands, Balearic Islands or Ceuta & Melilla for extension would significantly impair  French results cover only mainland France and do not include Corsica confidence levels in the ND(L)  Portuguese results only cover mainland Portugal and do not include Madeira research for the further territories or the Azores included  Greek results only cover mainland Greece and do not include the Greek  In some instances (e.g. Greek islands islands), shipment data is also insufficient for the purposes of this  UK results only cover Great Britain and Northern Ireland and do not include study the Channel Islands

Non-PMI  Empty pack survey results do not  In some instances, the volume of legal domestic consumption may be Not adjusted counterfeit identify non-PMI brand counterfeit overstated where domestic counterfeit variants are identified for packs − this may lead to minimal understatements of C&C volumes for non-PMI − only the manufacturer / brands trademark owner can confirm  Moreover, we cannot distinguish between non-PMI brand counterfeit (non- whether their brand pack is domestic variants) and contraband product, although this will not impact the genuine overall volume of C&C OTP  Empty pack surveys collect cigarette  However, based on extensive interviews, seizure data and analysis of other Not adjusted packs only available information, the scale of non-domestic OTP consumption is for − non-domestic consumption for believed to be limited when compared to manufactured cigarettes OTP cannot be measured via − the only significant non-domestic OTP inflow is into the UK, which is a empty pack survey results non-signatory to the Anti-counterfeit and Anti-contraband Agreement  Although precise C&C volumes for OTP are not available we have used domestic legal sales of OTP, seizure data and in market interviews to identify trends consistent with growth in C&C inflows − these trends are identified in the relevant country section where appropriate

Non-EU  In order to calculate consumption,  Net outflows besides Sweden are believed to be minimal, supported by Partially outflows we have assumed no outflows of anecdotal evidence from non-EU EPS surveys (including Switzerland and adjusted for LDS outside the EU, with the Turkey) exception of Sweden (see country-  Non-EU LDS outflows are not considered to be material due to the high specific refinements) prices relative to other parts of the world and duty free import restrictions  Potential minimal overstatement of EU consumption

Although there are Source Limitations

limitations to any source, Empty pack  In some geographies, the results may not be absolutely representative of total consumption because of the sample size, or, more we are comfortable that surveys likely, practical limitations to collection locations we have used the most − depending on the source of packs collected, either homes and workplaces or public spaces (in Germany) are not covered appropriate sources − the sample is more heavily weighted towards populous, urban areas and therefore may not be fully representative of available consumption habits in rural regions  Results from Germany are based on a monthly analysis of 10,000 packs collected at recycling centres and so are not directly comparable with the EPS results from other countries due to the difference in methodology  Empty pack surveys are only conducted at set periods and results may be influenced by seasonal factors such as tourist inflow − in some instances the timing of an EPS has changed between 2006 and 2007. In order to ensure comparability of results, monthly LDS figures, consumption trends and visitor data are all analysed and adjustments made where appropriate  Brand and market variant share can only be extrapolated with a degree of statistical accuracy for brands where a sufficiently large number of packs have been collected − EPS results are analysed to identify any outliers that may impact results, such as geographic concentrations of a specific brand or market variant. Brand specific data is also compared to known sales in the source market to identify whether results are credible − where data suggests a sampling or data capture error may have occurred at a specific location, results are adjusted and the remainder of the survey is re-weighted accordingly  In some specific instances, it is not possible to differentiate between duty-free and duty-paid variants from the empty packs collected as the tear tape on the packet is required in order to make the necessary distinction  However, EPS represents the most consistent source of non-domestic share across markets. We believe, especially at a total market level, that these results are credible and robust. Brand trends and analysis of country flows from EPS results further supports this conclusion  When allied to other methods of corroboration, such as consumption index modelling, we believe the results are fully fit for purpose Legal  Shipment data is the most reliable source for legal domestic sales in a market. However, in some markets it is not available. In the domestic absence of shipment data, we have used either AC Nielsen Retail Audit data or tax stamp data as available sales − in some cases tax stamp data may not correspond to the calendar year and may also be distorted by trade loading in advance of increases in taxation. In these instances we have used the LDS source considered by local PMI management to be the most representative of smoker consumption during the calendar year  AC Nielsen Retail Audit data is derived from retail sales information but may exclude particular sales channels or retailers − in markets where we have used Retail Audit data, PMI local management have calculated the appropriate uplift to derive total market sales, including volumes not accounted for in Retail Audit data  Slight timing variances may arise between the date the product was shipped and actual consumption but, following discussions with local management, this is not considered significant and the full year LDS information we have is considered to be a fair and accurate representation of full year 2007 sales in each market

Although there are Source Limitation

limitations to any source, ND(L)  As with any CATI-based market research approach, our samples may potentially exclude certain demographic segments, in particular, we are comfortable that those without a permanent home, registered address or telephone line we have used the most  The nature of the market research programme requires that people can recall, with a high degree of accuracy, trip and purchase volumes undertaken over the past year. However, pilot and roll-out results give us confidence that this is not a significant issue for appropriate sources respondents available  Respondents are asked to recall purchase volumes in packs and we assume 20 sticks per pack for our pack to stick conversion  To ensure that we record legal imports only, we have capped total individual purchases and applied a cap to imports from certain source countries where import restrictions apply  Due to the nature of the survey, market research does not capture non-domestic (legal) product arising from inbound tourism. However, these flows are likely to be limited in nature and, in many cases, can be adjusted within the ND(L) methodology through the use of corroborating sources  Duty-free and duty-paid variants are impossible to distinguish between in the ND(L) research due to the consumer confusion when buying cigarettes abroad, particularly in airports when travelling intra-EU. We have however attempted to estimate legal Duty Free purchases by using ND(L) inflows from non-EU markets as an approximation. This approach assumes that EU nationals purchase Duty Free variants when they travel to non-EU markets and buy cigarettes as measured by the ND(L) research − this assumption is predicated on the fact that Duty Free variants are typically available at a lower prices than legal tax-paid cigarettes in non-EU destination countries  There are also some specific country limitations: − Austria - no gender information was captured for respondents who agreed to be interviewed but were screened out because they either did not travel of did not purchase tobacco products abroad. In order to weight the data appropriately, an additional omnibus was conducted for 1,000 respondents to capture gender information; The sample for the omnibus survey was drawn from the same, fully random sample as the core programme − UK - the sample size in the UK was limited due to reluctance to participate in interview programme. Neither the gross respondent nor the net respondent targets were therefore reached in the UK. Given the low PMI brand share in the UK and non-signatory status, it was decided not to extend the overall project fieldwork phase for just this market

 Project overview and timing

 Methodology

 EU cigarette market

− 2007 market overview

− smoking trends

− OTP sector analysis

 2007 Project Star results

 Appendices

Legal sales of cigarettes in The eight largest EU markets account for 75% of total cigarette sales the EU totalled 677 billion  Of these, Italy, Germany and Spain each accounted for more than 85 billion sticks of legal domestic sales in 2007 sticks in 2007

Increasing prices and The pricing and legislative environment for cigarettes has tightened across the EU over the last several years as a result of domestic policy and smoking restrictions have pressure on new Member States to achieve a minimum level of cigarette taxation contributed to a decline in  Over the last 10 years, prices have increased in all 27 EU markets cigarette consumption  An increasing amount of legislation has been introduced, limiting smoking in public places and preventing manufacturers from advertising their products in a growing number of media formats As a result of these changes and other social trends, the majority of countries have experienced a decline in cigarette consumption  Over the last four years, smoking prevalence has declined in all EU member states and only six countries have seen an increase in average daily consumption

Cigarette sales have Total EU sales entered a period of steady decline in 2002 declined at a European  EU cigarette sales were relatively flat until 2002 but have subsequently declined at a rate of 3.3% per annum level, partly driven by declining consumption − the majority of this decline can be attributed to Germany and France, which legal sales have decreased by 9.0% and 7.4% per annum respectively between 2002 to 2007

However, differences in The relationship between changes in consumption and changes in sales differs substantially between Member States sales and consumption  Although the majority of markets have experienced a decline in consumption, the impact on legal sales differs markedly between Member States trends between countries highlight the existence of − markets such as Germany, France and the UK have seen substantial price increases and experienced more rapid sales declines than can be price-driven cross-border explained by decreases in consumption alone trade These differences point towards a significant and fluctuating cross-border trade in cigarettes  It is likely that markets where sales have declined more rapidly than consumption have experienced growing penetration of non-domestic cigarettes − in contrast, markets such as the Czech Republic and the Baltic States, which remain relatively low priced markets despite price increases, have experienced steady or growing sales despite declining consumption

EU legal domestic cigarette sales, 2007(1) Legal domestic cigarette sales, 2007(1) The eight largest markets account for three-quarters of total EU legal cigarette 700 Total, 677.4 Finland sales 5.0bn

600 Other, 170.0

Sweden Estonia 6.3bn 2.5bn 500 Romania, 31.1 Greece, 32.7 Latvia UK, 47.0 Denmark 4.7bn 400 Luxembourg 7.9bn France, 54.9 5.0bn Lithuania Netherlands 5.3bn 300 Poland, 69.9 14.5bn Belgium

Volume (billion sticks) Ireland UK 12.5bn Germany, 89.3 5.4bn 47.0bn Poland 200 Germany 69.9bn 89.3bn Spain, 89.5 Czech 100 Republic 24.0bn Slovakia Italy, 92.8 7.7bn Austria 0 France 13.7bn Hungary Legal domestic sales 54.9bn 16.5bn Slovenia Romania Italy 4.8bn 31.1bn 92.8bn

Bulgaria 18.1bn Spain Portugal 89.5bn 14.0bn Greece 32.7bn

Cyprus Malta 1.7bn 0.5bn Key:  More than 30bn sticks  10bn – 30bn sticks Source: (1) In Market Sales provided by PMI management  Less than 10bn sticks

Map denotes Marlboro price per 20 cigarettes Despite ongoing moves to 1 July 2007(1)(a) align cigarette prices and

taxes, retail prices vary Finland €4.30 markedly across the EU, Norway with a pack of Marlboro €8.16

costing from €1.33 in 1 Sweden 7 Estonia Lithuania to €8.04 in the €4.94 % €1.79 UK Latvia Russia Denmark €1.47 €0.86 Furthermore, prices are Luxembourg €4.24 €3.70 Lithuania even lower in many of the €1.33 markets which border the Netherlands €0.86 €4.11 Ireland EU, such as Ukraine and UK Belgium €7.05 €4.53 Russia €8.04 Poland €2.10 Germany €4.71 Czech Ukraine Republic €0.63 €2.46 Slovakia €2.35 Austria Moldova Hungary France Switzerland €3.90 €0.67 €2.39 €5.00 €3.82 Romania Italy Slovenia €1.66 €4.10 €2.50 BiH Serbia €1.58 €1.42 Croatia Bulgaria €2.73 €1.64

Spain Montenegro €1.50 Portugal €2.95 FYROM Turkey €3.15 Greece €1.95 €2.60 Albania €3.00 €1.62

Malta Cyprus  Key: Greater than or equal to €5 €3.50 €3.86  €4 to €4.99  €3 to €3.99  €2 to €2.99 Note: (a) Based on pack of Marlboro 20 King Size on 1st July 2007  €1 to €1.99 Source: (1) PMI management based on PMI standard exchange rates  Less than €1

 Project overview and timing

 Methodology

 EU cigarette market

− 2007 market overview

− smoking trends

− OTP sector analysis

 2007 Project Star results

 Appendices

Pack price breakdown of the MPPC Taxes form a large 2007(1)(a) proportion of the price of cigarettes, though the 100% 6% 6% 7% 4% 7% 4% nature of taxation varies 9% 15% 15% 15% 13% 10% 28% 29% 28% 26% 36% 35% 34% 35% 34% 36% 31% across countries 80% 40% 43% 40% 40% 52% 55% 50% 54% 39% 48% 58% 57% 51% 43% 43% 45% The proportion of tax paid 60% 27% 26% 15% Specific tax 34% 34% 28% 28% 25% 24% 14% 20% Ad valorem tax as share of the price of a 32% 23% 18% 22% 15% 40% 16% VAT 20% 16% 20% 16% 13% packet of cigarettes 17% 15% 16% 17% 18% 17% 16% 17% 16% 13% 16% 17% 18% 17% 17% 17% 16% 14% 15% Manufacturer & trade take 15% increased in 23 of the 27 20% 38% 27% 27% 27% 28% 28% 29% 30% 33% 20% 20% 20% 21% 21% 22% 22% 23% 23% 23% 24% 24% 25% 25% 25% 25% 26% member states in 2007 17% 0% UK

2004 and 2007 Accession Italy Spain Malta Latvia France Poland Ireland Austria Cyprus Finland Greece Estonia Bulgaria Portugal Sweden Belgium Hungary Slovenia Romania Lithuania States typically Germany Denmark Netherlands experienced the highest Luxembourg Czech Republic Slovak Republic growth in tax burden Change in tax burden of MPPC 2006-2007(1)(a)(b)

180% 159% 160% 140% 120% 100% 80% 60% 37% 40% 25% 24% 17% 13% 12% 10% 10% 9% 7% 7% 20% 6% 6% 5% 5% 5% 4% 4% 3% 3% 2% 1% 0% (0)% 0% (20)% (1)% (2)% UK Italy Spain Malta Latvia Poland Ireland France Austria Cyprus Finland Greece Estonia Bulgaria Portugal Sweden Belgium Hungary Slovenia Romania Lithuania Denmark Germany Netherlands Luxembourg Czech Republic Slovak Republic

Note: (a) MPPC is the Most Popular Price Category ` (b) MPPC prices in local currencies at 1st January, 2007 and 2008, and therefore may include some distortion from currency fluctuations Source: (1) Information supplied by UK Tobacco Manufacturers Association

Average annual price change in local currency Prices have increased 1997-2007(1)(a)(b) substantially across the EU, largely driven by 32% 29% increases in cigarette 28% taxes 24% 21% 20% 16% 14% 11% 10% 12% 9% 9% 7% 6% 6% 6% 8% 6% 5% 5% 5% 5% 5% 5% 5% 5% 4% 3% 2% 4% 2% 1% 1%

Annual average price change 0% (4)% (1)% UK Italy Spain Malta Latvia Poland Ireland France Cyprus Austria Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Hungary Slovenia Romania Lithuania Germany Denmark Netherlands Luxembourg Czech Republic

Note: (a) Price CAGR calculated on longest period of available data, 1997 – 2007 except for: Belgium, Ireland and Luxembourg (from 1998); Hungary and Malta (2000); Greece and Finland (2002); Cyprus (2003); Denmark, Estonia, Latvia, Lithuania, Romania and Sweden (2004); Bulgaria (2005) (b) The decrease in weighted average price for Cyprus reflects consumers trading-down to cheaper priced cigarettes rather than decreasing prices Source: (1) Weighted average price data supplied by PMI management

Smoking legislation time-line(1)(a) There has been significant legislation in recent years CURRENT FUTURE limiting smoking in public Belgium Denmark places Sweden This change is ongoing in Austria Slovakia Romania Finland Slovenia Netherlands many Member States, 1995 with further restrictions Cyprus Netherlands Malta Luxembourg Estonia Portugal Latvia already announced in many markets 2009+ 1995 2002 2003 2004 2005 2006 2007 2008

Hungary Greece UK France 1999 2002 Ireland Bulgaria Latvia

Poland Italy Czech Germany Slovakia Republic Spain Lithuania

Key: Minimal smoking restrictions – no smoking restrictions Partial smoking restrictions - no smoking in government buildings, public transport, schools Moderate smoking restrictions – partial smoking restrictions plus no workplace smoking Extensive smoking restrictions – moderate smoking restrictions plus limited or no smoking in restaurants, cafes and bars Denotes future smoking restrictions Note: (a) Restrictions shown are most stringent or most recent Sources: (1) Press articles; “European Trends towards non-smoking provisions”, European Network for Smoking Prevention, December 2007’; and European Public Health Alliance website

Change in cigarette smoking prevalence Cigarette smoking prevalence Smoking prevalence has 2007(1)(a) (percentage points) declined across all EU 2003-2007(1)(a)(b)(c)(d)(e)

Member States since 2003 Finland Belgium (5.4) 18% Sweden (4.9) Germany (4.4) Italy (4.4) 1 Netherlands (4.3) Sweden 7 Estonia 14% % 36% Czech Republic (4.1) Poland (4.0) Latvia Portugal (3.6) Denmark 37% 18% Romania (3.5) Lithuania Luxembourg n/a Lithuania (3.1) 34% France (2.9) Netherlands 12% Estonia (2.8) Ireland Belgium UK (2.5) UK 28% 18% 16% Poland Slovenia (2.1) Germany 39% Spain (1.9) 19% Hungary (1.2) Czech Cyprus (0.9) Republic 33% Austria (0.8) Slovakia 33% Ireland (0.7) Austria Hungary Latvia (0.5) France 29% 26% 35% Slovakia (0.1) Romania 44% (6.0) (5.0) (4.0) (3.0) (2.0) (1.0) 0.0 Italy Slovenia 21% 32% PP change in smoking prevalence Bulgaria (2003-2007) Greece 51% 30% Spain Notes: (a) GCTS data based on smokers 19 to 64 years of age smoking 3 or Portugal 32% more cigarettes a day 30% (b) Data for Luxembourg or Malta or historic data for Bulgaria is not available (c) There is no 2003 data available for Finland, Portugal, Spain and Slovenia. Therefore, data for 2004 (Finland) and 2005 (Portugal and Spain) and an average of 2002 and 2004 (Slovenia) has been Cyprus used Malta 34% (d) There is no 2007 data available for Ireland and Cyprus, so 2005 and n/a 2006 data has been used respectively (e) Denmark, Finland, Greece excluded due to change Key:  30% or higher in methodology  20% - 29% Source: (1) GCTS, PMI management  Less than 20%

Stick change in average daily cigarette Average daily cigarette consumption Average daily consumption 2007(1)(a)

consumption of cigarettes 2003-2007(1)(a)(b)(c)(d)(e) has also been declining in Finland the majority of Member France (2.0) 14.8 Germany (1.4) States UK (1.2) Austria (1.1) 1 Slovakia (1.1) Sweden 7 Estonia Cyprus (1.0) 12.8 % 14.9 Ireland (1.0) Denmark (0.9) Latvia Denmark Belgium 14.8 (0.8) 15.5 Italy (0.7) Luxembourg Lithuania Portugal (0.7) n/a 14.9 Netherlands Greece (0.6) 13.5 Hungary (0.4) Ireland UK Belgium Czech Republic (0.2) 16.9 14.6 15.4 Poland Spain (0.2) Germany 16.7 Netherlands (0.0) 15.3 Latvia 0.1 Czech Republic Poland 0.2 14.4 Slovakia Slovenia 0.4 13.4 Estonia 0.6 Austria France Hungary Romania 0.8 17.5 Romania 13.8 17.8 Lithuania 1.0 16.7 Italy (2.5) (2.0) (1.5) (1.0) (0.5) 0.0 0.5 1.0 1.5 13.6 Slovenia Stick change in average daily cigarette consumption 18.0 (2003-2007) Bulgaria Greece 19.1 23.2 Spain Portugal 16.2 Notes: (a) GCTS data based on smokers 19 to 64 years of age smoking 3 or more cigarettes a day 17.7 (b) Data not available for Luxembourg and Malta (c) There is no 2003 data available for Finland, Denmark, Portugal, Spain and Slovenia. Therefore, 2004 (Finland) 2005 (Denmark, Portugal, Spain) and an average of 2002 and 2004 Cyprus data (Slovenia) has been taken Malta 25.5 (d) There is no 2007 data for Ireland and Cyprus, therefore 2005 and n/a 2006 data has been used (e) Sweden and Finland have been excluded due to change in Key:  More than 20 sticks methodology  15 – 20 sticks Source: (1) GCTS, PMI management  10 – 15 sticks

Historic legal domestic cigarette sales for EU 27 Legal sales in the EU have 1997-2007(1)(a)(b) declined from 800 billion sticks in 1997 to 677 billion 1,000 sticks in 2007 800 804 810 800 789 793 771 800 729 The rapid fall in sales in 698 687 677 221 215 215 195 197 197 191 189 190 France and Germany has 600 185 187 CAGR (%) 1997-1999 1999-2002 2002-2007 been an important  2004 and 2007 accession countries (1.3)% (2.8)% (1.1)% 378 400 359 366 370 370 371 376 371  Other EU15 legal sales 1.5% 0.7% (1.7)% contributing factor to the 357 353 346  Germany legal sales 1.5% 0.5% (9.0)%

LDS (billion sticks) (billion LDS 200 declining EU legal sales 137 138 141 141 141 143 134  France legal sales 0.4% (1.3)% (7.4)% 114 97 93 89 83 84 84 83 83 81 70 Total 0.6% (0.4)% (3.3)% trend 0 55 55 56 55 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Percentage change in legal domestic cigarette sales for EU 27 2003 – 2007(1)(c )

100% 75% 80%

60%

40% 36% 24% 18%  2004 and 2007 Accession 20% 17% 10% 12% States 3% 7% 0%  Other EU (4)% (4)% (3)% (1)% (10)% (9)% (8)% (20)% (17)%(15)%(15)%(14)%(13)%(12)%(12)%(12)% (21)%(17)%

Percentage change in LDS (2003-2007) (40)% (34)% UK Italy Spain Malta Latvia Poland France Ireland Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Slovenia Hungary Romania Lithuania Denmark Germany Netherlands Luxembourg Czech Republic

Notes: (a) No data available for Latvia, Lithuania and Malta prior to 2000, therefore 2000 data has been used for 1997-1999. No data available for Romania prior to 2004 and so 2004 data has been used for 1997-2004. Similarly, no data available for Bulgaria prior to 2005, so 2005 data has been used for 1999-2005 (b) Other EU legal sales is total legal domestic sales in EU 15 less Germany and France (c) There is no data available for Romania and Bulgaria for 2003, so 2004 and 2005 data used respectively Source: (1) In Market Sales provided by PMI

 Project overview and timing

 Methodology

 EU cigarette market

− 2007 market overview

− smoking trends

− OTP sector analysis

 2007 Project Star results

 Appendices

EU legal domestic sales of OTP smoking tobacco EU OTP sales have remained relatively flat since 2004 EU OTP sales were 113 (1)(a)(b) 2003 – 2007  Fine cut tobacco accounts for the majority of EU OTP sales with sales of 90 billion billion stick equivalents in 114 112 113 120 109 stick equivalents in 2007 2007 and together 97 10 11 12 100 9 8 − fine cut tobacco sales include RYO (roll your own), MYO (make your own) and 9 5 9 11 represented 14.3% of 6 portions 80 overall EU legal domestic − it should be noted that OTP product definitions can be problematic due to 60 ambiguity of intended use. For example, it is estimated that 75% of pipe tobacco sales 94 97 93 90 40 83 tobacco is used for MYO/RYO sticks. 20 Despite the decline in EU legal domestic sales of manufactured cigarettes, OTP billion sticks (equivalents) 0 products still account for a small percentage of total tobacco consumption 2003 2004 2005 2006 2007  OTP trends across Europe vary significantly with large differences in product LDS categories, competitive environment and consumer behaviour in Member States cigarettes 771 729 698 687 677

CAGR (%) 2003-2007  Cigars/cigarillos 2.1%  Pipe tobacco 18.1%  Fine cut tobacco 8.4%

Share of EU legal domestic tobacco sales by category 2007(1)(a)(b) Pipe tobacco Cigars/cigarillos 1.4% 1.5% Fine Cut tobacco 11.4%

Cigarettes 85.7%

Notes: (a) Fine cut and smoking tobacco calculated based on 0.75 grams per stick (b) Analysis excludes smokeless and water pipe tobacco Source: (1) In Market Sales provided by PMI

EU leading markets for smoking tobacco (fine-cut and pipe) The five largest markets for smoking tobacco in the EU account for over 72% of EU OTP sales are by volume, 2007(1)(a) overall sales concentrated in a small 120  Germany and the Benelux countries have the highest sales of smoking tobacco in number of Member States 100.8 Europe Other EU 100 9.4 Hungary  France, Poland and the UK have significant sales of smoking tobacco, although 2.9 4.5 Spain incidence levels are significantly lower than in Germany and the Benelux countries 80 4.8 6.2 6.3 Luxembourg − it is understood that sales of smoking tobacco in Poland are inflated by high 60 10.0 UK sales along the German border region 10.3 Poland − small price differentials between Germany, the Benelux countries and France 40 14.4 Belgium are believed to limit cross border flows of smoking tobacco between those France

Stick equivalents (billion) countries 20 Netherlands 32.0 The five largest markets for cigars and cigarillos in the EU account for over 88% Germany of overall sales 0  Germany is the largest market and represents over 50% of total EU sales 2007 − the major driver for growth in Germany has been the filter cigarillos market EU leading markets for cigars and cigarillos by volume which represent approximately 75% of all cigar and cigarillos sticks sold in the 2007(1)(a ) country  France is next most significant market, although sales are less than a third of 14 those in Germany 11.9 12 0.6  Only six other markets had sales of over 100 million sticks in 2007 0.1 0.3 0.3 Other EU 10 0.5 0.8 Greece 1.1 Netherlands 8 1.7 Italy 6 Belgium

Sticks (billion) Sticks UK 4 6.4 Spain 2 France Germany 0 2007

Notes: (a) Fine cut and smoking tobacco calculated based on 0.75 grams per stick Source: (1) In Market Sales provided by PMI

EU tobacco sales by category The Benelux countries (1)(a)(b) 2007 Manufactured cigarettes Smoking tobacco Cigars/cigarillos have the largest share of OTP sales relative to the 100% 1% 1% 2% 0% 1% 1% 1% 0% 1% 1% 0% 1% 1% 0% 1% 1% 1% 1% 1% 0% 0% 0% 0% 2% 5% 3% 4% 3% 1% overall tobacco market in 8% 7% 6% 6% 5% 5% 5% 15% 13% 12% 11% Europe 90% 17% 15% 25% 80% 49% 48% 43% 70%

60%

50% 99% 99% 99% 100% 100% 100% 96% 96% 98% 99% 92% 92% 93% 93% 94% 95% 95% 85% 86% 87% 88% 40% 82% 82% 70% 30% 54% 50% 51% 20%

10%

UK Italy Malta Spain Latvia Finland France Poland Cyprus Austria Greece Ireland Estonia BelgiumGermany HungaryDenmark Sweden Portugal Slovenia LithuaniaRomaniaBulgaria NetherlandsLuxembourg Czech Republic Slovak Republic

Notes: (a) Fine cut and smoking tobacco calculated based on 0.75 grams per stick (b) Excludes smokeless and water pipe tobacco Source: (1) In Market Sales provided by PMI

Smoking tobacco accounts for approximately half of all tobacco sold in the Netherlands, Belgium and Luxembourg  Germany is the largest market for cigars and cigarillos which together account for 5% of tobacco sales in that country The degree of substitution between product types can be influenced by a number of factors such as relative price, historical/cultural factors, product availability and consumer smoking behavior  For the purposes of this comparison smoking tobacco sales volumes have been calculated at one stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick for stick basis − smokeless and water pipe tobacco has been excluded from the above analysis

Estimate of domestic versus non-domestic consumption Interviews support the view that OTP C&C is not a significant feature of Non-domestic for the Interval brand, 2007(1) other markets consumption of OTP is not  Feedback from Customs and PMI management suggests that they do not recognized as a priority Spain origin 13% consider smuggling of RYO to be a major issue issue in the majority of Belgium origin PMI Management has identified non-domestic flows of the Interval Member States product 4% brand into France Of the major OTP markets,  The Interval brand is a France specific product that also has strong sales in estimates of non-domestic the border areas of Belgium and Spain Domestic share of consumption are 83% − for the purposes of this analysis, it is assumed that all sales in Belgium only available for France and Spain are destined for the French market and the UK UK Customs’ estimate of 67% non-domestic incidence for rolling tobacco implies a net inflow of 8,000 tons  UK Customs believes that the three major sources of non-domestic product are France, Belgium and Duty Free UK Customs estimate of domestic consumption of rolling tobacco as a share of total, 2007(2) − domestic sales of smoking tobacco in France and Belgium are expected to be 7,700 tons and 7,500 tons respectively in 2007 − data for duty free sales of rolling tobacco are not available Domestic 33%  UK Customs estimate is based on a 56% share of consumption for illicit product and an 11% estimate for non-domestic (legal) flows

Non-domestic 67%

Source: (1) Interview with PMI management (2) Measuring indirect tax losses, HMRC, October 2007

Market size of smoking tobacco in selected EU and non-EU The six largest markets by OTP volume near the EU are Norway, Switzerland, Flows of non-domestic Markets, 2007(1)(a)(b) Russia, Turkey, Iceland and Egypt OTP products from outside  The only markets of comparable size to the EU Member States are Norway and the EU are likely to be EU Non-EU Switzerland 12,000 10,766 limited by the very small 10,000  Legal domestic sales in Russia and Turkey remain low, at 119 tonnes and 145 size of alternative source 7,735 7,478 tonnes respectively 8,000 markets such as Russia − a number of countries which act as source markets for manufactured 6,000 4,737 4,675 cigarettes do not have significant smoking tobacco sales; these countries 4,000 include Ukraine, Belarus, Morocco, Tunisia and Algeria Market size (tons) 2,000 997 542 − in Asia, Thailand and Pakistan have annual smoking tobacco sales of 119 145 8 4 0 approximately 20,000 tonnes each, although products sold in these countries tend to be highly tailored to local markets UK Egypt France Poland Russia Turkey Iceland Belgium Norway Netherlands Switzerland  The low share for PMI reflects the dominance of local brands in individual markets − there are a limited number of international brand. Only Golden Virginia, Drum Philip Morris share of smoking tobacco market by region and Old Holborn have a significant presence outside their ‘home’ markets 2007(1)(a )(b) − the introduction of cigarette brands into the RYO/MYO sectors has not been successful

10%  PMI OTP market share in the EU is expected to increase in 2008 due to the 8.1% recent acquisition of a number of brands from Imperial Tobacco, including the 8% leading French brands Interval, Bergerac, Santoya and Wervicq

6% To date, there have been no recorded seizures of counterfeit or contraband PMI OTP product in Europe 4% 3.0% % share sales % share 1.6% 2% 0.0% 0.0% 0% EU region Australia and Latin EEMA region Asia New Zealand America and Canada

Notes: (a) Smoking tobacco includes both fine cut and pipe tobacco (b) EEMA Region refers Eastern Europe Source: (1) In Market Sales provided by PMI

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

− consumption overview

− overall non-domestic inflows

 non-domestic (legal) inflows

 C&C inflows and country comparison

− seizures

 Appendices

Consumption of manufactured Legal domestic consumption accounted for 87.4% of consumption in 2007 cigarettes in the EU reached  The accession of Romania and Bulgaria results in a 51 billion stick addition to the EU consumption total versus 2006 720 billion sticks in 2007  Like-for-like comparison between the 2006 Member States shows that consumption of manufactured cigarettes declined by 1.9% in 2007

The volume of non-domestic Non-domestic cigarettes accounted for 12.6% of EU consumption versus 12.9% in 2006 inflows reached 91 billion in  Like-for-like comparison between the 2006 Member States shows that non-domestic volumes declined by 2.4% in 2007 2007 for the 27 Member States  Germany and France, which represent the two largest markets for non-domestic inflows, experienced a decline in volumes in 2007 − non-domestic volumes in Germany declined by 3.4 billion sticks to 19.3 billion sticks in 2007 − non-domestic volumes in France declined by 0.8 billion sticks to 15.3 billion sticks in 2007  The UK was the largest market to experience an increase in non-domestic inflows − non-domestic inflows to the UK increased by 2.3 billion sticks to 13.5 billion in 2007

Non-domestic (legal) volumes Total non-domestic (legal) accounted for 4.2% of total consumption in 2007 accounted for 30.3 billion  This represents a slight decline from 2006 when ND(L) represented 4.5% of total consumption sticks in 2007 Like-for-like comparison between the 2006 Member States shows that non-domestic (legal) volumes declined by 4.5% in 2007  Declining volumes in France and Germany have been partially offset by an increase in non-domestic (legal) volumes in the UK and Netherlands − France and Germany accounted for 14.2 billion sticks of EU non-domestic (legal) volumes in 2007, down from 18.2 billion sticks the previous year − the UK and Netherlands accounted for 6.5 billion sticks of EU non-domestic (legal) volumes in 2007, up from 4.2 billion sticks the previous year

Total EU C&C volumes Total C&C volumes accounted for 8.4% of total consumption in 2007 amounted to 60.6 billion sticks  Like-for-like comparison between the 2006 Member States shows that the consumption of C&C volumes declined by 0.6 billion sticks in 2007 to in 2007 56.2 billion sticks  Approximately half of C&C volumes can be accounted for by intra-EU flows  C&C volumes for Bulgaria and Romania were 4.4 billion sticks in 2007

Total EU cigarette EU consumption by type 2006-2007(1) consumption patterns 100% have not changed 8.3% 8.4% significantly in 2007 with 4.5% 4.2% 80% ND(L) and C&C volumes

stable as a proportion of 60% C&C total consumption ND(L) 87.1% 87.4% 40% LDC Consumption in the 2006 Member States fell by 20% 1.9% in 2007 to 669 billion 0% sticks EU 2006 EU 2007

EU consumption by type 2006-2007(1)

800 720.1 682.0 669.3 700 (12.7) 50.8 60.6 56.8 56.2 30.3 600 31.0 29.5 500 % change 2006-2007 6.7% 400  C&C  Non-domestic (legal) (2.1)% 300 594.2 583.6 629.2 5.9% 200  Legal domestic consumption Volume (billion sticks) Total consumption 5.6% 100 0 2006 Decline in 2007 Accession 2007 (25 2006 Member consumption (25 2006 Member of Romania (all Member States) States) States only) and Bulgaria

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

Legal domestic sales by country versus total EU non-domestic cigarette volumes Total non-domestic flows 2007(1) to the EU of 91 billion 100 92.8 sticks in 2007 are 90.9 89.5 89.3 90 comparable to legal 80 69.9 domestic sales in the 70 60.6 largest single market 60 54.9 47.0 50

40 32.7 31.1 30.3 30 24.0 Volume (billion sticks) 18.1 16.5 20 14.5 14.0 13.7 12.5 7.9 7.7 6.3 5.4 5.3 5.0 5.0 4.8 4.7 10 2.5 1.7 0.5 0.5 0 UK Italy Spain Malta Malta Latvia Poland France Ireland Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Portugal Belgium Sweden Hungary Slovenia Total ND Romania Lithuania Germany Denmark Total C&C Total ND(L) Netherlands Luxembourg Czech Republic Czech

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

Total cigarette consumption by type High priced northern 2007(1) European markets Volume (bn) 6 7 6 59 19 69 16 7 20 2 107 18 31 66 1 12 34 4 96 4 1 13 84 8 8 2 20 typically have higher C&C 100% 2% 0% 5% 4% 4% 3% 2% 2% 2% 1% 7% 7% 7% 6% 6% 5% 1% 3% 2% shares than other Member 11% 11% 10% 10% 1% 1% 4% 7% 15% 13% 13% 1% 1% 1% 5% 8% 17% 17% 16% 16% 2% 6% States 1% 2% 2% 26% 4% 8% 8% 9% 80% 6% 7% 12% 13% 1%

60%

99% 95% 96% 95% 96% Outflows 91% 92% 92% 93% 94% 91% 94% 91% 88% 89% 87% 89% 84% 87% 40% 82% 77% 77% 78% 78% C&C 73% 71% 72% ND(L)

LDC 20% Share total of consumption

0% (1)% (3)% (4)% (2)% (3)% (1)% (4)% (2)% (3)% (8)% (4)% (8)% (2)% (2)% (5)% (9)% (10)% (12)% (13)% (12)% (14)% (17)% (20)% (21)% (21)% (20)% (23)%

(422)% (430)%(40)% UK Italy Spain Malta Latvia Ireland France Poland Austria Cyprus Finland Greece Estonia Bulgaria Belgium Slovakia Sweden Portugal Hungary Slovenia Romania Lithuania Germany Denmark Netherlands Luxembourg Czech Republic

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

− consumption overview

− overall non-domestic inflows

 non-domestic (legal) inflows

 C&C inflows and country comparison

− seizures

 Appendices

The volume of non-domestic Non-domestic inflows accounted for 12.6% of EU consumption versus 12.9% in 2006 inflows amounted to 91 billion in  Like-for-like comparison between the 2006 Member States shows that non-domestic volumes declined by 2.4% in 2007 2007 for the 27 Member States  Germany and France, which represent the two largest markets for non-domestic inflows, experienced a decline in volumes in 2007 − non-domestic volumes in Germany declined by 3.4 billion sticks to 19.3 billion sticks in 2007 − non-domestic volumes in France declined by 0.8 billion sticks to 15.3 billion sticks in 2007  The UK was the largest market to experience an increase in non-domestic inflows − non-domestic inflows to the UK increased by 2.3 billion sticks to 13.5 billion in 2007

The proportion of inflows from In 2007, 75% of inflows to Estonia, Lithuania, Poland and Slovenia came from outside the EU outside the EU typically increases  Less than 20% of inflows to Denmark and Luxembourg came from outside the EU in 2007 with proximity to EU borders

Intra-EU flows increased to 62% of The increase in non-domestic inflows from EU markets was driven by the accession of Bulgaria and Romania to the EU non-domestic inflows in 2007 from  Cigarettes from Bulgaria and Romania accounted for 4% of total non-domestic inflows to EU markets in 2007 58% the previous year The share of inflows from Ukraine declined in 2007  Inflows from Ukraine declined to 9% of EU total non-domestic inflows in 2007, down from 12% the previous year

Total non-domestic inflows by destination Non-domestic volume by destination(1) Declines in non-domestic 2006-2007 (1) inflows to France and Destination 2006 2007 Germany were partially Germany 22.7 19.3 France 16.1 15.3 offset by increases to the 100% UK 11.1 13.5 10.0% UK and Poland in 2007 12.7% Netherlands 5.5 5.4 90% 2.2% Poland 4.4 5.3 Bulgaria and Romania 2.3% 3.9% 2.5% Italy 5.0 4.8 2.0% 2.6% together accounted for 80% 1.7% Spain 3.3 4.3 2.2% 2.8% 5.7% of non-domestic 3.8% 3.2% Romania n/a 2.9 4.7% Other EU Austria 1.9 2.5 inflows in 2007 70% 5.7% 5.3% Hungary Bulgaria n/a 2.3 5.0% Ireland Greece 1.5 2.3 5.8% 60% 6.2% Greece Ireland 1.7 2.1 5.9% Bulgaria Hungary 3.4 2.0 Lithuania 2.4 1.7 50% 12.7% Austria Belgium 1.7 1.6 14.8% Romania Spain Sweden 1.7 1.6 40% Italy Finland 1.4 1.5 18.3% Poland Denmark 0.5 0.5 30% 16.9% Netherlands Portugal 1.0 0.5 UK Latvia 0.5 0.4 France Estonia 0.6 0.3 20% Germany Slovakia 0.6 0.3 25.8% Slovenia 0.4 0.3 10% 21.2% Czech Republic 0.2 0.2 Cyprus 0.1 0.1

0% Luxembourg 0.1 0.1 2006 2007 Malta 0.0 0.0 Total ND volumes Total 87.8 90.9 87.8 90.9 (billion sticks): ND as a percentage 12.9% 12.6% of consumption

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

Share of non-domestic consumption by country Higher priced northern 2007(1) European markets tend to 35% have higher non-domestic 28.6% 30% 28.1%27.2% consumption relative to 25% 23.3%22.7% the rest of the EU 22.2%21.5% 20% 18.0% 16.1% Of the 2004 Accession 15% 13.1%12.7%12.6% 11.8%11.3%11.1% States, Lithuania has the 9.3% 9.3% 9.1% 10% 8.0% 7.8% 6.9% 6.2% 6.2% 5.1% 5.0% highest non-domestic 3.7% 3.6%

Share of total of Share consumption 5% share at 27.2%, down 1.1% 0% significantly from 42.6% UK Italy Total Spain Malta the previous year Latvia Ireland France Poland Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Sweden Portugal Belgium Hungary Slovenia Romania Lithuania Germany Denmark Netherlands Luxembourg Czech Republic

Percentage point change in share of non-domestic consumption by country 2006-2007(1)(a)

4.8 4.0 5 3.1 1.7 2.5 0.6 1.4 1.2 n/a 0.1 n/a 0.0 0.0 0.0 0.0 0 -0.6 -0.5 -0.6 -0.2 -0.1 -1.9 (5) -3.3 -2.6 -3.2 -4.4 -7.0 (10) Percentage points (15) -12.6 -15.3 (20) UK Italy Total Spain Malta Latvia France Poland Ireland Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Sweden Portugal Belgium Slovenia Hungary Romania Lithuania Germany Denmark Netherlands Luxembourg Czech Republic Note: (a) Percentage point change for “Total” measures the change in average non-domestic share of consumption between the 25 Member States in 2006 and the 27 Member States in 2007 Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

The proportion of inflows Share of total non-domestic inflows by source, EU vs. non-EU 2007(1) from outside the EU Volume (bn) 0.3 0.3 1.7 5.3 0.3 2.9 1.5 0.1 0.2 4.8 2.0 0.0 1.6 0.4 0.5 15.3 19.3 1.6 4.3 13.5 2.5 2.3 5.4 2.1 2.3 0.5 0.1 typically increases with 100% proximity to EU borders 12% 18% 90% 19% 22% 21% 25% 30% 28% 32% 31% 31% 38% 37% 80% 41% 44% 46% 46% 52% 52% 55% 70% 62% 61% 60% 66%

78% 77% 60% 85% 85%

50% Non-EU origin EU origin 88% 82% 40% 81% 78% 79% 75% 70% 72% 68% 69% 69% 62% 63% 30% 59% 56% 54% 54% 48% 48% 45% 20% 38% 39% 40% 34%

22% 23% 10% 15% 15%

0% UK Italy Total Spain Malta Latvia Poland France Ireland Cyprus Austria Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Slovenia Hungary Romania Lithuania Germany Denmark Netherlands Luxembourg Czech Republic

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

Non-domestic inflows by country of origin (1) Intra-EU flows accounted 2006-2007(1) Non-domestic inflows by country of origin for 62% of non-domestic Source 2006 2007 EU DF 8.6% 11.8% inflows in 2007, up from 100% Spain 12.4% 11.1% 58% the previous year 9.3% 11.7% Poland 11.3% 10.4% 90% Czech Republic 5.4% 4.6% This increase was driven 6.9% 6.1% Italy 2.4% 2.2% by the accession of 6.8% 80% 6.8% Luxembourg 2.2% 2.2% Bulgaria and Romania to Belgium 2.3% 1.7%

the EU 70% 12.0% 9.1% Greece 1.5% 1.6% Other non-EU France 1.0% 1.5% 4.3% 4.6% Russia Germany 2.1% 1.4% 60% 2.4% 4.0% WW DF Portugal 1.4% 1.4% Ukraine UK 0.6% 1.2% 50% 17.9% 17.8% PMI counterfeit Latvia 0.9% 1.0% Romania and Bulgaria Slovenia 0.7% 0.9% Other EU Netherlands 0.7% 0.9% 40% 2.4% 2.2% Other EU25 4.4% 4.1% 5.4% 4.6% Italy Czech Republic Total 2006 Member States 58.0% 58.0% 30% 10.4% 11.3% Poland Romania 2.2% 3.2% Spain Bulgaria 0.2% 0.7% 20% EU DF 11.1% Total 2007 Accession States 2.4% 4.0% 12.4% PMI counterfeit 4.6% 4.3% 10% 11.8% 8.6% Ukraine 12.0% 9.1% 0% WW DF 6.8% 6.8% 2006 2007 Russia 6.9% 6.1% Serbia 0.7% 1.1% Moldova 0.0% 1.1% Other non-EU 8.6% 9.5% Total non-EU 35.0% 33.7%

Total 100% 100%

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

− consumption overview

− overall non-domestic inflows

 non-domestic (legal) inflows

 C&C inflows and country comparison

− seizures

 Appendices

Non-domestic (legal) Total non-domestic (legal) accounted for 4.2% of total consumption in 2007 volumes accounted for 30.3  This represented a slight decline from 2006 when ND(L) represented 4.5% of total consumption billion sticks in 2007 Declining volumes in France and Germany have been partially offset by an increase in non-domestic (legal) volumes in the UK and Netherlands  France and Germany accounted for 14.2 billion sticks of EU non-domestic (legal) volumes in 2007, down from 18.2 billion sticks the previous year  The UK and Netherlands accounted for 6.5 billion sticks of EU non-domestic (legal) volumes in 2007, up from 4.2 billion sticks the previous year

Non-domestic (legal) Germany, France and the UK together accounted for almost 60% of EU ND(L) consumption in 2007 inflows as a proportion of  The Netherlands and Spain are the only other Member States that accounted for over 5% of the EU total total consumption are typically higher in northern The Netherlands has the highest level of non-domestic (legal) relative to consumption at 12.5% in 2007 states  Ireland is the only other Member State where non-domestic (legal) accounted for more than 10% of consumption, at 11.7% in 2007  ND(L) flows accounted for over 7.5% of total consumption in Sweden, Germany, France and Luxembourg

Flows from EU source The share of non-domestic (legal) accounted for by intra-EU flows increased to 82% in 2007, up from 74% the previous year countries account for the  Spain was the largest single source of non-domestic (legal) inflows, representing 19.3% of ND(L) flows in 2007 large majority of non- domestic (legal) consumption

The share of non-domestic Growth in PMI share of non-domestic (legal) flows is attributable to increased share of Marlboro and L&M (legal) inflows accounted  Marlboro share of total non-domestic (legal) increased to 32% in 2007 up from 29% the previous year for by PMI brands increased from 37% to 45%  L&M share of total non-domestic (legal) increased to 9% in 2007, up from 5% the previous year in 2007  Non-domestic (legal) accounted for by non-PMI product has fallen to 55% in 2007, down from 63% the previous year

Non-domestic (legal) inflows by destination countries (1) Non-domestic (legal) volume by destination Declining non-domestic 2006-2007(1) (legal) volumes in France Destination 2006 2007 Germany 10.6 8.9 and Germany were 100% France 7.7 5.3 partially offset by UK 3.4 4.1 13.4% 12.4% increases in the UK and Spain 1.2 2.4 1.9% Netherlands 0.9 2.4 Netherlands in 2007 2.0% 2.2% 2.3% 2.2% Ireland 1.0 0.8 80% 2.8% 2.7% 3.2% 2.8% Belgium 0.9 0.8 2.8% Romania n/a 0.7 4.0% 7.9% Other EU Sweden 0.7 0.7 10.9% 8.0% Poland Poland 0.6 0.6 60% Sweden Austria 0.2 0.6 Romania Italy 1.0 0.5 13.4% Greece 0.3 0.5 Belgium Finland 0.7 0.4 24.7% Ireland Denmark 0.4 0.4 Netherlands 40% Hungary 0.1 0.3 17.4% Spain Latvia 0.3 0.2 UK Slovakia 0.3 0.1 France Czech Republic 0.1 0.1 Germany Portugal 0.3 0.1 20% 34.0% Bulgaria n/a 0.1 29.2% Cyprus 0.1 0.1 Lithuania 0.1 0.1 Luxembourg 0.1 0.1 0% Estonia 0.1 0.1 2006 2007 Slovenia 0.1 0.0 Malta 0.0 0.0 Total 31.0 30.3 ND(L) total volumes 31.0 30.3 (billion sticks) ND(L) share of 4.5% 4.2% consumption: ND(L) share of non- 35.3% 33.4% domestic:

Source: [1) KPMG analysis based on EPS, LDS and ND(L) research

Non-domestic (legal) share of consumption by country The Netherlands had the 2007(1) highest non-domestic 14% (legal) share of 12.5% 11.7% consumption in 2007 at 12%

12.5% 10% 9.0% 8.3% 7.7% 7.6% 7.4% 8% 6.9% 6.5% 6.4% 6% 5.3% 4.4% 4.2% 3.6% 4% 2.9% 2.3% 2.1% Share of total of Share consumption 1.9% 1.6% 1.5% 2% 1.3% 1.3% 1.1% 1.0% 0.9% 0.7% 0.7% 0.5% 0% UK Italy Total Spain Malta Latvia Poland Ireland France Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Hungary Slovenia Romania Lithuania Denmark Germany Netherlands Luxembourg Czech Republic Percentage point change in non-domestic (legal) share of consumption by country 2006-2007(1)(a)

10 7.9 8

4 2.2 1.2 1.4 2 0.8 0.6 0.8 0.0 0.1 (0.0) 0.0 n/a n/a 0 (0.1) (0.1) Percentage points (0.3) (0.3) (0.4) (0.5) (2) (1.0) (0.8) (0.9) (1.2) (1.9) (1.6) (4) (3.1) (4.1) (6) (5.1) UK Italy Total Spain Malta Latvia Poland Ireland France Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Hungary Slovenia Romania Lithuania Denmark Germany Netherlands Luxembourg

Note: (a) Percentage point change for “Total” measures the change in average non-domestic (legal) share of consumption Czech Republic between the 25 Member States in 2006 and the 27 Member States in 2007 Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

ND(L) share of total consumption by Member State Non-domestic (legal) flows 2007(1) as a percentage of consumption are typically Finland higher in northern 6% Member States

Sweden 9% Estonia 2% Latvia Denmark 5% 4% Lithuania 1% Luxembourg 8%

Ireland Netherlands 12% UK 13% Poland 7% Germany 1% Belgium 8% 6% Czech Republic 1% Slovakia 2% Austria Hungary 4% France 2% 8% Romania Slovenia 2% 1% Italy Bulgaria 1% 1%

Portugal Spain Greece 1% 3% 1%

Cyprus Malta 7% 1%

Key:  More than 10%  5% to 10% Source: (1) KPMG analysis based on EPS, LDS and ND(L) research  Less than 5%

Major ND(L) flows over 0.5 billion sticks France, Germany, and the 2007(1) UK are the destination markets for the largest non-domestic (legal) flows by volume

Key: ND(L) flows over 0.5 billion sticks Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

Non-domestic (legal) flows by country of origin (1)(a) ND(L) flows between the 2006-2007(1)(a) Non-domestic (legal) inflows by country of origin 2006 Member States Source 2006 2007 increased to 82% of total 100% Spain 18.7% 19.3% Czech Republic 10.9% 7.8% ND(L) inflows in 2007, up 11.5% Poland 13.3% 6.5% from 74% the previous 90% 23.0% 1.1% France 2.4% 5.9% 1.7% year 2.8% 0.9% Belgium 3.1% 5.4% 80% Italy 3.3% 5.3% 0.9% Other non-EU 2.2% Luxembourg 4.7% 4.6% 0.5% United States Of America Germany 3.0% 4.5% 70% 31.8% Turkey UK 1.2% 3.8% 21.8% Switzerland Greece 2.8% 3.4% 60% Austria 0.8% 2.1% Romania and Bulgaria Portugal 2.1% 2.1% Other EU 50% 3.3% Netherlands 1.7% 2.1% 5.3% 3.1% Italy Hungary 0.3% 1.2% 2.4% 5.4% Sweden 0.8% 1.2% 40% Belgium Other EU25 4.4% 6.7% 13.3% 5.9% France Total 2006 Member States 73.5% 82.0% 30% 6.5% Poland Romania 0.2% 0.3% Czech Republic 10.9% 7.8% Bulgaria 0.3% 0.7% 20% Spain Total 2007 Accession States 0.5% 0.9%

Switzerland 0.0% 2.8% 10% 18.7% 19.3% Turkey 2.2% 1.7% United States Of America 0.9% 1.1%

0% Ukraine 0.4% 0.8% 2006 2007 Egypt 0.0% 0.8% Other non-EU 22.6% 9.9% Total non-EU 26.1% 17.0%

Total 100% 100%

Note: (a) ND(L) by origin shares are based on country of purchase rather than variant purchased. As a result, these shares include purchases of Duty Free variants made when travelling outside the EU Source: (1) KPMG analysis based on EPS, LDS, and ND(L) research

Non-domestic (legal) flows by brand Growth in non-domestic 2006-2007(1) Non-domestic (legal) inflows by brand(1) (legal) volumes of Brand 2006 2007 Marlboro and L&M 100% Marlboro 28.7% 31.7% increased overall PMI L&M 4.7% 8.5% share of legal flows to 45% 90% Chesterfield 1.2% 1.1% 27.0% Philip Morris 1.0% 1.0% in 2007 Parliament 0.1% 0.2% 80% 38.8% Other PMI 1.7% 2.2%

2.3% PMI total 37.3% 44.7% 70% 2.4% 3.3% Other non-PMI Camel 4.7% 6.2% 4.1% Lucky Strike Pall Mall 3.4% 5.1% 60% 2.3% 1.6% 4.8% John Player Special Benson & Hedges 4.0% 4.8% 2.6% 5.1% Gauloises West 5.5% 4.1% 5.5% 50% West Gauloises 2.6% 3.3% 4.0% 6.2% Benson & Hedges John Player Special 1.6% 2.4% 3.4% 3.3% 1.1% Pall Mall 40% 4.7% Lucky Strike 2.3% 2.3% 2.8% 8.5% Camel Prince 0.6% 1.6% 1.2% Other PMI Fortuna 0.3% 1.5% 30% 4.7% Chesterfield Silk Cut 1.2% 1.2% L&M Winston 1.1% 1.0% 20% Marlboro Superkings 1.0% 1.0% 31.7% 28.7% HB 0.9% 0.8% 10% Brooklyn #N/A 0.7% Dunhill 0.0% 0.7% Other PMI total #N/A 18.5% 0% Non-PMI total 62.7% 55.3% 2006 2007 Total 100% 100%

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

− consumption overview

− overall non-domestic inflows

 non-domestic (legal) inflows

 C&C inflows and country comparison

− seizures

 Appendices

Total EU C&C volumes Total C&C volumes accounted for 8.4% of total consumption in 2007 amounted to 60.6 billion  Like-for-like comparison between the 2006 Member States shows that the consumption of C&C declined by 0.6 billion sticks in 2007 to 56.2 billion sticks in 2007 sticks  Approximately half of C&C volumes can be accounted for by intra-EU flows  C&C volumes for Bulgaria and Romania were 4.4 billion sticks in 2007

Germany and France are the Germany and France accounted for a third of C&C inflows in 2007 two largest destination  A decline in C&C consumption in Germany of 1.7 billion sticks was partially offset by an increase in French C&C of 1.6 billion sticks from 2006 levels countries for C&C inflows  Lithuania remains the Member State with the highest C&C share of consumption, although it did see a significant decline in 2007

Approximately half of C&C Within the EU, Poland is the largest source country for C&C flows volumes can be accounted  Poland accounts for the largest share of intra-EU C&C flows by source country at 12.4% in 2007 for by intra-EU flows Russia and Ukraine are the key external source markets for EU C&C volumes  Inflows from Russia and Ukraine accounted for 22.3% of C&C inflows in 2007  Inflows from Ukraine declined to 8.1 billion sticks in 2007, down from 10.5 billion sticks the previous year

C&C volumes by destination country (1) C&C volume by destination Declining C&C volumes in 2006-2007(1) Germany and the Destination 2006 2007 Germany 12.1 10.4 Netherlands have been 100% France 8.5 10.0 9.6% 8.9% partially offset by UK 7.8 9.4 2.6% increases in France, the UK 4.1% Poland 3.8 4.7 2.8% and Poland 5.7% 3.0% Italy 4.0 4.3 80% 3.1% Netherlands 4.6 3.0 3.6% 3.3% 2.1% Romania n/a 2.2 3.6% 3.1% Bulgaria n/a 2.2 3.7% Other EU 8.1% Austria 1.7 2.0 4.9% Lithuania Greece 1.2 1.9 7.1% Hungary 60% 7.1% Spain 2.1 1.8 Spain Hungary 3.3 1.7 6.7% 7.7% Greece Lithuania 2.3 1.6 Austria Ireland 0.7 1.2 13.7% Bulgaria 15.5% Finland 0.6 1.1 40% Romania Sweden 0.9 0.9 Netherlands Belgium 0.9 0.8 Italy 14.9% Portugal 0.6 0.3 Poland 16.6% Estonia 0.4 0.2 UK 20% Slovenia 0.3 0.2 France Latvia 0.3 0.2 Germany Denmark 0.2 0.2 21.3% 17.2% Slovakia 0.3 0.1 Czech Republic 0.1 0.1 0% Luxembourg 0.0 0.0 2006 2007 Malta 0.0 0.0 Cyprus 0.0 0.0 C&C total Total 56.8 60.6 volumes 56.8 60.6 (billion sticks) C&C share of 8.3% 8.4% consumption: C&C share of 64.7% 66.6% non-domestic:

Source: (1) KPMG analysis based on EPS, LDS, and ND(L) research

C&C as a share of consumption by country Although Lithuania 2007(1) remains the Member State with the highest C&C 30% 25.7% share of consumption at 25% 25.7% in 2007, it had the 20% 16.9%16.9% greatest decline in C&C 15.8%15.6% 14.6% share of total consumption 15% 12.6%12.5% 11.1%10.7% 9.7% 9.5% 8.4% 10% 7.2% 7.1% 6.5% 6.2% 5.5% 5.1% 4.5% 4.1% 5% 3.7% Share of total of Share consumption 2.5% 2.2% 1.9% 1.7% 1.7% 0.3% 0% UK Italy Total Spain Malta Latvia Poland France Ireland Cyprus Austria Finland Greece Estonia Bulgaria Slovakia Portugal Sweden Belgium Slovenia Hungary Romania Lithuania Denmark Germany Netherlands Luxembourg Czech Republic

Percentage point change in C&C as a share of consumption by country 2006-2007(1)(a)

10 6.8 6.8

5 2.8 2.7 1.9 0.9 1.5 0.8 0.6 0.1 0.3 0.0 0.0 0.0 0 n/a n/a (0.9) (0.8) (0.5) (0.2) (2.0) (5) (2.3) (2.5) (2.8)

(10) (8.6) (8.5) (7.8) Percentage points

(15) (15.0) (20) UK Italy Total Spain Malta Latvia Ireland France Poland Austria Cyprus Finland Greece Estonia Bulgaria Slovakia Sweden Portugal Belgium Hungary Slovenia Romania Lithuania Denmark Germany Netherlands Luxembourg

Note: (a) Percentage point change for “Total” measures the change in average C&C share of consumption Czech Republic between the 25 Member States in 2006 and the 27 Member States in 2007 Source: (1) KPMG analysis based on EPS, LDS, and ND(L) research

C&C share of total consumption by Member State C&C as a share of 2007(1) consumption varies significantly across the EU Finland Member States 17%

Estonia Sweden 11% 13% Latvia Denmark 4% Luxembourg 2% 4% Lithuania Netherlands 26% 16% Ireland UK 17% 16% Poland Germany 7% Belgium 10% 6% Czech Republic 0% Slovakia 2% Austria Hungary 13% France 10% 15% Slovenia Romania Italy 5% 7% 5%

Bulgaria 11%

Portugal Spain 3% Greece 2% 5%

Cyprus 2% Malta Key:  More than 15%  10% to 15% 7%  5% to 10% Source: (1) KPMG analysis based on EPS, LDS and ND(L) research  Less than 5%

Major C&C flows greater than 0.5 billion sticks Russia and Ukraine remain 2007(1) source countries for C&C flows to multiple countries

Counterfeit

Key: Counterfeit

Key: C&C flows over 0.5 billion sticks Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

C&C inflows by country of origin Counterfeit and contraband inflows by country of origin(1)(a) The share of C&C 2006-2007(1)(a) accounted for by intra EU Source 2006 2007 Poland 10.3% 12.4% flows was broadly flat at 100% Spain 8.9% 7.0% 51% in 2007 11.4% Other EU25 30.4% 26.6% 19.8% 90% Total 2006 Member States 49.6% 46.0%

10.0% Romania 3.3% 4.7% 80% Bulgaria 0.1% 0.8% 8.9% Total 2007 Accession States 3.4% 5.5% 70% 18.5% PMI counterfeit 7.1% 6.5% 13.4% Other non-EU Ukraine 18.5% 13.4% 60% Russia 10.0% 8.9% 7.1% Russia 6.5% Other non-EU 11.4% 19.8% 3.4% Ukraine 50% 5.5% Total non-EU 39.9% 42.0% PMI counterfeit Total 100% 100% Romania and Bulgaria 40% Other EU 30.4% 26.6% Spain 30% Poland

20% 7.0% 8.9% 10% 12.4% 10.3%

0% 2006 2007

Note: (a) C&C by origin shares are based on variant purchased Source: (1) KPMG analysis based on EPS, LDS, and ND(L) research

Total Ukraine outflows (ND(L) and C&C) by destination The decline in outflows to Germany and Hungary was a key driver Total inflows from Ukraine 2006-2007(1) for the overall decline in Ukrainian outflows into the EU declined to 8.3  Flows to Germany fell to 2.3 billion sticks in 2007, from 4.1 billion billion sticks in 2007 sticks the previous year 12 10.5  Flows to Hungary fell to 0.5 billion sticks in 2007, from 2.5 billion the 10 1.4 previous year 8.3 Outflows from Russia into the EU remained relatively stable in 2007 8 2.5 0.9 Other EU 0.5  Outflows to Germany and Lithuania fell by 0.9 billion sticks in 2007, 0.9 0.9 Hungary 6 although this was partially offset by a 0.7 billion stick increase in 1.0 Romania outflows to Poland and Finland 4 4.1 2.3 Italy

Volume (billion sticks) Germany 2 2.6 Poland 1.7 0 2006 2007

Total Russian outflows (ND(L) and C&C) by destination 2006-2007(1)

Other EU 8 6.0 Estonia 5.5 6 1.1 Finland 0.2 0.8 0.6 0.2 Germany 4 0.9 1.7 1.1 Lithuania Volume (billion sticks) 2 Poland 1.6 1.2 0.8 1.2 0 2006 2007

Source: (1) KPMG analysis based on EPS, LDS and ND(L) research

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

− consumption overview

− overall non-domestic inflows

 non-domestic (legal) inflows

 C&C inflows and country comparison

− seizures

 Appendices

Seizure data is weighted Counterfeit product accounted for 83% of seizure volume in 2007 towards counterfeit product  Anecdotal evidence suggests that counterfeit is more likely to be transported in large shipments such as containers − average seizure size for counterfeit product was 1.0 million sticks in 2007, compared to under 0.1 million sticks for genuine product − in addition, smaller seizures of less than five cases and instances of individuals attempting to exceed import allowances may not be reported  Ready availability of cheaper genuine product in other markets means this can be moved in large numbers of smaller shipments which may be more difficult to detect and intercept

Counterfeit share of cigarette seizures (all EU PMI seizures) Counterfeit product accounted for the majority of seized cigarettes PMI counterfeit accounted 2006-2007(1) in the EU for 83.1% of PMI seizure  Despite changes in the number and location of seizures in 2007, the volumes in 2007 composition of seized volumes in terms of counterfeit and authentic 100% products has remained stable 18.9% 16.9% Counterfeit seizures tend to be larger than seizures of authentic 80% product

60%  The average size of counterfeit seizures was 1.0 million sticks in PMI Genuine 2007, down from 1.1 million the previous year PMI Counterfeit 40% 81.1% 83.1%  The average size of genuine seizures was 96 thousand sticks in 2007, down from 98 thousand sticks the previous year 20% Seizure data is weighted towards counterfeit product

0%  Smaller seizures of less than five cases and instance of individuals 2006 2007 attempting to exceed import allowances may not be reported  Anecdotal evidence suggests that counterfeit is more likely to be transported in large shipments such as containers  Ready availability of cheaper genuine product in other countries means that it can be moved in a large number of smaller quantities which may be more difficult to detect and intercept

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

 Appendices

− country results

− summary of country flow refinements

− sources

− scope of work

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Austrian consumption Non-domestic inflows to 2006-2007(1)(2)(b) Austria increased to 2.54 billion sticks in 2007 100% 11.7% 12.6% 1.3% 3.6% Inflows from Slovenia and 80% Hungary both increased 60% C&C following the relaxation of ND(L) 40% 87.0% 83.9% import allowances LDC 20% Outflows As a result of this change, 0% the share of consumption total of Share consumption (3.1%) (2.8%) accounted for by ND(L) (20%) more than doubled in 2007 2006 2007

Volume (bn sticks): 14.9 15.8

Total non-domestic inflows (ND(L) and C&C) by origin (1)(2)(b) Key:  Austria  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to larger flow 3.0 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2006 2.5

0.93 Proportion of total EU levels(1)(2) 2.0 Other countries

15.5% 0.69 Czech Republic 16% 14.8% 1.5 0.16 0.32 Hungary 13.4% 0.17 1.0 0.13 Slovenia 0.23 0.53 12% Volume (bn sticks) 0.5 Duty Free 0.68 0.60 8.2% 8% 0.0 2006 2007

Volume (bn sticks): 1.94 2.54 4%

Percentage of EU total 2.2% Note: (b) This report and the Project Star 2006 results use 13.4 billion sticks as estimated legal domestic sales (LDS) in Austria during 2006. However, PMI internal estimates now 0% suggest that LDS in 2006 was actually 13.8 billion sticks. Use of the higher 13.8 billion sticks estimate for LDS in 2006 gives implied total consumption of 15.3 billion sticks, Consum- ND ND(L) C&C Outflows which is more in line with 2007 results ption

Historic cigarette prices and legal domestic sales Updated data suggest that Legal domestic sales 1997-2007(1)(2)(a) legal domestic sales were (a) declined in 2005 in actually 13.8bn sticks in 2006 response to price 18 4.0 increases, but have 16 3.5 (Euro) per 20 sticks Price recovered slightly over the 14 0.4 3.0 last two years 12 2.5 10 2.0 8 15.1 15.6 15.4 15.2 15.4 15.1 13.6 14.5 13.4 13.7 1.5 6 13.0 Volume (bn sticks) 4 1.0 2 0.5 0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-1999 1999-2005 2005-2007  Legal domestic sales 7.1% (3.0)% 2.7%  Average pack price 1.9% 4.8% 1.5%

Note: (a) This report and the Project Star 2006 results use 13.4 billion sticks as estimated legal domestic sales (LDS) in Austria during 2006. However, PMI internal estimates now suggest that LDS in 2006 was actually 13.8 billion sticks Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI This report and the Project Star 2006 results use 13.4 billion sticks as estimated legal domestic sales in Austria during 2006. However, PMI internal estimates now suggest that LDS in 2006 was actually 13.8 billion sticks  Use of the higher 13.8 billion sticks estimate for LDS in 2006 suggests that LDS declined by 1.0% in 2007 versus 2006  Import restrictions from neighbouring lower-cost markets were relaxed in July 2007 − up until July 2007, Austrian residents were only able to legally bring in 25 cigarettes per person per trip from the 2004 EU Accession States, including neighbouring Slovenia and Hungary  Following the lifting of import restrictions to 800 sticks per trip from the 2004 EU Accession States, non-domestic inflows to Austria increased − the import limit was subsequently lowered to 200 sticks, effective from January 1st 2008

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Austria and neighbouring €1.33 lower-cost markets €0.86 narrowed slightly in 2007 n/a €4.71 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €2.73 €4.10 €1.58 €1.42 €1.64

Key:  Austria  Major source country Evolution of Marlboro price differential between Austria and key  Major destination country markets(2)(a)

40% 21% 21% 20%

(20%) July 2006 (40%) July 2007 (39)% (36)% (39)% (37)% (45)% (42)% (60%)

(80%) (83)% (84)%

Price premium/(discount) versus local (100%) Germany Slovenia Hungary Czech Ukraine Republic

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Austrian consumption is 2007(1)(2)(3)(4)(a)(b)(c) estimated at 15.8 billion Outflows Inflows 0.45 billion sticks 2.54 billion sticks sticks versus legal sales of 18 13.7 billion sticks, giving a

net inflow of 2.1 billion 16

sticks 1.24

0.10 0.19 14 0.19 0.15 0.01 0.17 0.30 0.13 0.28 0.23

“Other” includes Duty Free inflows 10 Outflows C&C

8 15.76 ND(L)

Volume (bn sticks) 13.66 13.22 6

0 LDS Germany Other LDC Slovenia Hungary Czech Republic Ukraine Other Consumption

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) (c) 2007 Slovenian inflows have been adjusted downward to reflect the fact that the EPS was conducted while higher import allowances were in effect (see ‘Austria: comparison of non-domestic sources’ for more detail) Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) In Market Sales supplied by PMI (4) Interviews with PMI Local Management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin The legal share of total 2006-2007(1)(2)(3) 2007(1)(2)(3) inflows from the 2004 EU 0.7 Accession States has been 0.6 0.56 adjusted upwards in 2007 Other countries 27% to reflect increases in 0.5 0.15 0.4 Slovenia import allowances from 0.05 Other countries 41% these markets and 0.3 0.13 Greece 0.20 Hungary 0.2 updated EPS results Volume (bn sticks) Slovenia 0.15 0.1 0.01 0.23 Greece

0.04 0.00 8% 0.0 2006 2007

ND(L) share of Hungary 1% 4% consumption: 24% ND(L) share of non- 10% 22% domestic:

Non-domestic (legal) by brand ND(L) analysis(a) 2006-2007(1)(2)(3) Ave. Total 0.7 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.6 0.56 Values x x x x = 0.56bn 0.5 2007 6.5m 59% 13% 4.8 230

0.4 0.30 Non-PMI

0.3 Other PMI EU rank 0.20 0.05 Marlboro 2007 15 8 24 6 9 11 0.2 Volume (bn sticks) 0.11 0.1 0.02 0.21 0.08 0.0 2006 2007

PMI share of ND(L): 47% 47%

Note: (a) Average purchases have increased from 82 sticks in 2006 to reflect changes in import restrictions from selected countries in 2007 including Hungary and Slovenia Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Total counterfeit and 2006-2007(1)(2)(3) 2007(1)(2)(3) contraband inflows to 1.98 2.0 Austria increased to 1.74 almost 2.0 billion sticks in 1.5 Other countries Slovenia 2007 1.24 15% 1.18 Ukraine 1.0 Czech Republic Hungary 0.10 Hungary 9% 0.15 Other countries Volume (bn sticks) 0.5 0.05 Slovenia 0.16 0.19 63% 0.12 0.30 Czech Republic 0.22 8% 0.0 Other countries 2006 2007 includes Duty Free inflows Ukraine C&C share of 12% 13% 5% consumption: C&C share of non- 90% 78% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4) EU Flows Model results for both 2006 and 2007 are comparable to the KPMG’s estimate of non- Customs estimated a Austrian Chamber of Commerce data for those years domestic consumption is range of 11% to 15% for non-domestic incidence  External data sources therefore support the trend for increased non- broadly in line with that of domestic consumption in 2007 20% 18.1% 17.4% the Austrian Chamber of 16.7% 18% 16.1% KPMG’s non-domestic estimate for Austria has been based on a Commerce in 2007 16% 13.0% downwards revision to the Austrian Chamber of Commerce EPS results 13.0% 13.0% 14%  EPS fieldwork was conducted between September and December and 12% therefore may not reflect the annual consumption trend 10% − the Austrian Chamber of Commerce EPS was conducted after import 8% allowances were relaxed in July 6% 4%  In order to build an accurate full-year picture of non-domestic Share of total of Share consumption 2% consumption, KPMG has adjusted Slovenian inflows downward to reflect the different import allowances for the first-half of the year versus the 0% second-half −

2006 Slovenian inflows have been adjusted downward by 0.25 billion sticks 2006 Gallaher 2005 2007 2006

Customs to 0.53 billion sticks (2006 inflows were 0.23 billion sticks). EU Flows EU Flows Commerce Commerce Commerce Model 2007 Model 2006 Chamber of Chamber of Chamber of − Slovenian adjustment was calculated based on the uplift in Slovenian LDS in the second half of 2007 versus 2006(5) and is consistent with the overall consumption trend in Slovenia

Sources: (1) KPMG EU Flows Model (2) Interview with Austrian Customs (3) Austrian Chamber of Commerce empty pack survey, 2005, 2006 and 2007. Full PMI results were available along with total non-domestic share (4) Gallaher EPS 2006 (5) Slovenia In Market Sales supplied by PMI

results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in smoking habits. However, GCTS data cannot be used to quantify total 18 remained reasonably consumption in a market due to exclusions of particular age cohorts and 16 stable since 2002 the problem of under-reporting of consumption 14  The KPMG consumption index is calculated using the incidence of 12 smoking and the average daily consumption (both from GCTS) and the 10 population of each market (from Euromonitor GMID) 8 15.8 − these figures are then indexed to the earliest year for which we have 6 both GCTS and LDS data available, 2002 for Austria

Volume (billion sticks) 4  If LDS trends differ from those shown by the consumption index, it 2 implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2002-2005 2005-2007 2002-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS (5.5)% 2.7% (2.3)% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (3.7)% 0.7% (2.0)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI present in the first analysed year (3) KPMG EU Flows Model  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Austria remained reasonably stable between 2002 and 2007  Consumption and LDS in Austria have declined at similar rates since 2002  This result suggests that non-domestic consumption in Austria has been reasonably stable since 2002 Consumption index modelling suggests that the decline in legal sales in 2005 and the recovery in 2006 were not driven by changes in consumption  This suggests that non-domestic consumption increased in 2005 and decreased in 2006 which is consistent with Chamber of Commerce EPS results − see “Comparison of sources for non-domestic estimates” page for more detail

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Belgian consumption Non-domestic inflows to 2006-2007(1)(2)(b) Belgium declined slightly 100% 6.6% 6.2% to 1.6 billion sticks in 2007 6.6% 6.5% 80%

60% C&C 87.3% 40% 86.8% ND(L) LDC 20% Outflows

Share of total of Share consumption 0% (15.6%) (12.6%) (20)% 2006 2007

Volume (bn sticks): 13.1 12.5

Total non-domestic inflows by origin (ND(L) and C&C) 2006-2007(1)(2)(c)

   Key: Belgium Major source country Major destination country 1.8 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 1.6 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 1.4 0.58 (2) Synovate ND(L) research, 2007 Other countries 1.2 0.90 Spain (1)(2) 0.16 Proportion of total EU levels 1.0 0.07 France 4% 3.3% 0.8 0.45 Duty Free 0.6 0.07 3% 0.15

2.7% Volume (bn sticks) Luxembourg 0.4 0.21 3% 0.2 0.47 0.25 2% 1.7% 1.7% 0.0 2006 2007 2% 1.3% Volume (bn sticks): 1.73 1.58 1% Notes: (b) 2006 legal domestic share of total consumption differs slightly from what was shown in the Percentage of EU total 1% 2006 Project Star Results presentation due to a technical change in approach to the calculation of consumption share (2006 LDC was shown as 86% in the 2006 Results presentation compared to 86.8% in this presentation) 0% (c) 2006 Luxembourg share of non-domestic was incorrectly labelled as 22% in the 2006 Project Consum- ND ND(L) C&C Outflows Star Results presentation. The correct 2006 share for Luxembourg is 26% i.e. 0.47 billion ption sticks divided by 1.73 billion sticks

Historic cigarette prices and legal domestic sales Sale of manufactured 1997-2007(1)(2) cigarettes have declined since 2003 16 14.3 14.3 5 13.7 13.6 13.4 13.0 13.4 13.4 12.5 14 12.3 (Euro) per 20 sticks Price 11.6 4 12

10 3 8 6 2

Volume (bn sticks) 4 1 2 0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2003 2003-2007  Legal domestic sales 5.9% 1.3% (3.3)%  Average pack price n/a 5.6% 4.8%

Sources: (1) Tax stamp data supplied by PM Local Management (2) Weighted average pack price supplied by PMI

Legal domestic cigarette sales declined by 6.7% in 2007, whilst weighted average prices increased by 9.5%  Prices rose in Q1 2007 as higher taxes on cigarettes came into force  Stricter anti-smoking legislation was introduced in 2007 − from January 2007 onwards, smoking is only permitted in restaurants in separate rooms that are completely closed off from the rest of the building and where no food is served − bars and cafes are exempt from the ban, although they must have adequate ventilation as well as a no-smoking zone

Marlboro price comparison There are significant price 1 July 2007(1)(2)(a) differentials between €4.30 Belgium and other EU €8.16 €1.79 member states €4.94 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.1 €2.10 €4.531

€3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66

€4.10 €1.64

€3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Belgium and key (2)(a) markets €3.86 Key:  Belgium €3.50 40% 25%  Major source country  Major destination country 20% 10% 0% 0% July 2006 (20%) (9)% (13)% (18)% July 2007 (40%) (31)% (35)% (60%)

(80%) (81)% (81)% (84)%

Price premium/(discount) versus local (100%) (86)% Spain Russia France Ukraine Netherlands Luxembourg

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Belgian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 12.5 billion Outflows Inflows 1.58 billion sticks 1.58 billion sticks sticks versus legal sales of 14 12.5 billion sticks, with inflows and outflows balancing each other out 12 0.73 0.77

0.73 0.41 0.15 0.12 0.25

8 Outflows C&C ND(L) 12.49 12.50 6 Volume (bn sticks) 10.92

0 LDS France Netherlands Other LDC Luxembourg France Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3) 2007(1)(2)(3)

inflows to Belgium 1.0 declined by 6% in 2007 0.86 0.81 PMI share of non-domestic 0.8 Luxembourg 0.24 31% 0.34 (legal) inflows increased 0.6 Other countries 0.13 Other countries markedly in 2007 Spain 41% 0.06 0.07 0.4 France 0.15 Luxembourg Volume (bn sticks) 0.2 0.43 0.25 0.0 2006 2007 France Spain 19% ND(L) share of 7% 6% 9% consumption: ND(L) share of non- 50% 51% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 1.0 Population Propensity Propensity Ave. no. purchases ND(L) 0.86 0.81 19+ to travel to purchase of trips (sticks) (sticks) 0.8 Values 0.08 x x x x = 0.81bn 2007 8.1m 70% 17% 8.3 107 0.42 0.6 0.26 Non-PMI Other PMI 0.4 EU rank 0.17 12 3 22 1 19 6 Marlboro 2007 Volume (bn sticks) 0.47 0.2 0.28

0.0 2006 2007

PMI share of ND(L): 51% 91%

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI local management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Belgium 2006-2007(1)(2)(3) 2007(1)(2)(3) decreased by 11% 1.0 Russia 0.86 between 2006 and 2007 9% 0.77 Ukraine 0.8 Other countries 5% 77% Other countries Bulgaria 0.6 5% Romania 0.72 0.59 Romania Bulgaria 0.4 4% Ukraine

Volume (bn sticks) 0.2 Russia 0.03 0.01 0.04 0.05 0.04 0.08 0.07 0.0 2006 2007

C&C share of 7% 6% consumption: C&C share of non- 50% 49% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI local management

Non-domestic market volume estimates(1)(2)(3)(a) NMA survey results show a faster decline in non-domestic consumption The trend of declining non- than EU Flows Model results domestic consumption  However, part of the decrease implied by NMA results may be driven by identified by the EU Flows 16% 14.8% seasonality in non-domestic consumption 13.2% Model is also broadly 14% 12.7%  The 2006 NMA Survey was conducted in Q2 and may therefore include 11.8% reflected in the results of 12% greater non-domestic consumption by visitors to Belgium and Belgian nationals returning from holiday. By comparison, the 2007 NMA Survey the NMA survey 10% 8.8% was conducted in December 8% − the PMI Empty Pack Survey was conducted in Q3 in 2007 6% KPMG’s estimate for non-domestic share of total consumption in 2007 is 4% higher than PMI EPS results due to an adjustment to inflows from Share of total of Share consumption 2% Luxembourg 0%  Inflows from Luxembourg were increased slightly in 2007 to control for NMA EU Flows PMI EPS EU Flows NMA changes to the EPS sample survey Model 2007 Model survey 2006 2006 2007 2007 − the 2007 EPS sample excluded the ‘Luxembourg’ region of Belgium

Note: (a) NMA survey results are for PMI brands only Sources: (1) KPMG EU Flows Model (2) NMA market survey, Q2 2006 and Q4 2007 (3) PMI EPS Q3 2007

results suggest that that  The rate of change in the GCTS smoking prevalence and average smoking outflows from Belgium volume figures provide a useful indication of observed changes in 16 smoking habits. However, GCTS data cannot be used to quantify total have become increasingly consumption in a market due to exclusions of particular age cohorts and 14 important relative to the problem of under-reporting of consumption 12 Consumption  The KPMG consumption index is calculated using the incidence of inflows since 1997 has declined 10 since 1997 smoking and the average daily consumption (both from GCTS) and the while LDS population of each market (from Euromonitor GMID) 8 have increased 12.5 − these figures are then indexed to the earliest year for which we have 6 both GCTS and LDS data available, 1997 for Belgium 4 Volume (billion sticks)  If LDS trends differ from those shown by the consumption index, it 2 implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 1997-2000 2000-2003 2003-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS 5.9% 1.3% (3.3)% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (3.9)% 0.0% (7.7)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Sources: (1) KPMG Consumption Index model (2) Tax stamp data supplied by PMI Local Management present in the first analysed year (3) KPMG EU Flows Model  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Belgium increased between 1997 and 2006  Consumption in Belgium has declined since 1997 while LDS have increased  This result suggests that outflows from Belgium have become increasingly important relative to inflows i.e. non-domestic consumption has declined and/or outflows have increased The narrowing of the gap between LDS and the KPMG consumption index in 2007 suggests that outflows became less important relative to inflows  This is consistent with EU Flows Model results which show a decline in net outflows (i.e. outflows net of inflows) from 0.3 billion sticks in 2006 to 0.0 billion sticks in 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1998-2007(1)(2)(3)(a) prices affect cigarette consumption in Belgium that changes in  Historical data from 1998 to 2006 has been used to estimate the strength consumption in Belgium of relationships between consumption and prices and income 20 are related to changes in Control period Estimate 18 − estimated incidence of smoking and average daily consumption are disposable income and 16 based on GCTS data. Population and disposable income are estimated using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices 14 12 Cigarette prices are based on the weighted average pack price supplied by PMI 10 8 KPMG analysis suggests that changes in consumption in Belgium are 6 related to changes in disposable income and cigarette prices(c) Volume (billion sticks) 4  Disposable income is negatively related to changes in consumption i.e. as 2 income rises consumption tends to decrease 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007  Cigarette prices are also negatively related to changes in consumption  It is important to note that factors other than price and income are likely to impact consumption levels in Belgium CAGR (%) 1998-2002 2002-2006 2006-2007 − examples of other factors likely to impact consumption levels include

 LDS 3.5% (1.9)% (6.8)% changes in smoking restrictions and changes in demographic profiles  Estimated consumption(b) (3.8)% (4.4)% (16.4)%  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Based on historical data from 1998 to 2006 considered a broad approximation (b) The estimated consumption line has been rebased so that it equals the 2006 Project Star estimate of consumption for Belgium (13.1 billion sticks) The relationships between changes in consumption and changes in (c) 82% of changes in consumption can be explained by changes in disposable income and cigarette prices disposable income and prices have been used to estimate 2007 Sources: (1) KPMG EU Flows Model consumption (2) Tax stamp data supplied by PMI Local Management (3) KPMG Consumption Regression Model  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is directionally consistent with EU Flows Model and consumption index modelling results. However, the magnitude of the estimated consumption decline appears overly large − this large decline in estimated consumption is driven by relatively high increase in average prices (10%) combined with high estimated price elasticity (1.12)

OTP sales in billion stick equivalents, 2003 to 2007(1)(a)(b) Sales of Other Tobacco Sales of OTP have declined at a lower rate than sales of manufactured cigarettes Products (OTP) decreased 15.0  Sales of OTP declined at an annual rate of 2.6% between 2003 to 2007 between 2003 and 2005 12.8 − by contrast sales of manufactured cigarettes declined at an annual rate 11.6 11.8 11.5 0.5 before increasing by 11.3% 12.0 0.5 0.6 0.5 10.5 of 3.3% over the same time period in 2006 0.5 OTP sales increased in 2006 driven by trade loading and stock build-up 9.0 before tax increases took effect in Q1 2007 OTP sales were 17.9% 6.0 12.2  Trade loading in 2006 is also a contributing factor to the relatively large lower in 2007 11.1 11.2 10.9 10.0 decrease in 2007 OTP sales

Stick equivalents (bns) 3.0 − OTP sales in 2007 were lower partly as a result of inventory overhang from 2006 0.0 Smoking tobacco is the largest category of OTP and accounted for 95% 2003 2004 2005 2006 2007 of OTP sales (excluding smokeless tobacco) in Belgium in 2007

Cigarettes 14.3 13.6 13.4 13.4 12.5 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 1.3% 0.3% (5.7)% (0.7)%  Smoking tobacco (0.8)% 11.8% (18.4)% (2.7)% Total OTP (0.7)% 11.3% (17.9)% (2.6)%

Manufactured cigarettes (3.2)% 0.0% (6.7)% (3.3)%

Notes: (a) Smoking tobacco volumes have been calculated at one stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick for stick basis (b) Smokeless tobacco has been excluded from this analysis Source: (1) OTP volumes supplied by PMI

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1) Total Bulgarian consumption Non-domestic inflows to 2007(1)(2) Bulgaria were 2.3 billion sticks in 2007, accounting 100% 11.1% 0.7% for 11.8% of total 80% consumption 60% C&C ND(L) The majority of inflows are 88.2% 40% LDC classified as C&C with Outflows 20% Serbia and the Duty Free 0% channel representing the Share of total consumption (3.4%) two main sources of non- (20%) domestic inflows in 2007 2007

Greece, France and the UK Volume (bn sticks): 19.7 are the major destination countries for outflows Total non-domestic inflows (ND(L) and C&C) by origin

2007(1)(2) Key:  Bulgaria  Top 3 source country  Top 2 destination country Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 3.0 (2) Synovate ND(L) research, 2007 2.5 Proportion of total EU levels(1)(2) 0.18 2.0 0.24 Other countries 4.0% 3.6% Serbia 3.5% 1.5 Duty Free 3.0% 2.7% 2.6% 1.0 1.90

2.5% Volume (bn sticks) 0.5 2.0% 1.4% 1.5% 0.0 2007 1.0%

Percentage of EU total 0.5% 0.5% Volume (bn sticks): 2.33 0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales in 2004 to 2007(1)(2) Bulgaria increased by 5.3% 25 4.0 in 2007 but remain below 21.5 20.5 3.5 (BGN) per 20 sticks Price 2005 levels 20 18.1 17.1 3.0

15 2.5 2.0

10 1.5

Volume (bn sticks) 1.0 5 0.5 0 0.0 2004 2005 2006 2007

CAGR (%) 2004-2007 2006-2007  Legal domestic sales (5.7)% 5.3%  Average pack price 29.1%(a) 1.3%

Note: (a) CAGR for average pack price is from 2005 to 2007 Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal sales grew by 5.3% in 2007 but remain 3.4 billion sticks below 2004 levels  Tax-driven price increases in January 2006 are understood to have been a contributing factor to the decline in LDS in 2006  The relatively small increase in weighted average price in 2007 partly reflects down-trading by Bulgarian consumers to lower priced brands − the most popular price category (MPPC) shifted from BGN 2.30 to BGN 2.00 during 2007 − the absence of a minimum excise tax has enabled growth in lower priced cigarettes  Until October 2006 retail selling prices were regulated and there was significant price differential between domestic and international brands Tax stamp data for 2007 suggests sales of 19.6 billion sticks  For the purposes of Project Star, tax stamp data has not been used due to the impact of trade loading by retailers in response to price increases  Local Philip Morris Management have observed that the 2008 price increase did not have an immediate impact on retail prices and that the implied stock build up of approximately two months sales at the end of 2007 explains the difference between the two sources

Marlboro price comparison, July 1st 2007(1)(2)(a) Price differentials between

Bulgaria and key €4.30 destination markets €8.16 €4.94 €1.79 narrowed in 2007 €0.86 €1.47 Serbia and Ukraine, the €4.24 €1.33 €0.86 main source markets for €7.05 n/a €4.71 inflows, became relatively €8.04 €4.11 €2.10 cheaper over the same €4.53 €3.70 €2.46 €0.63 time period €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Bulgaria and key Key:  Bulgaria (2)(a) €3.86 markets  Top 2 source country

 Top 3 destination country 600% 483% 500% 390% 400% 294% 300% July 2006 205% 200% July 2007 120% 83% 100%

0% (11)% (13)% Price premium/(discount) versus local (100%) (50)% (62)% UK France Greece Serbia Ukraine

Notes: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown, 2007(1)(2)(3)(4)(a)(b) Bulgarian consumption is estimated at 19.7 billion Outflows Inflows 0.7 billion sticks 2.3 billion sticks sticks in 2007 versus legal sales of 18.1 billion sticks, 20

giving a net inflow of 1.6 1.86

18 0.17 0.15 0.14 billion sticks 0.09 0.27 0.24 0.08

16 Outflows “Other” includes Duty Free inflows C&C 14 ND(L)

10 19.7

Volume (bn sticks) 18.1 17.4 8

0 LDS France Greece UK Other LDC Serbia Ukraine Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows accounted 2007(1)(2)(3) 2007(1)(2)(3) for only 6% of non- 0.2 domestic consumption in 0.14 Other countries 0.1 21% 2007 0.03 0.1 0.01 0.1 0.01 Other countries 0.1 Germany Germany 0.1 Greece 7%

Volume (bn sticks) 0.0 0.09 Duty Free Duty Free 0.0 Greece 62% 10% 0.0 2007

ND(L) share of consumption: 1% ND(L) share of non-domestic: 6%

Non-domestic (legal) by brand ND(L) analysis 2007(1)(2)(3) Ave. Total 0.2 0.14 Population Propensity Propensity Ave. no. purchases ND(L) 0.1 19+ to travel to purchase of trips (sticks) (sticks) 0.1 Values 0.06 6.0m x 22% x 20% x 2.6 x 210 = 0.14bn 0.1 2007 Non-PMI 0.1 Other PMI 0.1 0.04 EU rank Marlboro 2007 16 26 17 20 10 21 Volume (bn sticks) 0.0

0.0 0.04 0.0 2007

PMI share of ND(L): 56%

Source: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Duty Free variant and 2007(1)(2)(3) 2007(1)(2)(3)

Serbian origin products Other countries 2.5 6% accounted for the majority 2.19 0.13 of C&C consumption in 2.0 Serbia 0.24 11% 2007 1.5 Other countries

Serbia 1.0 1.82 Duty Free Volume (bn sticks) 0.5

0.0 2007

C&C share of consumption: 11% Duty Free 83% C&C share of non-domestic: 94%

Source: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(5)(a) Bulgartabac’s study defined “illegal cigarettes” as cigarette packs without a Bulgartabac’s estimate for tax sticker, those with a foreign tax sticker and those with a Duty Free label non-domestic incidence is  Bulgartabac's estimate is based on an Empty Pack Count carried out by the significantly higher than 25% Bulgarian Academy of Science (BAS) covering the 8 largest cities in Bulgaria

that of KPMG 20% − the study took place in July and August and analysed a total of 10,000 discarded packs A key reason for this 15% − of the 22.8% of packs classified as illegal, a total of 10.1% of packs 22.8% difference is likely to be 10% collected appeared to be of Bulgarian origin but did not have a tax sticker 15.6% the treatment of packs 14.0% and were therefore assumed to be illicit in nature 5% 10.5% 11.8% without a tax sticker  A total of 12.7% of non-domestic packs were identified by the study, which is Share of total of Share consumption 0% directly comparable with the EU Flows Model result of 11.8% TMA 2004 KPMG tax Bulgartabac EPS 2007 EU Flows − of packs identified by Bulgartabac, 2.8% had a foreign tax sticker, while loss 2007 Model 2007 9.5% were from duty free. A further 0.3% did not have any indication of estimate 2006 origin and were assumed to be “illegal” KPMG estimates suggest that non-domestic inflows account for 11.8% of Note: (a) TMA estimate is for smuggled cigarettes share of total consumption total Bulgarian cigarette consumption (b) An additional BGN 50 million in Government revenue is estimated to be lost through incorrect invoicing  The estimate is lower than PMI EPS results due to an adjustment made to Sources: (1) KPMG EU Flows Model (2) World Cigarette Production and Consumption Report – 2005 Edition, Tobacco Duty Free packs found during the survey Merchant Association, 3rd January 2006 (3) “BGN 150 million lost from illegal cigarette trade, PARI Daily, 6th June 2007 − Duty Free packs have been re-weighted to reflect actual PMI Duty Free (4) PMI EPS Q3 2007 (completed in September) sales in border shops both inside and outside Bulgaria (5) “Bulgaria loses BGN 180 M from illegal cigarettes, Trud Daily, 24th September 2007  KPMG’s estimated non-domestic share is consistent with estimated non- domestic share based on Bulgarian tax losses − media reports suggest that the Bulgarian government losses BGN 100 million from non-domestic cigarette sales(3)(b) − using this estimate of losses and tax rates in effect during 2006 and 2007 implies non-domestic incidence of between 7% and 14%  The TMA estimate for 2004 smuggled cigarette share of total consumption is in line with the 2007 EPS result  Non-domestic (legal) research conducted in Bulgaria suggests that there are limited legal inflows, meaning that the smuggled share of total consumption is similar to total non-domestic share of consumption

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Cypriot consumption by origin Non-domestic inflows to 2006-2007(1)(2)(b) Cyprus were stable in 2007

at 0.14 billion sticks 100% 1.5% 1.7% 7.6% 7.4% ND(L) accounted for 82% 80%

of non-domestic inflows, a 60% C&C relatively high share 90.9% 90.9% 40% ND(L) compared with other EU LDC 20% Member States Outflows 0% Share of total of Share consumption The UK remains the main (15.2%) (17.2%) destination country for (20%) 2006 2007 Cypriot outflows

Volume (bn sticks): 1.51 1.55

Total non-domestic inflows (ND(L) and C&C) by origin Key:  Cyprus  Major destination country 2006-2007(1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow (b) ND(L) and EPS research has not been updated for 2007 but corroborating 0.16 research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 0.14 (2) AC Nielsen ND(L) research, 2006 0.12 (1)(2) 0.06 0.06 Proportion of total EU levels 0.10 Other countries 0.6% 0.6% 0.08

0.5% 0.06 Duty Free

0.4% Volume (bn sticks) 0.04 0.08 0.08 0.4% 0.02 0.3% 0.00 0.2% 2006 2007 0.2% 0.2% Percentage of EU total 0.1% 0.0% Volume (bn sticks): 0.14 0.14

0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Although cigarette sales in 1997-2007(1)(2)(a) Cyprus decreased between 2001 and 2005, they have

2.0 2.5 Pound) (Cypriot per 20 sticks Price increased in each of the 1.68 1.62 1.67 1.57 1.52 1.57 1.53 1.53 last two years 1.47 1.49 1.50 2.0 1.5

1.5 1.0 1.0

Volume (bn sticks) 0.5 0.5

0.0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2005 2005-2007  Legal domestic sales 3.5% (2.9)% 5.6%

 Average pack price n/a n/a 0.4%

Note: (a) The decrease between 2001 and 2002 was due to a methodology change rather than a decrease in sales. The new methodology now covers approximately 100% of the market and measures actual shipment data for PMI brands and provides an estimate on competitor volumes Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

An increase in legal domestic sales occurred in 2007 while prices were broadly flat  Legal domestic sales increased by approximately 3.0% in 2007 − the weighted average price per pack increased by 0.6% versus 2006  Cypriot anti-smoking laws currently prevent smoking in all public places, including public transport − smoking is also banned in private cars carrying passengers under 16 years of age  Smoking is still allowed in workplaces but is heavily regulated

Marlboro price comparison Cyprus remains relatively 1 July 2007(1)(2)(a) less expensive versus the €4.30 UK, the main destination €8.16 €4.94 €1.79 for Cypriot cigarettes €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66

€1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Cyprus and key markets(2)(a) €3.49 €3.86

120% 108% Key:  Cyprus  Major destination country 100% 90%

80%

60% July 2006 40% 28% 30% July 2007 20%

(20%)

Price premium/(discount) versus local (22)% (40%) (28)% UK France Greece

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Cypriot consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 1.55 billion Outflows Inflows 0.27 billion sticks 0.14 billion sticks sticks versus legal sales of 1.8 1.67 billion sticks, giving a net outflow of 0.12 billion 1.6 sticks in 2007 0.25 0.02 0.11 1.4 0.02 Outflows C&C ND(L) 1.2

1.0

1.67 0.8 1.55 Volume (bn sticks) 1.40

0.6

0.4

0.2

0.0 LDS UK Other LDC Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin In countries with high 2007(1)(2)(3) 2007(1)(2)(3) levels of inbound tourism, 0.2 UK non-domestic (legal) 12% 0.11 inflows are likely to Greece include some consumption 0.1 5% from tourists home 0.10 Other countries countries Greece 0.1

Volume (bn sticks) UK These inflows reflect the 0.01 small volume of sticks 0.01 0.0 carried by tourists when 2007 Other countries they travel and consumed includes Duty Free Other countries ND(L) share of consumption: 7% 83% at their destination inflows ND(L) share of non-domestic: 82%

Non-domestic (legal) by brand ND(L) analysis 2007(1)(2)(3) Ave. Total 0.2 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.11 Values 0.6m x 64% x 41% x 1.8 x 420 = 0.11bn 0.1 2007 0.05 Non-PMI Other PMI EU rank 0.1 0.02 Marlboro 2007 25 7 2 27 2 22 Volume (bn sticks)

0.04

0.0 2007

PMI share of ND(L): 53% Note ND(L) and EPS research has not been updated for 2007, but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin 2007(1)(2)(3) 2007(1)(2)(3)

Ukraine 0.035 6%

0.030 0.026 0.025

0.020 Other countries Other countries 0.015 includes Duty Free 0.024 Ukraine inflows 0.010 Volume (bn sticks)

0.005

0.000 0.002 2007 C&C share of consumption: 2% Other countries C&C share of non-domestic: 18% 94%

Note ND(L) and EPS research has not been updated for 2007, but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 2003-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that that  The rate of change in the GCTS smoking prevalence and average smoking volume figures provide a useful indication of observed changes in smoking habits. outflows from Cyprus However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and the problem of under-reporting of have become increasingly 1.8 Consumption has decreased consumption 1.6 since 2003 while important relative to  The KPMG consumption index is calculated using the incidence of smoking and 1.4 LDS have inflows since 2003 increased the average daily consumption (both from GCTS) and the population of each 1.2 market (from Euromonitor GMID) 1.0 − these figures are then indexed to the earliest year for which we have both 2007 consumption GCTS and LDS data available, 2003 for Cyprus 0.8 1.5 modelling results suggest 0.6  If LDS trends differ from those shown by the consumption index, it implies that the an increase in net outflows relationships between inflows, outflows and domestic consumption have changed Volume (billion sticks) 0.4 over the analysed period which is consistent with 0.2 Please note that consumption modelling can only identify market trends and EU Flows Model results 0.0 cannot be used to quantify total consumption 2003 2004 2005 2006 2007  Comparing consumption and sales trends helps to understand market direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period − the gap between indexed consumption and LDS in the final year is therefore CAGR (%) 2003-2005 2005-2007 2003-2007 only a broad indication of the extent to which the market has changed over the  LDS 0.2% 5.6% 2.9% analysed period  KPMG consumption index (2.2)% 1.1% (0.6)% − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  KPMG consumption estimate n/a n/a n/a  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI The gap between LDS and the KPMG consumption index for Cyprus increased (3) KPMG EU Flows Model between 2003 and 2007  LDS in Cyprus have increased since 2003, while consumption has remained fairly stable  This result suggests that outflows from Cyprus have become increasingly important relative to inflows i.e. non-domestic consumption has declined and/or outflows have increased − this is consistent with EU Flows Model results showing an increase in net outflows from 0.09 billion sticks in 2006 to 0.12 billion sticks in 2007

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Czech consumption Non-domestic inflows to 2006-2007(1)(2)(b) the Czech Republic were stable in 2007 at 0.21 0.3% 0.3% 100% 0.7% 0.7% billion sticks

75% Outflows from the Czech C&C Republic declined slightly 50% 98.9% 98.9% ND(L) in 2007, with the UK and LDC 25% Germany remaining the Outflows

main destination countries 0% Share of total consumption (24.1)% (21.0)% (25)% 2006 2007

Volume (bn sticks): 19.8 20.0

Total non-domestic inflows (ND(L) and C&C) by origin Key:  Czech Republic  Major destination country 2006-2007(1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow (b) ND(L) and EPS research has not been updated for 2007 but corroborating 0.25 research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2006 0.20 0.01 0.01 0.02 Proportion of total EU levels(1)(2) 0.02 Other countries 0.04 0.04 10% 0.15 8.7% Slovakia

0.06 0.06 Germany 8% 0.10 Ukraine

Volume (bn sticks) Duty Free 6% 0.05 0.08 0.08

4% 0.00 2.8% 2006 2007

Percentage of EU total 2% 0.5% Volume (bn sticks): 0.21 0.21 0.2% 0.1% 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Part of the increase in legal sales in 2007 was driven by trade Legal domestic sales have (1)(2) 1997-2007 loading and stock build-up before increased from 2000 to tax increases took effect in 2008 2006 due in part to an 35 80 30.7 70 increase in cross-border 30 (CZK) per 20 sticks Price 23.9 24.3 60 sales 25 22.9 22.6 20.7 20.4 20.0 19.1 50 20 18.3 18.3 40 15 30 10 Volume (bn sticks) 20

5 10

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2006 2006-2007  Legal domestic sales (7.2)% 4.8% 26.4%  Average pack price 1.0% 1.8% 18.3%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales increased by 26.4% in 2007, while prices increased by 18.3%  Part of the increase in legal domestic sales in 2007 was driven by trade loading and stock build-up before tax increases took effect in January 2008  Legal domestic sales for 2007 have been adjusted downwards for the EU Flows Model to control for the impact of loading − adjusted volumes for 2007 have been calculated by applying the 2007 AC Nielsen growth rate to 2006 IMS volumes  Anti-smoking legislation came into effect in January 2006, banning smoking at bus, train and tram stops, schools, cinemas and theatres, sports halls and administrative buildings Average weighted prices have increased significantly following EU accession in response to tax increases  The Czech Republic is increasing taxes to bring them into compliance with the EU minimum of €64 per 1,000 cigarettes  In January 2007, the excise tax law was amended to establish forestalling regulations and new excise tax rates − there was a three month transition period from January 1st, during which cigarette manufacturers and importers could collect a limited number of fiscal stickers and sell cigarettes with old tax stickers. The regulation also included limits on inventories cigarettes with old tax stickers − these regulations also applied to OTP products  In December 2007, the Czech Republic joined the Schengen zone

Marlboro price comparison Relatively high prices in 1 July 2007(1)(2)(a) the UK and Germany make €4.30 these countries attractive €8.16 €4.94 €1.79 destination markets for €0.86 Czech cigarettes €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Czech Republic and key €3.86 (2)(a) markets Key: : Czech Republic : Major destination country

300% 247% 250% 227% 200%

150% 110% 91% July 2006 100% July 2007 50%

0% (2)%(4)% (50)% (20)%(15)% (45)% (40)% (63)% (65)% Price premium/(discount) versus local (100)% (70)%(74)% UK Latvia Russia Poland Ukraine Slovakia Germany

Notes: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Czech Republic 2007(1)(2)(3)(4)(a)(b) consumption is estimated Outflows Inflows 4.19 billion sticks 0.21 billion sticks at 20 billion sticks versus 30 legal sales of 24 billion sticks, giving a net outflow of 4 billion sticks 25

3.28 Outflows 0.01 C&C 0.29 0.16 20 0.47 0.06 0.15 ND(L)

Volume (bn sticks) 23.97

10 19.78 19.99

0 LDS Germany UK Poland Other LDC Ukraine Other Consumption

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Legal domestic sales for 2007 has been adjusted downwards for the EU Flows Model to control for the impact of loading. The adjusted volumes for 2007 have been calculated by applying the 2007 AC Nielsen growth rate to 2006 IMS volumes (4) Interviews with PMI Local Management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin At 0.7%, non-domestic 2007(1)(2)(3) 2007(1)(2)(3)

(legal) inflows account for 0.20 Other countries 3% a low share of total 0.15 Slovakia consumption compared 0.15 0.00 11% 0.02 with other EU Member Other countries States 0.10 0.04 Slovakia Germany Duty Free Volume (bn sticks) 0.05 Duty Free 0.08 Germany 56% 30% 0.00 2007

ND(L) share of consumption: 0.7% ND(L) share of non-domestic: 69.9%

Non-domestic (legal) by brand ND(L) analysis 2007(1)(2)(3) Ave. Total 0.20 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.15 Values 0.15 x x x x = 0.15bn 2007 8.3m 26% 18% 3.9 98 0.04 Non-PMI 0.10 0.01 Other PMI EU rank L&M 0.04 2007 11 24 20 12 22 19

Volume (bn sticks) 0.05 Marlboro 0.05

0.00 2007

PMI share of ND(L): 70.7%

Note: ND(L) and EPS research has not been updated in 2007, but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin At 0.3%, C&C inflows 2007(1)(2)(3) 2007(1)(2)(3) account for a low share of 0.08 Other countries total consumption 11% 0.063 compared with other EU 0.06 0.007 Member States Other countries The majority of the C&C is 0.04 Ukraine from Ukraine 0.057

Volume (bn sticks) 0.02

0.00 Ukraine 2007 89% C&C share of consumption: 0.3% C&C share of non-domestic: 30.1%

Non-domestic market volume estimates(1)(2)(a) KPMG’s estimate for non-domestic share is lower than the Czech Project Star results Customs estimate of 5% suggest a lower share of  Czech Customs’ estimate is for the level of total counterfeit in the Czech non-domestic product Republic and may therefore include product destined for other countries 6% than Czech Customs’ (i.e. product purchased in the Czech Republic but taken outside the Czech 5.0% Republic either for resale or consumption) estimates 5% − 3.3 billion sticks of Czech product were consumed in Germany in 2007 4%  There are no other known estimates of non-domestic consumption 3%

2% 1.1% 1.1% 1% Share of total consumption

0% Czech Customs EU Flows Model EU Flows Model 2006 2006 2007

Note: (a) EPS research has not been updated for 2007 but corroborating research has not indicated any significant changes Sources: (1) Interview with Czech Customs (2) KPMG EU Flows Model (3) PMI EPS Q3 2006

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 1997-2007(1)(2)(3)(a)(b) The corroborate the extent of non-domestic consumption in a market consumption results suggest that and LDS trend  The rate of change in the GCTS smoking prevalence and average smoking volume lines have figures provide a useful indication of observed changes in smoking habits. outflows from the Czech 30 diverged since 2004 However, GCTS data cannot be used to quantify total consumption in a market Republic have become due to exclusions of particular age cohorts and the problem of under-reporting of 25 consumption increasingly important  The KPMG consumption index is calculated using the incidence of smoking and 20 relative to inflows since the average daily consumption (both from GCTS) and the population of each 2004 market (from Euromonitor GMID) 15 − these figures are then indexed to the earliest year for which we have both 10 20.0 GCTS and LDS data available, 1997 for Czech Republic

Volume (billion sticks)  If LDS trends differ from those shown by the consumption index, it implies that 5 the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market direction and its likely inflow/outflow status. However, it will not identify any consumption CAGR (%) 1997-2004 2004-2006 2006-2007 gap that is present at the start of the period  LDS (0.2)% 3.7% (1.4)% − the gap between indexed consumption and LDS in the final year is therefore  KPMG consumption index (1.5)% (0.4)% (11.3)% only a broad indication of the extent to which the market has changed over the  KPMG consumption estimate n/a n/a n/a analysed period − in many markets, some level of inflows or outflows are likely to be present in Notes: (a) Consumption in 2000 has been estimated as the mid-point between the first analysed year consumption in 1999 and 2001 (b) Legal domestic sales for 2007 include a downward adjustment to control  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will for the impact of loading. also have an impact on the conclusions generated Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI The gap between LDS and the KPMG consumption index for the Czech Republic (3) KPMG EU Flows Model has increased since 2004  Consumption has decreased since 2004 while LDS have risen  This result suggests that outflows from the Czech Republic have become increasingly important relative to inflows i.e. non-domestic consumption has declined and/or outflows have increased − this is consistent with 2007 EU Flows Model results showing relatively large outflows from the Czech Republic (4.2 billion sticks) compared with inflows (0.2 billion sticks) Consumption modelling results show an increase in net outflows in 2007 which is not consistent with EU Flows Model results  However, the large implied consumption decline as measured by GCTS may not be realistic. 2008 CGTS data will be used to help corroborate the scale and direction of the 2007 consumption decline ©2008 KPMG LLP, the UK member firm of KPMG International, a Swiss cooperative. All rights reserved. KPMG and the KPMG logo are registered trademarks of KPMG International, a Swiss cooperative. 119 Czech Republic OTP market size and growth

OTP sales in billion stick equivalents Sales of Other Tobacco Between 2003 and 2006, OTP sales declined at an average rate of 19% 2003-2007(1)(a)(b) a year Products (OTP) decreased 1.8 1.70  By contrast, sales of manufactured cigarettes increased by 6% per year between 2003 and 2005, 0.04 1.6 over the same time period but increased in 2007 1.4 1.29  In 2007 total OTP sales increased by over 135%, while sales of 0.02 1.16 1.2 manufactured cigarettes increased by 26% Part of the increase in OTP 0.02 1.0 0.86 − these increases in sales include the impacts from trade loading and sales in 2007 was driven 0.72 0.8 0.03 1.66 stock build-up before tax increases took effect in January 2008 0.03 Cigars/ cigarollos by trade loading and stock 1.27 0.6 1.13 Smoking tobacco In 2007, smoking tobacco accounted for 98% of total OTP sales build-up before tax 0.83 Stick equivalents (bns) 0.4 0.69 (excluding smokeless tobacco) by volume increases took effect in 0.2 January 2008 0.0 2003 2004 2005 2006 2007

Cigarettes 20.4 22.6 23.9 24.3 30.7 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 11.8% 22.2% 21.2% 16.7%  Smoking tobacco (19.3)% (16.9)% 141.6% 6.9% Total OTP (18.7)% (15.7)% 136.1% 7.1%

Manufactured cigarettes 8.3% 1.6% 26.4% 10.8%

Notes: (a) Smoking tobacco volumes have been calculated at 1 stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick per stick basis (b) Smokeless tobacco has been excluded from this analysis Source: (1) OTP volumes supplied by PMI

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Danish consumption Non-domestic inflows to 2006-2007(1)(2)(b) Denmark were stable in 1.8% 1.9% 100% 2007 at 0.51 billion sticks 4.4% 4.4% 80% The Duty Free channel and 60% Sweden were the two 93.8% 93.8% C&C 40% main sources of non- ND(L) domestic inflows in both 20% LDC 2006 and 2007 Outflows 0% Share of total consumption (3.3)% (2.0)% (20)% 2006 2007

Volume (bn sticks): 8.32 8.28

Total non-domestic inflows (ND(L) and C&C) Key:  Denmark  Major source country 2006-2007(1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both 0.6 source and destination countries are coded according to the larger flow (b) ND(L) and EPS research has not been updated for 2007 but corroborating research has not indicated any significant changes to 2006 results 0.5 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 0.15 0.15 (2) Synovate ND(L) research, 2006 0.4 Other countries 0.03 0.03 Spain (1)(2) 0.04 Proportion of total EU levels 0.3 0.04 1.4% Poland 0.10 0.10 1.2% 1.1% 0.2 Sweden 1.2%

Volume (bn sticks) Duty Free 1.0% 0.1 0.20 0.20

0.8% 0.0 0.6% 2006 2007 0.6% 0.4% 0.4% 0.3% Volume (bn sticks): 0.52 0.51 Percentage of EU total 0.2%

0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales A reduction in excise tax in 1997-2007(1)(2) 2003 contributed to a 0.6 billion stick increase in the 9.0 35 legal market as consumers 8.0 8.1 8.1 7.9 8.0 7.4 7.0 7.1 7.1 7.2 30 (DKK) per 20 sticks Price 6.7 6.8 switched away from non- 7.0 25 domestic cigarettes 6.0 5.0 20

4.0 15 3.0 10 Volume (bn sticks) 2.0 5 1.0 0.0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2004 2004-2007  Legal domestic sales 1.6% 4.2% (0.5)%  Average pack price n/a n/a 1.5%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales decreased by 1.8% in 2007 while prices increased by 1.6%  Denmark introduced a smoking ban in August 2007 covering bars and restaurants. However, some exceptions are permitted including separate smoking rooms and bars smaller than 40 square metres

Marlboro price comparison The cigarette price 1 July 2007(1)(2)(a) differential versus Sweden €4.30 and Germany widened in €8.16 €4.94 €1.79 2007 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66

€4.10 €1.64

€3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Denmark and key markets(2)(a) €3.86

Key:  Denmark 40%  Major source country 17% 20% 11% 4% 7% 0% July 2006 July 2007 (20)%

(26)% (29)% (40)% (30)% (30)% (33)% (34)% (60)% (50)% (59)% (65)%

Price premium/(discount) versus local (80)% (72)% Sweden Germany Portugal Greece Spain Poland Latvia

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Danish consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 8.3 billion Outflows Inflows 0.17 billion sticks 0.51 billion sticks with a net outflow of 0.4 9 billion sticks in 2007 0.12 8 0.25 0.07 0.03 0.03 0.07 0.10 0.01

7 Outflows C&C ND(L) 6

8.3 4 7.9 7.8 Volume (bn sticks)

0 LDS Sweden Spain Other LDC Sweden Poland Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows accounted 2007(1)(2)(3)(a) 2007(1)(2)(3)(a) for 70% of total non- 0.4 0.36 domestic consumption in Sweden 2007 28% 0.3 0.22 0.2 Other countries Greece 0.04 Volume (bn sticks) 0.1 Sweden Other countries 61% Greece 0.10 11% 0.0 2007

ND(L) share of consumption: 4% ND(L) share of non-domestic: 70%

Non-domestic (legal) by brand ND(L) analysis(b) 2007(1)(2)(3)(a) Ave. Total 0.4 0.36 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.09 0.3 Values x x x x = 0.32bn(b) 2007 4.2m 68% 20% 4.1 138

0.2 Other non-PMI 0.22 EU rank Prince 2007 18 5 18 10 16 15 Volume (bn sticks) 0.1 Other PMI 0.00 Marlboro 0.05 0.0 2007

PMI share of ND(L): 15% Note: (a) ND(L) and EPS research has not been updated for 2007, but corroborating research has not indicated any significant changes to 2006 results (b) ND(L) analysis from Synovate excludes adjustments for inbound tourism inflows. Total ND(L) including these inbound tourism inflows in 0.36 billion sticks Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI local management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows account for a 2007(1)(2)(3)(a) 2007(1)(2)(3)(a) relatively low share of 0.15 0.16 total consumption Poland 21% compared with other EU 0.12 Member States ADD PMI 0.10 CF TO Other countries Poland is the largest 0.08 Latvia OTHER Portugal source of C&C inflows for Portugal 8% 0.01

Denmark Volume (bn sticks) 0.04 0.01 Poland Latvia 0.03 5% 0.00 2007 Other countries 66% C&C share of consumption: 2% C&C share of non-domestic: 30%

Note: (a) ND(L) and EPS research has not been updated for 2007, but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interview with PMI local management

Non-domestic market volume estimates(1)(2)(3)(a) Stable non-domestic share of consumption is reflected in Danish KPMG’s estimate for non- Customs’ estimate of inflow volumes in 2007 domestic share in 2006 8%  Danish Customs’ estimate is based on a 500 million stick inflow to the and 2007 is consistent 7% country 6.0% 6.2% 6.0% 6.2% with Danish Customs’ 6% − this estimate has remained unchanged since 2005 estimates for the period 5%  No other estimates of non-domestic share are currently available 4%

3% 2%

Share of total of Share consumption 1% 0% Customs 2006 EU Flows Customs 2007 EU Flows Model 2006 Model 2007

Note: (a) EPS research was not updated in 2007 but corroborating research has not indicated any significant changes Sources: (1) “Status over Graensehandel”, Ministry of Finance, June 2007 (2) KPMG EU Flows Model (3) PMI EPS Q3 2006

OTP sales in billion stick equivalents Total OTP volumes declined at an annual rate 13.3% per year between Sales of Other Tobacco 2003-2007(1)(a)(b) 2003 and 2007 Products (OTP) have 2.5  By contrast, legal domestic sales of manufactured cigarettes increased at declined since 2003 2.22 an annual rate of 1.7% over the same time period 0.12 2.0 1.82 Smoking tobacco accounted for almost 95% of OTP sales in 2007 0.11 1.59 1.44 1.5 0.10 0.09 1.25 0.07 1.0 2.10 1.71 1.49 1.34 1.18 Stick equivalents (bns) 0.5

0.0 2003 2004 2005 2006 2007

Cigarettes 7.4 8.0 8.1 8.1 7.9 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos (10.0)% (5.4)% (24.0)% (12.6)%  Smoking tobacco (15.8)% (10.0)% (11.9)% (13.4)% Total OTP (15.4)% (9.7)% (12.7)% (13.3)%

Manufactured cigarettes 4.5% (0.3)% (1.8)% 1.7%

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Estonian consumption Non-domestic inflows to 2006-2007(1)(2)(b) Estonia fell to 0.29 billion sticks in 2007 100% 10.7% 19.2% 2.3% Inflows from Latvia and 80% 6.4%

Lithuania were 60% C&C substantially down on 86.9% ND(L) 40% 74.4% 2006 levels LDC 20% Outflows The level of outflows 0% (30.0%) (23.3%) decreased in 2007, driven total of Share consumption (20%)

primarily by reduced flows (40%) to Finland 2006 2007

Volume (bn sticks): 2.16 2.23

Total non-domestic inflows (ND(L) and C&C) by origin Key:  Estonia  Major source country  Major destination country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 0.6 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2007 0.5 0.10

0.06 Proportion of total EU levels(1)(2) 0.4 Other countries Lithuania 1.2% 0.3 0.18 1.1% 0.03 0.01 0.02 Latvia 1.0% 0.2 Russia Volume (bn sticks) 0.23 0.8% 0.1 0.22

0.6% 0.0 0.4% 2006 2007 0.4% 0.3% 0.3% Volume (bn sticks): 0.55 0.29 0.2% Percentage of EU total 0.2% Note: (b) 2006 legal domestic share of total consumption differs slightly from what was shown in the 2006 Project Star Results presentation due to a technical change in approach to the 0.0% calculation of consumption share (2006 LDC was shown as 75% in the 2006 Results Consum- ND ND(L) C&C Outflows presentation compared to 74.4% in this presentation) ption

Historic cigarette prices and legal domestic sales(1)(2) Legal cigarette sales have 2000-2007 generally been trending upwards since 2000 despite rising prices 3.0 30 2.46

2.35 (EEK) per 20 sticks Price 2.5 2.25 2.25 2.26 25 2.10 2.09 1.96 2.0 20

1.5 15

1.0 10 Volume (bn sticks) 0.5 5

0.0 0 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 2000-2004 2004-2007  Legal domestic sales 3.5% 3.0%

 Average pack price n/a 5.5%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales increased by 9.4% in 2007 despite an increase in the average weighted price of 1.5%  Estonia is raising taxes into line with EU excise tax regulations, with prices rising at an annual average of 5.5% between 2004 and 2007 − although no tax changes occurred during 2007, relatively large increases are expected during 2008  Smoking bans in enclosed public places and workplaces were introduced in 2005 − wider-ranging restrictions, including a ban on smoking in the hospitality sector, came into force in June 2007

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Estonia and key source €4.30 and destination markets €8.16 €4.94 €1.79 remained relatively stable €0.86 in 2007 €1.47 €4.24 €1.33 €0.86 Key exceptions are Poland €7.05 n/a €4.71 which became more €8.04 €4.11 €2.10 expensive, and Latvia €4.53 €3.70 €2.46 €0.63 where the price gap €2.35 €3.90 €2.39 narrowed €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Estonia and key (2)(a) markets €3.86

Key:  Estonia  Major source country  Major destination country 200% 143%140% 150%

100%

50% 17% July 2006 July 2007 0% (1)% (18)% (50%) (26)% (32)% (52)% (27)% (55)% Price premium/(discount) versus local (100%) Finland Poland Latvia Lithuania Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Estonian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 2.3 billion Outflows Inflows 0.52 billion sticks 0.29 billion sticks sticks versus legal sales of 3.0 2.5 billion sticks, giving a net outflow of 0.2 billion sticks in 2007 2.5 Outflows 0.25 C&C 0.02 ND(L) 0.04 0.11 0.04 0.22 2.0 0.12 0.01

1.5

2.46 Volume (bn sticks) 2.23

1.0 1.94

0.5

0.0 LDS Finland Latvia Poland Other LDC Russia Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows to Estonia 2006-2007(1)(2)(3) 2007(1)(2)(3) decreased to 50 million 0.16 sticks in 2007 driven by 0.14 falling inflows from other 0.02 0.12 Other countries Baltic markets 0.01 Russia Other countries 36% 0.04 Latvia Latvia 0.08 0.05 Lithuania 43%

Volume (bn sticks) 0.04 0.02 0.06 0.01 0.02 0.00 0.00 2006 2007

ND(L) share of Russia 6% 2% consumption: 21% ND(L) share of non- 25% 18% domestic:

Non-domestic (legal) by brand 2006-2007(1)(2)(3) ND(L) analysis Ave. Total 0.16 0.14 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.12 Values x x x x = 0.05bn 0.06 Non-PMI 2007 1.0m 45% 20% 3.6 156 Other PMI 0.08 Marlboro 0.05 EU rank 0.05 2007 24 13 16 14 14 25 Volume (bn sticks) 0.04 0.02 0.02 0.02 0.01 0.00 2006 2007

PMI share of ND(L): 54% 59%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Estonia 2006-2007(1)(2)(3) 2007(1)(2)(3)

decreased by 42% 0.5 Other countries 7% between 2006 and 2007 0.42 0.4 0.08

0.3 0.14 0.24

0.2 Other countries 0.02 Latvia Volume (bn sticks) 0.1 0.20 0.22 Russia Russia 93%

0.0 2006 2007

C&C share of 19% 11% consumption: C&C share of non- 75% 82% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4) EU Flows Model results for 2006 are broadly consistent with alternative Estonian Customs’ estimates given timing differences estimate of non-domestic 30%  EU Flows Model result for 2006 was based on the PMI Empty Pack consumption was made in 25.7% Survey which was conducted in Q2 2006 25% 22.5% November 2006 and is 21.0% − Estonian Customs’ estimate for 2006 was based on full year data and consistent with the decline 20% would therefore reflect a decline in non-domestic incidence during the second half of the year in incidence over the 15% 13.1% − EIER estimates were published in April 2006 and are based on illicit period shown by the EU 10% inflows only, therefore excluding non-domestic (legal) inflows Flows Model 5%  No alternative 2007 estimates are currently available to corroborate our Share of total of Share consumption findings 0%

EU Flows EIER C&C Customs 2006 EU Flows Model 2006 2006 Model 2007

Sources: (1) Interview with Estonian Customs (2) KPMG EU Flows Model (3) “Consumption of and trade in illegal tobacco products in Estonia”, Estonian Institute of Economic Research, April 2006 (4) PMI EPS Q2 2006 and Q3 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 2000-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that that  The rate of change in the GCTS smoking prevalence and average smoking outflows from Estonia volume figures provide a useful indication of observed changes in have become increasingly 3.0 smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and important relative to LDS increased 2.5 relatively fast in the problem of under-reporting of consumption inflows since 2005 2007 while 2.0 consumption  The KPMG consumption index is calculated using the incidence of declined smoking and the average daily consumption (both from GCTS) and the 2007 consumption 1.5 population of each market (from Euromonitor GMID) modelling results suggest 2.2 1.0 − these figures are then indexed to the earliest year for which we have an increase in net outflows Volume (bn sticks) both GCTS and LDS data available, 2000 for Estonia 0.5 which is consistent with  If LDS trends differ from those shown by the consumption index, it EU Flows Model results 0.0 implies that the relationships between inflows, outflows and domestic 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 2000-2006 2006-2007  Comparing consumption and sales trends helps to understand market  LDS 2.4% 9.1% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index 1.0% (1.8)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Estonia was relatively stable between 2000 and 2006  This suggests that net outflows were relatively stable during this time frame Legal sales increased in 2007 while consumption declined suggesting an increase in net outflows  This is consistent with EU Flows Model results showing an increase in net outflows to 0.23 billion sticks in 2007, up from 0.09 billion sticks the previous year

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Finnish consumption Non-domestic inflows to 2006-2007(1)(2) Finland rose to 1.5 billion 100% 10.1% sticks in 2007, driven 16.9% 11.5% primarily by an increase in 80% 6.4%

consumption of Russian 60% product C&C 40% 78.4% 76.7% ND(L) 20% LDC Outflows 0% Share of total consumption (1.0)% (0.6)% (20)% 2006 2007

Volume (bn sticks): 6.4 6.4

Key:  Finland  Major source country Total non-domestic inflows (ND(L) and C&C) Note: (a) Map shows flows over 1% of consumption. Countries which are both 2006-2007(1)(2) source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 1.6 0.07 1.4 0.14 0.25 1.2 Proportion of total EU levels(1)(2) 1.0 0.43 0.29 2.0% 1.8% 0.8 1.6% 0.19 Other countries 1.6% 0.6 1.4% Estonia 0.89 Volume (bn sticks) 0.4 1.2% 0.62 Duty Free 0.2 0.9% Russia 0.8% 0.0 2006 2007

Percentage of EU total 0.4% 0.1% Volume (bn sticks): 1.4 1.5 0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales(1)(2) Legal domestic sales of 1997-2007 cigarettes increased between 2003 and 2006, but declined in 2007 6 5 4.93 4.93 5.08 5.05 4.95 4.80 4.80 4.84 4.80 (Euro) per 20 sticks Price 5 4.71 4.65 4

4 3 3 2 2 Volume (bn sticks) 1 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2005 2005-2007  Legal domestic sales 0.6% 1.2% (1.2)%  Average pack price n/a n/a 2.5%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Legal domestic sales declined by 2.0% in 2007 while prices have increased by 2.1%  In June 2007, anti-smoking legislation was introduced in Finland banning smoking in bars and restaurants. Some exceptions are permitted such as smoking rooms with a separate ventilation system − there is a two year transition period for bars and restaurants to install such areas

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Finland and key source markets were fairly stable €4.30 in 2007 versus 2006 €8.16

€4.94 €1.79 €0.86

€1.47 €4.24 €1.33 €0.86

n/a €4.71 €2.10

€3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €1.66

Evolution of Marlboro price differential between Finland and key Key:  Finland markets(2)(a)  Major source country

20% 15% 3% 0%

(20)%

(40)% July 2006 July 2007 (60)% (59)% (58)% (66)% (80)% (72)% (81)% (80)%

Price premium/(discount) versus local (100)% Sweden Estonia Latvia Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Finnish consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 6.4 billion Outflows Inflows 0.04 billion sticks 1. 49 billion sticks sticks versus legal sales of 7 5.0 billion sticks, giving a net inflow of 1.4 billion 0.15 0.18 sticks in 2007 0.02 6 0.10 0.15 0.00

0.81 Outflows C&C 5 0.02 0.02 0.08 ND(L)

6.4 3 Volume (bn sticks)

5.0 4.9

0 LDS Sweden Other LDC Russia Estonia Latvia Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows accounted 2006-2007(1)(2)(3) 2007(1)(2)(3) for 0.4 billion sticks of non- 0.80 0.73 domestic consumption in Other countries 25% 2007 0.17 0.60 Estonia 35% PMI brands accounted for 0.41 0.40 0.29 over 90% of ND(L) inflows 0.10 Other countries 0.08 in 2007 Russia

Volume (bn sticks) 0.12 0.20 0.09 Sweden

0.15 0.15 Estonia Russia 0.00 19% 2006 2007 Sweden ND(L) share of 21% 12% 6% consumption: ND(L) share of non- 53% 27% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.80 0.73 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.18 Values 0.60 4.1m x 57% x x x = 0.41bn 0.03 Non-PMI 2007 25% 4.8 148

0.16 0.41 Other PMI 0.03 0.40 Marlboro 0.03 EU rank L&M 0.15 2007 19 10 11 5 15 14 Volume (bn sticks) 0.20 0.36 0.20

0.00 2006 2007 PMI share of ND(L): 76% 91%

Note: The methodology for the 2007 ND(L) research differs to that used in 2006 due to the fact that Finland was included in the Pilot Study. ND(L) research for 2007 uses a larger sample size compared with 2006 Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI local management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows accounted for 2006-2007(1)(2)(3) 2007(1)(2)(3) 1.1 billion sticks of non- 1.20 1.08 Other countries domestic consumption in 16% 2007 1.00 0.17 0.64 0.10 0.80 Russia is the main source Estonia Other countries 0.04 9% of C&C inflows to Finland 0.60 Estonia 0.28 0.40 0.81 Russia Volume (bn sticks) 0.20 0.33

0.00 Russia 2006 2007 75% C&C share of 10% 17% consumption: C&C share of non- 47% 73% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI local management

Non-domestic market volume estimates(1)(2)(a) KPMG’s estimate for non-domestic share of total consumption is lower PMI EPS results for 2007 than PMI EPS results to control for a change in timing of the PMI EPS have been adjusted for 30%  PMI EPS results for 2007 show a 4 percentage point increase compared Project Star to reflect the 25.6% with 2006 which appears high given feedback from market participants 25% 23.3% impact of seasonal 21.6% and other corroborating sources fluctuations in non- 20%  Some of this apparent increase is likely to be driven by differences in timing of the 2007 survey compared with 2006 domestic incidence 15% − the 2007 empty pack survey was conducted in June and may 10% therefore capture greater numbers of people returning from holiday and more inbound tourist arrivals 5% Share of total consumption  KPMG has therefore used a lower estimate for non-domestic 0% consumption than suggested by PMI EPS results EU Flows Model PMI EPS 2007 EU Flows Model − the use of adjusted EPS results helps give a more realistic 2006 2007 consumption trend in Finland (unadjusted EPS results imply a 4% increase in overall consumption) − KPMG’s non-domestic estimate for 2007 still assumes that the Sources: (1) KPMG EU Flows Model absolute levels of inflows and non-domestic incidence have increased (2) PMI EPS research, Q3 2006 and Q2 2007 compared with 2006 No other estimates of non-domestic share are currently available

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 1998-2006(1)(2)(3)(a) corroborate the extent of non-domestic consumption in a market results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in remained fairly stable 7 smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and 6 since 1998 the problem of under-reporting of consumption 5  The KPMG consumption index is calculated using the incidence of 4 smoking and the average daily consumption (both from GCTS) and the

3 6.4 population of each market (from Euromonitor GMID) − these figures are then indexed to the earliest year for which we have 2 Volume (bn sticks) both GCTS and LDS data available, 1998 for Finland 1  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 1998 1999 2000 2001 2002 2003 2004 2005 2006 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1998-2002 2002-2006 1998-2006  Comparing consumption and sales trends helps to understand market  LDS 0.6% 0.6% 0.6% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.5)% 4.0% 1.2% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Note: (a) Consumption index modelling has only been completed up until 2006 due to a change in GCTS methodology in 2007. As a result of the change, GCTS changed over the analysed period consumption data for 2007 is not comparable to previous years Sources: (1) KPMG Consumption Index model − in many markets, some level of inflows or outflows are likely to be (2) In Market Sales supplied by PMI present in the first analysed year (3) KPMG EU Flows Model  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Finland remained reasonably constant between 1998 and 2006  Consumption and LDS in Finland have increased at similar rates since 1998  This result suggests that non-domestic consumption in Finland has been stable since 1998

OTP sales in billion stick equivalents Legal sales of OTP have declined since 2003 while sales of manufactured OTP categories account for 2003-2007(1)(a)(b) cigarettes have increased over the same period a relatively large share of 1.5  From 2003 to 2007, sales of OTP declined at an average rate of 4.2% per overall tobacco 1.33 1.26 1.22 year while sales of manufactured cigarettes increased at an average of 0.09 1.15 0.09 1.12 0.8% per year consumption in Finland 0.09 0.10 0.10 1.0 OTP categories account for a relatively large share of overall tobacco consumption in Finland  Total cigarette consumption is estimated at 6.4 billion sticks(2) versus 1.1 1.24 1.17 1.12 0.5 1.06 1.02 billion sticks for OTP (excluding smokeless tobacco)

Stick equivalents (bns)

0.0 2003 2004 2005 2006 2007

Cigarettes 4.8 4.9 5.1 5.1 5.0 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 4.0% 2.8% 2.0% 3.2%  Smoking tobacco (4.8)% (5.9)% (3.6)% (4.8)% Total OTP (4.2)% (5.2)% (3.1)% (4.2)%

Manufactured cigarettes 2.9% (0.5)% (2.0)% 0.8%

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total French consumption Non-domestic inflows to 2006-2007(1)(2) France declined to 15.3 billion sticks in 2007 100% 11.9% 14.6% Inflows from Romania and 80% 10.8% 7.7% Belgium were higher in 60% 2007 resulting in both C&C ND(L) countries accounting for 40% 77.3% 77.8% LDC greater than 1% of French 20% Outflows domestic consumption

Share of total of Share consumption 0 (1.3)% (1.9)% (20)% 2006 2007

Volume (bn sticks): 71.0 69.0

Total non-domestic inflows (ND(L) and C&C) (1)(2) Key:  France  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 20 (2) AC Nielsen ND(L) research, 2007

15 Other countries Proportion of total EU levels(1)(2) Italy 20% 7.3 5.7 17.4% Belgium 16.9% 16.6% 10 0.5 0.7 0.7 Luxembourg 0.5 0.9 1.0 15% 0.3 1.3 Romania

Volume (bn sticks) 5 2.8 2.9 Duty Free 9.6% 10% 3.4 3.3 Spain 0 2006 2007 5%

Percentage of EU total 2.7% Volume (bn sticks): 16.1 15.3 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales 1997-2007(1)(2) declined between 2002 and 2004, but have been 90 83.8 83.7 83.5 6 relatively stable since that 83.0 82.5 80.6 80 year 69.6 5 (Euro) per 20 sticks Price 70 60 54.9 54.8 55.8 54.9 4 50 3 40 30 2 Volume (bn sticks) 20 1 10 0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2004 2004-2007  Legal domestic sales 0.1% (13.0)% 0.0%  Average pack price 4.8% 15.6% 1.3%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI based on Altadis Distribution (3) “Cost of cigarettes increase in France”, Les Echos, 1 October 2007

Legal domestic sales decreased by 1.5% in 2007, whilst average weighted prices increased by 2.4%  The minimum sales price for a packet of 20 cigarettes was raised in August 2007 to 4.79 euros from 4.49 euros(3)  Anti-smoking regulations were tightened in 2007 − in February 2007, smoking was banned in workplaces, schools, shopping centres, public administration buildings, stations, airports, hospitals, sports facilities and entertainment venues − cafés and restaurants are exempt from this ban until 1st January 2008

Marlboro price comparison The price differential 1 July 2007(1)(2)(a) between France and key €4.30 source markets was fairly €8.16 €4.94 €1.79 stable in 2007 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.30(a) €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between France and key markets(2)(a)(b) €3.86 Key:  France 80%  Major source country 52%  Major destination country 48% 40%

0% July 2006

(15)% July 2007 (22)% (23)% (20)% (20)% (20)% (40)% (30)% (30)% (41)% (42)% (44)% (45)%

(80)% (75)% (69)% UK Price premium/(discount) versus local Italy Spain Portugal Belgium Romania Netherlands Luxembourg

Notes: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 except for France 2007. France 2007 price is for August rather than July and as a result includes the 30 cents price increase that occurred that month (equivalent to 6.7%) Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown French consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 69 billion Outflows Inflows 1.3 billion sticks 15.3 billion sticks sticks versus legal sales of 70 55 billion sticks, giving a 6.78 net inflow of 14 billion 2.32 60 sticks in 2007 0.55 0.73 1.30 1.41 0.40 0.38 1.84 0.00 0.35 0.31 0.28 Outflows 50 C&C ND(L)

68.95

Volume (bn sticks) 54.95 53.63

0 LDS UK Netherlands Spain Other LDC Spain Romania Luxembourg Belgium Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows decreased by 2006-2007(1)(2)(3) 2007(1)(2)(3) 7.7 31% in 2007 to 5.3 billion 8 sticks Spain 6 5.3 The main driver of this 4.1 35% Other countries decline was a reduction in Other countries 42% 4 2.2 Italy the propensity to travel of 0.7 Belgium 0.5 French consumers 0.5 Spain Volume (bn sticks) 2 0.7 2.3 1.8 0 2006 2007 Belgium Italy 14% 9% ND(L) share of 11% 8% consumption: ND(L) share of non- 48% 34% domestic:

Non-domestic (legal) by brand (a) 2006-2007(1)(2)(3) ND(L) analysis Ave. Total 7.7 8 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) Values 6 5.3 48.7m x 34% x 26% x 2.9 x 376 = 4.7bn(a) 4.3 2007 Non-PMI

4 Other PMI 3.1 Marlboro EU rank 0.5 2007 2 18 10 17 4 2

Volume (bn sticks) 2 0.2 2.9 2.0

0 2006 2007 PMI share of ND(L): 44% 42% Note: (a) ND(L) analysis from AC Nielsen excludes adjustments for inbound tourism inflows. Total ND(L) including these inbound tourism inflows in 5.29 billion sticks Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

A key driver of the decline in ND(L) inflows in 2007 was the ND(L) analysis comparison(1) A key driver of the reduction in propensity to travel reduction in non-domestic Ave. Total Population Propensity Propensity Ave. no. purchases ND(L)  The reduction in propensity to travel implies that a smaller (legal) volumes is the 19+ to travel to purchase of trips (sticks) (sticks) percentage of the French population visited foreign countries in decline in foreign travel by 2007 French nationals Values  The decline in outbound tourism as measured by the ND(L) x x x x = 4.7bn 2007 48.7m 34% 26% 2.9 376 consumer research has been corroborated by looking at outbound This decline is tourism data from Eurostat corroborated by available − Eurostat data suggests that French outbound tourist levels were over 13% lower (in absolute terms) in the first half of outbound tourism data Values 2007 compared with 2006 from Eurostat 2006 46.7m x 44% x 32% x 4.4 x 261 = 7.7bn − in percentage terms, the decline is slightly more marked and implies an 18% reduction in the propensity to travel given population growth French nationals made fewer trips in 2007 but purchased a higher number of cigarettes per trip when they did travel  These two changes have opposing impacts on ND(L) volumes and broadly cancel each other out French outbound tourists  This trend may reflect a behavioural change towards less frequent 2005-2007(2) shopping trips as consumers move to bulk purchases in response to rising transport costs

4 3.6 3.4 3.1 3

1 Outbound tourists (millions)

0 H1 2005 H1 2006 H1 2007

Sources: (1) AC Nielsen ND(L) research, 2006 and 2007 (2) Eurostat, retrieved 25th April 2008

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows increased by 2006-2007(1)(2)(3) 2007(1)(2)(3) 19% in 2007 to 10 billion 12 Spain sticks 10.0 14% 10 8.5 Other countries 8 Portugal Romania 6.4 13% 6 Luxembourg 6.3 Romania 4 0.4 Luxembourg Volume (bn sticks) 0.5 Spain 0.3 1.3 Other countries 5% 2 0.5 0.3 65% 1.1 1.4 Portugal 0 3% 2006 2007 Other countries includes Duty Free C&C share of inflows 12% 15% consumption: C&C share of non- 52% 66% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(a) BAT results since 2005 have consistently fallen in the range from 22% to KPMG’s non-domestic 26% of consumption estimate for 2007 is in line  BAT’s non-domestic estimate is based on a packet swap methodology with 2006 results, but 30% rather than an Empty Pack Survey which limits the comparability of the slightly below the 26.0% results 24.0% 25% 22.7% corresponding BAT 22.0% 22.2% PMI France’s non-domestic consumption estimates for 2005 and 2006 18.6% are lower than BAT and KPMG figures, but in line with a recent estimate 20% 16.5% 16.0% academic paper by C Ben Lakdhar 15%  PMI France 2005 and 2006 estimates are for cross border purchases only

10% and therefore may exclude specific flows from non-adjacent countries and Duty Free inflows 5% Share of total of Share consumption  C Ben Lakdhar’s estimates for 2006 range from 15.5% to 17.5% and are (a) 0% based on two different methodologies − analysing legal domestic sales of manufactured cigarettes across departments in order to measure cross border sales and tobacco 2005 2006 C Ben BAT 2007 BAT 2005 BAT 2006 smuggling EU Flows EU Flows PMI France PMI France Model 2006 Model 2007 Lakhdar 2006 − an empty pack survey approach collecting packs from waste collection centres Note: (a) The estimates by C Ben Lakdhar were given in tonnes of tobacco and have been converted into sticks using the 0.8g conversion factor provided in the research paper. The share of total consumption has been estimated using Project Star consumption figures. Sources: (1) “Un paquet de blondes sur cinq selon BAT/EPSY”, Revue de tabacs, March 2007 (2) Consumer purchasing habits, PMI France (3) C Ben Lakdhar, “Quantitative and qualitative estimates of cross-border tobacco shopping and tobacco smuggling in France”, Tobacco Control, 2008 (4) KPMG EU Flows Model (5) PMI EPS Q3 2006 and Q4 2007 (6) “Le chiffre”, Aujourd’hui, November 2007

results suggest non- The consumption and LDS trend lines  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption diverged in 2001 volume figures provide a useful indication of observed changes in 100 and have remained decreased in 2007, which apart since smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and is consistent with EU 80 the problem of under-reporting of consumption Flows Model results  The KPMG consumption index is calculated using the incidence of 60 smoking and the average daily consumption (both from GCTS) and the population of each market (from Euromonitor GMID) 40 69.0 − these figures are then indexed to the earliest year for which we have

Volume (billion sticks) 20 both GCTS and LDS data available, 1997 for France  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1997-2001 2001-2004 2004-2007  Comparing consumption and sales trends helps to understand market  LDS 0.1% (13.0)% 0.0% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index 2.8% (9.7)% (1.7)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated Consumption index modelling results suggest that the consumption and legal sales trend lines diverged in 2001 and have remained apart since  This suggests that non-domestic consumption in the years between 2001 and 2007 was higher relative to the years between1997 to 2000  Consumption appears to have declined faster than legal sales in 2007 suggesting that non-domestic consumption has decreased − this is consistent with EU Flows Model results showing a decline in net inflows from 15.2 billion sticks in 2006 to 14.0 billion sticks in 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1997-2007(1)(2)(3)(a) prices affect cigarette consumption in France that changes in  Historical data from 1997 to 2006 has been used to estimate the strength consumption in France are of relationships between consumption and prices and income 120 related to changes in Control period Forecast − estimated incidence of smoking and average daily consumption are disposable income and 100 based on GCTS data. Population and disposable income are estimated using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices 80 Cigarette prices are based on the weighted average pack price supplied by PMI Estimated consumption 60 KPMG analysis suggests that changes in consumption in France are results show a decrease in 40 related to changes in disposable income and cigarette prices(c)

non-domestic Volume (billion sticks) 20  Disposable income is positively related to changes in consumption i.e. as consumption in 2007 income rises consumption tends to increase 0 which is consistent with  Cigarette prices are, by contrast, negatively related to changes in 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 EU Flows Model results consumption i.e. as prices rises consumption tends to decrease  It is important to note that factors other than price and income are likely to CAGR (%) 1997-2001 2001-2006 2006-2007 impact consumption levels in France  LDS (0.3)% (8.4)% (2.1)% − examples of other factors likely to impact consumption levels include  Estimated consumption(b) 2.5% (4.5)% (4.4)% changes in smoking restrictions and changes in demographic profiles  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Relationships based on historical data from 1997 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 considered a broad approximation Project Star estimate of consumption for France (71.0 billion sticks) (c) 81% of changes in consumption can be explained by changes in disposable The relationships between changes in consumption and changes in income and cigarette prices disposable income and prices have been used to estimate 2007 Sources: (1) KPMG EU Flows Model consumption (2) In Market Sales supplied by PMI (3) KPMG Consumption Regression Model  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is consistent with EU Flows Model and consumption index modelling results  The decline in estimated consumption in 2007 is higher than the decline in legal sales suggesting that non-domestic consumption has decreased − this is consistent with EU Flows Model results

OTP sales in billion stick equivalents OTP categories account for OTP sales increased between 2003 and 2004 but have declined over the 2003 to 2007(1)(a)(b) last three years a relatively large share of 14  Total OTP sales declined at an annual rate of 1.0% between 2004 and overall tobacco 12.4 12.2 12.2 12.0 2007 while manufactured cigarettes sales were flat over the same time 12 11.1 consumption in France at 1.9 1.8 1.8 1.7 period 12 billion stick equivalents 10 1.7  Smoking tobacco has had the highest volume growth among the major in 2007 8 OTP categories since 2003 and now accounts for 86% of total OTP sales (excluding smokeless tobacco) 6 Cigars/ cigarellos 10.5 10.4 10.4 10.3 9.3 OTP categories account for a relatively large share of overall tobacco 4 Smoking tobacco consumption in France Stick equivalents (bns) 2  Total cigarette consumption is estimated at 69.0 billion sticks(2) versus 0 12.0 billion sticks for OTP (excluding smokeless tobacco) 2003 2004 2005 2006 2007

Cigarettes 69.6 54.9 54.8 55.8 54.9 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 3.4% (1.6)% (7.1)% (0.6)%  Smoking tobacco 5.5% 0.2% (0.8)% 2.5% Total OTP 5.1% (0.1)% (1.7)% 2.1%

Manufactured cigarettes (11.3)% 1.8% (1.5)% (5.8)%

Notes: (a) Smoking tobacco volumes have been calculated at 1 stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick per stick basis (b) Smoking tobacco has been excluded from this analysis Sources: (1) OTP volumes supplied by PMI (2) Project Star 2007 results

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total German consumption Non-domestic inflows to 2006-2007(1)(2) Germany fell to 19.3 billion sticks in 2007 100% 10.6% 9.7% 9.3% 8.3% Inflows from Ukraine fell 80% by 1.8 billion sticks in C&C 60% 2007, while inflows from ND(L) LDC 40% 80.1% 82.0% Russia also declined Outflows 20%

0% Share of total consumption (1.6%) (1.2%) (20%) 2006 2007

Volume (bn sticks): 114.0 107.3

Total non-domestic inflows (ND(L) and C&C) by origin 2006-2007(1)(2)(b) Key:  Germany  Major source country Note: (a) Map shows flows over 1% of consumption. Countries which are both 25.0 source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 20.0 5.0 Other countries (1)(2) 1.6 5.0 Proportion of total EU levels 15.0 1.7 Duty Free 60% 0.6 4.1 1.1 Russia 2.3 50% 10.0 Ukraine 3.8 3.3 Czech Republic Volume (bn sticks) 40% 5.0 Poland 6.9 29.2% 6.5 30% 0.0 21.2% 17.2% 2006 2007 20% 14.9%

Percentage of EU total 10% Volume (bn sticks): 22.7 19.3 2.7% Note: (b) In 2007, we have analysed both the Yellow Bag Survey results and PMI’s own Empty Pack 0% Survey (EPS) study. Total non-domestic inflows and their breakdown by origin and brand have Consum- ND ND(L) C&C Outflows been based on the Yellow Bag survey. However, the EPS study reported a higher level of PMI ption counterfeit incidence than found in the Yellow Bag Survey. In order to ensure consistency of approach for EU wide counterfeit estimates, the EPS result has been used to calculate counterfeit levels in Germany

Historic cigarette prices and legal domestic sales Legal domestic sales have 1997-2007(1)(2) been trending downwards since 2002 as prices have 160 5 141.2 140.8 141.1 143.5 risen 137.0 138.5 134.4

140 (Euro) per 20 sticks Price 113.6 4 120 96.9 93.2 89.3 100 3 80 60 2

Volume (bn sticks) 40 1 20 0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2002 2002-2005 2005-2007  Legal domestic sales 0.9% (12.3)% (4.0)%  Average pack price 2.9% 11.0% 3.7%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Legal domestic sales continued their decline in 2007, ending the year 4.1% lower versus 2006  A Federal ban on smoking in public transport, taxis, airports and government offices came into effect on September 1st 2007 − stricter state-level bans have also been introduced covering the majority of Germany’s population  Significant tax increases took place in March 2004, December 2004 and September 2005

Marlboro price comparison Prices in Germany remain 1 July 2007(1)(2)(a) higher than in neighboring €4.30 Eastern European markets €8.16 €4.94 €1.79 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Germany and key (2)(a) markets €3.86

Key:  Germany 0%  Major source country

(20%)

(40%) July 2006 (48)% July 2007 (60%) (52)% (55)% (62)% (80%) (83)% (82)% (86)% (87)%

Price premium/(discount) versus local (100%) Czech Poland Russia Ukraine Republic

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown German consumption is 2007(1)(2)(3)(4)(a)(b)(c) estimated at 107.3 billion Outflows Inflows 1.3 billion sticks 19.3 billion sticks sticks versus legal sales of 120 89.3 billion sticks, giving a net inflow of 18.0 billion 3.1 2.6 sticks in 2007 1.1 100 2.3 0.3 Outflows 3.0 C&C 0.29 3.6 1.01 3.3 ND(L)

107.3 Volume (bn sticks) 89.3 88.0 40

0 LDS Netherlands Other LDC Poland Czech Republic Ukraine Russia Other Consumption

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) (c) In 2007, we have analysed both the Yellow Bag Survey results and PMI’s own Empty Pack Survey (EPS) study. Total non-domestic inflows and their breakdown by origin and brand have been based on the Yellow Bag survey. However, the EPS study reported a higher level of PMI counterfeit incidence than found in the Yellow Bag Survey. In order to ensure consistency of approach for EU wide counterfeit estimates, the EPS result has been used to calculate counterfeit levels in Germany Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) In Market Sales supplied by PMI (4) Interviews with PMI Local Management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3) 2007(1)(2)(3) inflows to Germany were 12.0 8.9 billion sticks in 2007 10.6 Other countries 19% 10.0 8.9 Poland and the Czech 2.8 8.0 1.7 Poland Republic are the two 0.5 38% 0.9 largest source markets for 6.0 3.2 Other countries 3.0 Spain Spain non-domestic legal inflows 4.0 10% Czech Republic Volume (bn sticks) Poland 2.0 4.0 3.3

0.0 2006 2007

ND(L) share of Czech Republic 9% 8% consumption: 33% ND(L) share of non- 47% 46% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 12 Population Propensity Propensity Ave. no. purchases ND(L) 10.6 19+ to travel to purchase of trips (sticks) (sticks) 10 8.9 Values x x x x = 8.9bn 2007 67.1m 48% 25% 4.3 260 8 Non-PMI 7.8 5.0 6 Other PMI EU rank L&M 1 12 12 8 6 1 4 2007 0.3 Marlboro Volume (bn sticks) 0.2 1.6 2 0.5 2.1 2.1 0 2006 2007 PMI share of ND(L): 26% 44% Pall Mall ND(L) inflows were higher in 2007 at 1.2 billion sticks. West ND(L) inflows were lower in 2007 Sources: (1) KPMG EU Flows Model at 1.1 billion sticks (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Total C&C inflows to 2006-2007(1)(2)(3) 2007(1)(2)(3) Germany were 10.4 billion 14 sticks in 2007 12.1 12 10.4 Poland 10 3.8 34% 3.4 8 Other countries 1.7 Other countries Russia 33% 6 1.1 Ukraine 4.1 2.3 4

Volume (bn sticks) Poland 2 3.6 2.5 0 2006 2007 Russia C&C share of 11% 11% 10% Ukraine consumption: 22% C&C share of non- 53% 54% domestic:

Notes: (a) In 2007, we have analysed both the Yellow Bag Survey results and PMI’s own Empty Pack Survey (EPS) study. Total non-domestic inflows and their breakdown by origin and brand have been based on the Yellow Bag survey. However, the EPS study reported a higher level of PMI counterfeit incidence than found in the Yellow Bag Survey. In order to ensure consistency of approach for EU wide counterfeit estimates, the EPS result has been used to calculate counterfeit levels in Germany (b) C&C volumes for individual non-PMI brands is not available for 2006 Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4) KPMG estimates are consistent with the trends highlighted by the Yellow Project Star results for Bag survey results and other sources 2007 are broadly in line 25%  KPMG estimates are slightly higher than the PMI EPS results for 2007 with Yellow Bag Survey 19.9% 19.9% 20.3% 19.0% which is likely to reflect the geographic weighting of the survey 20% 18.0% results and recent 16.0% − EPS coverage in 2007 was weighted towards the more densely estimates by Reemtsma 15% populated western regions and away from regions bordering lower cost markets such as Poland and the Czech Republic Cigarettenfabriken 10%  This 2007 EU Flows model result is broadly consistent with comments from the CEO of Reemtsma Cigarettenfabriken that smuggled cigarettes 5% had a 20.3% share of total consumption in 2007 Share of total of Share consumption KPMG’s estimate for non-domestic consumption is based on Yellow Bag 0% survey results excluding the Herten recycling centre  In all EU countries we have sought to investigate data outliers and high 2007 2007 2006 2007 local concentrations of specific pack types to ensure the national picture is PMI EPS EU Flows EU Flows Reemtsma Yellow Bag Yellow Bag Model 2006 Model 2007 accurate. We have excluded or re-weighted results from particular cities or centres where non-domestic pack survey results were not thought to be reflective of that locality − Herten had an excessively high share of Ukrainian product which distorted the overall German picture Yellow Bag survey results 2005-2007(2) − excluding the Herten recycling centre makes Ukrainian outflows more 25% consistent with trends in Ukrainian legal domestic sales 22.0%22.2%22.6% 19.5%19.3%19.6%  Please note that results from the two remaining recycling centres in AC 20% 17.7% 18.4% 16.7% 16.9% Nielsen Region 20 (Lippe and Leverkusen) have been re-weighted to reflect 14.9%14.9% the exclusion of Herten 15%

10%

5% Non-domestic incidence

0% Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

2005 2006 2007 Sources: (1) KPMG EU Flows Model (2) “Yellow Bag” survey, an empty pack survey undertaken by the German Cigarette Industry Association (VDC), 2005 to 2007. Full PMI results were available along with total non-domestic market size (3) PMI EPS Q4 2007 (4) “Cigarette sales seen down 3.0 – 5.0 pct on smoking ban in Germany”, German News Digest, 11th April 2008

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 1999-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that non- Gap between LDS and consumption increased  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption up until 2004 but has volume figures provide a useful indication of observed changes in 160 since narrowed declined between 1999 smoking habits. However, GCTS data cannot be used to quantify total 140 consumption in a market due to exclusions of particular age cohorts and and 2004 but has been 120 the problem of under-reporting of consumption increasing over the last 100  The KPMG consumption index is calculated using the incidence of three years smoking and the average daily consumption (both from GCTS) and the 80 population of each market (from Euromonitor GMID) This is consistent with 60 107.3 − these figures are then indexed to the earliest year for which we have Yellow Bag survey results Volume (bn sticks) 40 both GCTS and LDS data available, 1999 for Germany showing non-domestic 20  If LDS trends differ from those shown by the consumption index, it consumption trending 0 implies that the relationships between inflows, outflows and domestic 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period higher between Q1 2005 Please note that consumption modelling can only identify market trends and Q4 2007 and cannot be used to quantify total consumption CAGR (%) 1999-2001 2001-2004 2004-2007  Comparing consumption and sales trends helps to understand market  LDS 0.0% (7.0)% (7.7)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.9)% (11.2)% (3.1)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Germany widened between 1999 and 2004 but has since narrowed  Between 1999 and 2004, consumption declined at a faster rate than LDS which suggests decreasing non-domestic consumption  Between 2004 and 2007, consumption declined at a slower rate than LDS suggesting an increase in non-domestic consumption − this result is consistent with Yellow Bag survey results showing non- domestic consumption trending upward between Q1 2005 and Q4 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1999-2007(1)(2)(3)(a) prices affect cigarette consumption in Germany that changes in  Historical data from 1999 to 2006 has been used to estimate the strength consumption in Germany of relationships between consumption and prices and income 200 − estimated incidence of smoking and average daily consumption are are related to changes in Control period Forecast 180 based on GCTS data. Population and disposable income are estimated disposable income and 160 using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices 140 Cigarette prices are based on the weighted average pack price supplied by PMI 120 100 KPMG analysis suggests that changes in consumption in Germany are (c) 80 related to changes in disposable income and cigarette prices 60  Cigarette prices are negatively related to changes in consumption i.e. as

Volume (billion sticks) 40 prices rises consumption tends to decrease 20  Disposable income is also negatively related to changes in consumption 0  It is important to note that factors other than price and income are likely to 1999 2000 2001 2002 2003 2004 2005 2006 2007 impact consumption levels in Germany − examples of other factors likely to impact consumption levels include changes in smoking restrictions and changes in demographic profiles CAGR (%) 1999-2003 2003-2006 2006-2007  LDS (1.4)% (11.5)% (4.1)%  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be (b)  Estimated consumption (3.5)% (8.7)% (3.4)% considered a broad approximation The relationships between changes in consumption and changes in Notes: (a) Relationship based on historical data from 1999 to 2006 disposable income and prices have been used to estimate 2007 (b) The estimated consumption line has been rebased so that it equals the 2006 consumption Project Star estimate of consumption for Germany (114.0 billion sticks) (c) 93% of changes in consumption can be explained by changes in disposable  Combining estimates for the strength of these relationships with actual income and cigarette prices 2007 values for disposable income and prices gives estimated 2007 Sources: (1) KPMG EU Flows Model consumption (2) In Market Sales supplied by PMI (3) KPMG Consumption Regression Model  Estimated 2007 consumption is lower than in 2006, which is consistent with EU Flows Model and consumption index modelling results  The decline in estimated consumption in 2007 is similar to the decline in legal sales, suggesting that non-domestic consumption was broadly flat in 2007. This compares to the Flows Model which shows a small decline in non-domestic share  Non-domestic consumption appears to have increased since 2003 in Germany. Prior to 2003, non-domestic consumption appears to have been decreasing

OTP sales in billion stick equivalents OTP sales have increased from 2003 levels while manufactured cigarette Smoking tobacco volumes 2003-2007(1)(a)(b) sales have declined over the same period were lower in 2007 versus 50  Total OTP sales grew at an average of 8.3% per year between 2003 and 2006, but remain above 44.4 41.2 2007 while manufactured cigarettes sales declined by 9.7% over the 2003 levels 3.7 38.4 same time period 40 36.3 4.9 3.6 6.4  Cigars and cigarillos had the highest growth among the three major OTP 27.9 30 categories and now account for 17% of total OTP sales by volume 2.9 Smoking tobacco remains the largest OTP category with a 83% share of 20 40.7 36.3 sales volume despite a sales decline of approximately 12% in 2007 32.7 32.0 25.0  Stick equivalents (bns) 10 In 2006, the EU ruled that tobacco portions (included in the smoking tobacco category) must be taxed at an equivalent rate to cigarettes, 0 leading manufacturers to cease production 2003 2004 2005 2006 2007 − Portions had experienced significant growth from their introduction in 2002 as a low-cost alternative to cigarettes

Cigarettes 134.4 113.6 96.9 93.2 89.3 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007

 Cigars/ cigarillos 12.7% 31.9% 31.5% 21.8%  Smoking tobacco 27.7% (10.9)% (11.8)% 6.4% Total OTP 26.2% (7.4)% (6.7)% 8.3%

Manufactured cigarettes (15.1)% (3.9)% (4.1)% (9.7)%

Smoking tobacco sales by type The Portions category experienced strong growth from 2003 to 2005, The elimination of the 2003-2007(1)(a)(c)(d) but sales have declined extensively in recent years tobacco portions category  In 2005, Portions accounted for 50% of smoking tobacco sales in stick has led to a decline in 50 equivalents 40.7 smoking tobacco sales  40 1.1 36.3 However, after a change in the tax treatment of Portions in 2006, sales despite strong growth in 32.7 1.2 32.0 declined rapidly 1.3 1.7 the MYO and RYO 30 25.0 20.4 10.9 0.0 Following the decline in the Portions category, sales of smoking tobacco 11.6 Pipe categories in 2007 1.2 14.2 products have increased 20 4.8 Portions 11.3  Sales of both RYO and MYO increased by 25% in 2007 from the previous 6.0 6.4 6.4 MYO year and account for the majority of smoking tobacco 10 16.1 RYO Stick equivalents (billions) 13.0 13.4 12.9 12.9 − RYO accounted for 50% of smoking tobacco sales in stick equivalents 0 in 2007 2003 2004 2005 2006 2007 − MYO accounted for 44% of smoking tobacco sales in stick equivalents in 2007 CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  However, growth in both the RYO and MYO categories has not offset the  RYO (0.6)% 0.0% 25.3% 5.5% overall impact of the elimination of Portions sales  MYO 3.7% 76.8% 25.3% 24.2%  Portions 106.9% (46.7)% (100.0)% (100.0)%

 Pipe (7.1)% 14.7% 37.5% 8.0% Total 27.7% (10.9)% (11.8)% 6.4%

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Greek consumption Non-domestic inflows to 2006-2007(1)(2) Greece increased to 2.3 100% 3.6% 5.5% billion sticks in 2007, 0.8% 1.3% equivalent to 6.8% of total 80%

consumption 60% C&C 95.6% 93.1% ND(L) 40% Outflows from Greece also LDC increased from 2006 levels, 20% Outflows with the UK remaining the 0% Share of total of Share consumption main destination country (4.0)% (4.5)% (20%) 2006 2007

Volume (bn sticks): 33.7 33.6

Total non-domestic inflows (ND(L) and C&C) by origin (1)(2) Key:  Greece  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 2.5 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2006 2.0 0.56

Proportion of total EU levels(1)(2) 0.15 5% 4.7% 1.5 Other countries 0.42 Bulgaria 1.0 0.01 4% 1.60 Duty Free 3.1% 3.1% Volume (bn sticks) 0.5 1.04 3% 2.5%

0.0 2% 1.5% 2006 2007

Percentage of EU total 1% Volume (bn sticks): 1.48 2.30 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Cigarette sales in Greece 1997-2007(1)(2) increased between 1997 and 2004 but have 40 3.0 declined for the last three 34.3 32.9 33.3 33.2 34.0 33.6 32.7 35 31.6 32.1 (Euro) per 20 sticks Price years 29.9 30.6 2.5 30 2.0 25 20 1.5 15 1.0

Volume (bn sticks) 10 0.5 5 0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2007(a) 2006-2007  Legal domestic sales 2.4% (0.1)% (2.4)%

 Average pack price n/a 2.3% 5.2%

Note: (a) Weighted average pack price CAGR is from 2002 to 2007 Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Country management

Legal domestic sales decreased by 2.4% in 2007 while prices increased by 5.2%  In July 2006 the Government increased the Minimum Excise Tax (MET) to 75% of the most popular price category. Following this increase, the lowest retail selling price increased to €1.60 and the price gap in the market narrowed  Greece’s anti-smoking legislation bans smoking in most public buildings. Bars and restaurants must allocate space for non-smokers

Marlboro price comparison Greece remains relatively 1 July 2007(1)(2)(a) less expensive than the UK €4.30 and Germany, the main €8.16 €1.79 destination markets for €4.94 €0.86 Greek cigarettes €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €2.60 €3.00

Evolution of Marlboro price differential between Greece and key €3.49 €3.86 markets(2)(a)

Key:  Greece 200% 168%  Major destination country 165% 150%

100% 60% 57% July 2006 50% July 2007

(50%) (45)% (46)% (45)% (55)% (57)% (55)% (100%) (77)% Price premium/(discount) versus local (79)% UK Germany Ukraine Romania Albania Bulgaria

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Greek consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 33.6 billion Outflows Inflows 1.49 billion sticks 2.30 billion sticks sticks versus legal sales of 35

32.7 billion sticks, giving a 1.64 0.76 0.08 0.08 0.06 0.28 0.38 net inflow of 0.9 billion 0.07 sticks in 2007 30 0.45

Outflows C&C 25 ND(L)

“Other” includes Duty Free inflows 20

33.6 32.7 31.2 Volume (bn sticks) 15

0 LDS UK Germany Other LDC Bulgaria Ukraine Albania Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3)(a) 2007(1)(2)(3)(a) inflows increased to 0.45 0.5 billion sticks in 2007 0.45 Germany 0.4 25%

0.26 0.3 0.26 Other countries 0.2 Bulgaria 0.07

Volume (bn sticks) 0.23 Germany 0.1 Other countries 0.11 59% Bulgaria 0.02 0.0 0.01 16% 2006 2007

ND(L) share of 1% 1% consumption: ND(L) share of non- 17% 20% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3)(a) Ave. Total 0.5 0.45 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.4 Values x x x x = 0.45bn 0.18 2007 8.7m 27% 33% 2.4 239 0.3 0.26 0.03 Non-PMI 0.2 0.10 EU rank Other PMI 2007 9 23 6 22 7 13 0.01 Volume (bn sticks) 0.25 Marlboro 0.1 0.14

0.0 2006 2007

PMI share of ND(L): 61% 61%

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 Note: (a) Non-domestic (legal) inflows to Greece in 2007 are estimated to have changed in (3) Interviews with PMI Local Management proportion to changes in total non-domestic inflows. As a result, ND(L) inflows from Bulgaria and Germany are assumed to have increased in 2007

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows increased to 2006-2007(1)(2)(3) 2007(1)(2)(3)

1.9 billion sticks in 2007 Ukraine 2.0 1.85 Bulgaria 4% 4%

1.5 1.22 Other countries

1.0 1.70 Bulgaria

Ukraine 1.21

Volume (bn sticks) 0.5

0.08 0.08 0.0 0.00 0.01 2006 2007 Other countries includes Duty Other countries C&C share of Free inflows 4% 6% consumption: 92% C&C share of non- 83% 80% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3) Differences between KPMG results and the estimate of Greek Customs KPMG results for non- are likely to reflect the timing of EPS surveys domestic share in Greece 8% 6.9%  Empty Pack Surveys for Greece were conducted in Quarter 2 for both are broadly in line with 7% 2006 and 2007 6% Greek Customs’ November 5.0%  Greek Customs’ 2006 non-domestic estimate of 5% of consumption lies 2006 estimate 5% 4.4% in between our 2006 and 2007 estimates and is consistent with a rise in 4% incidence levels over the period 3% The number of counterfeit cigarettes seized in Greece remained flat in 2% 2007

Share of total of Share consumption 1%  115 billion counterfeit cigarettes were seized in Greece in 2007, (4) 0% compared to 119 billion counterfeit cigarettes the previous year EU flows model Greek Customs EU Flows Model  Seizures of contraband cigarettes declined significantly in 2007 2006 2006 2007 − 36 billion contraband cigarettes were seized in 2007 compared to 105 billion sticks in 2006(4) Sources: (1) Interview with Customs Control (2) KPMG EU Flows Model  It should be noted that seizures of cigarettes are not necessarily destined (3) PMI EPS Q2 2006 and Q2 2007 for consumption in that country and may instead reflect the transit of (4) Information provided by Greek Customs, 27 May 2008, George Tsagarakis, E’ section of the 33rd division of Customs Law Enforcement product destined for other markets

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 2004-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in remained fairly stable 40 smoking habits. However, GCTS data cannot be used to quantify total 35 consumption in a market due to exclusions of particular age cohorts and since 2004 the problem of under-reporting of consumption 30  25 The KPMG consumption index is calculated using the incidence of smoking and the average daily consumption (both from GCTS) and the 20 population of each market (from Euromonitor GMID) 33.6 15 − these figures are then indexed to the earliest year for which we have Volume (bn sticks) 10 both GCTS and LDS data available, 2004 for Greece 5  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 2004-2007  Comparing consumption and sales trends helps to understand market  LDS (1.5)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (2.6)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Greece remained reasonably constant between 2004 and 2007  Consumption and LDS in Greece have declined at similar rates since 2004  This result suggests that non-domestic consumption in Greece has been reasonably stable since 2004

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Hungarian consumption Non-domestic inflows to 2006-2007(1)(2) Hungary fell by 1.4 billion in 2007 to just under 2 100% 9.5% 17.3% billion sticks 1.6% 80% 0.7% The decline in total non- 60% domestic consumption C&C 88.9% ND(L) was primarily driven by a 40% 82.0% LDC sharp reduction in inflows 20% Outflows from Ukraine 0 Share of total consumption (2.3)% (3.7)% Inflows from Romania (20)% increased in 2007 and it is 2006 2007

now classified as a major Volume (bn sticks): 18.8 17.8 source country

Total non-domestic inflows (ND(L) and C&C) Key:  Hungary  Major source country  Major destination country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 4.0 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2007 3.5 0.22 3.0 0.10 Proportion of total EU levels(1)(2) 0.61 2.5 3% 2.8% Other countries 2.5% 2.0 Romania 0.62 2.2% 1.5 Duty Free 2.48 0.35

2% Volume (bn sticks) 1.0 Ukraine 0.47 1.4% 0.5 0.53 0.0 0.9% 1% 2006 2007 Percentage of EU total

Volume (bn sticks): 3.40 1.97 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales 1997-2007(1)(2) increased by 3.7% in 2007, but remain significantly 25 600 below pre-2002 levels 22.2 22.4 22.7 21.2 21.0 20.3 500 (HUF) per 20 sticks Price 20 18.7 14.9 15.9 16.5 400 15 13.8 300 10 200 Volume (bn sticks) 5 100

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2003 2003-2005 2005-2007  Legal domestic sales (2.8)% (14.1)% 9.3%  Average pack price 10.2%(a) 11.3% 5.9%

Note: (a) Price CAGR based on 2000 to 2003 only Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal sales declined between 1999 and 2005 but have since partially recovered  Legal domestic sales grew by 3.7% in 2007 despite an increase in prices of 8.5%  Outflows to Austria are understood to have increased as a result of a relaxation of import restrictions of cigarettes from Hungary in 2007  Hungary has a 40 stick allowance limit from Ukraine and applied a 200 stick limit along the Romanian border in 2007 − It is possible to import between 40 and 200 sticks from Ukraine provided this volume is reported to Customs who then stamp the packs

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Hungary and key source €4.30 and destination markets €8.16 €4.94 €1.79 narrowed in 2007 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

€3.86 Evolution of Marlboro price differential between Hungary and key Key:  Hungary (2)(a) markets  Major source country

 Major destination country

150% 121% 97% 100% 83% 63%

50% July 2006 July 2007 0%

(50)% (37)% (31)%

(68)%

Price premium/(discount) versus local (74)% (100)% Germany Austria Romania Ukraine

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Hungarian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 17.8 billion Outflows Inflows 0.66 billion sticks 1.97 billion sticks sticks versus legal sales of 20 16.5 billion sticks, giving a

net inflow of 1.3 billion 18 sticks 0.9 0.3 0.32 0.2 16 0.12 0.5 0.22 Outflows C&C

14 ND(L)

17.8 16.5 Volume (bn sticks) 8 15.8

0 LDS Austria Germany Other LDC Ukraine Romania Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal flows increased to 2006-2007(1)(2)(3) 2007(1)(2)(3) 280 million sticks in 2007, 0.3 0.28 due to volume growth in Austria 25% the Romanian, Italian and 0.11 Other countries Austrian inflows 0.2 Other countries 38% 0.14 0.04 Romania Italy 0.1 0.06 0.11 Austria Volume (bn sticks)

0.07 0.02 0.01 Italy 0.01 0.0 21% 2006 2007 Romania 16% ND(L) share of 1% 2% consumption: ND(L) share of non- 4% 14% domestic:

Non-domestic (legal) by brand (a) 2006-2007(1)(2)(3) ND(L) analysis Ave. Total 0.3 0.28 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) Values x x x x = 0.15bn(a) 2007 7.9m 30% 16% 5.1 77 0.2 0.15 Non-PMI 0.14 Other PMI L&M EU rank 0.03 13 19 23 4 24 17 0.1 0.07 0.01 Marlboro 2007 Volume (bn sticks) 0.01 0.00 0.09 0.05 0.0 2006 2007 PMI share of ND(L): 47% 46%

Note: (a) ND(L) analysis provided by AC Nielsen excludes adjustment to inflows associated with inbound tourism. Total ND(L) (including inbound tourist flows) is equal to 0.28 billion sticks Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Hungary 20060-2007(1)(2)(3) 2007(1)(2)(3) decreased by 48% in 2007 4 driven by a substantial Other countries 3.3 25% decline in inflows from 0.1 Ukraine 0.1 3 32% Ukraine 0.6 Other countries Romania 2 1.7 Duty Free 0.4 2.5 Ukraine 0.3

Volume (bn sticks) 1 0.4

0.5 0 Romania 2006 2007 18% Duty Free C&C share of 17% 10% 25% consumption: C&C share of non- 96% 86% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(5)(6) KPMG results for non-domestic share is in line with EPS results (MDSZ) All external estimates of and with both Customs (VPOP) and a Finance Ministry estimate for 2007 non-domestic share for  A level of 11% non-domestic incidence also corresponds to a BAT 2007 are consistent with estimate for 2006 35% KPMG findings 30% KPMG results and external estimates suggest that non-domestic 30% 27% incidence has declined rapidly since 2005 25% 22%  Differences in individual estimates for 2005 and 2006 are likely to be due 18% to timing differences 20% 17% − 15% estimates based on a single point in time methodology such as an 11% 11% 11% 11% 10% empty pack survey or pack swap approach may not correlate with full 10% year estimates due to the rate change over a 12 month period

Share of total of Share consumption 5%

0% BAT 2005 BAT 2006 Finance Finance EU Flows EU Flows VPOP 2007 MDSZ 2005 Model 2006 MDSZ 2006 MDSZ 2007 Model 2007 Ministry 2005 Ministry 2007

Sources: (1) KPMG EU Flows Model (2) “Stricter checks boost customs office revenue”, MTI – EcoNews, 18th December 2006 (3) “BAT welcomes government efforts to clamp down on black market cigarette sales”, MTI – EcoNews, 27th December 2006 (4) “Hungary black market clampdown yields results”, Reuters News, 10th July 2007 (5) “Illicit Cigarette Market”, Empty Pack Surveys carried out by GFK for MDSZ (Hungarian Tobacco Manufacturers’ Association), 2005, 2006 and 2007 (6) “Black market for cigarettes contracts because of crackdown”, MTI – EcoNews, 24th April 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 1997-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption volume figures provide a useful indication of observed changes in 25 peaked in 2005 but has smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and been declining over the Consumption 20 has declined the problem of under-reporting of consumption last two years since 2005 while LDS has  The KPMG consumption index is calculated using the incidence of 15 increased smoking and the average daily consumption (both from GCTS) and the This provides relatively population of each market (from Euromonitor GMID) strong corroboration for 10 17.8 − these figures are then indexed to the earliest year for which we have

BAT, TMA and Finance Volume (billion sticks) 5 both GCTS and LDS data available, 1997 for Hungary Ministry estimates  If LDS trends differ from those shown by the consumption index, it showing a decline in non- 0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period domestic consumption Please note that consumption modelling can only identify market trends between 2005 and 2007 and cannot be used to quantify total consumption CAGR (%) 1997-2003 2003-2005 2005-2007  Comparing consumption and sales trends helps to understand market  LDS (2.8)% (14.1)% 9.3% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.0)% (2.8)% (0.3)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated Consumption modelling results suggest that non-domestic consumption peaked in 2005 but has subsequently been in decline  The gap between LDS and the KPMG consumption index for Hungary increased between 2001 and 2005 and has since decreased  Legal sales have increased by 9.3% per year since 2005 while consumption has declined slightly  This suggests that non-domestic consumption has decreased since 2005 which is consistent with BAT, TMA and Finance Ministry estimates showing a decline in non-domestic consumption between 2005 and 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 2000-2007(1)(2)(3)(a) prices affect cigarette consumption in Hungary that changes in  Historical data from 2000 to 2006 has been used to estimate the strength consumption in Hungary of relationships between consumption and prices and income 25 are related to changes in Control period Forecast − estimated incidence of smoking and average daily consumption are based on GCTS data. Population and disposable income are estimated disposable income and 20 using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices Cigarette prices are based on the weighted average pack price supplied 15 by PMI Estimated consumption KPMG analysis suggests that changes in consumption in Hungary are results show a decrease in 10 related to changes in disposable income and cigarette prices(c)

non-domestic Volume (billion sticks) 5  Cigarette prices are negatively related to changes in consumption i.e. as consumption in 2007 prices rises consumption tends to decrease 0 which is consistent with  Disposable income is, by contrast, positively related to changes in 2000 2001 2002 2003 2004 2005 2006 2007 EU Flows Model results consumption  It is important to note that factors other than price and income are likely to CAGR (%) 2000-2005 2005-2006 2006-2007 impact consumption levels in Hungary

 LDS (8.0)% 15.5% 4.0% − examples of other factors likely to impact consumption levels include  Estimated consumption(b) (1.9)% 1.0% (3.2)% changes in smoking restrictions and changes in demographic profiles  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Relationship based on historical data from 2000 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 considered a broad approximation Project Star estimate of consumption for Hungary (18.8 billion sticks) (c) 53% of changes in consumption can be explained by changes in disposable The relationships between changes in consumption and changes in income and cigarette prices disposable income and prices have been used to estimate 2007 Sources: (1) KPMG EU Flows Model consumption (2) In Market Sales supplied by PMI (3) KPMG Consumption Regression Model  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is consistent with EU Flows Model and consumption index modelling results Estimated consumption results suggest that non-domestic consumption peaked in 2005 but has declined during the last two years  This is consistent with both the EU Flows Model and consumption index modelling results

OTP sales in billion stick equivalents Legal sales of smoking tobacco have increased since 2003 and now Smoking tobacco sales in 2003-2007(1)(a)(b) represent a greater share of Hungarian consumption Hungary have been 2.91  Smoking tobacco sales increased at a annual rate of 24.7% between 2003 3.0 increasing since 2003 0.04 and 2007, whereas manufactured cigarettes sales declined at an annual 2.41 2.41 rate of 3.1% over the equivalent time period while legal sales of 2.5 0.01 0.01  Smoking tobacco accounts for virtually all OTP sales in Hungary manufactured cigarettes 2.0 1.57 The rapid growth of OTP products during 2003 to 2005 corresponded have fallen 0.01 1.5 1.20 2.87 with a significant decline in manufactured cigarette sales 0.01 2.40 2.40 Cigars/ cigarellos 1.0 Smoking tobacco  Total OTP sales increased at an average of 41.6% per annum over the 1.56 period 2003 to 2005, compared with a decline of 14.1% per annum for Stick equivalents (bns) 1.19 0.5 manufactured cigarettes 0.0  However, the growth in OTP over this period did not fully account for the 2003 2004 2005 2006 2007 fall in manufactured cigarette sales − In 2005 sales of OTP were 1.2 billion stick equivalents above the 2003 level, whereas sales of manufactured sales had declined by 4.9 billion Cigarettes 18.7 14.9 13.8 15.9 16.5 (bn sticks) sticks

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 4.1% (10.0)% 388.9% 47.8%  Smoking tobacco 41.8% 0.1% 19.3% 24.5% Total OTP 41.6% 0.0% 20.7% 24.7%

Manufactured cigarettes (14.1)% 15.1% 3.7% (3.1)%

Notes: (a) Smoking tobacco volumes have been calculated at 1 stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick per stick basis (b) Total OTP sales excludes smokeless tobacco volumes of approximately 0.3 tons per year Source: (1) OTP volumes supplied by PMI

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Irish consumption Non-domestic inflows to 2006-2007(1)(2) Ireland rose to 2.1 billion sticks in 2007 100% 10.2% 16.9% The overall growth of non- 80% 13.7% 11.7% domestic flows coupled 60% with a small decline in C&C ND(L) ND(L) volumes has led to 40% 76.2% 71.4% LDC an increase in the share of 20% Outflows consumption accounted

Share of total of Share consumption 0 for by C&C to 16.9% (1.6)% (3.2)% (20)% 2006 2007

Volume (bn sticks): 7.24 7.23

Total non-domestic inflows (ND(L) and C&C) Key:  Ireland  Major source country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 2.5 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 2.0 Proportion of total EU levels(1)(2) 1.01 3% 2.8% 1.5 Other countries 0.75 Latvia 2.3% 1.0 0.13 0.20 Poland 2.0% 0.33 0.37 2% Volume (bn sticks) Spain 0.5 0.51 0.49

1.0% 0.0 1% 2006 2007 0.5% Percentage of EU total

Volume (bn sticks): 1.72 2.07 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales 1997-2007(1)(2) decreased by 4.3% in 2007 while the average pack 6.9 6.8 6.7 6.8 7 6.4 8 price increased by 7.5% 6.2 6.2 5.4

6 5.5 5.6 5.6 (Euro) per 20 sticks Price 6 5

4 4 3

2 Volume (bn sticks) 2 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2004 2004-2007  Legal domestic sales 2.9% (7.2)% (0.9)%  Average pack price 6.8%(a) 16.9% 3.9%

Note: (a) Price CAGR is from 1998 to 2001 Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal sales decreased by an average of 7.2% per year between 2001 and 2004 as prices rose by nearly 17% per year  Legal sales have declined at a relatively slower rate since 2004  Ireland has a Minimum Retail Price policy which limits the price differential between brands to a maximum of 3%  This has effectively prevented the development of the low and super-low price segments which have gained market share in many other markets, including the UK  In 2007, Ireland banned the sale of packs of ten sticks

Marlboro price comparison Although Marlboro prices 1 July 2007(1)(2)(a) are higher in the UK than Key:  Ireland €4.30  Major source country €8.16 in Ireland, the minimum €4.94 €1.79 retail price means that €0.86 brands in the low and €1.47 €4.24 €1.33 super-low segments are €0.86

cheaper €7.05 n/a €4.71 €8.04 €4.11 €2.10 Price differentials with €4.53 €3.70 €2.46 €0.63 major source markets €2.35 €3.90 €2.39 remain substantial €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

€3.86 Evolution of Marlboro price differential between Ireland and key Comparison of UK and Ireland price range, highest cost brand markets(2)(a) versus lowest cost brand (3)

40% Ireland UK 17% 9.00 20% 14% 8.00 7.20 7.82 6.65 0% 7.00 6.00 (20)% July 2006 5.00 4.48 July 2007 (40)% 4.00 3.00 (60)% 2.00 (57)%(58)% 1.00 (80)% (70)% € price pack per (20 sticks) (73)% (76)% 0.00 (80)% (82)%(79)% Price premium/(discount) versus local (100)% High Low High Low UK Spain Poland Romania Latvia

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI (3) Based on 20 stick packs (all brands) July 1st 2007

Consumption breakdown Ireland consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 7.2 billion Outflows Inflows 0.23 billion sticks 2.07 billion sticks sticks versus legal sales of 8 5.4 billion sticks, giving a net inflow of 1.8 billion sticks in 2007 7 0.50 0.25 0.12 0.14 Outflows 0.20 6 C&C 0.37 0.00 0.04 ND(L) 0.00 0.14 0.45 0.09 5

7.2 Volume (bn sticks) 3 5.4 5.2

0 LDS UK Other LDC Spain Poland Latvia UK Romania Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows declined 2006-2007(1)(2)(3) 2007(1)(2)(3) slightly in 2007 1.2 0.99 Other countries Spain remained the main 1.0 0.85 30% source of ND(L) inflows in 0.8 0.35 0.25 2007 Other countries 0.6 0.14 0.14 UK Spain 53% 0.4 Spain Volume (bn sticks) 0.50 0.2 0.45

0.0 UK 2006 2007 17%

ND(L) share of 14% 12% consumption: ND(L) share of non- 57% 41% domestic:

Non-domestic (legal) by brand The top two non-PMI brands, Benson & Hedge and Silk Cut, (1)(2)(3) 2006-2007 accounted for over half of non- PMI ND(L) in both 2006 and ND(L) analysis ND(L) 1.2 2007 Ave. Total share of 0.99 Population Propensity Propensity Ave. no. purchases ND(L) consum 1.0 19+ to travel 0.85 to purchase of trips (sticks) (sticks) ption Values 0.8 x x x x = 0.85bn 12% 2007 3.0m 69% 23% 2.1 836 0.6 0.86 Non-PMI 0.70 Other PMI 0.4 EU rank

Volume (bn sticks) Marlboro 2007 20 4 15 24 1 5 2 0.2 0.01 0.03 0.11 0.11 0.0 2006 2007 PMI share of ND(L): 13% 17%

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows increased by 2006-2007(1)(2)(3) 2007(1)(2)(3) 0.5 billion sticks in 2007 1.4 1.22 1.2 Poland Other countries 30% 38% 1.0 0.47 Other countries 0.8 0.74 Ukraine 0.07 Romania 0.6 0.19 0.12 0.03 0.05 0.20 Latvia 0.4 0.13 Volume (bn sticks) Poland

0.2 0.33 0.37 0.0 2006 2007 Latvia Ukraine 17% C&C share of 6% Romania 10% 17% consumption: 9% C&C share of non- 43% 59% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3) ITL event collection data shows an increase in non-domestic incidence ITL event collection data in 2007 corroborates a trend for  ITL data is based on pack collections at sporting events and in bars and increased non-domestic 35% may therefore not be fully representative of the population consumption in Ireland in  The range of results from ITL collection waves in 2007 was between 30% 28.6% 2007 18.3% and 22.6% 25% 23.8% − by contrast the range in 2006 was 8.2% to 17.0%  In 2007, the average non-domestic incidence of packs collected by ITL 20% 18.0% 15.9% collection rose by seven percentage points which is directionally in line 15% with the EU Flows Model John Player & Sons estimate of 18% for 2007 refers only to smuggled 10% cigarettes and is therefore consistent with the Project Star C&C level of Share of total of Share consumption 17% 5%

0% Industry EPS EU Flows John Player & EU Flows 2006 Model 2006 Sons 2007 Model 2007

ITL event collection EPS results(4)

25% 22.6% 21.0% 19.8% 20% 18.3% 17.0% 15.9% 15% 13.5%

9.0% 10% 7.8% 8.2%

5% Share of total of Share consumption

0% May 07 April 07 June 05 June 07 Q1 2007 March 06 August 06 Jul/Aug 05 October 06 June/July 06 Sources: (1) KPMG EU Flows Model (2) Interview with Irish Customs (3) “Irish chains hit by tobacco sales laws”, The Grocer, 17th November 2007 (4) ITL Event Collection EPS, 2007

results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in 9 smoking habits. However, GCTS data cannot be used to quantify total increased since 2000 consumption in a market due to exclusions of particular age cohorts and 8 the problem of under-reporting of consumption 7 Consumption 2007 consumption has increased  The KPMG consumption index is calculated using the incidence of 6 since 2000 while modelling results suggest LDS has smoking and the average daily consumption (both from GCTS) and the 5 population of each market (from Euromonitor GMID) an increase in non- decreased 4 7.2 − these figures are then indexed to the earliest year for which we have domestic consumption 3 both GCTS and LDS data available, 2000 for Ireland

which is consistent with Volume (billion sticks) 2  If LDS trends differ from those shown by the consumption index, it EU Flows Model results 1 implies that the relationships between inflows, outflows and domestic 0 consumption have changed over the analysed period 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2000-2004 20004-2006 2006-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS (4.7)% 0.9% (4.3)% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index 3.7% (1.7)% 2.1% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be

Sources: (1) KPMG Consumption Index model present in the first analysed year (2) In Market Sales supplied by PMI  Any changes in GCTS (methodology, supplier, sample size and selection, (3) KPMG EU Flows Model etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Ireland increased between 2000 and 2007  The KPMG consumption index shows that during the period 2001 to 2004, when cigarette sales were declining, consumption actually increased − over this period Ireland experienced rising disposable incomes and a steadily increasing population  This suggests that non-domestic consumption in Ireland has increased since 2000  Consumption increased in 2007 while LDS decreased, suggesting that non-domestic consumption grew in that year − this is consistent with EU Flows Model results

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 2000-2007(1)(2)(3)(a) prices affect cigarette consumption in Ireland that changes in  Historical data from 2000 to 2006 has been used to estimate the strength consumption are related of relationships between consumption and prices and income 10 − estimated incidence of smoking and average daily consumption are to changes in disposable Control period Forecast based on GCTS data. Population and disposable income are estimated income and cigarette 8 using Euromonitor GMID and Oxford Economic Forecasting data. prices in Ireland Cigarette prices are based on the weighted average pack price supplied 6 by PMI Estimated consumption KPMG analysis suggests that changes in consumption in Ireland are (c) 4 related to changes in disposable income and cigarette prices results show an increase  Disposable income is negatively related to changes in consumption i.e. as

in non-domestic Volume (billion sticks) 2 income rises consumption tends to decrease consumption in 2007  Cigarette prices are, by contrast, positively related to changes in which is consistent with 0 consumption 2000 2001 2002 2003 2004 2005 2006 2007 EU Flows Model results − positive correlation between price and consumption is unexpected  It is important to note that factors other than price and income are likely to impact consumption levels in Ireland CAGR (%) 2000-2003 2003-2006 2006-2007  LDS (3.9)% (5.0)% (6.3)% − examples of other factors likely to impact consumption levels include changes in smoking restrictions and changes in demographic profiles  Estimated consumption(b) 4.2% (0.7)% 1.1%  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Relationship based on historical data from 2000 to 2006 considered a broad approximation (b) The estimated consumption line has been rebased so that it equals the 2006 The relationships between changes in consumption and changes in Project Star estimate of consumption for Ireland (7.2 billion sticks) (c) 86% of changes in consumption can be explained by changes in disposable disposable income and prices have been used to estimate 2007 income and cigarette prices consumption Sources: (1) KPMG EU Flows Model (2) In Market Sales supplied by PMI  Combining estimates for the strength of these relationships with actual (3) KPMG Consumption Regression Model 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is slightly higher than in 2006 but is in line with 2005. This is broadly consistent with EU Flows Model results which show stable consumption between 2006 and 2007  Estimated consumption increased in 2007 while LDS decreased, suggesting non-domestic consumption grew − this is consistent with EU Flows Model and consumption index modelling results

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Italian consumption Non-domestic inflows to 2006-2007(1)(2) Italy remained relatively stable in 2007 at 4.8 billion 100% 4.2% 4.5% sticks 1.0% 0.5% 80%

60% C&C 94.8% 95.0% ND(L) 40% LDC 20% Outflows

Share of total of Share consumption 0 (2.2)% (2.1)% (20)% 2006 2007

Volume (bn sticks): 96.7 95.6

Total non-domestic inflows (ND(L) and C&C) Key:  Italy  Major source country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 6 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 5

1.43 Proportion of total EU levels(1)(2) 4 1.56 Other countries 15% 0.28 13.3% 0.17 Poland 3 0.93 0.68 Romania 10% 2 0.89 1.00 Ukraine

7.1% Volume (bn sticks) 1 Duty Free 5.3% 1.49 4.2% 1.40 5% 0 1.7% 2006 2007

0% Percentage of EU total

Volume (bn sticks): 5.0 4.8 -5% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales in 1997-2007(1)(2) Italy remain below 2002 peak levels but are still higher than 10 years ago 120 6 103.0 100.4 101.6 101.2 98.9

95.9 (Euro) per 20 sticks Price 92.4 93.8 92.8 100 89.2 91.0 5

80 4

60 3

40 2 Volume (bn sticks) 20 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2002 2002-2005 2005-2007  Legal domestic sales 2.9% (3.6)% 0.2%  Average pack price 2.7% 8.5% 4.6%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales declined by 1.1% in 2007 as prices rose by 3.3%  Prices were increased by €0.10 per pack of 20 sticks in February 2007  The introduction of new fiscal measures in 2004, including a minimum excise tax, had the effect of amplifying the levels and the frequency of tax-driven price increases − in 2005, the government introduced a minimum reference price for cigarettes  A law banning smoking in enclosed public places came into force in Italy in January 2005 − all enclosed places open to the public are prohibited from allowing smoking on their premises unless a dedicated smoking area responding to stringent technical requirements is provided

Marlboro price comparison Cigarette prices in Ukraine 1 July 2007(1)(2)(a) and Romania, the two €4.30 main sources of non- €8.16 €4.94 €1.79 domestic inflows, remain €0.86 relatively low compared to €1.47 €4.24 €1.33 Italy €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €4.10 €1.58 €1.42 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Italy and key €3.86 (2)(a) markets Key:  Italy

 Major source country 0% (10)% (20)% (30)% (28)% (31)% (40)% July 2006 (39)% (50)% (44)% July 2007 (49)% (60)% (57)% (60)% (70)% (68)% (80)% Price premium/(discount) versus local (90)% (84)%(85)% Spain Slovenia Poland Romania Ukraine

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Italian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 95.6 billion Outflows Inflows 2.0 billion sticks 4.8 billion sticks sticks versus legal sales of 120 92.8 billion sticks, giving a net inflow of 2.8 billion sticks in 2007 100

2.64 0.50 0.45 1.06 1.00 0.67 Outflows 0.49 0.01 C&C

ND(L) 80

Volume (bn sticks) 95.6 92.8 90.8 40

0 LDS France Netherlands Other LDC Ukraine Romania Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows account for a 2006-2007(1)(2)(3) 2007(1)(2)(3) 0.98 relatively low share of 1.0 Spain non-domestic 24% consumption in Italy 0.8

0.6 0.85 0.51 Other countries Germany 0.4 0.34 Spain Germany

Volume (bn sticks) 9% 0.2 0.05 Other countries 0.03 0.10 0.12 67% 0.0 2006 2007

ND(L) share of 1% 1% consumption: ND(L) share of non- 20% 11% domestic:

Non-domestic (legal) by brand (a) 2006-2007(1)(2)(3) ND(L) analysis 0.98 Ave. Total 1.0 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.8 0.47 Values x x x x = 0.47bn(a) 2007 48.4m 20% 17% 2.6 109 0.6 0.51 Non-PMI 0.07 Other PMI 0.16 0.4 Marlboro EU rank 0.07 2007 3 27 21 19 18 12 Volume (bn sticks) 0.2 0.44 0.28

0.0 2006 2007 PMI share of ND(L): 52% 69%

Note: (a) ND(L) analysis from Synovate excludes adjustments for inbound tourism inflows. Total ND(L) including these inbound tourism inflows in 0.51 billion sticks Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interview with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows increased 2006-2007(1)(2)(3) 2007(1)(2)(3) slightly in 2007 5 4.31 Ukraine Ukraine and Romania are 4.04 23% 4 the two major source Other countries countries for C&C inflows 3 2.22 2.64 to Italy Romania 2 Other countries Ukraine 61% 0.93 0.67 Volume (bn sticks) 1 Romania 0.89 1.00 16% 0 2006 2007 Other countries includes Duty Free C&C share of inflows 4% 5% consumption: C&C share of non- 80% 89% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3) Italian Tobacconist Association data corroborates Italy’s status as a KPMG estimates of non- market with a stable and low share of consumption accounted for by domestic consumption are non-domestic products broadly consistent with  The Italian Tobacconist Association estimate does not include legal 6% 5.6% 5.6% inflows and therefore is directly comparable to the Project Star 2006 the Italian Tobacconist 5.2% 5.0% results for C&C incidence of 4.5% Association 5% 4.0% KPMG estimates suggest that non-domestic inflows account for 5.0% of 4% total cigarette consumption in 2007 compared with 5.2% in 2006 3%  KPMG’s estimate of non-domestic share in 2007 is lower than PMI EPS results due to an adjustment to inflows from San Marino 2% − inflows from San Marino to Italy have been reduced to provide a more

Share of total of Share consumption 1% accurate national consumption estimate. The reduced inflow level is also more comparable to San Marino LDS volumes 0% Italian PMI EPS EU Flows PMI EPS EU Flows Tobacconist (2006) Model (2007) Model Association (2006) (2007) (2006)

Sources: (1) KPMG EU Flows Model (2) Interview with Italian Tobacconist Association (3) PMI EPS Q2 2006 and Q2 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption index 1997-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market modelling results suggest  The rate of change in the GCTS smoking prevalence and average smoking that non-domestic volume figures provide a useful indication of observed changes in 120 consumption has declined smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and since 1997 100 Consumption the problem of under-reporting of consumption 80 has declined This is consistent with since 1997 while  The KPMG consumption index is calculated using the incidence of LDS have smoking and the average daily consumption (both from GCTS) and the reports of a large decrease 60 increased population of each market (from Euromonitor GMID) in inflows in response to 95.6 40 − these figures are then indexed to the earliest year for which we have

improved border controls Volume (billion sticks) both GCTS and LDS data available, 1997 for Italy 20 and enforcement  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1997-2002 2002-2005 2005-2007  Comparing consumption and sales trends helps to understand market  LDS 2.9% (3.6)% 0.2% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (3.9)% (4.8)% (4.3)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has

Sources: (1) KPMG Consumption Index model changed over the analysed period (2) In Market Sales supplied by PMI − in many markets, some level of inflows or outflows are likely to be (3) KPMG EU Flows Model present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Italy increased between 1997 and 2007  Consumption in Italy has decreased since 1997 while LDS have increased  This result suggests that non-domestic consumption has declined since 1997 − this is consistent with reports that substantial inflows (e.g. from Kosovo) during the start of this period have been reduced through improvements to border controls and enforcement

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1997-2007(1)(2)(3)(a) prices affect cigarette consumption in Italy that changes in  Historical data from 1997 to 2006 has been used to estimate the strength consumption in Italy are of relationships between consumption and prices and income 140 related to changes in Control period Forecast − estimated incidence of smoking and average daily consumption are 120 disposable income and based on GCTS data. Population and disposable income are estimated using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices 100 Cigarette prices are based on the weighted average pack price supplied 80 by PMI Estimated consumption 60 KPMG analysis suggests that changes in consumption in Italy are related results show a decline in (c) 40 to changes in disposable income and cigarette prices

non-domestic Volume (billion sticks)  20 Disposable income is negatively related to changes in consumption i.e. as consumption in 2007 income rises consumption tends to decrease 0 which is consistent with  Cigarette prices are also negatively related to changes in consumption 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 EU Flows Model results  It is important to note that factors other than price and income are likely to impact consumption levels in Italy CAGR (%) 1997-2002 2002-2006 2006-2007 − examples of other factors likely to impact consumption levels include

 LDS 2.8% (3.0)% (1.2)% changes in smoking restrictions and changes in demographic profiles  Estimated consumption(b) (2.3)% (4.3)% (5.6)%  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be considered a broad approximation Notes: (a) Relationship based on historical data from 1997 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 The relationships between changes in consumption and changes in Project Star estimate of consumption for Italy (96.7 billion sticks) (c) 97% of changes in consumption can be explained by changes in disposable disposable income and prices have been used to estimate 2007 income and cigarette prices consumption Sources: (1) KPMG EU Flows Model (2) In Market Sales supplied by PMI  Combining estimates for the strength of these relationships with actual (3) KPMG Consumption Regression Model 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is consistent with EU Flows Model results  Estimated consumption decreased at a faster rate relative to LDS in 2007 suggesting non-domestic consumption decreased − this is consistent with EU Flows Model modelling results of a small decline in overall non-domestic incidence

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Latvian consumption Non-domestic inflows to 2006-2007(1)(2) Latvia declined to 0.39 billion sticks in 2007 100% 6.5% 4.1% 6.1% 5.3% Inflows from Russia, the 75% largest source of non- C&C domestic inflows, were 50% ND(L) 87.4% 90.7% down on 2006 levels LDC 25% Outflows

0% Share of total of Share consumption (19.8%) (21.1%) (25%) 2006 2007

Volume (bn sticks): 4.2 4.2

Total non-domestic inflows (ND(L) and C&C) by origin (1)(2) Key:  Latvia  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 0.6 (2) AC Nielsen ND(L) research, 2006 0.5 0.09 Proportion of total EU levels(1)(2) Other countries 0.4 0.09 2.0% 1.8% 0.03 Lithuania 0.09 0.3 0.11 Estonia 1.5% 0.11 0.2 Russia Volume (bn sticks) 0.1 0.24 1.0% 0.17 0.7% 0.6% 0.0 0.4% 0.5% 2006 2007 0.3% Percentage of EU total

Volume (bn sticks): 0.53 0.39 0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal cigarette sales have 2000-2007(1)(2) trended upwards since 2000 despite increasing 5.0 4.7 0.8 prices 4.5 4.5 4.0 0.7 Price per 20 sticks (LVL) per 20 sticks Price 4.0 3.7 0.6 3.5 2.9 2.7 0.5 3.0 2.5 2.6 2.5 0.4

2.0 0.3 1.5 Volume (bn sticks) 0.2 1.0 0.5 0.1 0.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 2000–2003 2003–2007 2004–2007  Legal domestic sales 2.8% 37.6% 8.4%

 Average pack price n/a n/a 13.8%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales increased by 5.9% in 2007 despite an increase in average weighted prices of 26%  Latvia is raising taxes into line with EU excise tax regulations with prices rising at an annual average of 13.8% between 2004 and 2007 − taxes increased by 22%in January 2007 with a further 33% rise in July. Further increases are expected during 2008  The Latvian government is in the process of introducing new anti-smoking measures − smoking in educational establishments and restaurants was banned in July 2007

Marlboro price comparison The price differential 1 July 2007(1)(2)(a) between Latvia and €4.30 Estonia narrowed during €8.16 €4.94 €1.79 2007 €0.86 €1.47 Marlboro prices in €4.24 €1.33 Lithuania are now cheaper €0.86 €7.05 n/a €4.71 than in Latvia €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Latvia and key markets(2)a) Key:  Latvia  Major source country  Major destination country

600% 534%

500% 447% 444% 380% 400% 271% 300% 236% July 2006 July 2007 200%

100% 48% 22% 9% 0% (10)%

Price premium/(discount) versus local (100%) (33)% (41)% UK Ireland Sweden Estonia Lithuania Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Latvian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 4.2 billion Outflows Inflows 0.9 billion sticks 0.4 billion sticks sticks versus legal sales of 4.7 billion sticks, giving a 5 net outflow of 0.5 billion 5 0.33 sticks in 2007 0.01 0.20 0.07 0.06 0.21 4 0.23 0.16

0.01 4

Outflows 3 4.71 C&C ND(L) 4.22

Volume (bn sticks) 2 3.82

0 LDS UK Ireland Lithuania Sweden Other LDC Russia Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows decreased 2006-2007(1)(2)(3) 2007(1)(2)(3)

slightly between 2006 and 0.30 Other countries 2007 0.25 12% 0.25 0.22 Estonia and Lithuania 0.06 0.20 0.03 remain the main sources of non-domestic (legal) 0.15 0.09 0.09 Other countries Estonia Lithuania inflows 0.10 49%

Volume (bn sticks) Estonia 0.05 0.11 0.11 Lithuania 39% 0.00 2006 2007

ND(L) share of 6% 5% consumption: ND(L) share of non- 48% 56% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.30 Population Propensity Propensity Ave. no. purchases ND(L) 0.25 19+ to travel to purchase of trips (sticks) (sticks) 0.25 0.22 Values x x x x = 0.22bn 2007 1.8m 41% 37% 2.5 311 0.20 0.11 0.11 0.15 Non-PMI EU rank Other PMI 2007 22 15 3 21 5 16 0.10 0.09 0.04 Volume (bn sticks) Marlboro 0.05 0.05 0.07 0.00 2006 2007 PMI share of ND(L): 56% 49%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Latvia 2006-2007(1)(2)(3) 2007(1)(2)(3)

decreased by 37% from Other countries 0.30 2006 to 2007 0.27 6% 0.25 0.04

0.17 0.20 0.01 0.15 Other countries 0.23 0.10 Russia 0.16 Volume (bn sticks) 0.05

0.00 2006 2007 Russia C&C share of 94% 7% 4% consumption: C&C share of non- 52% 44% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2) Analysis of the implied consumption trend in Latvia suggests that PMI EU Flows Model results EPS results may understate actual non-domestic inflows were adjusted in 2007 in  Implied consumption in Latvia is 3% lower in 2007 compared with 2006 order to reflect inflow 14% 12.7% when unadjusted EPS results are used volumes from Estonia and 12%  KPMG consumption modelling by contrast, suggest consumption Lithuania increased by over 4% in 2007 (see next page for more detail) 9.3% 10% Analysis of pricing and legal sales trends in the Baltics suggest that

8% inflows from Estonia and Lithuania may be understated in 2007 6.7%  PMI EPS results suggest that inflows from Estonia and Lithuania 6% 5.0% decreased to almost zero in 2007

4%  These decreases are not consistent with changes in relative pricing in the

Share of total of Share consumption Baltics in 2007 2% − the price discount available in Estonia relative to Latvia narrowed 0% during 2007, but remains in excess of 20% PMI Baltics EU Flows PMI EPS 2007 EU Flows − Marlboro prices in Lithuania moved from being more expensive than 2006 Model 2006 Model 2007 Latvia to being less expensive KPMG’s estimate for non-domestic share is higher than PMI EPS results Sources: (1) KPMG EU Flows Model due to adjustments to inflows from Estonia and Lithuania (2) PMI EPS Q2 2006 and Q2 2007  Inflows from Estonia and Lithuania have been uplifted to bring them into line with 2006 EPS and ND(L) research results

results suggest outflows  The rate of change in the GCTS smoking prevalence and average smoking from Latvia have become volume figures provide a useful indication of observed changes in smoking habits. However, GCTS data cannot be used to quantify total 5 increasingly important consumption in a market due to exclusions of particular age cohorts and the problem of under-reporting of consumption relative to inflows and that 4 LDS have  The KPMG consumption index is calculated using the incidence of non-domestic inflows have increased since 2000 while smoking and the average daily consumption (both from GCTS) and the 3 declined consumption population of each market (from Euromonitor GMID) has been stable − 2 4.2 these figures are then indexed to the earliest year for which we have 2007 consumption both GCTS and LDS data available, 2000 for Latvia modelling results suggest Volume (billion sticks) 1  If LDS trends differ from those shown by the consumption index, it an increase in net outflows implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period which is consistent with 0 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends EU Flows Model results and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2000-2004 2004-2007 2000-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS 10.6% 8.4% 9.7% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index 1.5% (1.9)% 0.0% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be

Sources: (1) KPMG Consumption Index model present in the first analysed year (2) In Market Sales supplied by PMI  Any changes in GCTS (methodology, supplier, sample size and selection, (3) KPMG EU Flows Model etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Latvia increased between 2000 and 2007  Consumption has been stable since 2000 while LDS increased  This result suggests that outflows from Latvia have become increasingly important relative to inflows i.e. non-domestic consumption has declined and/or outflows have increased − this is consistent with 2007 EU Flows Model results showing relatively large outflows from Latvia (0.9 billion sticks) compared with inflows (0.4 billion sticks)  Consumption increased at a slower rate relative to legal sales in 2007 suggesting that net outflows increased − this is consistent with EU Flows Model results showing an increase in outflows from 0.30 billion sticks in 2006 to 0.50 billion sticks in 2007

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Lithuanian consumption Non-domestic inflows to 2006-2007(1)(2) Lithuania fell by 0.7 billion sticks in 2007 to 1.7 billion 100% sticks 25.7% 80% 40.8% The main driver for this 1.5% 60% decline was a reduction in 1.8% C&C ND(L) C&C inflows, most notably 40% 72.8% 57.4% LDC from Russia 20% Outflows

0% Share of total of Share consumption (16.0)% (11.9)% (20%) 2006 2007

Volume (bn sticks): 5.6 6.2

Total non-domestic inflows (ND(L) and C&C) by origin (1)(2) Key:  Lithuania  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 2.5 (2) Synovate ND(L) research, 2007 0.60 2.0 Proportion of total EU levels(1)(2) 0.15 3.0% 0.08 Other countries 2.6% 1.5 0.32 0.06 2.5% 0.07 Poland 1.0 Latvia 2.0% 1.9% 1.57 Russia

1.5% Volume (bn sticks) 1.24 1.5% 0.5

0.9% 1.0% 0.0 2006 2007 Percentage of EU total 0.5% 0.3% Volume (bn sticks): 2.40 1.70 0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal cigarettes sales have 2000-2007(1)(2) trended upwards since 2004 despite rising prices, but they remain below 8 7.2 4 7 1999 levels (LTL) per 20 sticks Price 5.7 6 5.4 5.3 3 4.8 5 4.1 3.9 4 2 3.1 3.3 3

Volume (bn sticks) 2 1

1 0 0 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1999-2004 2004-2007  Legal domestic sales (15.5)% 19.3%

 Average pack price n/a 1.6%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Legal domestic sales increased by 27% in 2007 despite an increase in average weighted prices of 7.8%  Legal sales increased by 19.3% per year between 2004 and 2007 but remain significantly below the 1999 peak level  Taxes were increased by 30% in March 2007, the first increase since 2005 − a similar increase is expected in March 2008  Tougher anti-smoking legislation was introduced in January 2007 including a smoking ban in bars, cafes and restaurants Tax Stamp data from the Lithuanian Ministry of Finance has not been used to determine annual legal domestic sales because annual trends do not provide a realistic view of sales to consumers due to the impact of trade loading on annual volumes  Trade loading occurs when retailers and/or wholesalers purchase in bulk prior to tax increases and then run down these stocks after the increase has been implemented

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Lithuania and key €4.30 destination markets €8.16 €4.94 €1.79 increased during 2007 €0.86 €1.47 Marlboro prices in €4.24 €1.33 €0.86 Lithuania are now cheaper €7.05 n/a €4.71 than in Latvia €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Lithuania and key markets(2)(a) Key:  Lithuania  Major source country  Major destination country

600% 505% 483% 500% 430% 400% 400% July 2006 300% July 2007 200%

58% 100% 35% 11% 0% (8)% (35)% Price premium/(discount) versus local (100%) (39)% UK Ireland Poland Latvia Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Lithuanian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 6.2 billion Outflows Inflows 0.7 billion sticks 1.7 billion sticks sticks versus legal sales of

5.3 billion sticks giving a 7 net inflow of 1.0 billion sticks in 2007 0.25 6 0.06 0.06 0.07 0.01 0.00

0.31 1.24 5 0.15 0.09 0.18 0.01 Outflows 4 C&C ND(L)

6.23 3 Volume (bn sticks) 5.27

4.53 2

0 LDS Poland UK Latvia Other LDC Russia Latvia Poland Other Consumption

Share of non-domestic (legal) by origin Non-domestic (legal) by origin Legal inflows account for a 2007(1)(2)(3) 2006-2007(1)(2)(3) small share of total non- 0.12 Germany domestic consumption at 0.10 12% 5% in 2007 0.10 0.09

0.08 Latvia 10% 0.07 0.06 0.07 Other countries UK 0.04 UK Latvia Volume (bn sticks) 0.01 7% 0.02 0.01 Germany 0.01 0.01 0.01 0.01 0.00 Other countries 2006 2007 71% ND(L) share of 2% 1% consumption: ND(L) share of non- 4% 5% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.12 Population Propensity Propensity Ave. no. purchases ND(L) 0.10 19+ to travel to purchase of trips (sticks) (sticks) 0.09 0.10 Values 0.02 x x x x = 0.09bn 2007 2.8m 44% 34% 2.0 111 0.08 0.02 0.04 0.06 EU rank Non-PMI 0.02 2007 21 14 5 26 17 23 0.04 Other PMI 0.07 Volume (bn sticks) Marlboro 0.02 0.03

0.00 2006 2007 PMI share of ND(L): 81% 57%

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Lithuania 2006-2007(1)(2)(3) 2007(1)(2)(3) decreased by 30% in 2007 2.5 2.30 Other countries 19% Russia remains the main source of C&C inflows 2.0 0.66 1.60 0.07 Other countries 1.5 0.30 Latvia 0.06 Latvia 4%

1.0 Russia 1.57

Volume (bn sticks) 1.24 0.5

0.0 Russia 2006 2007 77% C&C share of 41% 26% consumption: C&C share of non- 96% 95% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3) KPMG’s 2007 non-domestic estimate is lower than the EPS estimate to Estimates of non-domestic reflect the relatively rapid decline in non-domestic incidence apparent in consumption in Lithuania Lithuania since 2006 suggest a significant and  The reduction in non-domestic inflows helps give a more realistic implied continuing decline from 70% consumption increase for Lithuania 57.6% 2005 levels 60%  Lithuanian Customs has not been able to quantify the extent of non- 50% domestic consumption in the country, but believes that anecdotal 42.6% evidence supports the downwards trend shown by Project Star results 40% 32.5% Non-domestic incidence based on PMI EPS results may be overstated 27.2% 30% due to timing issues 20%  In 2007, EPS research was conducted in April-May versus September- October in 2006 Share of total of Share consumption 10% − as a result, EPS results only reflect 7 months of trends 0% House of EU Flows PMI EPS 2007 EU Flows  Therefore, the EU Flows Model has used adjusted EPS results based on Prince 2005 Model 2006 Model 2007 an assumption of a continuation of the trend over the full 12 month period Analysis of the implied consumption trend in Lithuania suggests that

Sources: (1) KPMG EU Flows Model PMI EPS results may overstate actual non-domestic inflows (2) House of Prince, quoted from “Appearance of Cheap cigarettes  Implied consumption in 2007 would be 23% higher than 2006 levels using facilitates growth of tobacco market”, ELTA, 30 March 2006 (3) PMI EPS Q3 2006 and Q2 2007 unadjusted EPS results showing a non-domestic incidence of 32.5% − this increase in consumption is driven by a 28% increase in legal sales, combined with lower outflows and a reduction in inflows of 7%  Although GCTS data and estimated consumption analysis both show increases in consumption in 2007, the magnitude of change is significantly lower − see following two pages for more detail

results suggest non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in 9 smoking habits. However, GCTS data cannot be used to quantify total been declining since 2004 consumption in a market due to exclusions of particular age cohorts and 8 the problem of under-reporting of consumption 7 Gap between 2007 consumption LDS and  The KPMG consumption index is calculated using the incidence of 6 consumption smoking and the average daily consumption (both from GCTS) and the modelling results show a increased up 5 population of each market (from Euromonitor GMID) continuation of this trend until 2004 but 4 has since − narrowed these figures are then indexed to the earliest year for which we have and therefore support EU 3 6.2 both GCTS and LDS data available, 1999 for Lithuania

Flows Model results Volume (billion sticks) 2  If LDS trends differ from those shown by the consumption index, it 1 implies that the relationships between inflows, outflows and domestic 0 consumption have changed over the analysed period 1999 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 1999-2004 2004-2007 1999-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS (15.5)% 19.3% (0.3)% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index 2.7% (2.4)% 0.5% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be

Sources: (1) KPMG Consumption Index model present in the first analysed year (2) In Market Sales supplied by PMI  Any changes in GCTS (methodology, supplier, sample size and selection, (3) KPMG EU Flows Model etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Lithuania widened between 1999 and 2004, but has since narrowed  Between 1999 and 2004, consumption increased while LDS decreased suggesting strong growth in non-domestic inflows during this period  Between 2004 and 2007, consumption decreased while LDS increased. This suggests than non-domestic consumption has declined over the last three years − this is consistent with PMI EPS results which show a 10 percentage point decline in non-domestic share of total consumption between Q3 2006 and Q2 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1999-2007(1)(2)(3)(a) prices affect cigarette consumption in Lithuania that changes in  Historical data from 1999 to 2006 has been used to estimate the strength consumption in Lithuania of relationships between consumption and prices and income 8 are related to changes in Control period Forecast − estimated incidence of smoking and average daily consumption are 7 disposable income and based on GCTS data. Population and disposable income are estimated 6 using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices 5 Cigarette prices are based on the weighted average pack price supplied by PMI Estimated consumption 4 KPMG analysis suggests that changes in consumption in Lithuania are results show a decline in 3 related to changes in disposable income and cigarette prices(c) 2

non-domestic Volume (billion sticks)  Disposable income is positively related to changes in consumption i.e. as consumption in 2007 1 income rises consumption tends to increase 0 which is consistent with  Cigarette prices are, by contrast, negatively related to changes in 1999 2000 2001 2002 2003 2004 2005 2006 2007 EU Flows Model results consumption  It is important to note that factors other than price and income are likely to CAGR (%) 1999-2004 2004-2006 2006-2007 impact consumption levels in Lithuania  LDS (15.1)% 16.1% 28.0% − examples of other factors likely to impact consumption levels include  Estimated consumption(b) 2.5% (2.3)% 4.5% changes in smoking restrictions and changes in demographic profiles  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Relationship based on historical data from 1999 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 considered a broad approximation Project Star estimate of consumption for Lithuania (5.6 billion sticks) (c) 55% of changes in consumption can be explained by changes in disposable The relationships between changes in consumption and changes in income and cigarette prices disposable income and prices have been used to estimate 2007 Sources: (1) KPMG EU Flows Model consumption (2) In Market Sales supplied by PMI (3) KPMG Consumption Regression Model  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is higher than in 2006 which is consistent with EU Flows Model results  Estimated consumption increased at a slower rate relative to LDS in 2007 suggesting a decline in non-domestic consumption − this is consistent with EU Flows Model modelling results  Non-domestic consumption also declined in 2006 after increasing between 2001 and 2005

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Luxembourg consumption Luxembourg has low 2006-2007(1)(2)(b) inflows and high outflows 3.7% 3.7% 100% due to its low prices 7.5% 7.6% relative to surrounding 80% countries 60% C&C Non-domestic inflows to 88.8% 88.7% ND(L) 40% Luxembourg were stable LDC in 2007 at 0.11 billion 20% Outflows

sticks total of Share consumption 0% (407)% (442)% Outflows from -20%(450)% Luxembourg increased in 2006 2007

2007 and are equivalent to Volume (bn sticks): 0.96 0.95 more than four times domestic consumption Total non-domestic inflows by origin (ND(L) and C&C)

Key:  Luxembourg  Major destination country 2006-2007(1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 0.14 (b) ND(L) and EPS research has not been updated for 2007 but corroborating research has not indicated any significant changes to 2006 results 0.12 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2006 0.10 0.031 0.031 Proportion of total EU levels(1)(2) 0.08 Other countries 10% 0.012 0.012 8.7% Duty Free 0.06 0.015 0.015 8% Germany

Volume (bn sticks) 0.04 0.018 0.018 France 6% Belgium 0.02 0.032 0.032

4% 0.00 2006 2007

Percentage of EU total 2% Volume (bn sticks): 0.11 0.11 0.1% 0.1% 0.2% 0.1% 0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Changes to external 1997-2007(1)(2) demand drivers (e.g. price 8 4 differential versus 7 6.5 neighboring countries) 6.0 5.8 (Euro) per 20 sticks Price 6 5.6 5.6 3 have driven fluctuations in 5.2 5.2 5.0 5.0 4.8 4.8 legal domestic sales in 5

Luxembourg 4 2

3 Volume (bn sticks) 2 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2004 2004-2007  Legal domestic sales 5.2% 1.9% (8.0)%  Average pack price 5.6%(a) 6.2% 7.0%

Note: (a) Average pack price CAGR is from 1998 - 2000 Sources: (1) Tax stamp data supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales have increased by 5.7% from 2006 to 2007, despite prices increasing by 11.4% over the same period  Stricter anti-smoking legislation was introduced to Luxembourg in September 2006 which covered smoking in public places including restaurants − some exceptions are permitted such as smoking rooms in restaurants (if these account for less than 25% of the total area of the venue) and cafes which have a ban in place during dining hours

Marlboro price comparison Prices in Luxembourg are 1 July 2007(1)(2)(a) lower than in neighboring €4.30 countries €8.16 €1.79 €4.94 €0.86

€1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10 €4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66

€4.10 €1.64

€3.15 €2.95 €3.00 €2.60 Evolution of Marlboro price differential between Luxembourg and key markets(2)(a)(b)

Key:  Luxembourg 160%  Major destination country 117% 120% 112% 91% 81% 80% 43% 35% 28% 40% 27% 22% July 2006 14% 14%11% July 2007 0%

(20)% (40)% (21)% (43)% (51)% (80)% Price premium/(discount) versus local UK Spain Ireland France Poland Belgium Germany Netherlands

Notes: (a) Based on pack of 20 Marlboro King Size on 1st July 2006 and 2007 (b) Marlboro prices increased in France by 30 cents in August 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling prices supplied by PMI

Consumption breakdown Luxembourg consumption 2007(1)(2)(3)(4)(a)(b) is estimated at 1.0 billion Outflows Inflows 4.2 billion sticks 0.1 billion sticks sticks versus legal domestic sales of 5.0 6 billion sticks in 2007

5 Outflows

0.9 C&C ND(L)

4 0.38

0.29 0.25

5.0 Volume (bn sticks)

2 2.24

1 0.04 0.10 0.03

0.8 0.95

0 LDS France Germany UK Belgium Consumption Other LDC Belgium Other Consumption adjustment (see next page) 0.04

Consumption estimate The scale of outflows from Luxembourg and the small size of the Our estimate of total 2006(1)(2)(3)(b) domestic market make an exact measurement of domestic consumption consumption in very challenging

Luxembourg uses the 1.2  The estimation of outflows from Luxembourg requires the measurement same methodology as of flows from Luxembourg into all other countries 1.0 0.11 11% non- − these inflows are particularly difficult to measure as they are likely to 2006 domestic 0.8 0.36 be concentrated geographically within markets and are relatively small compared to the domestic markets of the destination countries It is based on survey data 0.01 0.6 0.06 and the non-domestic  As a result, this approach is likely to underestimate the volume of flows 0.90 out of Luxembourg and hence overstate consumption share of the market 0.4 0.58 Volume (billion sticks) We have used consumer survey data to adjust our estimate of relative to neighbouring 0.2 consumption in 2006 countries is consistent 0.0  We have used consumer survey data to estimate domestic consumption with the market dynamics Luxembourg Cross-border International Under- Total by residents, commuters and visitors to Luxembourg residents commuters tourists reporting consumption of these different countries uplift − we estimated the extent of under-reporting in the survey data at 40% by performing the same calculation for Belgium and the Netherlands, Total Domestic Non-domestic and comparing the result to our own consumption estimates  Applying an uplift for under-reporting implies total consumption of 1.01 billion sticks Consumption calculation (2006)(b) We have applied the results of the empty pack survey in Luxembourg to Luxembourg Belgium Netherlands our revised consumption estimate

Population 18+ (million)(1) 0.4 8.6 13.2  The empty pack survey results for Luxembourg imply a non-domestic market share of 11%, equivalent to 110 million sticks Smoking incidence(1) 29% 20% 19%  There are no external estimates of the size of the non-domestic market in Luxembourg Avg. daily consumption(1) 15.4 14.7 13.8 − however, this figure is substantially below that for Belgium and the Netherlands, which is consistent with its low price point relative to Implied consumption 0.7 8.7 12.0 (a)(2) neighbouring countries and it being a low priority target for smugglers, (billion sticks) who are likely to favour larger, higher priced markets KPMG consumption n/a 13.2 19.0 (billion sticks)(3)

Implied understatement 36% 37% 35% Notes: (a) Implied consumption for Luxembourg includes 0.07 billion sticks attributed to international commuters and tourists (b) This analysis has not been updated for 2007 Consumption estimate 1.01 n/a n/a Sources: (1) Attitudes of Europeans towards tobacco, Eurobarometer, 2006 (billion sticks) (2) http://ec.europa.eu (3) KPMG EU Flows Model

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows account for 2007(1)(2)(3) 2007(1)(2)(3) around two thirds of non- 0.08 0.073 Other countries domestic consumption in 10% 0.07 0.007 Luxembourg in 2007 0.06 0.015 0.05 Other countries Germany 0.04 0.018 Belgium Germany 21% 44% 0.03 France

Volume (bn sticks) 0.02 Belgium 0.032 0.01 0.00 2007 France ND(L) share of consumption: 8% 25% ND(L) share of non-domestic: 67%

Non-domestic (legal) by brand (1)(2)(3) 2007 ND(L) analysis Ave. Total 0.08 0.073 Population Propensity Propensity Ave. no. purchases ND(L) 0.07 19+ to travel to purchase of trips (sticks) (sticks) Values 0.06 x x x x = 0.073bn 2007 0.4m 88% 54% 2.0 208 0.05 0.045

0.04 Non-PMI EU rank 0.03 26 1 1 25 11 24 0.007 Other PMI 2007 Volume (bn sticks) 0.02 Marlboro 0.01 0.021

0.00 2007

PMI share of ND(L): 38%

Note: ND(L) and EPS research was not updated in 2007 but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interview with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows account for a 2007(1)(2)(3) 2007(1)(2)(3) relatively low percentage 0.04 of total consumption in 0.035 Spain Luxembourg 28% 0.03

Spain is a key source of 0.022 Other countries C&C inflows to 0.02 Poland Luxembourg 0.004 Spain

Volume (bn sticks) 0.01

0.010 Poland 0.00 10% Other countries 2007 62% C&C share of consumption: 4% C&C share of non-domestic: 33%

OTP sales in billion stick equivalents OTP sales decreased between 2003 and 2006 but increased in 2007 Smoking tobacco accounts 2003-2007(1)(a)(b) 4.9  Total OTP sales increased by 24.1% in 2007 for a relatively large 5 4.1 0.1 proportion of tobacco 4.1 4.0 3.9  Smoking tobacco has had the highest volume growth among the major OTP categories since 2003 and now accounts for 99% of total OTP sales 4 0.1 0.1 0.1 sales in Luxembourg at 4.9 0.1 (excluding smokeless tobacco)

billion stick equivalents in 3 OTP categories account for a relatively large share of overall tobacco 2007 4.8 sales in Luxembourg 2 4.0 4.0 4.0 3.9  Total cigarette sales are estimated at at 5.0 billion sticks(2) versus 4.9 billion sticks for OTP (excluding smokeless tobacco)

Stick equivalents (bns) 1

0 2003 2004 2005 2006 2007

Cigarettes 5.6 6.5 5.2 4.8 5.0 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos (1.5)% (24.7)% 17.9% 3.7%  Smoking tobacco (0.4)% (2.2)% 24.2% 4.7% Total OTP (0.4)% (2.7)% 24.1% 4.6%

Manufactured cigarettes (3.8)% (8.3)% 5.7% (2.6)%

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Maltese consumption Non-domestic inflows to 2006-2007(1)(2)(b) Malta were stable in 2007 at 0.04 billion sticks 100% 7.3% 6.5% 0.5% 1.3% ND(L) increased from 0.5% 80%

to 1.3% of consumption 60% C&C while C&C declined 92.2% 92.2% ND(L) 40% slightly LDC 20% Outflows Outflows from Malta 0% Share of total consumption increased in 2007, with the (6.5)% (7.5)% UK and Italy remaining the (20%) 2006 2007 main destination countries

Volume (bn sticks): 0.542 0.536

Total non-domestic inflows (ND(L) and C&C) by origin Key:  Malta  Major destination country 2006-2007(1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow (b) ND(L) and EPS research has not been updated for 2007 but corroborating 0.05 research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 0.04 (2) Synovate ND(L) research, 2006 0.008 0.008 Proportion of total EU levels(1)(2) 0.005 0.005 0.03 Other countries 0.10% 0.08% Italy 0.02 0.08% 0.07% 0.030 0.030 Duty Free Volume (bn sticks) 0.06% 0.01 0.06% 0.05%

0.04% 0.00 2006 2007 0.02% 0.02% Percentage of EU total Volume (bn sticks): 0.042 0.042 0.00% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Cigarette sales in Malta 1997-2007(1)(2) increased slightly in 2007 but remain below 2004 0.7 2.0 levels 0.635 Lira) (Maltese per 20 sticks Price 0.580 0.580 0.580 0.580 0.570 0.560 0.573 0.6 0.550 0.535 0.535 1.5 0.5

0.4 1.0 0.3

0.2 Volume (bn sticks) 0.5 0.1

0.0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2003 2003-2007  Legal domestic sales (0.9)% (0.7)%

 Average pack price n/a (0.2)%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Country management Legal domestic sales increased slightly 2007 while prices decreased by 1.8% on an average weighted pack basis  Coinciding with Malta’s accession to the EU in 2004, the government abolished the discriminatory excise tax regime for imported versus locally produced cigarettes  New anti-smoking legislation was introduced in 2005 that requires bars, restaurants and cafes to allocate designated smoking areas on their premises

Marlboro price comparison Cigarette prices in Malta 1 July 2007(1)(2)(a) are significantly lower €4.30 than the UK, the main €8.16 €1.79 destination market for €4.94 €0.86 cigarette outflows €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Malta and key markets(2)(a) €3.49 €3.86

150% 130% Key:  Malta 112%  Major destination country 100%

50% July 2006 15% 17% July 2007 0%

(20)% (14)% (50%)

(78)% (75)% Price premium/(discount) versus local (100%) UK Italy Greece Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Maltese consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 0.54 billion Outflows Inflows 0.04 billion sticks 0.04 billion sticks sticks versus legal sales of 0.6 0.54 billion sticks, implying outflows equivalent to

inflows in 2007 0.03 0.03 0.5 0.01 0.00 0.00

Outflows C&C ND(L) 0.4

0.3

0.54 0.54 0.49 Volume (bn sticks)

0.2

0.1

0.0 LDS UK Other LDC Italy Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin In 2007, non-domestic 2006-2007(1)(2)(3) 2007(1)(2)(3) (legal) flows reached 7 0.008 million sticks 0.007 0.007 Other countries 22% 0.006 0.001 0.005 0.001 0.004 0.003 Other countries 0.003 0.005 Duty Free Duty Free Volume (bn sticks) 0.002 0.001 8% Italy 0.001 0.001 0.001 0.000 Italy 2006 2007 70%

ND(L) share of 1% 1% consumption: ND(L) share of non- 7% 16% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.008 Population Propensity Propensity Ave. no. purchases ND(L) 0.007 0.007 19+ to travel to purchase of trips (sticks) (sticks) Values 0.006 x x x x = 0.007bn 2007 0.3m 30% 8% 4.6 196 0.005 0.004 0.006 Non-PMI 0.003 Other PMI EU rank 0.003 Marlboro 2007 27 21 27 7 12 27

Volume (bn sticks) 0.002 0.002 0.000 0.001 0.000 0.001 0.001 0.000 2006 2007

PMI share of ND(L): 20% 20%

Note: ND(L) inflows to Malta are higher in 2007 compared with 2006 due to a change in treatment for inflows from Italy. In 2007, all inflows from Italy are assumed to be legal to reflect the fact cigarettes prices in Italy are higher than Malta Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows were 35 2006-2007(1)(2)(3) 2007(1)(2)(3) million sticks in 2007 0.05 Other countries 11% 0.040 Russia 0.04 0.035 6% 0.008 0.004 0.03 0.002 0.002 Other countries

0.02 Russia 0.029 0.029 Duty Free Volume (bn sticks) 0.01

Duty Free 0.00 83% 2006 2007 C&C share of 7% 7% consumption: C&C share of non- 93% 84% domestic:

Note: C&C inflows to Malta are lower in 2007 compared with 2006 due to a change in treatment for inflows from Italy. In 2007, all inflows from Italy are assumed to be legal to reflect the fact cigarettes prices in Italy are higher than Malta Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(a) KPMG estimate of non-domestic consumption is stable in 2007 and in KPMG estimates of non- line with that of Interbrands domestic incidence are Interbrands 12% provided a range  Interbrands’ 2007 estimate is for C&C volumes only and is comparable broadly consistent with 10.0% of 5% to 6% with a level of C&C consumption of 6.5% from the EU Flows Model 10% those of Interbrands (PMI − Interbrands are PMI’s distributor in Malta distributor 7.8% 7.8% Malta’s local distributor) 8% 5.5% − the estimate provided in 2007 was a range of 5-6% for contraband 6% cigarettes only

4% The number of cigarettes seized remained flat at 26.2 million sticks in 2007(3)

Share of total of Share consumption  Counterfeit cigarettes accounted for 21.3 million sticks of the total seizure 0% volume(3) Interbrands EU Flows Interbrands EU Flows 2006 Model 2006 2007 (C&C Model 2007 only)

Note: (a) EPS research was not updated in 2007 but corroborating research has not indicated any significant changes to 2006 results Sources: (1) KPMG EU Flows Model (2) Interviews with Interbrands (PMI distributor) (3) Interview with Fiscal and Customs officials

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Dutch consumption Non-domestic inflows to 2006-2007(1)(2) the Netherlands were broadly flat in 2007 at 5.4 100% 15.6% billion sticks 24.2% 80% 12.5% ND(L) increased from 4.6% 4.6% 60% C&C to 12.5% of total ND(L) 40% consumption while C&C 71.2% 71.9% LDC

declined significantly 20% Outflows

0% Share of total consumption (3.0%) (4.3%) (20%) 2006 2007

Volume (bn sticks): 19.0 19.1

Total non-domestic inflows by origin (ND(L) and C&C) Key:  Netherlands  Major source country  Major destination country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 6.0 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 5.0

Other countries (1)(2) 4.0 2.78 2.65 Proportion of total EU levels 7.9% 8% Duty Free 3.0 Italy 5.9% 0.83 0.80 Spain 6% 2.0 0.45

4.9% Volume (bn sticks) 0.65 Belgium 0.72 1.0 0.66 4% 0.54 0.73 0.0 2.6% 2006 2007 1.7% 2%

Percentage of EU total Volume (bn sticks): 5.5 5.4

0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales A major price increase in 1997-2007(1)(2) 2004 resulted in a significant drop in 20 6 domestic sales 17.0 16.6 16.5 16.7 16.3 16.9 15.6 5 (Euro) per 20 sticks Price 16 14.8 14.1 14.5 13.5 4 12 3 8 2 Volume (bn sticks) 4 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2003 2003-2005 2005-2007  Legal domestic sales 0.4% (10.8)% 3.5%  Average pack price 4.6% 10.6% 1.0%

Sources: (1) Tax stamp data supplied by PMI is used for historic data; IMS data is used for 2007 (2) Weighted average pack price supplied by PMI Legal domestic sales have increased by 2.8% in 2007 compared to 2006, despite increases in weighted average prices of 1.7%

Marlboro price comparison Marlboro prices in Belgium 1 July 2007(1)(2)(a) in 2007 were higher than €4.30 in the Netherlands €8.16 €4.94 €1.79

€1.47 €4.24 €1.33 €0.86

€7.05 €4.11 n/a €4.71 €8.04 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66

€4.10 €1.64

€3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Netherlands and key markets(2)(a) €3.50 €3.86 25% 30% 22% Key:  Netherlands  Major source country 10%  Major destination country 10% 0% 0% July 2006 July 2007 (10%) (0)% (10)% (13)% (30%) (28)% (31)%

(50%) (49)%

(57)%

Price premium/(discount) versus local (70%) Italy Spain Poland France Belgium Luxembourg

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Dutch consumption is 2007(1)(2)(3)(4)(a)(b)(c) estimated at 19.1 billion Outflows Inflows 0.8 billion sticks 5.4 billion sticks sticks versus legal sales of 20 14.5 billion sticks, giving a net inflow of 4.6 billion 18 2.11 sticks

16 0.99 0.35 0.40 0.46 0.06 0.26 0.26 14 0.17 0.73 0.380.0

12 “Other” includes Duty Free inflows 10 Outflows 19.1 C&C ND(L) 8 Volume (bn sticks) 14.5 13.7 6

0 LDS France UK Other LDC Belgium Spain Italy France Other Consumption

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) (c) Inflows from Belgium have been adjusted downwards by 0.3 billion sticks to 0.7 billion sticks based on analysis and consumption, legal sales and pricing trends in the two countries. Inflows from “Other countries” have been increased by an equivalent amount of 0.3 billion sticks. As a result, total non-domestic levels remain the same Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) IMS data supplied by PMI (4) Interviews with PMI Local Management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin In 2007, non-domestic (1)(2)(3)(a) 2006-2007 2007(1)(2)(3)(a) (legal) inflows were 2.4 3.0 billion sticks 2006 ND(L) 2.38 Other countries 2.5 including overseas 21% Belgium Non-domestic (legal) visitor adjustment 0.50 2.0 would have been Other countries 30% inflows have increased in 1.33 billion sticks 0.25 UK 1.5 0.26 2007 partly as a result of a Spain 0.29 0.87 change in the treatment of 1.0 Germany 0.35 UK overseas visitor inflows Volume (bn sticks) 0.42 France 11% 0.5 0.01 0.20 0.73 Belgium 0.08 0.04 0.0 0.11 2006 2007 Spain France 11% 15% ND(L) share of Germany 5% 13% consumption: 12% ND(L) share of non- 16% 44% domestic:

Non-domestic (legal) by brand (b) 2006-2007(1)(2)(3)(a) ND(L) analysis 3.0 Ave. Total Population Propensity Propensity Ave. no. purchases ND(L) 2.38 2.5 19+ to travel to purchase of trips (sticks) (sticks)

Non-PMI Values 2.0 12.7m x 73% x 35% x 2.2 x 233 = 1.7bn(b) 1.12 Other PMI 2007 1.5 Marlboro 0.87 0.15 1.0 EU rank 8 2 Volume (bn sticks) 2007 4 23 8 4 0.42 1.11 0.5 0.06 0.39 0.0 2006 2007

PMI share of ND(L): 52% 53%

Notes: (a) ND(L) inflows have increased in 2007 partly as a result of changes in treatment of inbound tourism flows. In 2006, Sources: (1) KPMG EU Flows Model had the same methodology been applied, ND(L) including overseas inbound visitor flows would have been (2) Synovate ND(L) research, 2007 approximately 1.33 billion sticks (3) Interviews with PMI Local Management (b) ND(L) analysis from Synovate excludes adjustments for inbound tourism inflows. Total ND(L) including these inbound tourism inflows in 2.38 billion sticks

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to the 2006-2007(1)(2)(3)(a) 2007(1)(2)(3)(a) Netherlands stood at 3.0 5 4.59 billion sticks in 2007 4 Duty Free 2.61 2.97 21% 3 Other countries Other countries Poland 43% 1.29 2 0.19 Italy 0.20 0.63 Spain

Volume (bn sticks) 0.40 1 0.45 0.46 Duty Free Spain 0.70 0.62 16% 0 2006 2007 C&C share of 24% 16% Italy consumption: “Other countries” includes 0.88 billion Poland 13% C&C share of non- 7% 84% 56% sticks from non-EU domestic: countries

Note: (a) C&C inflows have decreased in 2007 partly as a result of changes in treatment of inbound tourism flows. In 2006, had the same methodology been applied, C&C inflows would have been approximately 4.13 billion sticks Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interview with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(a) Results of Netherlands Customs’ empty pack surveys in late 2005 KPMG estimates of non- showed non-domestic consumption of 23% domestic consumption are 35%  The estimate is based on an incidence of non-domestic packs of 18% plus broadly in line with those 28.8% a further 5% share accounted for by contraband product 30% 28.1% of Netherlands Customs − Early findings in 2008 suggest a decline in counterfeit incidence based 23.0% 25% on pack collections in football stadiums and schools 20%  There are no other external estimates of non-domestic incidence for the 15% Netherlands

10%

Share of total of Share consumption 5%

0% Customs 2005 EU flows model EU Flows Model 2006 2007

Note: (a) Netherlands Customs 2005 estimate revised based on feedback following interview with relevant parties in April 2008 Sources: (1) KPMG EU Flows Model (2) PMI EPS Q3 2006 and Q3 2007 (3) KPMG interview with Netherlands Customs

Consumption  The KPMG consumption index is calculated using the incidence of This is consistent with EU 15 has declined smoking and the average daily consumption (both from GCTS) and the Flows Model results and since 2005 while LDS have population of each market (from Euromonitor GMID) estimated consumption 10 19.1 increased − these figures are then indexed to the earliest year for which we have

analysis Volume (billion sticks) 5 both GCTS and LDS data available, 1997 for Netherlands  If LDS trends differ from those shown by the consumption index, it

Sources: (1) KPMG Consumption Index model changed over the analysed period (2) In Market Sales supplied by PMI − in many markets, some level of inflows or outflows are likely to be (3) KPMG EU Flows Model present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated Consumption index modelling results suggest that the consumption and legal sales trend lines were broadly similar up until 2005  This result suggests that non-domestic consumption was reasonably stable between 1997 and 2005 However, since 2005 legal sales have been increasing while consumption has continued to decline  This suggests that non-domestic consumption has declined since 2005 which is consistent with EU Flows Model results and estimated consumption analysis (see next page for more detail)

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1997-2007(1)(2)(3)(a) prices affect cigarette consumption in the Netherlands that changes in  Historical data from 1997 to 2006 has been used to estimate the strength consumption in the of relationships between consumption and prices and income 30 − estimated incidence of smoking and average daily consumption are Netherlands are related to Control period Forecast based on GCTS data. Population and disposable income are estimated changes in disposable 25 using Euromonitor GMID and Oxford Economic Forecasting data. income and cigarette Cigarette prices are based on the weighted average pack price supplied 20 by PMI prices 15 KPMG analysis suggests that changes in consumption in the Netherlands are related to changes in disposable income and cigarette prices(c) Estimated consumption 10  Disposable income is positively related to changes in consumption i.e. as

results show a slight Volume (billion sticks) 5 income rises consumption tends to decrease decline in non-domestic  Cigarette prices are, by contrast, negatively related to changes in consumption in 2007 0 consumption 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 which is consistent with  It is important to note that factors other than price and income are likely to impact consumption levels in the Netherlands EU Flows Model results − examples of other factors likely to impact consumption levels include CAGR (%) 1997-2001 2001-2006 2006-2007 changes in smoking restrictions and changes in demographic profiles  LDS (1.2)% (3.2)% 2.6%  Given that we have not assessed the impact on consumption from factors  Estimated consumption(b) 2.2% (3.8)% 2.0% other than price and income, estimated consumption can only be considered a broad approximation

Notes: (a) Relationship based on historical data from 1997 to 2006 The relationships between changes in consumption and changes in (b) The estimated consumption line has been rebased so that it equals the 2006 disposable income and prices have been used to estimate 2007 Project Star estimate of consumption for the Netherlands (19.0 billion sticks) consumption (c) 88% of changes in consumption can be explained by changes in disposable income and cigarette prices  Combining estimates for the strength of these relationships with actual Sources: (1) KPMG EU Flows Model 2007 values for disposable income and prices gives estimated 2007 (2) In Market Sales supplied by PMI consumption (3) KPMG Consumption Regression Model  Estimated 2007 consumption is higher than in 2006 which is consistent with EU Flows Model results  Estimated consumption increased at a slightly slower rate than LDS in 2007 suggesting non-domestic consumption declined − this is consistent with EU Flows Model modelling results showing a slight fall in non-domestic consumption  Estimated consumption results show relatively stable non-domestic share between 1997 and 2007 which is consistent with consumption index modelling results up until 2005

OTP sales in billion stick equivalents Legal sales of OTP have declined by 4.8% per year since 2003 Sales of smoking tobacco 2003-2007(1)(a)(b)  Legal sales of manufactured cigarettes declined by 3.9% per year over have been declining since the same time period 2003 and stood at 14.6 20.0 17.9 0.4 16.5 Smoking tobacco accounts for 98% of OTP volumes (excluding billion stick equivalents in 0.5 15.2 15.0 14.6 smokeless tobacco) in the Netherlands 15.0 0.5 0.5 0.3 2007  Cigars and cigarillos have declined by almost 40% in 2007

10.0 17.4 16.1 14.7 14.5 14.4 5.0 Stick equivalents (bns)

0.0 2003 2004 2005 2006 2007

Cigarettes 17.0 14.8 13.5 14.1 14.5 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos 2.0% 4.1% (39.2)% (9.9)%  Smoking tobacco (8.0)% (1.5)% (1.2)% (4.7)% Total OTP (7.7)% (1.3)% (2.4)% (4.8)%

Manufactured cigarettes (10.8)% 4.2% 2.8% (3.9)%

Notes: (a) Smoking tobacco volumes have been calculated at one stick per 0.75 grams, while cigars and cigarillos have been calculated on a stick for stick basis (b) Smokeless tobacco is excluded from this analysis Source: (1) OTP volumes supplied by PMI

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Polish consumption Non-domestic inflows to 2006-2007(1)(2) Poland increased to 5.3

billion sticks in 2007, with 100% 5.7% 7.1% 0.9% 0.9% the majority of the 80% increase attributed to C&C 60% C&C flows ND(L) 93.4% 92.0% 40% LDC Inflows from Ukraine and Outflows 20% Russia, the two largest 0 sources of non-domestic Share of total consumption (14.9)% (14.4)% product, increased by a (20)% total of 1.3 billion sticks in 2006 2007 2007 Volume (bn sticks): 66.9 65.7

Total non-domestic inflows (ND(L) and C&C) by origin Key:  Poland  Major source country  Major destination country 2006-2007(1)(2) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 6 (2) AC Nielsen ND(L) research, 2007 5 1.14 Proportion of total EU levels(1)(2) 4 0.31 1.58 Other countries 10% 1.23 3 Lithuania 0.34 7.9% Russia 2 0.82

Volume (bn sticks) Ukraine 5.9% 2.58 1 4.9% 1.67 5% 0 2006 2007 2.6% 1.7% Percentage of EU total Volume (bn sticks): 4.4 5.3

0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal cigarettes sales have 1997-2007(1)(2) been trending down in Poland since 2005 as 100 92.1 7 prices have risen 90.2 88.0

6 (PLN) per 20 sticks Price 73.1 74.4 75.4 72.8 73.9 80 72.0 72.4 69.9 5

60 4

40 3 2 Volume (bn sticks) 20 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2005 2005-2007  Legal domestic sales (7.4)% 0.2% (2.7)%

 Average pack price 15.5% 5.6% 9.8%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI (3) “Taking advantage of cigarette sales”, Polish News Bulletin, 9th December 2005 (4) “Poland will enact fuel and tobacco excise hikes in January 2007”, PAP Market Insider, 29th November 2006

Legal domestic sales decreased by 3.5% in 2007 while prices increased 14.7%  Legal sales in Poland remain well below 1999 levels  Excise taxes on cigarettes have been increased several times following Poland’s entry into the EU in 2004 − excise tobacco taxes rose by 16% and 13% in January 2006 and January 2007(3)(4)

Marlboro price comparison Price differentials between 1 July 2007(1)(2)(a) Poland and key €4.30 destination markets, €8.16 €4.94 €1.79 including the UK and €0.86 Germany, narrowed €1.47 €4.24 €1.33 slightly in 2007 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

€3.86 Evolution of Marlboro price differential between Poland and key Key:  Poland €3.50 (2)(a) markets  Major source country

 Major destination country 400% 333% 283% 300% 271% 236%

200% 161% July 2006 124% July 2007 100%

Price premium/(discount) versus local (54)% (100)% (63)% (70)% (59)% UK Ireland Germany Ukraine Russia

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Polish consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 65.7 billion Outflows Inflows 9.5 billion sticks 5.3 billion sticks sticks versus legal sales of 80 69.9 billion sticks, giving a net outflow of 4.2 billion 70 sticks in 2007 6.85 1.02 1.22 1.02 0.43 0.37 1.23 2.46 60 0.01 0.12 Outflows

C&C

50 ND(L)

69.9 65.7 30 60.4 Volume (bn sticks)

0 LDS Germany UK Ireland Other LDC Ukraine Russia Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows account for a 2006-2007(1)(2)(3) 2007(1)(2)(3)

relatively low share of 0.7 Germany non-domestic 0.61 0.6 0.57 26% consumption in Poland at 0.5 1% in 2007 0.29 Other countries 0.4 0.43 Lithuania Other countries 0.3 0.01 Ukraine 50% 0.2 0.12 Germany Volume (bn sticks) 0.11 0.1 0.14 0.04 Ukraine 0.03 0.0 22% 2006 2007 Lithuania 2% ND(L) share of 1% 1% consumption: ND(L) share of non- 14% 11% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.7 Population Propensity Propensity Ave. no. purchases ND(L) 0.61 19+ to travel to purchase of trips (sticks) (sticks) 0.6 0.57 Values x x x x = 0.57bn 0.5 2007 30.4m 28% 13% 5.3 99 0.21 0.35 Non-PMI 0.4 Other PMI 0.01 0.3 Marlboro EU rank 0.00 0.15 6 22 25 3 21 10 L&M 2007 0.2 0.06 Volume (bn sticks)

0.1 0.20 0.20

0.0 2006 2007

PMI share of ND(L): 43% 63%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Poland 2006-2007(1)(2)(3) 2007(1)(2)(3) increased by 23% from Other countries 5 4.7 15% 2006 to 2007 3.8 0.7 4 Ukraine and Russia are the 0.3 1.2 Other countries Lithuania two main sources of C&C 3 1.2 Lithuania 6% 0.2 inflows to Poland Russia 2 0.8 Ukraine

Volume (bn sticks) 2.5 1 Ukraine 1.6 53%

0 2006 2007 Russia 26% C&C share of 6% 7% consumption: C&C share of non- 86% 89% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4) A PMI consumer research study undertaken in 2007 suggested a non- KPMG’s estimate of non- domestic incidence of 9% domestic consumption in 16% 15.0% 15.0%  This study was based on researchers viewing the pack of cigarettes 2007 has been 14% currently being smoked by the respondent corroborated by a 12% − of the 2,583 packs shown by consumers during the 2007 consumer (5) consumer survey 10% 9.0% research, 232 packs lacked Polish tax stickers 8.0% undertaken by PMI 8% 6.6% Non-domestic consumption is higher in areas bordering lower-cost 6% cigarette markets 4%  A recent survey conducted by PMI and focusing on areas bordering

Share of total of Share consumption 2% Russia found 51% non-domestic consumption compared to 8% nationally (EU Flows Model) 0% − a similar study in areas bordering Ukraine found 37% non-domestic consumption(6) 2006 TMA 2006 EU Flows EU Flows  The results of this survey have been weighted and included in the EU Model 2006 Model 2007 Dr A Hempel Survey 2007 Flows Model PMI Consumer Other external estimates of non-domestic consumption may include flows destined for consumption in other countries Sources: (1) Tobacco Merchants Association, Smuggling Trends, 2006 (2) KPMG EU Flows Model (3) PMI EPS Q3 2006 and Q3 2007 (4) PMI consumer research (5) Interview with representatives of the Ministry of Finance, 15 January 2008, (6) PMI EPS Border Focus Surveys, Q3 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption index 1997-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market modelling results suggest  The rate of change in the GCTS smoking prevalence and average smoking that non-domestic volume figures provide a useful indication of observed changes in 120 consumption declined smoking habits. However, GCTS data cannot be used to quantify total consumption in a market due to exclusions of particular age cohorts and between 2002 and 2005 100 Consumption the problem of under-reporting of consumption but has been rising over 80 has increased since 2005 while  The KPMG consumption index is calculated using the incidence of the last three years LDS have smoking and the average daily consumption (both from GCTS) and the 60 declined population of each market (from Euromonitor GMID)

40 − these figures are then indexed to the earliest year for which we have 65.7

Volume (billion sticks) both GCTS and LDS data available, 1997 for Poland 20  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1997-2000 2000-2005 2005-2007  Comparing consumption and sales trends helps to understand market  LDS (7.4)% 0.2% (2.7)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.2)% (3.7)% 4.6% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Note: (a) Consumption in 2001 has been estimated as the mid-point between 2000 and 2002 consumption changed over the analysed period Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI − in many markets, some level of inflows or outflows are likely to be (3) KPMG EU Flows Model present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Poland increased between 2005 and 2007  Consumption in Poland has increased since 2005 while LDS have declined  This result suggests that inflows from Poland have become increasingly important relative to outflows i.e. non-domestic consumption has increased and/or outflows have decreased − this is consistent with EU Flows Model results showing an increase in non-domestic consumption in 2007

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1997-2007(1)(2)(3)(a) prices affect cigarette consumption in Poland that changes in  Historical data from 1997 to 2006 has been used to estimate the strength consumption are related of relationships between consumption and prices and income 100 to changes in disposable Control period Forecast − estimated incidence of smoking and average daily consumption are income and cigarette 80 based on GCTS data. Population and disposable income are estimated using Euromonitor GMID and Oxford Economic Forecasting data. prices in Poland 60 Cigarette prices are based on the weighted average pack price supplied by PMI Estimated consumption 40 KPMG analysis suggests that changes in consumption in Poland are results show an increase related to changes in disposable income and cigarette prices(c)

in non-domestic Volume (billion sticks) 20  Disposable income is negatively related to changes in consumption i.e. as consumption in 2007 income rises consumption tends to decrease which is consistent with 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007  Cigarette prices are also negatively related to changes in consumption EU Flows Model results  It is important to note that factors other than price and income are likely to impact consumption levels in Poland CAGR (%) 1997-2002 2002-2006 2006-2007 − examples of other factors likely to impact consumption levels include  LDS (3.7)% (0.9)% (3.3)% changes in smoking restrictions and changes in demographic profiles  Estimated consumption(b) 0.0% (5.5)% (3.2)%  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be considered a broad approximation Notes: (a) Relationship based on historical data from 1997 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 The relationships between changes in consumption and changes in Project Star estimate of consumption for Poland (66.9 billion sticks) (c) 56% of changes in consumption can be explained by changes in disposable disposable income and prices have been used to estimate 2007 income and cigarette prices consumption Sources: (1) KPMG EU Flows Model (2) In Market Sales supplied by PMI  Combining estimates for the strength of these relationships with actual (3) KPMG Consumption Regression Model 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is consistent with EU Flows Model results  Estimated consumption decreased at a slightly slower rate than LDS in 2007 suggesting non-domestic consumption increased − this is consistent with EU Flows Model results

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Portuguese consumption Non-domestic inflows to 2006-2007(1)(2)(b) Portugal decreased to 0.48

billion sticks in 2007, 100% 4.5% 2.5% 2.3% 1.1% driven by falling flows 80% C&C from Spain ND(L) 60% LDC 96.4% 93.2% Outflows 40%

20%

0 Share of total consumption (8.9)% (9.7)%

(20)% 2006 2007

Volume (bn sticks): 14.0 13.2

Total non-domestic inflows (ND(L) and C&C) by origin 2006-2007(1)(2) Key:  Portugal  Major source country  Major destination country Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 1.2 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2006 1.0

(1)(2) 0.8 0.26 Proportion of total EU levels Other countries 5% 0.6 Spain

4% 0.4 0.70 0.28 Volume (bn sticks) 3% 2.7% 0.2 0.20 1.8% 0.0 2% 2006 2007

Percentage of EU total 1% 0.5% 0.5% 0.5% Volume (bn sticks): 0.95 0.48

Note: (b) 2006 legal domestic share of total consumption differs slightly from what was shown in the 0% 2006 Project Star Results presentation due to a technical change in approach to the calculation of Consum- ND ND(L) C&C Outflows consumption share (2006 LDC was shown as 94% in the 2006 Results presentation compared to ption 93.2% in this presentation)

Historic cigarette prices and legal domestic sales Legal cigarette sales have 1997-2007(1)(2) trended downward since 2001 as prices have 20 5 increased 17.6 17.2 17.4 17.0 16.3 16.8 16.5 16.0 15.6 (Euro) per 20 sticks Price 16 14.3 14.0 4

12 3

8 2 Volume (bn sticks) 4 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2007 2004-2007  Legal domestic sales 2.4% (3.7)% (5.3)%

 Average pack price 3.6% 7.6% 9.3%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI Legal domestic sales have been declining since 2001 and fell by 2.1% in 2007. Prices have been increasing over the same period in response to increases in taxes  Annual increases in the consumption tax on cigarettes have helped drive relatively large increase in prices since 2004 − the average weighted pack price has increased by 30% since 2004  Since 2005, new brands have been introduced into the lower price categories of the market in a response to decreasing affordability − there has also been some repositioning of existing brands into lower-priced categories  Portugal approved new anti-smoking laws during 2007 which included further restrictions on smoking in public places, restrictions on media advertising and limits on access of minors to vending machines − the new laws came into force on January 1st 2008

Marlboro price comparison Price differentials with 1 July 2007(1)(2)(a) France and Italy narrowed €4.30 slightly in 2007, but €8.16 €4.94 €1.79 remain significant €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of price differential between Portugal and key markets(2)(a) €3.86

180% Key:  Portugal 156% 155% 160%  Major source country 140%  Major destination country 120% 100% July 2006 80% 72% 59% July 2007 60% 38% 40% 30% 20% 0 Price premium/(discount) versus local (20)% (5)% (6)% UK France Italy Spain

Note: (a) Based on pack of 20 Marlboro King Size Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Portuguese consumption 2007(1)(2)(3)(4)(a)(b) is estimated at 13.2 billion Outflows Inflows 1.3 billion sticks 0.5 billion sticks sticks versus legal sales of 16 14.0 billion sticks, giving a net outflow of 0.8 billion 14 sticks in 2007 0.51 0.48 0.21 0.290.0 0.13 0.07 0.07 12 Outflows C&C ND(L)

14.0 13.2 Volume (bn sticks) 12.7 6

0 LDS UK France Other LDC Spain Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows to Portugal 2006-2007(1)(2)(3) 2007(1)(2)(3)

decreased by 55% 0.4 between 2006 and 2007, 0.32

driven by falling flows 0.3 0.06 from Spain 0.01 Other countries Other countries 0.2 France 43% 0.14 Spain Spain 0.26 51% 0.06 Volume (bn sticks) 0.1 0.01 0.07 0.0 2006 2007 France ND(L) share of 2% 1% 6% consumption: ND(L) share of non- 34% 30% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.4 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.32 Values 0.3 x x x x = 0.14bn 2007 8.3m 25% 29% 4.0 58 0.15 Non-PMI 0.2 Other PMI 0.14 EU rank 0.04 Marlboro 2007 10 25 8 11 26 20 0.06 Volume (bn sticks) 0.1 0.14 0.02 0.06 0.0 2006 2007

PMI share of ND(L): 55% 58%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Contraband and counterfeit by origin Share of C&C by origin C&C inflows to Portugal 2006-2007(1)(2)(3) 2007(1)(2)(3) decreased by 47% from 0.7 0.63 2006 to 2007 0.6 0.19 Other countries Spain 0.5 62% 38%

0.4 0.33 0.3 0.21 Other countries 0.2 0.44 Volume (bn sticks) Spain 0.1 0.13 0.0 2006 2007 C&C share of 4% 3% consumption: C&C share of non- 66% 70% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3) KPMG’s estimate of non-domestic consumption in 2006 was KPMG’s estimate of non- comparable to the Ministry of Finance estimates domestic share for 2006  Ministry of Finance estimates were based on a rough approximation of a 8% was in line with the 7.4% 6.8% 1 billion stick inflow 7% Ministry of Finance view  KPMG’s estimate for non-domestic share is consistent with PMI EPS 6% results in 2006 and 2007 5% − the Ministry of Finance has not updated its estimates for 2007 4% 3.6% 3% 2% Share of total of Share consumption 1%

0% Ministry of Finance EU Flows Model EU Flows Model 2006 2006 2007

Sources: (1) KPMG EU Flows Model (2) Interview with Ministry of Finance – Antifraud services, Information division (3) PMI EPS Q3 2006 and Q2 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 2005-2007(1)(2)(3)(a) corroborate the extent of non-domestic consumption in a market results suggest that  The rate of change in the GCTS smoking prevalence and average smoking outflows from Portugal volume figures provide a useful indication of observed changes in 18 have become increasingly smoking habits. However, GCTS data cannot be used to quantify total 16 consumption in a market due to exclusions of particular age cohorts and important relative to Consumption 14 has declined at a the problem of under-reporting of consumption faster rate than inflows since 2005 12 LDS since 2005  The KPMG consumption index is calculated using the incidence of 10 smoking and the average daily consumption (both from GCTS) and the

8 population of each market (from Euromonitor GMID) 13.2 6 − these figures are then indexed to the earliest year for which we have

Volume (billion sticks) 4 both GCTS and LDS data available, 2005 for Portugal 2  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 2005-2007  Comparing consumption and sales trends helps to understand market  LDS (5.2)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (7.2)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Note: (a) A change in consumption methodology occurred in 2005. Prior to 2005 changed over the analysed period information was collected via phone interviews on a monthly basis. Post 2005, the research supplier changed and interviews were conducted on a quarterly − in many markets, some level of inflows or outflows are likely to be basis and via face-to-face interviews. Data prior to 2005 is believed to be less reliable and hence has not been used in the model present in the first analysed year Sources: (1) KPMG Consumption Index model  Any changes in GCTS (methodology, supplier, sample size and selection, (2) In Market Sales supplied by PMI (3) KPMG EU Flows Model etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Portugal increased between 2005 and 2007  Consumption in Portugal has declined at a faster rate than LDS since 2005  This result suggests that outflows from Portugal have become increasingly important relative to inflows i.e. non-domestic consumption has declined and/or outflows have increased − this is consistent with EU Flows Model results showing an increase in net outflows from 0.3 billion sticks in 2006 to 0.8 billion sticks in 2007

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Romanian consumption Non-domestic inflows to 2007(1)(2) Romania were 2.9 billion 100% 7.2% sticks in 2007 2.1% 80% C&C accounted for the 60% C&C majority of non-domestic 90.7% ND(L) 40% inflows and 7.2% of LDC domestic consumption 20% Outflows

0% Hungary, Italy and France total of Share consumption (9.5%)

are the major destination (20%) countries for Romanian 2007 outflows

Volume (bn sticks): 31.1

Total non-domestic inflows (ND(L) and C&C) by origin (1)(2) Key:  Romania  Major source country  Major destination country 2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 3.5 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2007 3.0 0.10 (1)(2) 2.5 0.60 Proportion of total EU levels Other countries 2.0 8% Moldova 0.90 1.5 Ukraine 6.1% Duty Free 6% 1.0 Volume (bn sticks) 1.29 4.3% 0.5 3.7% 4% 3.2% 0.0 2007 2.2% 2% Percentage of EU total Volume (bn sticks): 2.89

0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales, 2004 to 2007(1)(2) Legal domestic sales increased slightly in 2007, 40 37.4 6 but remain below 2004 34.2

35 31.1 (RON) per 20 sticks Price levels 30.5 5 30 4 25 20 3 15 2

Volume (bn sticks) 10 1 5 0 0 2004 2005 2006 2007

CAGR (%) 2004-2007 2006-2007  Legal domestic sales (5.9)% 2.1%  Average pack price 21.4% 19.4%

Sources: (1) In Market Sales supplied by PMI (2) Average retail selling price supplied by PMI Local Management

Legal domestic sales grew by 2.1% in 2007, despite an increase in average pack price of 19.4%  Romania is raising taxes into line with EU excise tax regulations with prices rising at an annual average of 21.4% between 2004 and 2007 − 2006 price increases were driven by the introduction of a €10 per thousand sticks “vice tax” combined with an excise increase of €4 per thousand sticks − the excise yield was increased by a further €8 per thousand sticks in 2007  Anti-smoking legislation in Romania includes a smoking ban in public places and a ban on sales to people under 18 years of age  EU Member States including the UK, Germany and Hungary, restrict legal imports of cigarettes from Romania to 200 sticks per trip

Marlboro price comparison, July 1st 2007(1)(2)(a) Price differentials between

Romania and key €4.30 destination markets €8.16 €4.94 €1.79 narrowed slightly in 2007 €0.86 €1.47 Ukraine and Moldova both €4.24 €1.33 €0.86 became relatively less €7.05 n/a €4.71 expensive versus Romania €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €0.67 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between Romania and key markets(2)(a) €3.86

Key:  Romania 350%  Major source country 294%  Major destination country 300%

250% 215% 201% 200% July 2006 147% 150% July 2007

100% 59% 44% 50% 0% (50%) (49)% (50)% (100%) (60)% (62)% Price premium/(discount) versus local France Italy Hungary Moldova Ukraine

Notes: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown, 2007(1)(2)(3)(4)(a)(b) Romanian consumption is estimated at 31.1 billion Outflows Inflows 2.9 billion sticks 2.9 billion sticks sticks versus legal sales of 35 31.1 billion sticks, implying equal levels of inflow and 1.30 0.73 30 0.66 outflow volumes in 2007 0.68 0.60 0.35 0.89 0.61 0.00 0.00

Outflows C&C 31.1 31.1 ND(L) 15

Volume (bn sticks) 28.2

0 LDS France Italy Hungary Other LDC Ukraine Moldova Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2007(1)(2)(3) 2007(1)(2)(3)

consumption was 0.7 0.7 0.66

billion sticks in 2007 0.6 0.17 Other countries 0.5 26% 0.07 0.4 Other countries Italy 0.3 Duty Free 0.2 0.42 Volume (bn sticks)

0.1 Italy Duty Free 0.0 10% 64% 2007

ND(L) share of consumption: 2% ND(L) share of non-domestic: 23%

Non-domestic (legal) by brand 2007(1)(2)(3) ND(L) analysis Ave. Total 0.7 0.66 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.6 Values 17.6m x 30% x 30% x 4.2 x 101 = 0.66bn 0.5 0.34 2007

0.4 Non-PMI Other PMI 0.3 EU rank 0.10 Marlboro 2007 7 20 7 9 20 7 0.2 Volume (bn sticks)

0.1 0.23

0.0 2007 PMI share of ND(L): 49%

Source: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Ukraine and Moldova are 2007(1)(2)(3) 2007(1)(2)(3)

the main sources of C&C 3.0 inflows to Romania 2.5 2.22 Other countries 33% There is also a significant 2.0 0.73 Other countries Duty Free inflow, although 1.5 Moldova in 2007 this was impacted 0.60 1.0 Ukraine Ukraine by the temporary closure 40% Volume (bn sticks) of a large number of 0.5 0.89 border shops during the Other countries 0.0 includes Duty Free variants first half of the year 2007

C&C share of consumption: 7% Moldova C&C share of non-domestic: 77% 27%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(5)(6)(7) 2007 estimates of non-domestic consumption will have been substantially Significant differences impacted by the closure of Duty Free shops in the region during the first between non-domestic half of the year 18% 16.0% 16.0% 15.1% estimates may reflect the 16% 14.5%  The PMI Empty Pack Survey was conducted in September 2007 14% 12.0% impact of the closure of 11.8% KPMG estimates of non-domestic consumption in 2007 are slightly lower 12% 10.5% 9.5% 9.3% than those of JTI and the Ministry of Economy and Finance and below Duty Free shops in the first 10% 7.9% 7.1% 6.3% BAT’s estimate half of 2008 8% 6%  4% JTI estimates are believed to be based on Ministry of Economy and Finance 2% data Share of total of Share consumption 0%  KPMG’s estimate of non-domestic share reflects the closure of Duty Free shops on Romania’s borders during the first half of 2007

JTI 2007  EPS results for 2007 have therefore been uplifted to reflect annual changes to BAT 2007 BAT 2006 BAT 2005 2007 Duty Free sales in the border regions for the full year EPS Q3 2007 EPS EPS Q3 2006 EPS Q4 2006 EPS EPS Q3 2005 EPS Q4 2005 EPS Dept 2006

EU Flows Model − at the time of data collection in September a number of high volume Duty and Finance 2007 Ministry Economy Fraud Investigation Free outlets either remained closed or had recently opened  Although we have adjusted our findings to estimate the average non- Sources: (1) KPMG EU Flows Model domestic incidence over the entire year, timing of research for other (2) Rompress, 30th January 2007 estimates means there are likely to be differences versus our measurement (3) Rompress, 7th February 2007 (4) Rompress, 27th June 2007 Estimates for non-domestic share of total consumption in 2006 and 2007 (5) PMI EPS Q3 2005 , Q4 2005, Q3 2006, Q4 2006, and Q3 2007 are generally higher than estimates for 2005 (6) Nine O’Clock, 28th June 2008 (7) Interview with Custom Control  This relative high-point in terms of non-domestic consumption corresponds to a relative low-point in legal domestic sales and a year of relatively rapid price increases (see “Historic sales and pricing trends“ page for more detail)  This analysis suggests that there was a shift from legal consumption in 2006 to non-domestic consumption and that this shift was partially reversed in 2007 KPMG estimates of non-domestic consumption are net of outflows to other countries  Project Star estimates consumption before outflows to be 34 billion sticks  This is in line with the local market estimates for the total Romania market of 35 billion sticks

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 2004-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in 40 Consumption increased in Romania has been stable smoking habits. However, GCTS data cannot be used to quantify total 35 since 2004 while consumption in a market due to exclusions of particular age cohorts and since 2004 LDS have the problem of under-reporting of consumption 30 declined 25  The KPMG consumption index is calculated using the incidence of smoking and the average daily consumption (both from GCTS) and the 20 population of each market (from Euromonitor GMID) 15 31.1 − these figures are then indexed to the earliest year for which we have 10 Volume (billion sticks) both GCTS and LDS data available, 2004 for Romania 5  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 2004-2007  Comparing consumption and sales trends helps to understand market  LDS (5.9)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index 0.0% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Romania increased between 2004 and 2006  Consumption in Romania was stable between 2004 and 2006 while LDS declined  This result suggests that inflows have become increasingly important relative to outflows from Romania i.e. non-domestic consumption has increased and/or outflows have decreased

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Slovakian consumption Non-domestic inflows to 2006-2007(1)(2) Slovakia decreased to 0.28 100% 4.5% 1.7% billion sticks in 2007, 3.6% 1.9% driven by declines in both 80% legal and illicit flows C&C 60% ND(L) 91.9% 96.3% LDC 40% Outflows 20%

0 Share of total consumption (7.7)% (4.7)%

(20)% 2006 2007

Volume (bn sticks): 7.0 7.7

Total non-domestic inflows (ND(L) and C&C) by origin 2006-2007(1)(2) Key:  Slovakia  Major source country  Major destination country Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 0.6 (2) AC Nielsen ND(L) research, 2007 0.5 Proportion of total EU levels(1)(2) 0.24 0.4 Other countries 1.2% 1.1% 0.3 0.04 Czech Republic 0.08 Ukraine 0.2 0.7% 0.05 0.8% Volume (bn sticks) 0.29 0.1 0.15 0.5% 0.0 0.4% 0.3% 2006 2007 0.2% Percentage of EU total Volume (bn sticks): 0.57 0.28

0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Slovakian legal cigarette 1997-2007(1)(2) sales declined by 27% in 2003, but have partially 12 70 recovered over the last 9.7

9.6 (SKK) per 20 sticks Price four years 10 60 8.3 8.4 8.5 8.5 7.7 50 8 7.1 7.0 6.3 6.3 40 6 30 4 20 Volume (bn sticks)

2 10

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2002 2002-2003 2003-2007  Legal domestic sales (2.3)% (26.6)% 5.4%

 Average pack price 3.4% 23.8% 4.9%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Slovakian legal cigarette sales increased in 2007 by 10.3% despite a 3.6% increase in prices that year  LDS growth since 2005 has been partially driven by the increasing popularity of the “super-low” price segment − the super-low price segment (brands sold for less than 55 SKK) now account for over 40% of sales by volume  Despite strong growth in 2007, Slovakian LDS remain below pre-2002 levels  An import limit of 25 sticks on the Ukrainian border is in force to limit non-domestic inflows from this lower-cost market  Slovakia joined the Schengen Zone on the 21st December 2007

Marlboro price comparison The price differential 1 July 2007(1)(2)(a) between Slovakia and €1.33 Ukraine, the key source €0.86 market for non-domestic n/a €4.11 €4.71 inflows, increased in 2007 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €2.73 €4.10 €1.58 €1.42 €1.64

Key:  Slovakia Evolution of Marlboro price differential between Slovakia and key  Major source country markets(2)(a)  Major destination country

150% 115% 100% 100%

50% July 2006 3% 5% 2% July 2007 0 (3)% (50)%

(69)% (63)% (100)% (73)% (63)% Price premium/(discount) versus local Germany Czech Hungary Ukraine Russia Republic

Note: (a) Based on pack of 20 Marlboro King Size as of 1 July 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Slovakia consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 7.7 billion Outflows Inflows 0.36 billion sticks 0.28 billion sticks sticks versus legal sales of 9 7.7 billion sticks, suggesting outflow 8 0.01 0.09 volumes were similar to 0.06 0.12 0.20 0.13 inflow volumes in 2007 7 Outflows C&C ND(L) 6

0.02

4 7.74 7.66 7.38 Volume (bn sticks)

0 LDS Hungary Germany Other LDC Ukraine Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3) 2007(1)(2)(3) consumption in Slovakia 0.30 was 0.15 billion sticks in 0.25 0.25 2007 Czech Republic 32% 0.20 Other countries Other countries 0.15 Austria 42% 0.15 0.19 Ukraine 0.06 0.10 Czech Republic

Volume (bn sticks) 0.01 0.01 0.02 0.05 0.01 0.04 0.05 0.00 Ukraine 2006 2007 Austria 16% ND(L) share of 10% 4% 2% consumption: ND(L) share of non- 44% 52% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.30 Population Propensity Propensity Ave. no. purchases ND(L) 0.25 19+ to travel to purchase of trips (sticks) (sticks) 0.25 Values x x x x = 0.15bn 0.07 2007 4.2m 53% 12% 7.6 74 0.20 0.15 Non-PMI 0.15 Other PMI EU rank 0.13 Marlboro 2007 17 11 26 2 25 18 0.10 0.10 Volume (bn sticks) 0.05 0.02 0.05 0.02 0.00 2006 2007

PMI share of ND(L): 72% 29%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with Local PMI Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows to Slovakia 2006-2007(1)(2)(3) 2007(1)(2)(3) Other countries declined by 58% between 0.40 4% 2006 and 2007 0.35 0.32

0.30 0.04 Ukraine remains the main 0.25 source of C&C inflows Other countries 0.20 Ukraine 0.13 0.15 0.28 0.01

Volume (bn sticks) 0.10

0.13 0.05 Ukraine 0 96% 2006 2007

C&C share of 5% 2% consumption: C&C share of non- 56% 48% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(5) Imperial Tobacco and local PMI Management estimates of non-domestic EU Flows Model results for consumption were 10% and 8% respectively in 2006 2006 were in line with  These estimates corroborate the EU Flows Model result of 8.1% non- Imperial Tobacco and PMI 20% 18.0% domestic consumption 18% Management estimates  Slovakian Customs estimate for 2006 was significantly higher than either 16% PMI or Imperial Tobacco estimates but also includes product transiting 14% Updated research Slovakia for consumption overseas 12% 10.0% suggests a significant Non-domestic consumption is higher in areas bordering lower cost 10% 8.0% 8.1% decline in non-domestic 8% cigarette markets 6% incidence in 2007 3.7%  A recent survey conducted by PMI and focusing on the East of the 4% country found 11.4% non-domestic consumption Share of total of Share consumption 2%  The results of this survey were included in the EU Flows Model to give a 0% national 3.7% non-domestic incidence Customs Imperial PMI EU Flows EU Flows 2006 Tobacco Slovakia Model Model Updated external estimates of non-domestic consumption for 2007 have 2006 2006 2006 2007 not been identified

Sources: (1) Interviews with Slovakian Customs (2) Imperial Tobacco Pack Count Study, April 2006 (3) Interview with PMI Slovakia (4) KPMG EU Flows Model (5) PMI EPS Q1 and Q3 2006 (Focus and Main survey) and Q3 2007 (Focus and Main survey)

results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in 12 smoking habits. However, GCTS data cannot be used to quantify total declined over the last four consumption in a market due to exclusions of particular age cohorts and years 10 Consumption the problem of under-reporting of consumption has continued to decline between  The KPMG consumption index is calculated using the incidence of 8 2007 consumption 2003 and 2007 smoking and the average daily consumption (both from GCTS) and the while LDS have population of each market (from Euromonitor GMID) modelling results suggest 6 increased after a large decline − these figures are then indexed to the earliest year for which we have a decline in non-domestic 4 7.7 both GCTS and LDS data available, 2000 for Slovakia

consumption which is Volume (billion sticks)  If LDS trends differ from those shown by the consumption index, it 2 consistent with EU Flows implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period Model results 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 1997-2003 2003-2006 2006-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS (6.8)% 3.9% 10.3% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (4.2)% (1.4)% (3.8)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI present in the first analysed year (3) KPMG EU Flows Model  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Slovakia remained reasonably constant between 1997 and 2002  This result suggests that non-domestic consumption in Slovakia was reasonably stable during this period However, in 2003 LDS declined at a much higher rate relative to consumption implying that non-domestic consumption increased  Legal sales have increased since 2003 while consumption has continued to decline, which suggests a decrease in non-domestic consumption  Consumption modelling results for 2007 show a relatively large decline in non-domestic consumption which is consistent with EU Flows Model results

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total Slovenian consumption Non-domestic inflows to 2006-2007(1)(2)(b) Slovenia fell to 0.26 billion sticks in 2007 100% 7.4% 5.1% 1.4% 1.0% Outflows from Slovenia 80% increased in 2007 60% C&C following the relaxation of ND(L) 91.2% 93.8% import restrictions by the 40% LDC Outflows Austrian government 20%

0 Share of total consumption (14.6)% (19.7)% (20)% 2006 2007

Volume (bn sticks): 4.27 4.21

Total non-domestic inflows (ND(L) and C&C) Key:  Slovenia  Major source country  Major destination country (1)(2)(b) Notes: (a) Map shows flows over 1% of consumption. Countries which are both 2006-2007 source and destination countries are coded according to the larger flow (b) ND(L) research has not been updated for 2007, but legal inflows are 0.4 believed to have fallen in line with the overall decline in non-domestic incidence Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) AC Nielsen ND(L) research, 2006 0.3 0.16 Other countries Proportion of total EU levels(1)(2) Duty Free 0.08 Croatia 2.0% 0.2 0.03 1.7% 0.03 0.04 Bosnia and Herzegovina 0.02

1.5% Volume (bn sticks) 0.1 0.15 0.12

1.0% 0.0 2006 2007 0.6% 0.5% 0.4% 0.3% Percentage of EU total Volume (bn sticks): 0.38 0.26 0.1%

0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales 1997-2007(1)(2) increased by over 5% in 2007 despite rising prices 4.78 5 4.55 4.57 4.52 2.5 4.37 4.34 4.44 Increased outflows to 4.21 4.23 4.17 4.20 Price per 20 sticks (Euro) per 20 sticks Price Austria was a contributing 4 2.0

factor to the increase in 3 1.5 legal sales 2 1.0 Volume (bn sticks) 1 0.5

0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2003 2003-2007  Legal domestic sales (0.1)% 1.1% 2.5%  Average pack price 9.9% 15.3% 7.4%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales increased by 5.8% in 2007, despite an increase in weighted average prices of 5.9%  Import restrictions into Austria were relaxed in July 2007 − up until July 2007, Austrian residents were only able to legally bring in 25 cigarettes per person per trip from the 2004 EU Accession States, including Slovenia  Following the lifting of import restrictions to 800 sticks per trip, outflows from Slovenia to Austria increased − the import limit was subsequently lowered to 200 sticks, effective from January 1st 2008  In January 2007, Slovenia’s national currency converted from Tolars to the Euro  New anti-smoking legislation took effect in August 2007, which banned smoking in all public places including hotels and restaurants except where they have installed special isolated smoking rooms  In December 2007, Slovenia joined the Schengen zone

Marlboro price comparison Cigarette prices in 1 July 2007(1)(2)(a) Slovenia are lower than those in Italy and Austria €4.24 €1.33 €0.86

n/a €4.71 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35

€3.90 €3.82 €2.39 €2.50 €1.66 €2.73

€1.58 €1.42 €4.10 €1.50(b) €1.64 €1.95 €1.62 €3.00 €2.60

Evolution of Marlboro price differential between Slovenia and key Key:  Slovenia (2)(a) markets  Major source country

 Major destination country 98% Note: (b) Kosovo 100% 88% 78% 80% 64% 64% 56% 60%

40% July 2006 22% July 2007 20% 9%

(20)% (40)% (30)% (37)% (43)%

Price premium/(discount) versus local (60)% (50)% Germany Italy Austria Croatia Bosnia Serbia

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Slovenian consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 4.2 billion Outflows Inflows 0.83 billion sticks 0.26 billion sticks sticks versus legal sales of 6 4.8 billion sticks, giving a net outflow of 0.6 billion sticks 5 Austria is the main destination market for 0.53 Outflows 0.01 C&C 0.14 0.09 outflows from Slovenia 0.07 0.01 ND(L) 4 0.080.0 0.12 0.03

3 Volume (bn sticks)

4.8

4.2 2 4.0

0 LDS Austria Italy Germany Other LDC Bosnia and Croatia Other Consumption Herzegovina

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3) 2007(1)(2)(3) consumption was 0.04 0.08 Croatia billion sticks in 2007 23% 0.06 0.06

0.04 Other countries 0.04 0.04 Croatia

0.03

Volume (bn sticks) 0.02

0.02 0.01 0.00 Other countries 77% 2006 2007

ND(L) share of 1% 1% consumption: ND(L) share of non- 16% 17% domestic:

Non-domestic (legal) by brand 2006-2007(1)(2)(3) ND(L) analysis Ave. Total 0.08 Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 0.06 Values 0.06 x x x x = 0.04bn 2007 1.6m 59% 26% 3.1 56 0.04 0.03 Non-PMI 0.04 Other PMI EU rank 0.02 Marlboro 0.00 2007 23 9 9 16 27 26

Volume (bn sticks) 0.02 0.00 0.03 0.02

0.00 2006 2007 PMI share of ND(L): 48% 48%

Note: ND(L) research was not updated in 2007, but legal flows are believed to have declined in line with the overall decline in non-domestic incidence Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Counterfeit and 20060-2007(1)(2)(3) 2007(1)(2)(3) contraband consumption 0.4 declined by a third to 0.2 0.32 billion sticks in 2007 0.3

0.16 0.22 Other countries 0.2 Other countries Bosnia and Herzegovina 46% 0.10 Bosnia and Herzegovina

Volume (bn sticks) 0.1 54% 0.15 0.12

0.0 Other countries 2006 2007 includes Duty Free inflows C&C share of 7% 5% consumption: C&C share of non- 84% 83% domestic:

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to 2007 consumption 2000-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market

modelling results suggest  The rate of change in the GCTS smoking prevalence and average smoking an increase in outflows volume figures provide a useful indication of observed changes in 6 smoking habits. However, GCTS data cannot be used to quantify total and a decrease in non- consumption in a market due to exclusions of particular age cohorts and 5 the problem of under-reporting of consumption domestic consumption, Consumption has declined  The KPMG consumption index is calculated using the incidence of both of which are 4 since 2000 while LDS have smoking and the average daily consumption (both from GCTS) and the consistent with EU Flows population of each market (from Euromonitor GMID) 3 increased Model results − these figures are then indexed to the earliest year for which we have 2 4.2 both GCTS and LDS data available, 2000 for Slovenia

Volume (billion sticks)  If LDS trends differ from those shown by the consumption index, it 1 implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2000-2005 2005-2007 2000-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS 1.7% 2.3% 1.9% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (3.1)% 1.2% (1.9)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI present in the first analysed year (3) KPMG EU Flows Model  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for Slovenia increased between 2000 and 2007  LDS in Slovenia have increased since 2000, while consumption has declined  This result suggests that outflows from Slovenia have become increasingly important relative to inflows i.e. outflows have increased and/or non-domestic consumption has decreased  The increase in the gap in 2007 suggests a decline in non-domestic consumption and/or an increase in outflows, both of which are consistent with EU Flows Model results

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Spanish consumption The share of consumption 2006-2007(1)(2)(b) accounted for by counterfeit and 100% 2.4% 2.2% contraband was 2.2% in 1.5% 2.9% 2007, representing one of 80% the lowest counterfeit and 60% C&C contraband rates in the EU 96.1% 94.9% ND(L) 40% LDC 20% Outflows

Share of total of Share consumption 0 (11.4)% (12.1)%

(20)% 2006 2007 Volume (bn sticks): 84.4 83.7

Total non-domestic inflows (ND(L) and C&C) (1)(2)(b) Key:  Spain  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow 5 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management (2) Synovate ND(L) research, 2007 4 (1)(2) 1.44 Proportion of total EU levels Other countries 3 25% 0.31 France 20.9% 1.83 0.43 2 Italy 20% 0.44 Canary Islands 0.22 Volume (bn sticks) 1 0.17 0.00 Duty Free 15% 1.65 11.6% 1.07 0 10% 8.0% 2006 2007

4.7% Volume (bn sticks): 3.3 4.3

Percentage of EU total 5% 3.0%

0% Note: (b) Inflows have increased in 2007, partly as a result of changes in the treatment of Canary Islands and Andorra variant cigarettes. In the 2006 Project Star results, Canary Islands and Consum- ND ND(L) C&C Outflows Andorra variant cigarettes were classified as Spanish legal domestic. In 2007, Canary Islands ption and Andorra variant cigarettes have been classified as non-domestic legal

Historic cigarette prices and legal domestic sales Legal domestic sales have 1997-2007(1)(2) trended downwards since 2004 as prices have 100 93.1 93.9 93.3 3.0 increased 90.8 91.2 90.6 89.5 86.9 86.7 88.5

78.2 2.5 (Euro) per 20 sticks Price 80

2.0 60 1.5 40 1.0 Volume (bn sticks) 20 0.5

0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2000 2000-2004 2004-2007  Legal domestic sales 4.2% 1.5% (1.6)%  Average pack price 6.2% 5.9% 4.8%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI (3) “Spain increases minimum tax on cigarettes”, El Pais, 11 November 2006 Legal domestic sales declined by 1.2% in 2007 compared to 2006, while average weighted prices increased by 10.4% in the same period  In November 2006, the government increased the minimum tax to €70 per thousand sticks from €55. The increased minimum tax affected cigarettes priced below €2.17 a pack(3)  In January 2006, Spain introduced a smoking ban covering all workplaces − in addition, smoking is restricted in bars and restaurants larger than 100m2 (although separate smoking rooms are allowed)

Marlboro price comparison There is a significant price 1 July 2007(1)(2)(a) differential between Spain €4.30 and other EU markets €8.16 €4.94 €1.79 including the UK, France €0.86 and Germany €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

€3.50 €3.86 Evolution of Marlboro price differential between Spain and key Key:  Spain (2)(a) markets  Major source country

 Major destination country

200% 169% 173%

150%

100% 82% 69% 63% 60% 45% July 2006 50% 39% July 2007 0%

(50)% (29)% (38)% (44)% (54)% Price premium/(discount) versus local (100)% UK France Germany Italy Poland Romania

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Spain consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 83.7 billion Outflows Inflows 10.1 billion sticks 4.3 billion sticks sticks versus legal sales of 100 89.5 billion sticks in 2007

4.22

3.25 1.85 0.92 1.25 80 1.71 0.44 0.43 0.31 Outflows C&C ND(L)

89.5 83.7

Volume (bn sticks) 40 79.4

0 LDS UK France Germany Other LDC Canary Italy France Other Consumption Islands

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) (c) Inflows have increased in 2007, partly as a result of changes in the treatment of Canary Islands and Andorra variant cigarettes. In the 2006 Project Star results, Canary Islands and Andorra variant cigarettes were classified as Spanish legal domestic. In 2007, Canary Islands and Andorra variant cigarettes have been classified as non-domestic legal Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) In Market Sales supplied by PMI (4) Interviews with PMI Local Management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal inflows were 2.4 2006-2007(1)(2)(3)(a)(b) 2007(1)(2)(3)(a)(b) 2.43 billion sticks in 2007 2.5 Canary Islands 18% 2.0 0.87 Other countries Other countries 35% 1.5 1.23 0.31 France

0.38 Duty Free 1.0 Italy 0.71 Italy 18% Volume (bn sticks) 0.43 Canary Islands 0.5 0.08 0.37 0.44 0.0 0.07 2006 2007 France Duty Free 13% 16% ND(L) share of 1.5% 2.9% consumption: ND(L) share of non- 37% 57% domestic: Non-domestic (legal) by brand (c) 2006-2007(1)(2)(3)(a)(b) ND(L) analysis 2.43 2.5 Ave. Total Population Propensity Propensity Ave. no. purchases ND(L) 19+ to travel to purchase of trips (sticks) (sticks) 2.0 1.09 Values Non-PMI 32.9 x 35% x 18% x 3.9 x 78 = 0.6bn(c) 1.5 2007 1.23 Other PMI 0.19 0.10 L&M 1.0 0.47 Marlboro EU rank 0.11 5 16 Volume (bn sticks) 2007 19 13 23 9 0.5 0.02 1.05 0.63

0.0 2006 2007 PMI share of ND(L): 62% 55%

Notes: (a) Inflows have increased in 2007, partly as a result of changes in treatment of Canary Islands Sources: (1) KPMG EU Flows Model and Andorra variant cigarettes. In the 2006 Project Star results, Canary Islands and Andorra (2) Synovate ND(L) research, 2007 variant cigarettes were classified as Spanish legal domestic. In 2007, Canary Islands and Andorra (3) Interview with PMI Local Management variant cigarettes have been classified as non-domestic legal (b) ND(L) inflows from Italy, France and Spain have been adjusted upward to reflect inflows associated with inbound tourism (c) ND(L) analysis provided by Synovate excludes adjustments to inflows from Canary Islands and Andorra, and inbound tourism inflows from Italy, France and Netherlands . Total ND(L) (including Canary Islands, Italy, France and Netherlands inflow) calculated at 2.43 billion sticks in 2007

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows declined by 2006-2007(1)(2)(3) 2007(1)(2)(3) 10% in 2007 compared 2.5 with 2006 Other countries 2.06 29% 2.0 1.85

1.5 0.52 Other countries 1.27 0.05 Poland 1.0 Duty Free 0.10

Volume (bn sticks) 1.27 Duty Free 0.5 Poland 68% 0.70 3%

0.0 2006 2007 C&C share of 2.4% 2.2% consumption: C&C share of non- 63% 43% domestic:

Sources: (1) KPMG EU Flows Model (2) Synovate ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2) The apparent increase in non-domestic share when comparing EU Flows A change in the treatment Model results for 2006 and 2007 reflects changes in the treatment of the of Canary Islands and Canary Islands and Andorra variant packs found in Spain Andorran product has 7%  In the 2006 Project Star results, Canary Islands and Andorra variant cigarettes were classified as Spanish legal domestic. In 2007, Canary resulted in a small 6% 5.7% 5.1% Islands and Andorra variant cigarettes have instead been classified as non- increase in non-domestic 5.1% 5% domestic legal inflows in 2007 3.9% 4% Like for like comparison of EPS results suggests a slight decline in non- domestic incidence in Spain in 2007 3%  This trend is consistent with a reduction in the non-domestic share of 2% consumption observed by ASM Grupo

Share of total consumption − 1% see ‘Comparison of external sources for non-domestic estimates (2 of 2)’ for ASM Grupo’s estimate for contraband consumption in Spain 0% PMI EPS 2006 EU Flows PMI EPS 2007 EU Flows Model 2006 Model 2007

Sources: (1) KPMG EU Flows Model (2) PMI EPS Q3 2006 and Q2 2007

Spain contraband estimates Results from the ASM 2002-2007(1) Grupo pack swap survey in 2007 also support the view that Spain has very low levels of C&C inflows 3.0% 2.4% 2.5%

2.0% 1.7% 1.8% 1.5% 1.5% 1.3% 1.3% 1.4% 1.3% 1.5% 1.2% 1.2% 1.2% 1.1% 0.8% 0.8% 0.9% 1.0% 0.6% 0.6% 0.7% 0.6% 0.7% 0.5% 0.5% 0.4% 0.5% Percentage contraband of 0.0% Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

2002 2003 2004 2005 2006 2007

Year average (%) 2002 2004 2005 2006 2007  Contraband estimate 1.9% 1.3% 1.3% 0.8% 0.5%

Source: (1) “Control Coincidental del Consumo de Tabaco Rubio”, ASM Grupo, 2007

The ASM Grupo estimate for contraband in 2007 was 0.5%, which compares to a Project Star estimate for C&C of 2.2% of consumption  KPMG estimates and the results from the ASM Grupo survey both support the view that Spain has one of the lowest rates of C&C consumption in the EU − C&C levels across the Member States vary from 1% to 27%, with an EU average level of 8% Differences between the two estimates for contraband are likely to be driven by survey methodology  The ASM Grupo estimate is based upon a quarterly survey of 13,500 people regarding their smoking habits − contraband incidence is determined based on the results of the pack swap that occurs on completion of the questionnaire

Cigarette flows from Spain to UK and Republic of Ireland (1)(2)(3)(4) Adjustments have been made to flows out of 6 PMI estimates Outflow adjustment calculation Spain to the UK and 5.19 ROI Republic of Ireland 0.05 ROI 0.19 UK 5 0.02 4.70 0.23 UK

0.45 ROI

4 Estimated EPS shortfall

1.67 UK 2.58

5.10 5.20 4.70 4.40 2 Volume (billions of sticks)

2.58 1 2.12

0 PMI planning 'in' PMI planning Spain LDS Expat domestic Tourist LDS outflows of ND(L) outflows C&C outflows EPS estimate (2005) 'out' estimate consumption consumption in UK brands (2005) Spain

As in 2006, adjustments were made to outflows from Spain to UK and Republic of Ireland (ROI)  Multiple alternative sources suggest that outflows from Spain to the UK and Republic of Ireland were understated by empty pack survey results  The EPS shortfall was estimated at 2.58 billion sticks in 2007, compared to an estimate of 3.10 billion sticks in 2006  Outflows to the UK and Ireland were therefore adjusted upwards by 2.31 billion sticks and 0.27 billion sticks respectively to reflect for sales of British brands in Spain of 5.2 billion sticks, net of: − estimated consumption by permanent UK and Irish expatriate residents in Spain − tourist consumption whilst in Spain Note: (a) Expatriate numbers for the UK (162,000) supplied by PMI Spain management Sources: (1) “European Union Industry Size" PMI Lausanne, 2006 (2) Spain In Market Sales, PMI management Spain (3) Spain EPS 2007 (4) GCTS Spain and UK

Cigarette flows from Spain to France (1)(2)(3) Adjusted volume of 3.25bn sticks in 2007 Further analysis identifies compared to 3.42bn stick in 2006 is a 10 reduction of 5%. This reduction is based on a country-to-country the decline in French outbound travel 9.2 shortfall in the EPS 8 outflows to France EPS EPS shortfall shortfall 6

4 3.42 3.25 3.7 3.5 1.5 1.3 2 3.0 Volume (billions sticks) of 2.3 1.9 1.8 2.0 0 PMI planning 'out' PMI planning 'in' Customer Customer NDL(2006) EPS (2006) ND(L) (2007) EPS (2007) estimate (2005) estimate (2005) purchasing habits purchasing habits estimate (2005) estimate (2006)

Sources: (1) "European Union Industry Size" PMI Lausanne, 2006 (2) France EPS 2006 and 2007 (3) AC Nielsen ND(L), 2006 and 2007

As in 2006, adjustments were made to outflows from Spain to France  Both PMI France and consumer purchasing habits imply a higher inflow from Spain than EPS results suggest  Alternative estimates by PMI and Consumer Purchasing Survey results imply an inflow from Spain in the region of 3.0 to 9.2 billion sticks, compared to EPS results of 1.95 billion sticks and ND(L) results of 1.84 billion sticks  PMI Planning ‘in’ estimate of 3.7 billion sticks in 2005 appears in line with 2005 purchasing habits estimate of 3.5 billion sticks  The 2006 reduction in purchasing estimates to 3.0 billion sticks reflects the closure of border shops  In 2007, the adjusted value of outflows from Spain to France is estimated at 3.25 billion sticks − this is a 5% reduction of the 2006 adjusted value of 3.42 billion sticks − this is based on a decline in trips to Spain reported by the French ND(L) research and further corroborated by Eurostat data showing a decline in outbound travel

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption index 2000-2007(1)(2)(3)(a) corroborate the extent of non-domestic consumption in a market

modelling results suggest  The rate of change in the GCTS smoking prevalence and average smoking that consumption has volume figures provide a useful indication of observed changes in 100 smoking habits. However, GCTS data cannot be used to quantify total declined since 2000 while Implied outflow equalled 10.7 consumption in a market due to exclusions of particular age cohorts and legal sales have increased billion sticks in the problem of under-reporting of consumption 80 2006  The KPMG consumption index is calculated using the incidence of smoking and the average daily consumption (both from GCTS) and the This implies that outflows 60 Consumption has declined population of each market (from Euromonitor GMID) from Spain have become since 2000 while 84.4 − these figures are then indexed to the earliest year for which we have an increasingly important 40 legal sales have increased both GCTS and LDS data available, 2000 for Spain

driver of legal domestic Volume (billion sticks) 20  If LDS trends differ from those shown by the consumption index, it sales implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 2000 2001 2002 2003 2004 2005 2006 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2000-2004 2004-2006 2000-2006 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS 1.5% (1.7)% 0.4% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (3.0)% 0.9% (1.7)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Note: (a) Consumption index modelling has only been completed up until 2006 due present in the first analysed year to a change in GCTS methodology in 2007. As a result of the change, GCTS consumption data for 2007 is not comparable to pervious years  Any changes in GCTS (methodology, supplier, sample size and selection, Sources: (1) KPMG Consumption Index model etc.) will also have an impact on the conclusions generated (2) In Market Sales supplied by PMI (3) KPMG EU Flows Model The gap between LDS and the KPMG consumption index for Spain increased between 2000 and 2006  LDS in Spain have increased since 2000, while consumption has decreased  This result suggests that outflows from Spain have become increasingly important relative to inflows i.e. outflows have increased and/or non- domestic consumption has decreased  Consumption index modelling suggests that there was an outflow of 10.7 billion sticks from Spain in 2006 which is consistent with the EU Flows Model estimate of 9.7 billion sticks for 2006

Estimated consumption based on its relationship with prices KPMG has investigated how changes in disposable income and cigarette KPMG analysis suggests and income, 1998-2007(1)(2)(3)(a) prices affect cigarette consumption in Spain that changes in  Historical data from 1998 to 2006 has been used to estimate the strength consumption are strongly of relationships between consumption and prices and income 120 related to changes in Control period Forecast − estimated incidence of smoking and average daily consumption are disposable income and 100 based on GCTS data. Population and disposable income are estimated using Euromonitor GMID and Oxford Economic Forecasting data. cigarette prices in Spain 80 Cigarette prices are based on the weighted average pack price supplied by PMI 60 KPMG analysis suggests that changes in consumption in Spain are 40 strongly related to changes in disposable income and cigarette prices(c) Volume (billion sticks) 20  Disposable income is positively related to changes in consumption i.e. as income rises consumption tends to increase 0  Cigarette prices are by contrast, negatively related to changes in 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption  It is important to note that factors other than price and income are likely to CAGR (%) 1998-2004 2004-2006 2006-2007 impact consumption levels in Spain

 LDS 0.1% (3.1)% (2.2)% − examples of other factors likely to impact consumption levels include  Estimated consumption(b) (1.1)% 0.7% (5.7)% changes in smoking restrictions and changes in demographic profiles  Given that we have not assessed the impact on consumption from factors other than price and income, estimated consumption can only be Notes: (a) Relationship based on historical data from 1998 to 2006 (b) The estimated consumption line has been rebased so that it equals the 2006 considered a broad approximation Project Star estimate of consumption for Spain (84.4 billion sticks) (c) 93% of changes in consumption can be explained by changes in disposable The relationships between changes in consumption and changes in income and cigarette prices disposable income and prices have been used to estimate 2007 Sources: (1) KPMG EU Flows Model consumption (2) In Market Sales supplied by PMI (3) KPMG Consumption Regression Model  Combining estimates for the strength of these relationships with actual 2007 values for disposable income and prices gives estimated 2007 consumption  Estimated 2007 consumption is lower than in 2006 which is directionally consistent with EU Flows Model results. The magnitude of the estimated consumption decline is, however, greater than expected

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a) Total Swedish consumption by origin Non-domestic inflows 2006-2007(1)(2) remained reasonably flat 100% at 1.6 billion sticks in 2007 11.9% 12.5% 80% 8.9% 9.0% Norway accounts for the 60% majority of outflows while C&C

some incidental flows to 40% 79.1% 78.5% ND(L) Denmark are also evident LDC 20% Outflows

0% Share of total of Share consumption (8.1)% (8.0)%

(20)% 2006 2007

Volume (bn sticks): 7.95 7.32

Total non-domestic inflows (ND(L) & C&C) (1)(2) Key:  Sweden  Major source country  Major destination country 2006-2007 Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 2.0 (2) AC Nielsen ND(L) research, 2007 1.8 Proportion of total EU levels(1)(2) 1.5 0.30 0.48 Other countries 1.3 0.17 2.5% Denmark 2.2% 1.0 0.26 0.07 0.11 Ukraine 0.16 0.15 2.0% 0.8 Russia 1.7% 0.27 0.23

1.5% Volume (bn sticks) 0.5 Poland 1.5% 1.2% 0.3 0.50 0.53 Duty Free 1.0% 1.0% 0.0 2006 2007

Percentage of EU total 0.5% Volume (bn sticks): 1.66 1.58

0.0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal cigarettes sales have 1997-2007(1)(2) been trending downwards in Sweden since 2002 as 7.7 8 7.4 50 7.3 7.1 7.2 prices have risen 6.9 7.1 7.0 7 6.3 (SEK) per 20 sticks Price 6.0 5.7 40 6

5 30 4

3 20

Volume (bn sticks) 2 10 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2002 2002-2004 2004-2007  Legal domestic sales 5.2% (3.9)% (3.8)%  Average pack price n/a n/a 5.3%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales decreased by 9.4% in 2007 while prices increased by 10.9%  Sweden introduced a smoking ban in June 2005 in bars and restaurants, although separately ventilated smoking rooms are permitted

Marlboro price comparison The price differential 1 July 2007(1)(2)(a) between Sweden and €4.30 Denmark increased in 2007 €8.16 €4.94 €1.79 €0.86 Norway became relatively €1.47 less expensive in 2007 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60 Evolution of Marlboro price differential between Sweden and key markets(2)(a) €3.86 Key:  Sweden  Major source country  Major destination country 100% 88% 65% 75% 50%

25% July 2006 0% July 2007 (3)% (25)% (4)% (14)% (13)% (50)% (37)% (40)% (61)% (75)% (57)%

(82)% (83)%(85)%

Price premium/(discount) versus local (100)% (87)% Norway Finland Denmark Spain Poland Russia Ukraine

Note: (a) Based on pack of 20 Marlboro King Size as of July 1st 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown Swedish consumption is 2007(1)(2)(3)(4)(a)(b)(c)(d) estimated at 7.3 billion Outflows Inflows 0.58 billion sticks 1.58 billion sticks sticks versus legal sales of 8 6.3 billion sticks, giving a net inflow of 1.0 billion 7 0.46 sticks in 2007 0.63

0.37 0.11 Outflows 6 0.15 0.10 0.20 0.12 0.00 0.00 C&C 0.03 ND(L)

4 7.32

Volume (bn sticks) 6.33 3 5.74

0 LDS Norway Denmark Other LDC Poland Russia Ukraine Other Consumption

Notes: (a) LDS – Legal Domestic Sales (b) LDC – Legal Domestic Consumption (i.e. legal in-market sales that are also consumed in the same market) (c) The outflow to Norway is an estimate based on EPS and LDS from Norway (d) The level of PMI counterfeit product has been increased by a factor of 2 to reflect analysis from local management, suggesting counterfeit levels were between 0.04 to 0.07 billion sticks in 2007. Local management’s estimates have been confirmed by local customs. Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) In Market Sales supplied by PMI (4) Interviews with PMI local management

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Non-domestic (legal) 2006-2007(1)(2)(3) 2007(1)(2)(3) inflows to Sweden were Germany 0.8 13% 0.7 billion sticks in 2007 0.71 0.7 0.66 0.6 Denmark 11% 0.5 0.39 0.45 0.4 Other countries 0.3 0.04 Greece Greece 0.10 9%

Volume (bn sticks) Denmark 0.2 0.06 0.07 Germany Other countries 0.1 0.18 0.08 67% 0.0 2006 2007

ND(L) share of 9% 9% consumption: ND(L) share of non- 43% 42% domestic:

Non-domestic (legal) by brand ND(L) analysis 2006-2007(1)(2)(3) Ave. Total 0.8 Population Propensity Propensity Ave. no. purchases ND(L) 0.71 19+ to travel to purchase of trips (sticks) (sticks) 0.7 0.66 Values x x x x = 0.66bn 0.6 2007 7.0m 64% 24% 3.4 183 0.37 0.5 0.37 0.4 EU rank Non-PMI 0.3 0.04 2007 14 6 13 15 13 8 0.04 Other PMI

Volume (bn sticks) 0.2 0.29 Marlboro 0.25 0.1 0.0 2006 2007 PMI share of ND(L): 48% 44%

Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI local management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin Counterfeit and 2006-2007(1)(2)(3)(a) 2007(1)(2)(3)(a)

contraband inflows to 0.95 1.0 0.92 Sweden were 0.9 billion Poland Other countries sticks in 2007 0.8 0.27 22% 50% 0.46 0.6 Other countries 0.26 Ukraine 0.4 0.11 Russia 0.16 0.15

Volume (bn sticks) Poland Russia 0.2 16% 0.26 0.20 0.0 2006 2007 Ukraine C&C share of 12% 13% 12% consumption: C&C share of non- 57% 58% domestic:

Note: (a) The level of PMI counterfeit product has been increased by a factor of 2 in 2007 to reflect analysis from local management, suggesting counterfeit levels were between 0.04 to 0.07 billion sticks in 2007. Local management’s estimates have been confirmed by local customs. Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2007 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(a) Swedish Customs estimate for 2006 refers only to consumption of illicit KPMG’s estimate of non- cigarettes domestic consumption 25%  In 2006, the EU Flows model estimated the share of consumption was higher than that 20.9% 21.5% accounted for by illicit product to be 11.9% provided by Swedish 20%  Customs estimate of 7.5% is based on the mid point of a 5% to 10% Customs 15.7% range which is believed to be representative 15% KPMG’s 2007 non-domestic incidence estimate for Sweden is higher This was partly due to the 7.5% than that suggested by PMI EPS results 10% exclusion of non-domestic  2007 PMI EPS results indicate a lower non-domestic incidence of 15.7% (legal) flows in Custom’s 5% versus 2006 results of 20.9% which may not be realistic given flow- Share of total of Share consumption through impacts on implied consumption 2006 estimate 0% − using unadjusted EPS results as inputs for the EU Flows Model gives Customs 2006 EU Flows PMI EPS 2007 EU Flows a decline in Swedish consumption of 14% which appears high relative Model 2006 Model 2007 to other available estimates  In addition, corroborating sources suggest that non-domestic Note: (a) Swedish customs estimate was for counterfeit and contraband cigarettes consumption in Sweden has been stable and may have actually increased Sources: (1) KPMG EU Flows Model in 2007, particularly towards the end of the year (2) PMI EPS research Q3 2006 and Q2 2007 (3) Interview with Swedish Customs  As a result, a stable share for genuine non-domestic consumption has been assumed for 2007 − use of the adjusted non-domestic incidence helps give a more realistic consumption decline of 8.3% − PMI counterfeit consumption is assumed to have increased based on PMI Empty Pack Survey results and feedback from market participants

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to 2007 consumption 2000-2007(1)(2)(3)(a) corroborate the extent of non-domestic consumption in a market

modelling results suggest  The rate of change in the GCTS smoking prevalence and average smoking an increase in non- volume figures provide a useful indication of observed changes in smoking habits. However, GCTS data cannot be used to quantify total 9 domestic consumption consumption in a market due to exclusions of particular age cohorts and 8 which is consistent with the problem of under-reporting of consumption 7 EU Flows Model results Consumption  The KPMG consumption index is calculated using the incidence of 6 has declined smoking and the average daily consumption (both from GCTS) and the faster than LDS 5 population of each market (from Euromonitor GMID) since 2000 4 7.3 − these figures are then indexed to the earliest year for which we have 3 both GCTS and LDS data available, 2000 for Sweden Volume (bn sticks) 2  If LDS trends differ from those shown by the consumption index, it 1 implies that the relationships between inflows, outflows and domestic consumption have changed over the analysed period 0 2000 2001 2002 2003 2004 2005 2006 2007 Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption  Comparing consumption and sales trends helps to understand market CAGR (%) 2000-2004 2004-2007 2000-2007 direction and its likely inflow/outflow status. However, it will not identify any consumption gap that is present at the start of the period  LDS 0.1% (3.8)% (1.6)% − the gap between indexed consumption and LDS in the final year is  KPMG consumption index (2.1)% (3.3)% (2.6)% therefore only a broad indication of the extent to which the market has  KPMG consumption estimate n/a n/a n/a changed over the analysed period − in many markets, some level of inflows or outflows are likely to be Note: (a) Smoking incidence and average daily consumption data are not robust present in the first analysed year enough prior to 2000 to enable reliable modelling Sources: (1) KPMG Consumption Index model  Any changes in GCTS (methodology, supplier, sample size and selection, (2) In Market Sales supplied by PMI etc.) will also have an impact on the conclusions generated (3) KPMG EU Flows Model The gap between LDS and the KPMG consumption index for Sweden increased between 2000 and 2007  Consumption in Sweden has declined at a faster rate than LDS since 2000  This result suggests that outflows from Sweden have become increasingly important relative to inflows i.e. outflows have increased and/or non-domestic consumption has decreased However, consumption decreased at a slower rate than LDS in 2007 suggesting non-domestic consumption increased  This is consistent with the increase in non-domestic consumption from 20.9% to 21.5% in 2007 based on EU Flows Model results

OTP sales in billion stick equivalents Legal sales of smoking tobacco have declined at a faster rate than OTP, particularly 2003-2007(1)(a)(b) manufactured cigarettes since 2003 smokeless tobacco, 1.5 1.32  Smoking tobacco declined at an average of 7% per year between 2003 1.25 accounts for a significant 1.15 0.05 and 2007, whereas manufactured cigarettes declined at an average of 4% 0.05 1.10 0.05 a year over the same time period share of tobacco sales in 0.05 1.0 Sweden at 0.9 billion stick 0.86 0.04 equivalents in 2007 1.21 1.28 1.10 0.5 1.05 0.82 Stick equivalents (bns)

0.0 2003 2004 2005 2006 2007

Cigarettes 7.4 7.1 7.2 6.9 6.3 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos (2.4)% 5.2% (12.4)% (3.2)%  Smoking tobacco 7.8% (17.6)% (22.1)% (7.1)% Total OTP 7.4% (16.9)% (21.7)% (6.9)%

Manufactured cigarettes (1.6)% (3.2)% (9.4)% (4.0)%

Smokeless tobacco sales Smokeless tobacco is the largest segment within OTP market in Smokeless tobacco is the 2003-2007(1) Sweden largest segment within  Legal sales of smokeless tobacco have declined at an average of 0.7% OTP market in Sweden 8,000 6,914 6,937 6,930 6,938 6,715 per year since 2003

6,000

4,000 Tons

2,000

0 2003 2004 2005 2006 2007

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Smokeless tobacco 0.1% 0.1% (3.2)% (0.7)%

Source: (1) OTP volumes supplied by PMI

 Austria  Latvia

 Belgium  Lithuania

 Bulgaria  Luxembourg

 Cyprus  Malta

 Czech Republic  Netherlands

 Denmark  Poland

 Estonia  Portugal

 Finland  Romania

 France  Slovakia

 Germany  Slovenia

 Greece  Spain

 Hungary  Sweden

 Ireland  UK

 Italy

Key source and destination markets(1)(a)(b) Total UK consumption Non-domestic inflows to 2006-2007(1)(2)(b) the UK increased to 13.5 billion sticks in 2007, 100% driven by an increase in 13.0% 15.8% 5.7% both legal and illicit flows 80% 6.9%

60% C&C The growth in inflows is in ND(L) part attributable to Spain 40% 81.3% 77.3% LDC Outflows and the Duty Free channel, 20% but is primarily driven by a 0 Share of total consumption large number of flows (0.8)% (1.8)% from a wide variety of (20)% 2006 2007 source countries Volume (bn sticks): 59.7 59.4

Total non-domestic inflows (ND(L) and C&C) Key:  UK  Major source country  Major destination country 2006-2007(1)(2)(b) Note: (a) Map shows flows over 1% of consumption. Countries which are both source and destination countries are coded according to the larger flow (b) ND(L) research was not updated in 2007, but corroborating research has not indicated any significant changes to 2006 results 14 Sources: (1) KPMG EU Flows Model and interviews with PMI Local Management 12 (2) AC Nielsen ND(L) research, 2006 5.34 Proportion of total EU levels(1)(2) 10 Other countries 4.13 15.5% Portugal 16% 14.8% 8 0.51 0.31 0.76 13.4% 0.68 Greece 6 1.02 1.40 1.61 Poland 12% 0.56 4 Duty Free Volume (bn sticks)

8.2% 2 4.07 4.22 Spain 8% 0 2006 2007

Percentage of EU total 2.2% Volume (bn sticks): 11.1 13.5

0% Consum- ND ND(L) C&C Outflows ption

Historic cigarette prices and legal domestic sales Legal domestic sales have 1997-2007(1)(2) been trending downwards since 1997 as prices have 80 75.5 6 risen 69.9

70 (GBP) per 20 sticks Price 62.1 5 55.2 60 52.4 54.2 53.6 52.0 50.5 49.0 47.0 4 50

40 3 30 2

Volume (bn sticks) 20 1 10 0 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

CAGR (%) 1997-2001 2001-2002 2002-2007  Legal domestic sales (8.8)% 3.5% (2.8)%  Average pack price 7.9% 1.8% 3.7%

Sources: (1) In Market Sales supplied by PMI (2) Weighted average pack price supplied by PMI

Legal domestic sales decreased by 4% in 2007 while the average pack price increased by 10%  Tougher anti-smoking legislation was introduced throughout the UK in 2007 − Smoking bans in public buildings were introduced on the 2nd April in Wales, the 30th April in Northern Ireland and the 1st July in England. Scotland introduced a similar ban on the 26th March 2006 − the legal buying age for tobacco was increased from 16 to 18 years of age on October 1st 2007

Marlboro price comparison Marlboro prices in the UK 1 July 2007(1)(2)(a) remain significantly higher €4.30 than in the rest of the EU €8.16 €4.94 €1.79 €0.86 €1.47 €4.24 €1.33 €0.86

€7.05 n/a €4.71 €8.04 €4.11 €2.10

€4.53 €3.70 €2.46 €0.63 €2.35 €3.90 €2.39 €5.00 €3.82 €2.50 €1.66 €1.58 €1.42 €4.10 €1.64 €1.62 €3.15 €2.95 €3.00 €2.60

Evolution of Marlboro price differential between UK and key markets(2)(a) €3.86 Key:  UK  Major source country  Major destination country 0%

(12)% July 2006 (20)% (14)% July 2007

(33)% (40)% (38)% (46)% (49)% (60)% (61)% (61)% (63)% (63)% (62)% (63)%

Price premium/(discount) versus local (74)% (80)% (77)% Ireland Netherlands Greece Poland

Note: (a) Based on pack of 20 Marlboro King Size on July 1st 2006 & 2007 Sources: (1) KPMG EU Flows Model (2) Marlboro retail selling price supplied by PMI

Consumption breakdown UK consumption is 2007(1)(2)(3)(4)(a)(b) estimated at 59.4 billion Outflows Inflows 1.05 billion sticks 13.47 billion sticks sticks versus legal sales of 70 47.0 billion sticks, giving a net inflow of 12.4 billion

sticks in 2007 60

5.01

1.94 0.46 0.34 1.02 50 0.30 0.17 2.55 0.00 Outflows 1.67 0.49 0.25 0.14 0.16 C&C ND(L)

30 59.4 Volume (bn sticks)

47.0 45.9 20

0 LDS France Netherlands Ireland Other LDC Spain Poland Greece Portugal Other Consumption

Non-domestic (legal) by origin Share of non-domestic (legal) by origin Legal flows are estimated 2006-2007(1)(2)(3) 2007(1)(2)(3) to have increased to 4.1 4.5 4.08 billion sticks in 2007 4.0 3.37 3.5 1.73 3.0 Spain 1.09 Other countries Other countries 41% 2.5 43% 0.27 0.30 Greece 2.0 0.41 0.38 France 1.5 Spain Volume (bn sticks) 1.0 1.61 1.67 0.5 0.0 2006 2007 Greece France ND(L) share of 7% 9% 6% 7% consumption: ND(L) share of non- 30% 30% domestic: The top three non- PMI brands are Non-domestic (legal) by brand Benson & Hedges 2006-2007(1)(2)(3) (1.3bn), John Player Special (0.3bn) and ND(L) analysis Superkings (0.3bn) 4.5 4.08 Ave. Total 4.0 Population Propensity Propensity Ave. no. purchases ND(L) 3.37 19+ to travel to purchase of trips (sticks) (sticks) 3.5 Values 3.0 46.8m x 34% x x 2.7 x = 4.08bn Non-PMI 2007 24% 401 2.5 3.15 Other PMI 2.0 2.61 Marlboro 1.5 EU rank

Volume (bn sticks) 2007 4 17 14 18 3 3 1.0 0.04 0.03 0.5 0.73 0.89 0.0 2006 2007 PMI share of ND(L): 23% 23%

Note: ND(L) research was not updated in 2007, but is assumed to have increased in proportion with the increase in non-domestic incidence i.e. ND(L) share of non-domestic is assumed to equal 30% in both 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) AC Nielsen ND(L) research, 2006 (3) Interviews with PMI Local Management

Counterfeit and contraband by origin Share of counterfeit and contraband by origin C&C inflows increased to 20060-2007(1)(2)(3) 2007(1)(2)(3) 9.4 billion sticks in 2007 10 9.39

7.77 8

5.35 Other countries Spain 6 3.50 27% Greece Other countries 4 0.41 Poland 0.46 57% 1.39 1.02 Spain Volume (bn sticks) 2 2.46 2.55 0 Other countries Poland 2006 2007 includes Duty Free 11% inflows C&C share of 13% 16% Greece consumption: 5% C&C share of non- 70% 70% domestic:

Note: ND(L) research was not updated in 2007, but is assumed to have increased in proportion with the increase in non-domestic incidence i.e. ND(L) share of non-domestic is assumed to equal 30% in both 2006 and 2007. Therefore C&C share of non-domestic is assumed to equal 70% in both 2006 and 2007 Sources: (1) KPMG EU Flows Model (2) AC Nielsen, ND(L) research, 2006 (3) Interviews with PMI Local Management

Non-domestic market volume estimates(1)(2)(3)(4)(5)(a) KPMG’s estimate for non-domestic share is in line with PMI EPS results KPMG’s estimates of non- but is higher than those provided by Customs domestic consumption is  However, Customs estimates are based on illegal cigarettes inflows only in line with those of UK and exclude non-domestic (legal) inflows 35% Customs but below that of 29.0%  KPMG estimates suggest that counterfeit and contraband cigarettes 30% the Tobacco accounted for 15.8% of total consumption in 2007 and 13.0% of total consumption in 2006. Manufacturer’s 25% 22.0% 22.7% 19.0% 18.7%  This estimate is in line with the range of estimates provided by UK Association 20% 18.0% Customs for both 2005 and 2006 15% − for 2006, the range provided by Customs for illicit cigarettes was 10% between 8% to 18%

Share of total of Share consumption 5% − for 2005, the range provided by Customs for illicit cigarettes was between 10% to 19% 0% KPMG estimates are also in line with the results of Gallaher’s 2006 study  The Gallaher study was based on a pack swap approach Customs Customs TMA 2006 2004/2005 2005/2006 EU Flows EU Flows

Model 2006 Model 2007 The TMA estimate of 29% reflects a consensus view of members rather Gallaher 2006 than a customer or pack analysis based methodology

Note: (a) A range of estimates was provided by Customs in ‘Measuring Indirect Tax Losses’. For 2004/2005 the range was from 10%-19%; for 2005/2006 the range was from 8% to 18% Sources: (1) Measuring Indirect Tax Losses, HMRC, October 2007 (2) Estimates of non-UK Duty paid cigarette consumption, Tobacco Manufacturer’s association, 2006 (3) Estimates of counterfeit and non-UK duty paid cigarette consumption based on Gallaher monthly pack swap data, Tobacco Manufacturer’s association, February 2006 (4) KPMG EU Flows Model (5) PMI EPS Q3 2006 and Q3 2007

Consumption index modelling Historic GCTS data can provide a helpful indicator of market trends to Consumption modelling 1997-2007(1)(2)(3) corroborate the extent of non-domestic consumption in a market results suggest that non-  The rate of change in the GCTS smoking prevalence and average smoking domestic consumption has volume figures provide a useful indication of observed changes in 90 increased since 1997 smoking habits. However, GCTS data cannot be used to quantify total 80 consumption in a market due to exclusions of particular age cohorts and 70 the problem of under-reporting of consumption 60 Consumption  The KPMG consumption index is calculated using the incidence of 50 has declined at a smoking and the average daily consumption (both from GCTS) and the slower rate than 40 LDS since 1997 population of each market (from Euromonitor GMID) 30 59.4 − these figures are then indexed to the earliest year for which we have

Volume (billion sticks) 20 both GCTS and LDS data available, 1997 for the UK 10  If LDS trends differ from those shown by the consumption index, it 0 implies that the relationships between inflows, outflows and domestic 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 consumption have changed over the analysed period Please note that consumption modelling can only identify market trends and cannot be used to quantify total consumption CAGR (%) 1997-2001 2001-2004 2004-2007  Comparing consumption and sales trends helps to understand market  LDS (8.8)% (0.2)% (3.3)% direction and its likely inflow/outflow status. However, it will not identify  KPMG consumption index (1.6)% (1.7)% (5.4)% any consumption gap that is present at the start of the period  KPMG consumption estimate n/a n/a n/a − the gap between indexed consumption and LDS in the final year is therefore only a broad indication of the extent to which the market has Sources: (1) KPMG Consumption Index model (2) In Market Sales supplied by PMI changed over the analysed period (3) KPMG EU Flows Model − in many markets, some level of inflows or outflows are likely to be present in the first analysed year  Any changes in GCTS (methodology, supplier, sample size and selection, etc.) will also have an impact on the conclusions generated The gap between LDS and the KPMG consumption index for the UK increased between 1997 and 2007  Consumption in the UK has declined at a slower rate than LDS since 1997  This result suggests that non-domestic consumption has increased over the period

OTP sales in billion stick equivalents, 2003 to 2007(1)(a)(b) OTP sales account for a OTP sales increased between 2003 and 2007 while manufactured cigarette sales decreased relatively small share of 8.0  Total OTP sales increased at an average of 6.4% per year between 2003 tobacco sales in the UK 7.0 and 2007, while manufactured cigarettes sales decreased by an average 5.9 0.8 of 3.2% per year over the same time period OTP sales increased in 6.0 5.5 5.5 5.6 0.9  Smoking tobacco had the highest growth among the major OTP 0.9 2007 to 7.0 billion 1.0 0.9 categories and now accounts for 89% of total OTP sales(b) by volume equivalent sticks 4.0 6.2 Cigars/ cigarelleos 4.6 4.7 5.1 Smoking tobacco 2.0 4.5 Stick equivalents (bns)

0.0 2003 2004 2005 2006 2007

Cigarettes 53.6 52.0 50.5 49.0 47.0 (bn sticks)

CAGR (%) 2003-2005 2005-2006 2006-2007 2003-2007  Cigars/ cigarillos (6.4)% (3.7)% (6.3)% (5.7)%  Smoking tobacco 2.7% 7.4% 23.1% 8.7% Total OTP 1.1% 5.7% 18.9% 6.4%

Manufactured cigarettes (2.9)% (3.0)% (4.1)% (3.2)%

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

 Appendices

− country results

− summary of country flow refinements

− sources

− scope of work

We have made specific Country Rationale Description Impact modifications where required to refine our Austria  EPS inflows from Slovenia may be  The increase in Slovenian legal sales in the  Inflows from Slovenia and total non- overstated due to timing of EPS research second half of 2007 is assumed to have domestic level reduced by 0.25 results at a country level been driven by increased outflows to billion sticks − EPS was conducted in Q4 2008 which Austria following the change to import is after import restrictions from  Post adjustment inflows from Slovenia were relaxed in June allowances Slovenia assumed to have increased  Inflows from Slovenia reduced by 0.25 from 0.23 billion sticks in 2006 to billion sticks to reflect changes in 0.53 billion sticks in 2007 Slovenian legal sales in the second half of 2007

Belgium  EPS inflows from Luxembourg may be  Inflows from Luxembourg uplifted by 0.11  Inflows from Luxembourg and total understated due to changes in sample billion sticks to bring them more into-line non-domestic level uplifted by 0.11 selection with 2006 EPS and ND(L) research results billion sticks − Luxembourg region of Belgium was  Post adjustment inflows from not included in 2007 EPS Luxembourg assumed to have decreased from 0.47 billion sticks in 2006 to 0.25 billion sticks in 2007

Bulgaria  Implied PMI Duty Free inflows to Bulgaria  PMI Duty Free inflows have been reduced  Duty Free inflows and total non- are higher than actual PMI Duty Free to reflect actual PMI Duty Free sales in domestic level reduced by 0.50 sales in border shops in Bulgaria and border shops in Bulgaria and neighbouring billion sticks neighbouring markets markets − Non-PMI Duty Free inflows reduced by an equivalent percentage  Duty Free inflows reduced by 0.50 billion sticks

Denmark  Non-domestic Marlboro brand share  Marlboro brand share re-calculated based  Difference moved to non-PMI brand appears overstated on average from brand share methodology share one and brand share methodology two

We have made specific Country Rationale Description Impact modifications where required to refine our Finland  Apparent increase in EPS inflows  2007 EPS results include non-domestic  Total non-domestic level reduced by reflects change in EPS timing consumption from a higher numbers of 0.20 billion sticks results at a country level rather than actual increase in non- inbound tourists and Finns returning from domestic consumption holiday − implied consumption is 4% − 2007 EPS was conducted in May and higher in 2007 when June whereas the 2006 EPS was unadjusted EPS results are completed in September used. This increase does not  Inflows to Finland reduced by 12% to give a appear realistic given historical more stable year-on-year consumption trend consumption trends and feedback from market participants

France inflows  Implied EPS inflows from Spain  Inflows from Spain uplifted by 1.29 billion  Non-domestic duty free reallocated appear low given PMI and Local sticks to 3.25 billion sticks to Spanish, Luxembourg, and Italian Customs non-domestic − alternative estimates by PMI and  Implied EPS inflow from Customer Purchasing Survey results − inflows from Spain and Luxembourg appears low given imply outflows from Spain in the region of Luxembourg are assumed to be ND(L) results and relative pricing 3.0 billion sticks to 3.7 billion sticks 5% lower in 2007 compared with compared with EPS results of 2.0 billion 2006, based on a decline in  Implied EPS inflow from Italy appears low given ND(L) results sticks and ND(L) results of 1.8 billion French outbound tourism as sticks measured by Eurostat  Inflows from Luxembourg uplifted by 0.59  Total non-domestic level unchanged billion sticks to 0.95 billion sticks  Inflows from Italy uplifted by 0.16 billion sticks to 0.50 billion sticks to bring into line with observed legal sales trend in French and Italian border regions

France brand  Non-domestic Marlboro brand  Brand share methodology two used which is  Methodology two brand shares share share appears overstated more in line with ND(L) results and domestic applied to total non-domestic volume brand share for key brands  Adjustments made to “Other PMI” and “Other non-PMI” to compensate for changes to key brands

We have made specific Country Rationale Description Impact modifications where required to refine our Germany  Very high share of Ukrainian product at  Herten centre results excluded from  Non-domestic inflows reduced by 3.2 Herten recycling centre distorts analysis, which makes Ukrainian outflows billion sticks to 20.5 billion sticks results at a country level national consumption picture consistent with legal sales trend − remaining recycling centres in the Nielsen region containing Herten re- weighted to reflect regional population share

Ireland  Spanish legal sales results for UK/Irish  See Spain for detail  Non-domestic duty free reallocated to inflows brands suggest there is a shortfall in Spanish non-domestic  Inflows from Spain uplifted by 0.27 billion outflows to Ireland as measured by the sticks  Total non-domestic level unchanged Irish EPS

Ireland brand  Non-domestic Marlboro brand share  Brand share methodology two used which  Methodology two brand shares share appears overstated is more in line with ND(L) results and applied to total non-domestic volume domestic brand share for key brands  Adjustments made to “Other PMI” and “Other non-PMI” to compensate for changes to key brands

Italy  Unusually high incidence of San Marino  Inflows from San Marino reduced by 0.54  Inflows from San Marino and total packs in two cities distorts national billion sticks to provide a more accurate non-domestic level reduced by 0.54 consumption picture national consumption estimate billion sticks − reduced flow level is more consistent with legal sales trend in San Marino

We have made specific Country Rationale Description Impact modifications where required to refine our Latvia  Analysis of the implied consumption trend  Inflows from Estonia and Lithuania  Total non-domestic level suggests EPS results understate non- have been uplifted to bring them into increased by 0.20 billion sticks results at a country level domestic inflows line with 2006 EPS and ND(L) research results − implied consumption is 3% lower in 2007 when unadjusted EPS results are used. − inflows from Estonia uplifted by KPMG consumption modelling by 0.11 billion sticks to 0.11 billion contrast, suggest a consumption increase sticks of over 4% − inflows from Lithuania uplifted by − decline in flows from Estonia and 0.09 billion sticks to 0.09 billion Lithuania is not consistent with changes sticks in relative pricing in the Baltics in 2007

Lithuania  Analysis of the implied consumption trend  Inflows for 2007 reduced by a greater  Total non-domestic level suggests that PMI EPS results may overstate amount than suggested by EPS results reduced by 0.50 billion sticks actual non-domestic inflows in order to give a more stable overall  Non-domestic share of total consumption trend − implied consumption is 23% higher in consumption assumed to 2007 when unadjusted EPS results are  Total non-domestic level reduced by decrease from 43% to 25% used. However, GCTS data and 0.50 billion sticks to 1.74 billion sticks rather than pre-adjustment estimated consumption modelling estimate of 33% − all inflows reduced by same suggest relatively stable consumption proportion levels

− 2007 EPS was conducted in May/April and will overstate full-year non-domestic incidence in an environment of rapidly declining non-domestic consumption

Luxembourg  Outflows of legal domestic sales are not  Increase in legal sales in Luxembourg  Net reduction in legal domestic accurately captured in destination market in 2007 reflects higher outflows rather consumption EPS resulting in unrealistic implied than increased domestic consumption consumption trend  Net outflow (modelled to outside of the EU) uplifted from 1.96 billion sticks in 2006 to 2.24 billion sticks in 2007

Malta  Implied large increase in outflows to the UK  Outflows to the UK reduced to 2006  Total consumption 1.0% lower does not appear realistic given changes to level in 2007 rather than 7.5% decline relative prices and tourism flows and impact using original outflow level − outflows to UK reduced from 77 on implied consumption million sticks to 27 million sticks

We have made specific Country Rationale Description Impact modifications where required to refine our Netherlands  Large increase in inflows from Belgium  Inflows from Belgium reduced by 0.30  Belgian non-domestic reallocated does not appear realistic given implied billion sticks to 0.74 billion sticks to “other not-specified” non- results at a country level consumption trends, legal sales domestic; total non-domestic changes and relative changes in pricing level unchanged  Post adjustment inflows from Belgium assumed to have increased from 0.54 billion sticks in 2006 to 0.74 billion sticks in 2007

Poland  Non-domestic Marlboro and L&M  Marlboro brand share re-calculated based brand share appears overstated on average from brand share methodology one and brand share methodology two  L&M brand share re-calculated based on average from brand share methodology one and brand share methodology two

Romania  EPS Duty Free inflows do not reflect  Duty Free inflows increased by 0.80 billion  Duty Free inflows and total non- full year picture in Romania sticks to 1.29 billion sticks to reflect full domestic level increased by 0.80 year Duty Free sales in border regions billion sticks − EPS was conducted in September when a number of Duty Free outlets remained closed (due to changing regulations) or had only recently reopened

Spain inflows  Implied EPS inflow from Italy appears  Inflows from Italy reduced by 0.16 billion  Italian non-domestic reallocated too high given ND(L) results and sticks to 0.43 billion sticks to Duty Free non-domestic; total relative pricing non-domestic level unchanged

Spain outflows  Analysis of Spanish consumption and  Outflows to the UK and Ireland increased  Spanish outflows increased outflows suggest there is an EPS to reflect sales of UK/Irish brands in Spain, resulting in a net reduction in shortfall to France, Ireland and the UK net of: consumption and therefore, total C&C volume − estimated consumption by UK and Irish nationals resident in Spain and tourist consumption whilst in Spain  Outflows to France increased as above

We have made specific Country Rationale Description Impact modifications where required to refine our Spain brand  Non-domestic Marlboro brand share  Marlboro brand share re-calculated  Difference moved to non-PMI share appears overstated based on average from brand share results at a country level methodology one and brand share methodology two

Sweden inflows  Analysis of the implied consumption  Inflows for 2007 uplifted by a greater  Total non-domestic level uplifted by trend suggests EPS results understate amount than suggested by EPS 0.54 billion sticks non-domestic inflows results in order to give a more stable  Genuine non-domestic share of total overall consumption trend − implied consumption is 14% higher consumption assumed to be stable at in 2007 when unadjusted EPS  Genuine non-domestic inflows uplifted 20.8% results are used. KPMG to give a stable share of total consumption modelling by contrast, consumption compared with 2006 suggest consumption was relatively  PMI counterfeit inflows uplifted stable − higher PMI counterfeit estimate is − decrease in EPS inflows is not in line with estimates from consistent with feedback from Swedish Customs market participants showing an increase in non-domestic consumption

Sweden  Norway EPS data provides evidence of  Outflow to Norway estimated at 0.37  Net reduction of Swedish legal outflows significant inflows of Swedish product billion sticks in 2007 using Norwegian domestic consumption EPS and LDS data

UK inflows  Spanish IMS results for UK/Irish brands  See Spain for detail  Non-domestic duty free reallocated to suggest there is a shortfall in outflows Spanish non-domestic  Inflows from Spain uplifted by 2.31 to the UK as measured by the UK EPS billion sticks  Total non-domestic level unchanged

UK brand share  Non-domestic Marlboro brand share  Brand share methodology two used  Methodology two brand shares appears overstated which is more in line with ND(L) applied to total non-domestic volume results and domestic brand share for key brands  Adjustments made to “Other PMI” and “Other non-PMI” to compensate for changes to key brands

A limited number of Non-domestic (legal) results validation process adjustments to ND(L) Nature of adjustment Impact results have been made on Corroboration with total non-  Romania: reallocation of 898 million sticks from ND(L) to C&C the basis of corroborating domestic volumes for both brand and − ND(L) research overstated flows from more expensive EU countries. Volumes capped at total non- evidence country domestic inflows from those countries  Small discrepancies may exist  Sweden: reallocation of 200 million sticks from ND(L) to C&C The net impact of these between the ND(L) data and total adjustments in 2007 is a non-domestic volumes which leads − ND(L) research overstated flows from Finland and Denmark. Volumes capped at total non-domestic inflows from those countries 0.8 billion stick to a negative C&C level  Belgium: reallocation of 122 million sticks from ND(L) to C&C reallocation from C&C to − ND(L) research overstated flows from Spain. Volumes capped at total non-domestic inflow for Spain ND(L) Review of key indicators for specific  Czech Republic to Germany flow: reallocation of 1.3 billion sticks from C&C to ND(L) flows − Average packs per trip brought by smokers identified by research appears understated at nine packs given  Results for a small number of flows savings available (€2.25 per pack of Marlboro) into Finland and Germany suggested − Average packs per trip increased to 20 in line with local market observations and observed purchasing under reporting of pack purchases behaviour on other key shopping routes with similar price differentials between markets given the significant economic incentives to buy  Poland to Germany flow: reallocation of 762 million sticks from C&C to ND(L) − Average packs per trip brought by smokers identified by research appears understated at nine packs given savings available (€2.51 per pack of Marlboro) \ − Average packs per trip increased to 20 in line with local market observations and observed purchasing behaviour on other key shopping routes with similar price differentials between markets  Russia to Finland flow: 71 million stick reallocation from C&C to ND(L) − average packs purchased by smokers very low at one pack per trip given significant price savings available (€3.47 per pack of Marlboro) − Average packs per trip increased to ten in line with import allowance

Cross referencing with tourist and  France to Luxembourg flow: 276 million stick reallocation from C&C to ND(L) border crossing data − visitor numbers reported during research programme understated versus actual data  Number of trips made is a key driver − Flow volume recalculated based on actual visitor numbers and average packs per trip data from survey of ND(L) volumes, particularly where there is a large differential between  Spain to Ireland flow: 189 million stick reallocation from C&C to ND(L) cigarette pricing and stringent − visitor numbers reported during research programme understated versus actual data import restrictions between neighbouring countries − Flow volume recalculated based on actual visitor numbers and average packs per trip data from survey

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

 Appendices

− country results

− summary of country flow refinements

− sources

− scope of work

Our assessment Other sources incorporates analysis from several sources  Attitudes of Europeans towards tobacco, Eurobarometer, 2006  Austrian Chamber of Commerce empty pack survey, 2005, 2006 and 2007  “BAT welcomes government efforts to clamp down on black market cigarette sales”, MTI – EcoNews, 27th December 2006 (Hungary)  “BGN 150 million lost from illegal cigarette trade, PARI Daily, 6th June 2007 (Bulgaria)  “Black market for cigarettes contracts because of crackdown”, MTI – EcoNews, 24th April 2007 (Hungary)  “Bulgaria losses BGN 180 M from illegal cigarettes”, Trud Daily, 24th September 2007  C Ben Lakdhar, “Quantitative and qualitative estimates of cross-border tobacco shopping and tobacco smuggling in France”, Tobacco Control, 2008  “Cigarette sales seen down 3.0 – 5.0 pct on smoking ban in Germany”, German News Digest, 11th April 2008  “Consumption of and trade in illegal tobacco products in Estonia”, Estonian Institute of Economic Research, April 2006  Consumer purchasing habits, PMI France, 2006  “Control Coincidental del Consumo de Tabaco Rubio”, ASM Grupo, 2007 Spain)  “Cost of cigarettes increases in France”, Les Echos, 1st October 2007  Estimates of counterfeit and non-UK duty paid cigarette consumption based on Gallaher monthly pack swap data, Tobacco Manufacturer’s Association, February 2006  Estimates of non-UK Duty paid cigarette consumption, Tobacco Manufacturer’s Association, 2006  Estimates of Member State population, Euromonitor Global Market Information Database (GMID), retrieved April 1st 2007  Estimates of Member State population, Oxford Economic Forecasting, retrieved April 1st 2007

Our assessment Other sources Other sources incorporates analysis from several sources  European Commission website, http://ec.europa.eu  “Poland will enact fuel and tobacco excise hikes in January 2007”, PAP th  “European Trends towards non-smoking provisions”, European Market Insider, 29 November 2006 Network for Smoking Prevention, December 2007  Smuggling Trends, Polish Tobacco Merchants Association, 2006  “European Union Industry Size" PMI Lausanne, 2006  Rompress, 27th June 2007 (Romania)  Gallaher empty pack survey 2006 (Austria)  Rompress, 30th January 2007 (Romania)  “Appearance of Cheap cigarettes facilitates growth of tobacco  Rompress, 7th February 2007 (Romania) market”, ELTA, 30th March 2006 (Lithuania)  Seizure data supplied by PMI Management based on notifications  “Hungary black market clampdown yields results”, Reuters News, received from local law enforcement agencies as of 28th March 2008 10th July 2007  Smoking restriction information from the European Public Health Alliance  “Illicit Cigarette Market”, Empty Pack Surveys carried out by GFK website, http://eupha.org for MDSZ (Hungarian Tobacco Manufacturers’ Association), 2005,  “Spain increases minimum tax on cigarettes”, El Pais, 11th November 2006 and 2007 2006  Imperial Tobacco Pack Count Study, April 2006 (Slovakia)  “Status over Graensehandel”, Denmark Ministry of Finance, June 2007  “Irish chains hit by tobacco sales laws”, The Grocer, 17th  “Stricter checks boost customs office revenue”, MTI – EcoNews, 18th November 2007 December 2006 (Hungary)  ITL Event Collection EPS, 2007 (Ireland)  “Taking advantage of cigarette sales”, Polish News Bulletin, 9th  “Le chiffre”, Aujourd’hui, November 2007 (France) December 2005  Measuring Indirect Tax Losses, HMRC, October 2007 (UK)  Tax breakdown/cigarette prices for EU Member States, UK Tobacco th  National Manufacturers Association survey results 2006 and 2007 Manufacturers Association, 13 March 2008 (Belgium)  Tourism data from the Eurostat website, http://ec.europa.eu/eurostat  Nine O’Clock, 28th June 2008 (Romania)  Un paquet de blondes sur cinq selon BAT/EPSY”, Revue de tabacs,  PMI Poland consumer research, October 2007 March 2007 (France)  World Cigarette Production and Consumption Report – 2005 Edition, Tobacco Merchant Association, 3rd January 2006 (Bulgaria)  “Yellow Bag” survey, an empty pack survey undertaken by the German Cigarette Industry Association (VDC), 2005 to 2007

 Project overview and timing

 Methodology

 EU cigarette market

 2007 Project Star results

 Appendices

− country results

− summary of country flow refinements

− sources

− scope of work

1. Through this study we will update individual research reports covering the estimated 5 Interviews and data from external sources will be obtained on a best efforts basis. size and composition of the total cigarette market (including counterfeit and We will work with PMI to identify and contact key customs and Manufacturer’s contraband products), as detailed below, for each of the 25 EU member states Associations members. We will require access to identified PMI personnel covered in the first phase of Project Star; i.e. the 25 member states as at September throughout this project and our ability to deliver this scope depends on this access 2005. being made available 2. In addition we will cover the cigarette market in the two new member states of 6 We will design and develop a multi-year multi-country database EU flows model. Romania and Bulgaria from year two. This shift from a spreadsheet-based tool to a database approach will have significant 3. The findings from the 27 member states will be used to produce an overall view of benefits in: comparability of year-on-year results; ease of adding in new countries the total EU market. We will also comment on counterfeit and contraband cross (e.g. Croatia) or incorporating other market developments; and speed of analysis border flows on a pan-European basis. given the large volumes of data involved. 4. Information from several independent sources will be combined assisting 7 We will communicate our final findings to OLAF and the member states as set out in representativeness and reliability. These sources will include: Appendix 3. Provision for more extensive communication of findings to OLAF or • Tobacco industry research and statistics the member states is not covered by this engagement letter. - Sales statistics, consumer surveys and empty pack surveys 8 PMI will be responsible for the development and maintenance of country dashboards provided by PMI and/or Tobacco Manufacturers’ for each of the member states. Associations. - In year two, empty pack surveys will be conducted in the following countries; Austria, Belgium, Bulgaria, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, UK. - In years three and four empty pack surveys will also be conducted in Croatia. • Existing public studies and statistics - Research and data published by government agencies (including Ministries of Finance), health bodies, customs authorities, market researchers and academics • Independent research - New research designed to calibrate and augment existing data. - For year two, surveys to analyse the flows of non-domestic (legal) sales will be undertaken in the following countries; Belgium, Bulgaria, Estonia, Finland, France, Germany, Ireland, Italy, Lithuania, Netherlands, Poland, Romania, Slovakia, Spain, Sweden • Third party research will be directed by KPMG, but will be contracted directly by PMI. Where agreed, data gained through this research will be passed directly to KPMG, and will not be released to PMI, for example data regarding the prevalence of smoking among the juvenile segment of the population • Expert opinions and expert panel data - Structured interview programmes designed to capture and quantify the opinions of relevant expert groups including, among others, customs and law enforcement officials