USF Libraries
USF Digital Collections

The single market and pharmaceutical industry in the European Union

MISSING IMAGE

Material Information

Title:
The single market and pharmaceutical industry in the European Union is there any evidence of price convergence?
Physical Description:
Book
Language:
English
Creator:
Timur, Aysegul
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Market integration
European pharmaceutical market
Unit root
Hedonic regression
Laspeyres
Paasche
Dissertations, Academic -- Business Administration -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: During the last two decades, the European Union (EU) has experienced closer market integration through the removal of trade barriers, the establishment of a single market, and the reduction of exchange rate volatility. In addition, there have been several structural reforms in product markets designed to increase competition, monitor cross-country price differences and increase transparency. One anticipated effect of market integration is price convergence, because of the reduced potential for price discrimination across the EU. This dissertation explores market integration and price convergence in the European pharmaceutical market, which is the fifth largest industry in the EU. Since 1985, many EU directives have been adopted to achieve a single EU-wide pharmaceutical market, with the aim of enhancing the quality of life for European citizens and the European pharmaceutical industry's competitiveness and research and development capability. Using annual 1994--2003 data from five EU countries on prices of drugs used to treat cardiovascular disease, this dissertation explains how the integration process has affected cross-country drug price dispersion in the EU. The results show strong evidence of price convergence in the pharmaceutical market, with long term price differences arising from country fixed effects.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Aysegul Timur.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 154 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001926313
oclc - 191864149
usfldc doi - E14-SFE0001911
usfldc handle - e14.1911
System ID:
SFS0026229:00001


This item is only available as the following downloads:


Full Text

PAGE 1

The Single Market and Pharmaceutical Industry in the European Union: Is There Any Evidence of Price Convergence? by Aysegul Timur A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Economics College of Business Administration University of South Florida Major Professor: Gabriel Picone, Ph.D. Donald M. Bellante, Ph.D. Jeffrey S. DeSimone, Ph.D. Mark G. Herander, Ph.D. Bradley P. Kamp, Ph.D. Date of Approval: December 14, 2006 Keywords: Market integration, Europ ean pharmaceutical market, unit root, hedonic regression, Laspeyres, Paasche Copyright 2007 Aysegul Timur

PAGE 2

Dedication This dissertation is dedicated to special fr iends, an institution, and most especially to my family, my son and my husband. To my dear friends, my mentors, Susan Casey, Katherine Dew and Barbara Caldwell: from th e first day I started this journey, you have never stopped encouraging and helping me I could not have made it without you. To International College and especially, Dr Frederick Nerone: I am thankful for your belief in me, your continuing support and lead ership. To my family: even though we are many miles apart on different continents, your spirit was with me and helped me accomplish this goal. Finally, to my son Ef ehan and my husband Mete: I would not be where I am today without your never ending love, support and patience. You both gave me strength to move forward and reach my goa l. We dreamed and did it together. I love you so much.

PAGE 3

Acknowledgments I am truly thankful to my disserta tion committee, Bradley P. Kamp, Ph.D., Donald M. Bellante, Ph.D., Jeffrey S. DeSi mone, Ph.D., and Mark G. Herander, Ph.D. for their helpful comments and advice, and mo st especially I would like to thank my dissertation committee chair, Gabriel Picone Ph.D., for his tremendous help, patience and advice. I could not have finished this dissertation w ithout their support. I also gratefully acknowledge financial support from the Kenneth O. Johnson Endowed Professorship Fund at the Kenne th Oscar Johnson School of Business, International College, and the Gaiennie Foundation at the College of Business Administration, University of South Florida. These two funds helped me purchase the data used in this dissertation. I also appreciated the helpful comments and suggestions from Dr. Naci Bitik, Turgan Gurmen, Cihangir Topkar, Mine Kurt ay, M.D., Pierre Wertheimer and Jacques Perrotto at the beginning of my research that led me to the right resources. Lastly, I would like to tha nk IMS Health, Barbara Doyle, IMS Account Manager, and Brigitte Baker, Senior Market Research Analyst, for providing me with unlimited assistance in understanding the data set. W ithout their support, I would not have been able to utilize the data.

PAGE 4

i Table of Contents List of Tables................................................................................................................. ....iii List of Figures................................................................................................................ ....vi Abstract....................................................................................................................... .....viii Chapter 1 Introduction......................................................................................................... 1 1.1 The Formation of the European Union..............................................................1 1.2 The Impact of European Inte gration on Price Convergence: Theory................................................................................................................2 1.3 The Single Pharmaceutical Market ...................................................................4 Chapter 2 The Background of the EU Pharmaceutical Industry..........................................8 2.1 Distinguishing Characteristics of the Pharmaceutical Industry.........................8 2.2 Regulating Pharmaceuticals in the EU..............................................................9 2.2.1 National Health Care Systems..........................................................11 2.2.2 Regulating European Pharmaceutical Prices and Reimbursement................................................................................ 11 2.2.2.1 Supply-Side Regulations........................................................12 2.2.2.2 Demand-Side Regulations.....................................................14 2.2.3 Parallel Importing in Pharmaceuticals in the EU..............................15 2.2.4 EU Pharmaceutical Background Summary......................................18 Chapter 3 Literature Review..............................................................................................19 3.1 Market Integration and Pr ice Convergence in the EU.....................................19 3.2 Impact of Parallel Trade in the EU..................................................................23 3.3 Cross Country Price Differences for Pharmaceuticals.....................................25 Chapter 4 Research Design................................................................................................28 4.1 Objectives and Hypothesis...............................................................................28 4.2 Description of Data..........................................................................................29 4.2.1 Definition of Drug and Characteristics of IMS Health Data...............................................................................29 4.2.2 Sample Construction ........................................................................32 4.3 Methodology ...................................................................................................33 4.3.1 Price Indexes.....................................................................................33 4.3.1.1 Specification of the Indexes..................................................35

PAGE 5

ii 4.3.2 Quasi-Hedonic Price Regressions ...................................................37 4.3.2.1 Specification of the Quasi-Hedonic Price Models.........................................................................38 4.3.2.2 Description of Variables.......................................................39 4.3.3 Price Convergence Regressions ......................................................41 4.3.3.1 Specification of the Price Convergence Models...................42 Chapter 5 Research Results ..............................................................................................45 5.1 Unadjusted Bilateral Sta ndard Unit Price Differences....................................45 5.1.1 Standard Unit Price Differences for Bilaterally Matched Molecules...........................................................................46 5.1.2 Standard Unit Price Differences for Global Molecules .............................................................................51 5.1.3 Country Price Differences for All and Global Molecules, Relative to 1994..................................................55 5.2 Quality Adjusted Standard Unit Price Differences .........................................55 5.2.1 Drug Quality and Market (C ompetition) Characteristics..................56 5.2.2 Quality Adjusted Standard Unit Price Differences for All Molecules....................................................................................58 5.2.3 Quality Adjusted Standard Unit Price Differences for Global Molecules .............................................................................62 5.3 Price Convergence Results..............................................................................65 Chapter 6 Conclusions.......................................................................................................71 6.1 Main Findings..................................................................................................71 6.2 Limitations.......................................................................................................75 6.3 Future Research...............................................................................................76 References..................................................................................................................... .....79 Bibliography................................................................................................................... ...91 Appendices..................................................................................................................... ....92 Appendix A: Tables...............................................................................................93 Appendix B: Figures............................................................................................143 About the Author...................................................................................................End Page

PAGE 6

iii List of Tables Table 1 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Germany.........................................................47 Table 2 Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Germany....................................................52 Table 3 Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Molecules Overall.................................................56 Table 4 Pharmaceutical Drug Quality and Market (Competition) Characteristics for Global Molecules Overall............................................57 Table 5 Quasi-Hedonic Price Regression Results for All Molecules, Relative to Germany.........................................................60 Table 6 Quality Adjusted (by RE) Standard Unit Price Differentials for All Molecules, Relative to Germany...............................61 Table 7 Quasi-Hedonic Price Regression Results for Global Molecules, Relative to Germany....................................................63 Table 8 Quality Adjusted (by RE) Standard Unit Price Differentials for Global Mol ecules, Relative to Germany.........................64 Table 9 Results for Price Convergence Estimations for All Molecules (Adjusted by FE)................................................................67 Table 10 Results for Pri ce Convergence Estimations for Global Molecules (Adjusted by RE)....................................................70 Table A.1 National Controls for Pharmaceutical on the Supply-Side........................93 Table A.2 Summary of Approaches in the Regulation of Pharmaceutical Prices by On-Patent and Off-Patent Drugs (2003)....................................94 Table A.3 Demand-Side Policies in the Member States.............................................95 Table A.4 ATC Therapeutic Categori es for Cardiovascular Disease.........................96

PAGE 7

iv Table A.5 Largest Pharmaceutical Markets in the World...........................................97 Table A.6 Balanced Sample ATC/Molecule and Country Availability for 1994-2003........................................................................98 Table A.7 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Spain.............................................................101 Table A.8 Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Spain.......................................................105 Table A.9 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to 1994..............................................................109 Table A.10 Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to 1994.........................................................111 Table A.11 Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Molecules by Country.........................................112 Table A.12 Pharmaceutical Drug Quality and Market (Competition) Characteristics for Global Molecules by Country...................................114 Table A.13 Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Molecules by Country by Year...........................116 Table A.14 Pharmaceutical Drug Quality and Market (Competition) Characteristics for Global Molecules by Country by Year......................126 Table A.15 Quality Adjusted (by RE) Standard Unit Price Differentials for All Molecules, Relative to Spain..................................136 Table A.16 Quality Adjusted (by FE) Standard Unit Price Differentials for All Molecules, Relative to Germany.............................136 Table A.17 Quality Adjusted (by FE) Standard Unit Price Differentials for All Molecules, Relative to Spain..................................137 Table A.18 Quality Adjusted (by RE) Standard Unit Price Differentials for Global Mol ecules, Relative to Spain.............................137 Table A.19 Quality Adjusted (by FE) Standard Unit Price Differentials for Global Mol ecules, Relative to Germany.......................138

PAGE 8

v Table A.20 Quality Adjusted (by FE) Standard Unit Price Differentials for Global Mo lecules, Relative to Spain.............................138 Table A.21 Results for Price Convergence Estimations for All Molecules (Adjusted by RE)..............................................................139 Table A.22 Results for Price Convergence Estimations for All Molecules (Unadjusted).....................................................................140 Table A.23 Results for Price Convergence Estimations for Global Molecules (Adjusted by FE)........................................................141 Table A.24 Results for Price Convergence Estimations for Global Molecules (Unadjusted)...............................................................142

PAGE 9

vi List of Figures Figure 1 Quality Adjusted (by RE) Standard Unit Price Differentials for All Molecules, Relative to Germany....................................................62 Figure 2 Quality Adjusted (by RE) Standard Unit Price Differentials for Global Molecules, Relative to Germany..............................................65 Figure B.1 Summary of EU Pharmaceutical Background.........................................143 Figure B.2 Pharmaceutical Production in the European Union.................................144 Figure B.3 Total Pharmaceutical Sales in the European Union.................................145 Figure B.4 Bilateral Price Differences for All Molecules Between 1994-2003 by Laspeyres Index, Relative to Germany.............................146 Figure B.5 Bilateral Price Differences for All Molecules Between 1994-2003 by Paasche Index, Relative to Germany................................146 Figure B.6 Bilateral Price Differences for All Molecules Between 1994-2003 by Laspeyres Index, Relative to Spain..................................147 Figure B.7 Bilateral Price Differences for All Molecules Between 1994-2003 by Paasche Index, Relative to Spain......................................147 Figure B.8 Bilateral Price Differences by for Global Molecules Between 1994-2003 Laspeyres Index, Relative to Germany..................................148 Figure B.9 Bilateral Price Differences for Global Molecules Between 1994-2003 by Paasche Index, Relative to Germany................................148 Figure B.10 Bilateral Price Differen ces for Global Molecules Between 1994-2003 by Laspeyres Index, Relative to Spain..................................149

PAGE 10

vii Figure B.11 Bilateral Price Differen ces for Global Molecules Between 1994-2003 by Paasche Index, Relative to Spain......................................149 Figure B.12 Country Price Ch anges for All Molecules by Laspeyres Index, Relative to 1994......................................................150 Figure B.13 Country Price Changes for Global Molecules by Laspeyres Index, Relative to 1994......................................................151 Figure B.14 Quality Adjusted (by RE) Standard Unit Price Differentials for All Molecules, Relative to Spain..................................152 Figure B.15 Quality Adjusted (by FE) Standard Unit Price Differentials for All Molecules, Relative to Germany.............................152 Figure B.16 Quality Adjusted (by FE) Standard Unit Price Differentials for All Molecules, Relative to Spain..................................153 Figure B.17 Quality Adjusted (by RE) Standard Unit Price Differentials for Global Mo lecules, Relative to Spain.............................153 Figure B.18 Quality Adjusted (by FE) Standard Unit Price Differentials for Global Mol ecules, Relative to Germany.......................154 Figure B.19 Quality Adjusted (by FE) Standard Unit Price Differentials for Global Mo lecules, Relative to Spain.............................154

PAGE 11

viii The Single Market and Pharmaceutical Industry in the European Union: Is There Any Evidence of Price Convergence? Aysegul Timur ABSTRACT During the last two decades, the European Union (EU) has experienced closer market integration through the removal of tr ade barriers, the esta blishment of a single market, and the reduction of exchange rate vol atility. In addition, there have been several structural reforms in product markets desi gned to increase competition, monitor crosscountry price differences and increase transp arency. One anticipated effect of market integration is price convergence, because of the reduced potential for price discrimination across the EU. This dissertation explores mark et integration and price convergence in the European pharmaceutical market, which is the fifth largest industry in the EU. Since 1985, many EU directives have been a dopted to achieve a single EU-wide pharmaceutical market, with the aim of enha ncing the quality of life for European citizens and the European pharmaceutical indus trys competitiveness and research and development capability. Using annual 1994 da ta from five EU countries on prices of drugs used to treat cardiovascular di sease, this disserta tion explains how the integration process has affected cross-count ry drug price dispersion in the EU. The results show strong evidence of price conve rgence in the pharmaceutical market, with long term price differences aris ing from country fixed effects.

PAGE 12

1 Chapter 1 Introduction 1.1 The Formation of the European Union After World War II, there was a stro ng belief among a number of European leaders that the only way to secure a lasti ng peace between their countries was to unite them economically and politically. This effo rt began in 1951 when the Treaty of Paris created the European Coal a nd Steel Community, consisting of Belgium, West Germany, Luxembourg, France, Italy and the Netherlands To integrate other sectors of their economies, these six countries signed the Treaties of Rome in 1957, which created the European Atomic Energy Community and th e European Economic Community (EEC). The task of the latter was by estab lishing a common market and progressively approximating the economic policies of Memb er States, to promote throughout the Community a harmonious development of ec onomic activities (Wertheimer 2003). The member states set about removing trade ba rriers between them to form a common market, and in 1967 the institutions of the three Communities were merged, establishing a single Commission. Denmark, Ireland and the United Kingdom joined in 1973, followed by Greece in 1981 and Spain and Portugal in 1986. In 1992, the Treaty of Maastricht formed the European Union (EU) with the introduction of the Single Market Program (S MP), which created a common market with a free flow of goods, services, labor and capital Austria, Finland and Sweden joined in

PAGE 13

2 1995, and the European Monetary Union (EMU) replaced national currencies in 11 of the member countries (all except Denmark, the Ne therlands, Sweden and UK) by the single currency euro on January 1, 2002. The EU th en welcomed ten new countries in 2004, Cyprus, the Czech Republic, Estonia, H ungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia, raising total membership to 25 countries. Bulgaria and Romania expect to follow in 2007, and Croatia and Tu rkey began membership negotiations in 2005 (EUROPA 2005;Fontaine 2003). 1.2 The Impact of European Integration on Price Convergence: Theory The SMP was originally es tablished in a 1985 White Pa per with the goals of eliminating targeted trade ba rriers in product markets and forming policies aimed to ensure that integration brought more competition (Flam 1992). A 1988 European Commission report on costs of non-integr ated Europe, commonly known as the Cecchini report, estimated that the comple tion of a single European market would generate microeconomic gains of 4.3.4 percen t of GDP, with an additional 2.5 percent if supplemented by appropriate macroeconomic policies, a reduction in consumer prices of 6.1 percent and an employment increas e of 1.8 million (Peck 1989;Smith and Wanke 1993). Theory suggests that the SMP, al ong with the Economic and Monetary Union (EMU), have led the continuous European In tegration process since the 1990s. The SMP has been an important driver of change in product markets through the elimination of trade barriers, whereas the EMU supports the in ternal market by facilitating cross-border transactions and making markets more transp arent with the use of common notes and coins across the euro area. However, it is im possible to make a clea r distinction between

PAGE 14

3 the effects of the SMP and EMU on product mark ets because their different elements are continuously and dynamically interacting. The analytical framework suggests three di stinct effects of Eu ropean integration on product markets (European Commission 2002). First, integration provides firms easier access to each others markets, thus increasing competitiveness by eliminating trade barriers. Second, integration increases market size by raising efficiency, because of resource reallocation or economies of scale, and reducing transaction costs. Third, integration and the use of a single currency fa cilitate price comparisons across markets, making consumers more aware and responsiv e to price differences, producers more aware of their competitors responses, and multinational firms less able to segment national markets and maintain profit margins. A result of these three effects should be arbitrage among the member countries which ultimately leads to price convergence. According to th e law of one price, the price of a specific good should not differ signifi cantly across geogra phic locations, beyond differences arising from transport costs, ta x differences, and other systematic locationspecific factors. Price convergence is expect ed in an integrated market, and thereby provides important evidence regarding produc t market integration. However, the relationship between integr ation and price convergence depends on many structural, behavioral, and policy factors that influence pr ice trends. Since these factors vary across markets, price convergence patterns may deviate across industries. The absolute version of the law of one pr ice states that, in absence of transfer costs, identical traded products should sell for the same price in different countries when expressed in a common currency. The intuiti on is that international arbitrage should

PAGE 15

4 operate until prices are aligned. A less extreme version of the law of one price is the relative version, which holds that common curre ncy prices for a particular product should change in the same way over time in diffe rent countries, but allows a stable price differential across markets. 1.3 The Single Pharmaceutical Market The European Commission recognizes that the impact of market integration on the pharmaceutical industry is very complex. Different health care regimes in member states cause difficulties in achieving a single market and cause price variations between countries for the same product. This is unlik e price differentials in other sectors which are the result of market forces, and has brought about the issue of pa rallel trade. The Commission has viewed distortions as barr iers to establishing a single pharmaceutical market and has tried different strategies to address the issues (Permanand and Mossialos 2004). Since 1985, several Community level directiv es were adopted to achieve a single, EU-wide market for pharmaceuticals, guided in parts by the Treaty of Rome objectives, particularly Articles 8, 30, 36, 85, 86 and 92 (Burstall 1991). Formation of the single market (1) gives patients access to the medicine s they need at affordable prices, and (2) creates incentives for in novation and industrial devel opment (European Commission 2005). In response to concerns that the EU pharmaceutical industry was losing its competitiveness due to market divergence, in May 1998 the Internal Market Council recommended that Community policy be aimed at moving further towards a single

PAGE 16

5 market (Communication Commission 1998a 1998b). The Third Round Table of December 1998 established one goal in this effo rt as the development of a single price for pharmaceuticals (Huttin 1999). Among other ba rriers, many member governments have monopsony power (Read 1998). Although subs equent communications and further evidence revealed awareness of such problems,1 little progress on the policy side was made until 2001, when the Commission set up a new High Level Group on Innovation and the Provision of Medicines (G10 Medici nes 2002). In 2002, the new group issued fourteen specific recommendations in five di fferent areas regarding the attainment of a single pharmaceutical market (European Commission 2003). The most relevant recommendation for this dissertation is the sixth, which addresses the issue of the functioning and evaluation of the single market in pharmaceuticals. Specifically, it covers moving toward a competitive market structure for over-the-counter (OTC) products limiting the member states regulatory authority to only medicines purchased or reimbursed by the state. The implementation of this recommendation could be seen as the beginning of EU pharmaceutical market liberalization (P ollard 2002). Even though the single EU market was fo rmally completed at the end of 1992, the pharmaceutical industry has lagged behind because of its unique structural and regulatory components (Kanavos 2000). This makes it quite uncertain whet her price convergence has actually taken place in the EU pharmaceutical market. 1 The report on global competitiveness in pharmaceuti cals concluded that Europ e as a whole is lagging behind in its ability to generate, organize, and sustain innovation processes that are increasingly expensive and organizationally complex (Gambardella, A., Orsenigo, L., and Pammolli, F. 2000. "Global Competitiveness in Pharmaceuticals: A European Perspective," Report pr epared for the Directorate General Enterprise of the European Commission, pp. 1-100.).

PAGE 17

6 Two pertinent developments that occurred during this time frame, i.e. the 1994 2003 sample period, concern medicinal product licensing and patenting. In 1995, two new Community licensing procedures were introduced (Liikanen 2004). One, a centralized procedure, involves applying directly to the European Medicines Evaluation Agency (EMEA) for community-wid e marketing approval. The other is a mutual recognition procedure whereby applications are made to particular member states and approvals are mutually recognized by national marketing authorities (European Commission 2000b) although evaluation of produc t safety, efficacy and quality is still coordinated by EMEA (EFPIA 2002). Both procedures are designed to achieve a European rather than country-specific decision (Jefferys 1995). Implementation of these procedures has had a significant impact on pharmaceutical companies, particularly with regard to the structure of regulatory affairs departments. On the intellectual property side, patent protection is currently provided by both national patent systems and the European Pate nt Systems. Patents granted by the latter become a bundle of patents enforceable in the designated states, which are each subject to national rules. In 2000, the Commission proposed a Community Patent, aiming to establish a single patent that is valid thr oughout the EU. However, EU agreement on the Community Patent is still pending (Europ ean Commission 2000a). On the other hand, the pharmaceutical industry is offered up to 15 years of effective protection from the date of first authorization in the Community, slightly more than the 14 year maximum in the US. The licensing process further protects the data used for license applications for 6 years, compared to a maximu m of 5 years in the US.

PAGE 18

7 The 2010 industry goal of na tional market pricing for all medicines does not necessarily mean a community-wide single pri ce. Different prices may be negotiated with different customers to maximize access and affordability, but free movement of goods will ensure that there is no artificial segmentation of the market. In addition, the industry wants immediate access to all national markets afte r licensing for all medicines (EFPIA 1998). The results of this study provide strong evidence of drug pric e convergence, with half lives of drug price shocks of 3-5 years. Long term price differentials between member states persist despite the removal of trade barriers, because prices are primarily determined by each countrys distinct health care system and pharmaceutical pricing regulations. However, these price differences have recently been undermined by parallel trade, in which traditionally high price countries import lo wer priced goods from other EU countries. In addition, attempts to estab lish a single overall mark et in the EU over the last decade have reduced price differe ntials in the pharmaceutical industry. The remainder of this dissertation is orga nized as follows. Chapter 2 provides a detailed look at the background of the pha rmaceutical industry. Chapter 3 reviews the literature on market integration and price conve rgence in the EU, the impact of parallel trade in the EU, and cross country pharmaceutic al price differences. Chapter 4 describes the data and empirical strategy. Chapter 5 presents and describes the results of the analysis. Finally, chapter 6 summarizes th e main findings, limitations and areas for further research.

PAGE 19

8 Chapter 2 The Background of the EU Pharmaceutical Industry This chapter provides an overview of the European Union pharmaceutical industry. It discusses various important characteristics, in cluding market structure, national health systems, pri ce regulations and reimburseme nts, and parallel imports. 2.1 Distinguishing Characteristics of the Pharmaceutical Industry The pharmaceutical industry is one of the worlds most research-intensive industries, generating new drugs that satisfy vital consumer needs in health care by saving lives and significantly increasing quality of life. The industry is a crucial component in delivering health care (Scher er 2000). A defining difference from other industries is that several third parties, besides the manufactur er and consumer, are involved on both the demand and supply sides. Physicians, not the consumer, usually determine what drugs to purchase. Pharmaci sts usually follow physicians instructions on what to dispense, but their decisions can be influenced by payment methods when multisourced products are availabl e (Kanavos 2001). The consumer rarely pays the full price of the drug, with subsidies coming from govern ments, health insurance funds, and private insurance companies. They take part in pric ing and reimbursement decisions. Because of these unique characteristics of the pharmaceu tical industry and their interactions with

PAGE 20

9 regard to size and intensity, the drug market fails to meet the criteria for a perfectly competitive market (Mossialos et al. 2004). In the pharmaceutical industry, there are market imperfections on both the demand and supply sides. On the demand side, the demand does not reflect the marginal benefit to consumers, because the demand mi ght to some extent reflect the doctors preferences rather than the consumers. Als o, if the patient has in surance coverage, the price the patient pays is lower than the market price. Even if market price is equal to marginal cost (MC), therefore, there is no guara ntee that these equal the marginal benefit. Further, there is often monopsony power, particularly when ther e is a national health care system. On the supply side, supply is charac terized by economies of scale because drugs have a high proportion of cost in R&D, pate nt restrictions allo w producers to act as monopolists, and governments impose marke ting restrictions through the approval processes and testing requirements (Capri and Levaggi 2005;2004). As a result, the pharmaceutical industry is among the most h eavily regulated industries (Folland et al. 2004). Few aspects of the industry are unaffected by regulatory controls. 2.2 Regulating Pharmaceu ticals in the EU Regulating the pharmaceutical industry is a particularly difficult challenge for policy makers, who seek low health care co sts and affordable drugs, but also want accessibility to the highest quality medicines and more generally a successful industry (Permanand and Mossialos 2004). Despite th e earlier described EU-level movement toward the single pharmaceutical market and the European Commissions expanding role

PAGE 21

10 in this area, pharmaceutical polic y is still primarily determin ed at the national level by differences in health care systems, pricing and reimbursement regulations, demographics, cultures and attitudes towards consuming me dicine (Norris 1998). Pharmacoeconomics, parallel importing, and generic substituti on also impact policy d ecisions (Seget 2003). Despite four decades of Community leve l attempts at convergence, European pharmaceuticals remain 25 separate national ma rkets rather than a single internal market. The Commission has never propos ed legislative measures to address pharmaceutical price controls and reimbursement regulations at the EU level, considering this to be primarily a national concern. Ev en though the industry continues requesting that the Commission remove certain forms of national price regulation, the industry cannot be wholly protected from inter-brand competition from generics following patent expiration or intra-brand competition created by parallel importing from lower priced EU countries (Mossialos et al. 2004) However, some national governments have been taking me-too approaches for regulating drug pr ices and controlling re imbursement (Redwood 1994), imitating policies of other governments despite the limited effectiveness and evaluation of many of the measures adopted. This is described in the literature as the penguin effect (Guillen and Cabiedes 2003). Consequently, both i ndustrial policy and regulation of the pharmaceutical s remain responsibilities sh ared between the EU level and the member states. The Commission ha s no power to determine national prices, reimbursement regulations or profit controls but attempts to ensure that national procedures are efficient, transparent and fair.

PAGE 22

11 2.2.1 National Health Care Systems A common political belief in Europe is that governments should ensure that medical care is available to everyone. Each Me mber State has its own health care system to protect public health, provi de patient access to safe and effective medicine, maintain quality of care, and establish measures to m eet these expectations. EU actions are limited by the member states own decisions in managing pharmaceutical goals and budgets (Huttin 1999). The current EU national health care sy stems are diverse in both funding and delivery of health care. Table A.1 shows the di fferences in health care systems in the five major pharmaceutical markets in Europe. National governments have implemented many different measures, from controls and incenti ves to directly influe nce supply and demand or indirectly reduce expenditure s, which in terms of overall health have grown faster than GNP in all European countries over the last 20 years (Ess et al. 2003). To slow this growth, some countries have emphasized dir ect price controls a nd supply side cost containment regulations, while others have em phasized demand side financial incentives, quantity controls and physician educational initiatives (Kanavos 2001;Mossialos et al. 2004). 2.2.2 Regulating European Pharmaceu tical Prices and Reimbursement Regulating pharmaceuticals involves both s upply and demand side regulations, various aspects of which are explained in the next two sections.

PAGE 23

12 2.2.2.1 Supply-Side Regulations Supply regulations and cost containment measures consist primarily of direct fixed price controls, profit (or ra te of return) controls and reference pricing. Tables A.1 and A.2 show how pharmaceutical prices are re gulated in the five major EU markets, which are Germany, France, Italy, Spain and United Kingdom. Direct price controls incl ude negotiated prices, price-caps (fixed maximum price), cost-plus prices, price comparison to other countries or similar products within the same country, price cuts or freezes or price-volum e agreements. Almost all EU countries except Germany and the UK apply direct price c ontrols to on-patent drugs. In the UK new patent drugs can be freely launched but the prices are indirectly controlled by the rate of return controls that ensure pharm aceuticals are not realizi ng excessive profits. The permitted rate of return on capital is ar ound 17-21% (Ess et al. 2003). France also introduced free pricing in 2003, but only for products defined as innovative by the National Transparency Commission with some limitations. In some countries such as France, Italy, Spain, Austria, and Portuga l, prices are direc tly controlled through negotiations. In others, prices are fixed by national authorities according to a list of factors that depend on whether the main objectiv e is to achieve the lowest possible price or a price that balances pr ofitability with cost contai nment. Many countries have additionally applied cuts and freezes to the ma ximum fixed prices, often in an attempt to meet short-term budget constraints. In France, pr ices are set initially fo r 5 years. In most countries, price cuts have been the norm. Some countries, like Spain, reward companies that cont ribute to the economy or invest in research and development. In addition to using a cost-plus formula, Spain

PAGE 24

13 considers therapeutic value. Price comparis ons are another common measure. However, there is concern about ac curacy of the comparisons because of methodological difficulties and differences across countries in strength, formulation and pack sizes available (Mossialos et al. 2004). Another measure, reference pricing, has gain ed popularity over th e years in places like Germany, Italy, France, and Spain. Refere nce pricing is a syst em where the buying agent decides on a reimbursement price and th en the user/patient or insurer pays the difference if the chosen medicine is mo re expensive (Lopez-Casasnovas; Puig-Junoy 2000). Reference pricing aims to control pharmaceutical expenditures by defining a fixed amount to be paid by th e government or other third pa rty payer and can effectively eliminate price gaps between therapeutically similar pr oducts and improve market transparency by increasing patient and phys ician awareness of actual price levels (Dickson 1992). The latter can br ing about switches to cheape r drugs that lead to price decrease for the more expensive version (M ossialos et al. 2004), which encourages downward price convergence (LopezCasasnovas; Puig-Junoy 2000). Current EU price control sy stems limit the returns to any added therapeutic value of the drug. Reimbursement levels refl ect negotiations between the pharmaceutical company (a monopolist with respect to a new drug) and the government or insurer (a monopsonist). A number of countries have started to incorporate further economic evaluations into the decision-making process, e ither as an additional tool to determine the reimbursement price (e.g. Finl and) or as a mechanism to guide prescribers (e.g. the UK National Institute of Clinical Excellence). Ho wever, there has been little consistency in EU reimbursement regulations.

PAGE 25

14 2.2.2.2 Demand-Side Regulations Key elements of demand-side policies in member states are shown in table A.3. One is to influence the doctors who prescribe medicines for patients (Caves et al. 1991). This can be done through positive and nega tive lists, issuing guidelines to which medications can be prescribed for certai n conditions, and monitoring prescribing practices (doctors act as a gatekeeper). Additionally, budgets are imposed to force doctors to take costs into c onsideration when selecting between alternative treatments (e.g. individual doctor or group practice budgets in the UK, budgets for all doctors in a region in Germany). The second key element is restrictive lists that all member states operate in various ways. Regulatory approval that is n ecessary before a drug can be marketed does not imply that the drug will be covered by the h ealth care system; in principle, drugs that are less effective or more expensive than s ubstitutes should not be reimbursed. These lists operate in three different ways. In some countries, the drug must be on the positive list to be reimbursed, while inclusion on the negative list implies no government reimbursement. In others, only one list is use d. In positive list counties, the drug must be on the list to be reimbursed. In negative list countries, only drugs on the negative list are not reimbursable. Some countries (e.g. France) have been altering the system of paying physicians, moving from a fee-for-service and access to any physician/specialis t regime to more restrictive gatekeeper systems. An alternat ive way to regulate demand is regulating what products pharmacists can sell, who may sell prescription medicines, what they can dispense, how prescriptions are written a nd substitution procedures (Kanavos 2001).

PAGE 26

15 Another option is patient cost-sharing th rough paying some combination of a proportion of the total price, a fixed ch arge per prescription, and an a nnual deductible (Noyce et al. 2000). For example, in Spain co-payments are 40% of the sales pric e, while in France the majority of the population pays less than 5% of retail prices out of pocket. Finally, the size of the generic market has grown recently in several EU countries. For example, in the UK, the use of generic dr ugs has increased from 16% of prescriptions in 1977 to 54% in 1994 (Ess et al. 2003). Tabl e A.2 shows off-patent drug regulations. The two EU approaches to regulating generic drug prices are limiting the generic price to a fixed percent less than the or iginator product or the cheapest generic equivalent, and to apply a reference price scheme (Mossialos et al. 2004). In some countries (e.g. Germany), generic substitution has been a su ccessful short-term cost containment policy (Ess et al. 2003). 2.2.3 Parallel Importing in Pharmaceuticals in the EU Ganslandt and Maskus (2004) define parall el imports as leg itimately produced goods imported legally into a country without the authorization of a trademark, copyright, or patent holder. The legal foundation is the principle of the free movement of goods, while the legal doctrine governing the permissi bility of parallel imports is exhaustion. Patent distribution rights are exhausted ove r a pre-defined area upon first sale, after which the patent holder can no longer restrict the circulation of the product. Parallel trade is thus permitted in the geographical ar ea where the rights to control distribution have been exhausted but not from regions or countries outside (Ganslandt and Maskus 2004). Under national exhausti on the right holder may preven t such importation. In the

PAGE 27

16 EU, the exhaustion regime is the Community (Maskus and Chen 2004), and the Commission recognizes the manuf acturers right to control th e use of the brand but not necessarily of the product. Specifically, the manufacturers indus trial and commercial property right cannot be used to prevent th e parallel import of a medicinal product that has already been lawfully placed on th e market in another member state. The purpose of parallel importing is arbitr age between countries with different prices. Seget (2003) explai ns that parallel importing is the transportation of a pharmaceutical product from its original market, where it was sold directly by its manufacturer or marketing partner, to a di fferent market for resale by the importer. Parallel importing only occurs where there is sufficient difference in the price of the product in two markets to cover the importers costs and generate some profit to the importing company. In most industries wher e it happens, parallel importing had led to the convergence of prices. Pa rallel imports emerge where in ternational price differences exceed the costs of transportation and se lling the product across borders, and hence would not exit without pharmaceutical price differences between member states. According to the Commission, a parallel import must (1) have been granted a marketing authorization in the origin c ountry and (2) be sufficiently similar to a product that has already received marketing authoriz ation in the destination country. Recently, parallel imports have been grow ing as a fraction of EU pharmaceutical sales, as arbitrage opportunities have persis ted despite the goal of a single market, while EMEA harmonization of regulatory requirement s for registration (dosage requirements, labeling) has reduced repackaging costs. The major supplying c ountries are Belgium, France, Italy, Greece and Spain, while the main importers are Denmark, Germany, the

PAGE 28

17 Netherlands and the UK (Chaudhry and Wals h 1995). As of 1999, parallel imports composed 15% of the prescrip tion drug market in the Netherlands, 10% in Denmark and 7% in the UK. Overall, parallel imports made up 8% of the EU market in 2001 and are forecast to rise to 10% by 2006. The inco me loss from being undercut by parallel imports is estimated at $5.5.6 billion in 2001 (Arfwedson 2003). Parallel imports have complex effect s on markets (Maskus 2001), but should cause prices to fall in high pr ice countries, and may reciproca lly cause prices to rise in low price countries (OECD 2000). Accordi ng to Danzon (1998), parallel trade reduces economic welfare by undermining price differe ntials between markets. Theoretically, pharmaceutical R&D, which accounts for roughly 30% of total pharmaceutical costs, is a global joint cost of serving all consumers worl dwide. Optimal pricing to cover joint costs, i.e. Ramsey pricing, requires setting lo wer prices in markets with higher demand elasticities. However, parallel trade tends to force price convergen ce Moreover, in the long run, uniform prices might reduce drug development by limiting returns to R&D (Darba and Rovira 1998;Vogel 2004). The economic literature on parallel tr ade is limited because of the data availability. Most studies finds ambiguous welfare effects (Maskus and Chen 2004). However, these studies provide insights that are useful for frami ng policy and are thus summarized in the s ubsequent chapter.

PAGE 29

18 2.2.4 EU Pharmaceutical Background Summary As shown in Figure B.1, all member countri es have different health care systems and pharmaceutical price controls that creat e market distortions, resulting in price differences. This creates the opportunity for parallel trade, which in combination with the EU single market principle calling for the free movement of goods could lead to price convergence (Danzon 1997b). As discussed in chapter 1, solutions to this conflict between national price regulations, open competitive markets, and reasonable profits for R&D are being debated by the EC. Current ly, EU pharmaceutical prices are relatively low, i.e. 45 percent below US prices. One potential solution, a single European administered price control, is politically inf easible, and it is further unlikely that member nations would give up their market power while having to maintain the responsibility for their health care budgets (Pollard 2003). Therefore, the G10 recommends coordination of national results, not of the underlying regulations themselves. An important question is how this co nflict impacts price convergence among the member states. This dissertation helps enlighten this issue for the EU pharmaceutical industry by looking at empirical evidence of price convergence between 1994 and 2003.

PAGE 30

19 Chapter 3 Literature Review This chapter provides an overview of the literature relating to three different areas: market integration and price convergence in the EU, the impact of parallel trade on prices, and cross country price di fferences in pharmaceuticals. 3.1 Market Integration and Price Convergence in the EU The European Commission recognizes th e importance of price competition and price convergence through a si ngle market program: The removal of barriers and the freedom of supply which businesses will enjoy as a result of the single market program should lead, through increased competitive pr essure, to some downward convergence of prices of benefit to the customer. From the point of view of producers, the competitive pressure will be exerted firs t and foremost on price-cost margins, particularly in those sectors in which they held a certain monopoly power or position. Producers will also be induced urged on by pressure on their margins to become more efficient and thus cut their production and distribution costs. The in creased pressure which will be brought to bear in this way on costs and price-cost margins will be a powerful means of causing prices to converge on levels more consiste nt with economic and technical efficiency (DRI 1996). Furthermore, the EU Treaty also requires the Commission and the European

PAGE 31

20 Central Bank to report on convergence at th e macroeconomic level because convergence will ensure that a single EMU interest rate is appropriate for all participants. Moreover, when the EMU is hit by a macroeconomic shock, a high degree of convergence limits asymmetric economic developments at the country level, which can no longer be addressed by adjusting the exchange rate (European Commission 2004). Finally, price convergence indicates the evoluti on of product market integration. DRI (1996) showed a trend towards pr ice convergence in the EU-12 during 1980 1993 using price indexes for detailed pr oduct and service categories collected by Eurostat. This trend was more pronounced for consumer and equipment goods than for energy, services, and construc tion. The convergence in cons umer products and services has accelerated since the single market progr am. Convergence has been comparatively greater for the three 1989 entrants (Greece, Portugal and Spain) than the EU-9, which may reflect a catch-up effect of integrati on. The product categories with the greatest convergence were in highly traded (more open) industries. Four products and services related to health care had the highest price disparities in 1993. Th e study concluded that 78 product/services categories, represen ting 60% of EU private consumption expenditures, had significant price convergence, compared to only ei ght cases of price divergence (DRI 1996). Using Eurostat panel data for 1975 on monthly consumer price indexes for 12 EU countries, Sosvilla-River o and Gil-Pareja (2004) exam ined how European market integration has affected cross-country price dispersion in the EU. They used the Levin and Lin (1992) convergence equation to test fo r unit roots with panel data, with Germany as the benchmark country based on its central role in the EMS. The estimated speed of

PAGE 32

21 convergence ( ) for the general CPI was .006, which im plies a half life of a shock of 115 months. The highest estimated was for fruits (.073) with a half life of 9 months, and the lowest estimated was for recreation (.001) w ith a half-life of 693 months. The study concluded that there was empirical evidence of price convergence, especially for traded goods. The study failed to obtain su ch evidence in the cas es of non-tradable goods or goods that are subject to special taxes or regulations. A similar study by Gil-Pareja and Sosvill a-Rivero (2004) examined the degree and recent evolution of e xport-price dispersion between 1988 and 2001 among seven EU countries (Belgium-Luxembourg, France, Germ any, Italy, the Netherlands, Spain and the UK) for a number of eight-digit products using the Eurostat data. The data set was based on the annual free on board (f.o.b.) value and quantity of exports to selected OECD countries. It is expected th at relatively fixed exchange rates established by EMS would result in price convergence by imposing pri ce discipline among its members. As a measure of export-price dispersion, they used the coefficient of price variation, i.e. the ratio of the standard deviation to the mean, which is invarian t to changes of scale. To assess price convergence, the time series of variation for each source country-product pair were regressed on a constant and a linea r time trend. The results of the study showed that export-price dispersion was usually lower in the sample than across OECD countries. There was little evidence of convergence, but this was also str onger across the EU countries. The conclusion was that al though monetary stability may aid price convergence, it does not necessarily lead to complete convergence. Two other studies, Rogers (2001) and Roge rs et al. (2001), found direct evidence of price convergence in Europe, using Europ ean price indices from actual prices of 168

PAGE 33

22 goods and services, in 26 cities, in 18 c ountries, between 1990 and 1999. Prices became less dispersed in the euro area and convergence was evident for traded goods, more in the first half of the 1990s than the second half By some measures, traded goods price dispersion across the euro area was close to that across US cities. European Commission (2004) also conc luded that there is a continuing convergence of prices of recent entrants towa rds the considerably higher levels in the EU-15, again due to the catching-up process and price deregulation. As a result, price convergence in the new member states ha s been faster than in the EU-15. Camarero et al. (2000) examined price and inflation convergence between three European countries (Italy, Spain and the UK) The results rejected long-run convergence in all cases but found that prices ca tch up with the European average. There have been several studies of pr ice convergence in sp ecific EU product markets, particularly the car market (G aulier and Haller 2000;Goldberg and Verboven 2001, 2004;Verboven 1996). Goldberg and Verboven (2005) investigated the relationship between in tegration and price convergence us ing panel data on car prices between 1970 and 2000. Approximately 150 ve hicles per year and five markets [Belgium (the benchmark country), France, Germany, Italy and the UK] were included. They found strong evidence of bot h the absolute and relative ve rsions of the Law of One Price. Hedonic regressions were estimated to control for possible variations in characteristics of models across countries, a nd these quality-adjusted prices were used to form the dependent variable for the Levin and Lin (1992) convergence equations. The relative version of the Law of One Price implied half lives of shocks between 1.3 and 1.6 years.

PAGE 34

23 Kerem et al. (2005) estimated the convergen ce of health care expenditures in the EU using , convergence for the 1992 period. The study demonstrated that even though economic integration has facilita ted economic growth, the EUs enlargement process has not brought about ha rmonization of health care ex penditures in the EU-8 new member states (Czech Republic, Estonia, H ungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia). It would take approximately three years for health care expenditures as a share of GDP in the EU-8 countries to move halfway to the EU-15 average. Lastly, Ratfai (2006) examined pri ce convergence among geographically close locations that share the same currency, usi ng a sample of highly disaggregated product level prices of very narrowly defined homogeneous items in Hungary. Employing a series of panel data unit root tests using the Levin et al. (2002) procedure, the results showed that price differentials fading away quickly, with an es timated half-life of between 2.2 and 12.0 months and a medi an half-life of 4.0 months. Price convergence in pharmaceutical market s has not been subject to empirical investigation. However, the pharmaceutical industry was placed high among the most sensitive sectors of the single ma rket program (Allen et al. 1998). 3.2 Impact of Parallel Trade in the EU Parallel imports, also called gray-mar ket imports, is the process whereby goods protected by an intelle ctual property right (i.e. patent, tr ademark or copyright) are placed into circulation in one market, and then re -imported into a sec ond market without the authorization of the local owner of the intelle ctual property right. Ther e is debate as to whether parallel trade leads to lower prices for consumers or whether it undermines

PAGE 35

24 intellectual property protection or both and therefore weakens the incentives to invest in R&D, which may in turn harm the cons umer in various ways (Arfwedson 2003). A countrys law concerning the territorial ex haustion of these rights is an important component of how it regulates and limits thei r use. The EU pursues community (regional) exhaustion but excludes parallel trad e coming from nonmembers (Maskus 2001). Parallel imports of pharmaceuticals dramati cally increased during the last decade. PI were estimated at 4,265 million in the EU in 2003, which represents 5% of the pharmacy market value (at ex-f actory prices) (EFPIA 2005). Economic theory predicts that parallel tr ade (imports) forces price convergence (Danzon 1998;Ganslandt and Maskus 2004;Huttin 1999;Maskus 2001;Towse 1998). However, there is limited empirical work on th e impact of parallel imports on prices. Ganslandt and Maskus (2004) develop a m odel using bi-weekly pharmaceutical product data from Sweden. They use an OLS specifi cation in which prices are affected by the number of parallel importers, the poten tial for parallel import competition and a time trend. The Swedish market provided a natu ral test of parallel trade because it was prohibited until 1995, when Sweden entered the EU and adopted the EU exhaustion principle. They found that the prices of drugs subject to competition from parallel imports fell relative to those of other drugs between 1994 and 1999, concluding that parallel imports significan tly reduced prices of manu factured products by 12%. Chaudhry et al. (1994) interviewed 36 multinational pharmaceutical managers about their expectations of para llel trade in the EU. A major ity expected parallel imports to continue to exist, but opinions were mixe d regarding whether future parallel imports threatened the industry. Pharmaceutical firm s have been considering ways of reducing

PAGE 36

25 the impact of drugs entering EU markets by wa y of parallel trade, especially from Spain (Chaudhry et al. 1997). 3.3 Cross Country Price Diffe rences for Pharmaceuticals International price comparisons in th e pharmaceutical industry have been the subject of several empirical st udies. Schut and Van Bergeijk (1986) compared the prices of identical packages of pharmaceutical products for 32 countries during 1975 and examined whether factors including GDP pe r capita, volume of consumption, population, volume of consumption per capita, patent prot ection, indirect price controls, and direct price controls contributed to price differences. From the OLS estimate, a 10% increase in per capita GDP was associated with 8% highe r drug prices. In addition, direct price control measures resulted in a 20% pr ice reduction. Bulk purchasing through a centralized government agency, promotion of the use of generics and excluding patent protection were also successful in lowering pharmaceutical prices. To date, most studies have focused on pharmaceutical price differences for the US, Europe and Japan as a measure of industry competitiveness. Recently, cross-national price comparisons have been used for drawi ng conclusions about di fferences in average prices, evaluating regulatory sy stem performance, and setting domestic prices as a regulatory policy. Danzon and Kim (1998) argued that prev ious international price comparison studies were biased due to unrepresentative a nd small samples. They analyzed IMS data for the sales of cardiovascular products in the seven countries lis ted below during the year October 1991 to September 1992, defi ning the drug by molecule-therapeutic

PAGE 37

26 category and using units of one tablet, one capsu le, and five ml of a liquid as proxies for a single dose. They reported Laspeyres, Paasche a nd Fisher price indices relative to the US. The Laspeyres differentials of Japan 19.1% Canada 16.6%, Germany .8%, Sweden 12.9%, Italy .6%, UK .4 and France .8% im plies that other studies overstated US price differentials. A general conclusion was that international price comparisons are extremely sensitive to choices made about samp le selection, price and quantity units, the relative weight given to cons umption patterns in different countries, and the use of exchange rates. Danzon and Chao (2000b) investigated cros s-country price differences using IMS data on all molecules for the October 1991 to September 1992 period. They examined the contribution of various product and market characteristics to the dispersion of relative prices in the same countries as the prev ious study, using both price indices and the hedonic regression model. The conclusion wa s that the countries with strict price regulation, France, Italy and Ja pan, have systematically lower prices for older molecules and global products, relative to less-regulated regimes such as the US and UK. In addition, generic competition pr ovides more effective price control in less regulated regimes such as the US. With the same data, Danzon and Chao (2000a) tested the hypothesis that the regulation of manufactur er prices and retail pharmacy margins undermines price competition. They found that price competition between generic competitors is significant in unregulated or less regulated markets (the US, the UK, Canada and Germany) but that regulation undermines generic competition in strict regulatory systems (France, Italy, Japan). Earlier, Danzon (1997a) found that the most

PAGE 38

27 stringent regulatory regimes (F rance and Italy) have perfor med relatively poorly in terms of innovation, while Japan has produced many new drugs but few global drugs. Recently, Danzon and Furukawa (2003) compared the average prices of pharmaceuticals in eight countries (Canada, Chile, France, Germany, Italy, Japan, Mexico and the UK) to those in the US using post-1992 IMS data for 249 leading molecules. The results showed Japans prices to be 27% higher than US prices, and other countries prices ranging from 6% (the UK) to 33% (Canada) lower than US prices. They also concluded that income differential s contribute, both direc tly and indirectly, to price differentials. Garattini et al. (1994) analyzed the differences between the pharmaceutical markets of Italy, the UK, Germany and Fr ance from both the supply and demand sides, taking into account public policy differences that affected public expenditure and industry turnover. The sample included eight drugs that were top sellers in all four countries in 1992. In most cases, retail pri ces were lowest in France and highest in Germany. The analysis concluded that on both the demand and suppl y sides, sectoral differences across in the four countries were striking. Price regulation was one of several variables involved in pharmaceutical policy, wi th demand side regulations being equally important. Ess et al. (2003) showed that, as a consequence of pharmaceutical market fragmentation, prices varied across Europe. Prices were substantially lower in Greece, Spain and France but higher in Belg ium, Switzerland, the UK and Denmark.

PAGE 39

28 Chapter 4 Research Design This chapter focuses on the data and methods used in this dissertation. The first section describes the objectives and hypotheses tested. The second section describes the data analyzed. The third section descri bes the methodologies applied, including the variables and mode l specifications. 4.1 Objectives and Hypothesis As previously outlined, it is expected that market integration of sourcing, retailing and distribution contributes to price c onvergence among EU countries (Coopers 1996). However, because the relationship between European integration and price convergence depends on many different struct ural, behavioral, and policy factors that differ across markets, price convergence patterns are expe cted to vary by indus try. The objective of this dissertation is to test the hypothesis of price converg ence in the EU pharmaceutical industry, which has not yet been studied. Th e methods of controlling for the quality and market characteristics and th e model specifications are adopted from Goldberg and Verboven (2005) and Danzon and Chao (2000b). The analysis also examines bilateral price differences using Laspeyres and Paasche indexes without these adjustments.

PAGE 40

29 4.2 Description of Data The data for this dissertation represent selected cardiovascular pharmaceuticals. They come from IMS Health, which collects and reports sales and price data at the level within the pharmaceutical market supply and distribution chain that provides the most accurate information for a country (IMS 2005a). Products are classified by the Anatomic Therapeutic Category (ATC) system, which is similar to the World Health Organization (WHO) system (WHO 2002), and is devel oped and maintained by EphMrA (EPHMRA 2004). Products are categorized in the sales, medical and prom otional audits according to the EphMrA/PBIRG Anatomical Classification System, the main principle of which is that there is only one Anatomi cal Classification code allocat ed to a product/pack. This allows each product to be clas sified consistently in all c ountries (EphMrA/PBIRG 2005). 4.2.1 Definition of Drug and Charac teristics of IMS Health Data Prior to Danzon (1999), cross country studies compared the price for a single pack in the base country, but this pack may not be typical or even available in other countries (Berndt 2000). Danzon recognized that samp les using only comparison packs with the same ingredient, manufacturer, brand name, dosage form, pack size and strength in each country will exclude generic and OTC products These are likely close substitutes for originator and prescription drugs, respectivel y, so their omission w ill potentially result in unrepresentative samples (Hellerstein 1998), a problem that might be exacerbated by not including all forms and strengths (Ellis on et al. 1997;Schere r 1993;Scherer 2000). Danzon therefore defined the drug by active i ngredient, i.e. molecule (MOL), and ATC without regard to manufacturer and brand name. All forms of a given molecule,

PAGE 41

30 including generics and licensed products, are combined to form a weighted average price per MOL/ATC. Differences between products with identical MOL and ATC are ignored, but this is presumed to be much less problema tic than failing to sample substitutes and all forms and strengths. The IMS Health measure that meets the crit eria of being available for all dosage forms and strengths is the IMS Standard Un it (SU), which defines a single dose as one tablet or capsule, five milliliters of a liquid (i.e. one teaspoon), or one ampoule or vial of an injectable product (IMS 2002, 2005b). Aggreg ation of all dosage forms, strengths, and packs minimizes sample selection bias Danzon and Kim (1998) found that this strategy permits over 90% of sales to be in cluded for the US, the UK and Canada, and encompasses over two-thirds of sa les for most other countries. The ATC divides products into four hier archical groups according to anatomical site of action, their indications therapeutic use, composition and mode of action, etc. in the Anatomical Classification Sy stem, with levels beyond the 1st identifying therapeutic and pharmacological subgroups. For instance, the following scheme illustrates the complete structure of the data for one of th e ATC categories (C10) used in this research: C CARDIOVASCULAR SYSTEM (1st level, anatomical main group) C10 LIPID-REGULATING/ANTI-ATHEROMA PREPERATIONS (2nd level, therapeutic main group) C10A CHOLESTEROL & TRIGLYCERI DE REDUCTION PREPERATIONS (3rd level, pharmacological/therapeutic subgroup) C10A1 STATINS (HMG-CoA reductase inhibitors) (4th level, chemical/pharmaco logical/therapeutic subgroup).

PAGE 42

31 Consider the example of Zocor (M erck), a well-known drug in the C10A category used to reduce the risk of heart atta ck and stroke in patients with multiple risk factors for heart disease such as high cholestero l and high blood pressure. It is difficult to compare the price of Zocor across countries because strengths, forms and pack sizes vary greatly. In the UK, Zocor is available in strengths of 10, 20, 40, and 80 MG, pack sizes of 10, 28, 30, and 100 tablets, and 20 different forms. In Germany, strength levels are 5, 20, 40 MG and MGFT (e.g. FT=Forts means strong), pack sizes are 28, 30, 50, 100, and there are more than 50 different forms. Moreover, many other products have the exact same MOL, Simvastatin, and are thus close substitutes. And Lipitor (Pfizer) serves the same purpose and is thus unde r the same therapeutic category (C10A) as Zocor, but has a different MOL, Atorvastati n, as well as different pack sizes, strengths and forms. To avoid the above problem, the main un it of analysis in this study is the molecule-indication, defined by a single MOL and three-digit (3rd level) ATC (although results change little if the AT C requirement is dropped and the drug is defined simply as the MOL). A countrys SU price for a MOL/ ATC is its volume-weighted average price per dose over all presentations, including gene rics, licensed, OTC, a nd parallel imported products (Danzon and Furukawa 2003). Multip le molecule drugs are excluded because the relative mix of active ingred ients varies across countries. Because of data availability and cost, this study is limited to retail sales of drugs for cardiovascular disease (CVD), which is am ong the top three causes of death in OECD countries. CVD treatment has significant h ealth policy implications, because once CVD patients begin drug therapy, it must continue fo r the remainder of their lives. Of the 29

PAGE 43

32 IMS Health three-digit CVD categories, this dissertation samples the eight studied by Dickson and Jacobzone (2003). Listed in ta ble A.4, these categories cover a wide range of both newer and older innovati ons that form the core of pharmacotherapy for CVD. The five countries in the sample are Ge rmany, the UK, Italy, Spain, and France. These countries have the largest pharmaceu tical production (Figure B.2) and sales (Figure B.3) in the EU-15 (OECD 2003) a nd represent the five largest pharmaceutical markets in the world after the US and Japan (Table A.5) (Pammolli et al. 2004). Restricted as described above, the IMS H ealth data include 658 molecules (119 for France, 177 for Germany, 135 for Italy, 119 for Spain and 108 for the UK) for the 1994 2003 period. 4.2.2 Sample Construction Different data sets are constructed for different parts of the analysis. Index calculations are based on bila teral matches (to Germany and Spain) across countries for each year. Depending on the benchmark country, an average of 50% of retail sales do not match bilaterally and thus are not included. A similar index calcu lation data set (for global molecules) is constructed based on matc hes for all five countries for each year. The match rate varies from 21% (Germany) to 32% (Spain) depending on the benchmark country. As Danzon and Chao (2000b) comm ent, This heterogeneity in product-mix across countries implies that even the price indexes, which start with the universe of sales, may be unavoidably biased. Quasi-hedonic price regressions are base d on two separate samples, a balanced panel of all 379 molecules and the 38 globa l molecules over the 10 years and five

PAGE 44

33 countries, yielding sample sizes of 3,790 and 1,900, respectively. Table A.6 provides details for this data set by ATC/MOL and country. To form the dependent variable in the pr ice convergence regre ssions, two distinct bilateral matched samples are constructe d, with 3,210 and 2,940 observations depending on whether Germany or Spain, respectively, serv es as the base country. Another sample is constructed to form the dependent variab le measured in deviations from the cross country average. This sample includes all 379 molecules and has 3,790 observations. A final sample for global molecules contains 38 molecules and has 1,900 observations. 4.3 Methodology The first part of the analysis calculat es price indexes by examining bilateral standard unit price differen ces for all molecules and global molecules using selected benchmark countries. The second part estimat es quasi-hedonic regressions in order to generate price measures that control for variation in charact eristics of the drugs. The third part uses the residuals from these regre ssions as the prices in the price convergence regressions. The next three se ctions describe the methods and model specifications used in each component of the analysis. 4.3.1 Price Indexes The differences across sample countries in average drug prices, unadjusted for quality characteristics, are calculated using weighted price indexes. Previous studies showed that indexes using weighted aver ages are more accurate than those using unweighted averages (Gilles 1940). The most common price indexes, both generally and

PAGE 45

34 in this literature, are the La speyres and Paasche indexes. Griliches and Cockburn (1994) used these indexes to show significant diffe rences in effects of patent expiration and generic entry estimated using various price i ndexes. For one of the drugs studied, the standard price index rose by 14% over 45 m onths following patent expiration, while the preferred alternative index fell by 48%. A commentary response to their study by Feenstra (1997) supported thei r approach. Berndt et al. (1993) found that the BLS drug price index grows approximately 50% more ra pidly than an altern ative index using the IMS aggregate price data that includes generic drugs. Be rndt et al. (1996) estimated weighted Laspeyres and Paasche price indexe s using the IMS univers e of antidepressant drug prices and found large differences between the fixed-weight a nd average-weighted versions. Berndt et al. (1999) reviewed the conceptual and measurement issues underlying the construction of US medical car e consumer and producer price indexes. Recently, Berndt et al. (2002) estimated Laspeyres, Paasche and Fisher indexes that showed a small rise in price and quant ity for treatments of depression during 1991 1996. Kokoski (1992) had earlier used these inde xes to make inter-a rea cost of living comparisons. In sum, the literature suggest s these indexes are pow erful tools that are useful for making bilateral and multilater al price, output, input and productivity comparisons but can come to differen t conclusions (Caves et al. 1982). In this study, Laspeyres and Paasche pr ice and quantity indexes are calculated both across areas and over time for individual country, using samples of all molecules and only global molecules, to examine quality -unadjusted price differences. In addition, the Bortkiewicz decomposition formula is us ed to examine the ratio of Paasche and Laspeyres index differentials.

PAGE 46

35 4.3.1.1 Specification of the Indexes The set of prices and quantities at time t are denoted pt and qt, respectively, and weights are fixed at time zero. The Laspeyres, i.e. base-weighted, pr ice (quantity) index weights the price (quantity) of each good by the quantity (price) of that good at time zero: 0 0 0 0q p q p Pt t (Laspeyres Price Index) 0 0 0 0q p q p Qt t (Laspeyres Quantity Index) The Laspeyres price index, for example, is a weighted average of individual good price changes, with the weights equaling the e xpenditure share for each good in the base period. It compares the price of a base peri od basket of goods with the price of the same basket in the current period. In contrast, th e Paasche price index co mpares the price of a current basket of goods with the price of the same basket in the base period: t t t tq p q p P0 0 (Paasche Price Index) 0 0q p q p Qt t t t(Paasche Quantity Index) Because prices are weighted by current period quantities (and vice versa for the quantity index), the weights change for each period in which a Paasche index is calculated (Allen 1975). Additionally, because Laspeyres uses base weights while Paasche uses current weights, the two indexe s are generally different even for the same period, but usually are similar when the periods being compared are not too far apart (as

PAGE 47

36 in this study). The Paasche pr ice index tends to be lower th an the Laspeyres when prices are increasing and higher wh en prices are decreasing. The ratios of the Paasche and Laspeyres price and quantity indexes are the same. This ratio, P/L, depends on the dispersion in pr ice and quantity relatives to their means. Specifically, denoting price and quantity with the subscripts p a nd q, the Bortkiewicz decomposition formula, Pp/Lp=Pq/Lq=1+(r Vp Vq), shows that P/L differs from one based on the coefficient of correlation between price and quantity ( r ) and the coefficients of variation of price (Vp) and quantity (Vq), all measured relative to the respective Laspeyres index (Jonas and Sardy 1970). The diverg ence between P and L increases with the correlation between price and quantity and their individual dispersions. Because Vp and Vq are positive, P > L if r > 0, i. e. prices and quantities tend to move in the same direction between years 0 a nd t, while P < L if r < 0, i.e. prices and quantities tend to move in opposite directions. This is known as the Gerschenkron effect (Gerschenkron 1947, 1955) that price and quantity indexes change in di fferent industries. The characteristics of Gerschenkron effect is a negative coefficient of correlation between price and quantity, which accounts fo r the direction of th e divergence between Laspeyres and Paasche indexes. The typi cal economic case of P > L is a market dominated by suppliers, so that the reaction to a price incr ease is an outward shift in supply. Examples include exporters selling on a large international market and suppliers selling both domestic and imported goods in a market. The typical economic case of L > P is a demand-dominated market where consum ers purchases vary inversely with price movements. The leading example is the market for consumer goods (Allen 1975). In

PAGE 48

37 addition, L > P might simply reflect a subs titution effect, with each country using relatively more of those products that are relatively cheap in that country. 4.3.2 Quasi-Hedonic Price Regressions When making international price compar isons, international differences in product specifications pose a problem (Kravi s and Lipsey 1969). Products serving the same purpose might not only be highly diffe rentiated across domestic producers, but might further have considerable differences in characteristics across countries. For example, a Honda Accord might have differe nt horsepower in different countries, or come with automatic transmission in some c ountries and manual transmission in others. More relevant for this study, drugs come in different forms, pack sizes, and strength levels in different countries. This study uses hedonic regressions to a ddress this problem. Framing a product in terms of the characterist ics that affect its value, hedonic regression estimates the marginal contribution of each characteristic (Sirmans et al. 2005), thus explaining the price of a good in terms of these characteristics (Wooldri dge 2003). Here, hedonic price regressions estimate the value of observed char acteristics of the drugs/molecules. Drug characteristics and prices al so differ across countries becau se of their regulatory and reimbursement environments. Because market (competition) variables are used as explanatory variables along w ith drug quality characteristic s, and pure hedonic models control for only the latter (D iewert 2003, 2005), the correspon ding models in this study are more appropriately termed quasi-he donic price regressions (Danzon and Chao 2000b).

PAGE 49

38 4.3.2.1 Specification of the Quasi-Hedonic Price Models The model is a semi-log model, t j k k t j j t t j k t j ku c X P, , , ,ln where lnPk,j,t is the log price per SU for mol ecule k in country j and year t, X is a vector of quality and market characteristics for that molecule, country and year t, t is year indicators, j is country indicators, j,t is an interactions between indicators for country j and year t, ck is an indicator for molecule k, and uk,j,t is the remaining error. The main goals of this regression are to obtain consistent estimates of j,t, which reflects the pattern of bilateral country price diffe rences over time when the molecule and observable quality and market characteristics are held constant, and uk,j,t which represents cross country price differences that cannot be explained by observable drug charac teristics, specific molecules available and average year-specific price differences. This model is estimated using panel da ta methods that account for timeand country-invariant unobserved hetero geneity associated with each specific molecule. Both fixed and random effect models are estimated. These vary according to their treatment of the unobserved molecule-specific effect ck which is called a random effect when treated as a random variable and a fixed e ffect when treated as a parameter to be estimated for each molecule. Both models re quire zero conditional mean to hold, i.e. E( uk,j,t | Xk,j, ck) = 0, in order to generate consistent parameter estimates. Because the random effect model implicitly places ck in the error term, for consistency it further requires zero correlation between the obser ved explanatory variab les and the unobserved effect, i.e. Cov( Xk,j,t, ck) = 0. In contrast, the fixed eff ect model allows arbitrary correlation between ck and Xk,j,t. The fixed effect model is therefore more robust than the

PAGE 50

39 random effect model. The tradeoff is that th e effects of time-constant variables can be estimated in the latter but not the former, because there is no way to distinguish the effects of time-constant obs ervables and unobservables (B altagi 2005;Wooldridge 2001). Lastly, the Hausman (1978) test is used to check whether the random effects model gives similar results to the fixed effects model and is therefore valid. Most measures of molecule quality and market competition in the data are timevarying, as are the country/year interactions. In principle, drug quality also encompasses therapeutic value and convenience, characterist ics that are of intrin sic value but are not observable. If these are time-invariant and mo lecule-specific, the fixed effect model is appropriate. However, in order to include ATC3 indicators to proxy for market and regulatory factors that might di ffer across ATC3 categories, a random effect model is also estimated. One time-invariant market co mpetition variable, therapeutic substitute molecule entry lag, is also included as an e xplanatory factor in the random effect model. 4.3.2.2 Description of Variables Price (Leusuprice) is the average price per standard unit for each ATC/molecule, defined as the volume-weighted average reta il price over all forms and packs. Local currency prices are converted to euros by IM S Health using constant exchange rates, which minimize effects of exchange rate fluctuations. Molecule Age (Molage) is the number of years since the first product launch of molecule k in country j, and is the same for all products in a molecule. Strength (Strengthg) is the mean grams of active ingredient per standard unit, averaged over all packs. Standard units are multiplied by the different strength levels,

PAGE 51

40 and this quantity is divided by the total standa rd units. This variable was complicated to calculate due to the various measurements in the reported strength levels for packs in each molecule for the sample countries. All reported units, including international units, micrograms, percent, milligrams, pints, and qua litative units such as strong, weak, extra strong and mild, were converted to grams in consultation with IMS Health experts. Form Code (Formcode) represents the num ber of different formulations of the products in each molecule, and is included as a measure of the choice and convenience available to patients. Forms include different types of tabl ets (e.g. film, chewable, gel), capsules, ampoules, powders, drops, syrups, sy ringes, and liquids, along with different strengths and pack sizes. For example, the molecule metoprolol (in C7A) was available in 118 different formulations/prese ntations in Germany in 2003. Pack Size (Packsize) is the average number of standard units over all packs in a molecule. Pack sizes were converted to IM S standard units acco rding to guidelines provided by IMS Health (IMS 2006), multiplied by standard units per molecule and divided by the total standard units in the molecule. Global Penetration-Diffusion (G lobpenet) is the number of sample countries (i.e. between 0 and 5) in which the molecule is avai lable, as a measure of therapeutic value. Generic Competitors (Gencompet) is the number of manufacturers of the products in the molecule, including originators, licensees, parallel imports and generics. Therapeutic Substitute Molecules (Thsubsmol) is the number of therapeutic competitors that are chemically distinct but used to treat the same indication, i.e. the number of molecules in the ATC3.

PAGE 52

41 Therapeutic Substitute Molecule Entry Lag (Thsubsmolentlag) is the number of years between this molecules launch date and the launch of the first molecule in the ATC3, to control for any first mover advantage, and is time-invariant. 4.3.3 Price Convergence Regressions The earlier described quasi-hedonic pri ce regressions (without time, country and time-country interaction indicato rs) are used to form the depe ndent variable in the price convergence equations. In particular, the resi duals capture price vari ation that cannot be attributed to observable charac teristics, trends in price di fferences specific to country pairs, or specific molecules. These thus fo rm the relevant prices in the price convergence regressions. Goldberg and Verboven (2005) used a similar method to test price convergence in the EU car market. A common approach to examin ing price convergence is to apply a unit root test to determine whether price differential series are stationary, i.e. have mean and variance that do not vary systematically over time and are t hus stable. The rejection of the unit root hypothesis implies that relative prices have st ationary time series and will thus converge in the long run. Failure to reject the unit root hypothesis implies that relative prices follow a random path, so that any deviation from a single price becomes permanent (Fan and Wei 2006). Moreover, presen ce of a unit root implies that a shock today has a long lasting impact, determining whether a process has a unit root is of interest in its own right. Unit root tests for a single time series such as the often used Augmented-Dickey Fuller test, have low power in the sense that they too often reject statio narity. Levin et al. (2002) showed that use of a unit root test for panel data can significantly increase test

PAGE 53

42 power (Maddala 1999;Maddala and Wu 1999). Their model assumes that each panel unit shares the same AR(1) coeffici ent, but allows for individual and time effects and a time trend. This test maybe viewed as a pooled Dickey-Fuller test or an Augmented DickeyFuller test when lags are included, with the null hypothesis of nonstationary (Bornhorst and Baum 2001). 4.3.3.1 Specification of the Price Convergence Models The following convergence equations are ba sed on the panel data unit root test developed by Levin and Lin (1992;Levin and Lin 1993) and Le vin et al. (2002): Model 1: t j k l t j k L l l t j k t j kp p p, , 1 1 , , Model 2: t j k l t j k L l l t j k j k t j kp p p, , 1 1 , , Model 3: t j k l t j k L l l t j k j k t j kp t p p, , 1 1 , , In model 1, the null hypothesis H0: = 0 is tested against the alternative H1: < 0. In model 2, individual molecule/coun try fixed effects are added, and H0: = 0 and j,k = 0 is tested against H1: < 0 and j,k 0. In model 3, a time trend is added, and H0: =0 and =0 is tested against H1: <0 and 0. These three models serve different purposes. The first is used to test the absolute version of the Law of One Price, while the second model is used to test the relative version of the Law of One Price. Model 3 is not preferred in the literature but it is estimated for both adjusted and unadjusted pr ice convergence estimations to compare the

PAGE 54

43 results. For series that have clear time tre nds, the process must be modified to test for unit roots. A trend stationary process can be mistaken for unit root process if we do not control for a time trend. The convergence equation depende nt variable represents th e first difference in the log-price of molecule k in count ry j. This first difference is specified relative to a base country in one specification, and relative to the cross country average in another. In other words, if q represents log price, the two dependent variable specifications are 1) 1 , , , t j k t j k t j kp p p, where t y BaseCountr k t j k t j kq q p, , , and p k,BaseCountry,t = 1; 2) 1 , , , t j k t j k t j kp p p with t j ryaverage Crosscount t j k t j kq q p, , , The test for unit roots relates the first difference to the log price of the previous period; if the coefficient of the previous pe riods price is negative, price differentials across countries become smaller over time ( <0) and the hypothesis of a unit root is rejected. Therefore, denotes the speed of convergence. Under the null hypothesis of no convergence, = 0 and a shock to pk,j,t is permanent (i.e. has a unit root). If 0, the price differential is non-stationary, impl ying persistent price divergence. If < 0, prices converge. The coefficient estimate is tested according to the critic al values reported in Levin and Lin (1992, 1993) and tstar statistics (adjus ted t statistics, tabulated in Levin et al. 2002) are also reported. After transforma tion by factors provided by Levin and Lin (1992, 1993), the t-star statis tic is distributed standard normal under the null hypothesis of nonstationarity. The half -life of a shock to the price differential is ) 1 ln( ) 2 ln( Model 2, used to test the relative version of the Law of One Price, is the primary focus of this study. The k,j capture price differences that ar e specific to country pairs and

PAGE 55

44 molecules. Large values would indicate ma rket segmentation. Even in integrated markets, some permanent cross-country price dispersion might remain, reflecting local factors that cannot be arbitr aged away. If this dispersion fully explains any price differentials, the relative vers ion of the Law of One Price holds. If instead prices of identical products are equal ac ross countries, the absolute version of the law of one price holds (Goldberg and Verboven 2004). Dividing these fixed effects by yields the longterm systematic price differentials ac ross countries. For brevity, the average j,k across molecules by country is reported. The lags (L) l t j kp, are used to account for possibl e serial correlation in the error term. Because of the limited number of years available in the data set, the estimations include lags of zero, one or two years, consistent with previous studies. In addition to these, Campbell and Perrons topdown approach (Pammolli et al. 2004) is also used to find optimum amount of lag order when the equations are estimated. In this approach, the lag order is set to a maximum of two lags for each molecule/country estimation. If the absolute value of the t-statistic of 2 is less than 1.96, the lag order is set to one lag and the equation is re-estimated. If this t-statistic is less than 1.96, no lags are included.

PAGE 56

45 Chapter 5 Research Results This chapter describes the estimation results. The first section discusses the price indexes, the second section discusses the qua si-hedonic price regressions and the last section discusses price convergence. 5.1 Unadjusted Bilateral Standa rd Unit Price Differences This section reports the estimated La speyres and Paasche price and quantity indexes, as well as the decomposition of P/L di fferentials, with Spain and Germany as the benchmark countries. Spain typically has the lowest drug prices in Europe and the most pharmaceutical regulations, while Germany has high prices and fewer regulations. The standard unit is the volume measure and the fixed euro is the monetary measure. The Laspeyres index uses benchmark country weig hts, whereas the Paasche index uses own country weights. The indexes are measured for both bilateral matched molecules and global molecules. In addition, to investigate country specific tempor al fluctuations, year by year price differentials are measured for each country.

PAGE 57

465.1.1 Standard Unit Price Differences for Bilaterally Matched Molecules Table 1 shows SU prices relative to Ge rmany. In 1994, Laspeyres prices are higher by 7.5% in the UK and 14.9% and Fran ce, but lower by 4.3% in Italy and 27.5% in Spain. In 2003, prices are higher in all four other countries, by 41.4% in the UK, 35.3% in France, 28.5% in Italy and 0.25% in Spain. The Paasche index generally shows smaller price differentials. In 1994 prices ar e lower by 1.7% in Ital y, 2.3% in the UK, 11.7% in France and 27.8% in Spain. In 2003, prices are higher by 39.4% in the UK, 12.8% in Italy and 2.7% in France, but lower by 9.0% in Spain. Thus, both the magnitude and rank ordering of the price diff erentials depend on which weights are used. The Laspeyres index may be most relevant from the German perspective since it uses weights for Germany, and can thus be interp reted as an estimate of how much Germany might save by adopting another countrys prices, although it is a lower-bound savings estimate because it assumes no change in German consumption patterns. The Paasche index provides an upper-bound estimate of pot ential savings because it assumes that while Germany would adopt the other countrys consumption patterns, changes in prices and quantities would not affect R&D. Fi gures B.4 and B.5 show these bilateral Laspeyres and Paasche price differences. The Laspeyres quantity index shows less consumption than in Germany for all countries except France, and the Paasche qua ntity index indicates consistently less consumption than in Germany for all count ries. Indexes normalized by population size show that all countries have lower per capit a consumption than Germany. The results reveal large cross-national differenc es in per capita drug consumption.

PAGE 58

47 Table 1 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Germany 1994 1995 1996 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 81 73 74 89 78 73 72 89 78 74 73 85 Laspeyres Price Index (GR weighted) 0.7248 1.0754 1.1489 0.9571 0.7727 1.1250 1.2020 0.9334 0.8115 1.1745 1.2150 0.9986 Paasche Price Index (Own weighted) 0.7220 0.9775 0.8836 0.9834 0.7632 1.0861 0.9184 0.9053 0.7956 1.1186 0.9341 0.9643 Laspeyres Quantity Index 0.4342 0.5293 1.2233 0.6214 0.4393 0.5041 1.1681 0.6145 0.4294 0.5159 1.0794 0.5992 Paasche Quantity Index 0.4326 0.4811 0.9408 0.6385 0.4338 0.4867 0.8925 0.5960 0.4210 0.4913 0.8298 0.5786 Normalized Laspeyres Quantity Index by Relative Population Size 0.2095 0.3760 0.8657 0.4343 0.2120 0.3582 0.8273 0.4284 0.2070 0.3663 0.7643 0.4163 Normalized Paasche Quantity Index by Relative Population Size 0.2087 0.3418 0.6658 0.4462 0.2093 0.3459 0.6322 0.4155 0.2029 0.3489 0.5876 0.4020 Population Ratio (Comparison/GR) 0.4825 0.7105 0.7077 0.6989 0.4825 0.7106 0.7083 0.6972 0.4819 0.7101 0.7081 0.6948 Pp/Lp=Pq/Lq 0.9961 0.9090 0.7691 1.0275 0.9876 0.9654 0.7641 0.9700 0.9804 0.9525 0.7688 0.9657 r -0.0067 -0.0045 -0.1098 0.0123 -0.0205 -0.0171 -0.3023 -0.0144 -0.0270 -0.0328 -0.0572 -0.0142 Vp 0.4035 0.4017 0.4584 0.4985 0.4053 0.3800 0.4241 0.5403 0.4403 0.3861 0.4103 0.5460 Vq 1.4224 50.7645 4.5873 4.4643 1.4956 5.3272 1.8404 3.8708 1.6528 3.7541 9.8455 4.4123 r.Vp.Vq -0.0039 -0.0910 -0.2309 0.0275 -0.0124 -0.0346 -0.2359 -0.0300 -0.0196 -0.0475 -0.2312 -0.0343

PAGE 59

48 Table 1 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Germany 1997 1998 1999 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 77 78 74 84 78 79 76 82 77 78 75 83 Laspeyres Price Index (GR weighted) 0.8754 1.2225 1.2590 1.0863 0.9154 1.2921 1.2574 1.1234 0.9304 1.2972 1.2361 1.1557 Paasche Price Index (Own weighted) 0.8383 1.1818 0.9617 1.0062 0.8651 1.2234 1.0275 1.0437 0.8647 1.2586 0.9778 1.0746 Laspeyres Quantity Index 0.4380 0.5353 1.0988 0.5740 0.4596 0.5733 0.9813 0.5849 0.4774 0.6156 1.0189 0.5934 Paasche Quantity Index 0.4194 0.5175 0.8393 0.5317 0.4343 0.5429 0.8019 0.5434 0.4436 0.5973 0.8060 0.5517 Normalized Laspeyres Quantity Index by Relative Population Size 0.2111 0.3801 0.7786 0.3981 0.2220 0.4080 0.7300 0.7238 0.2316 0.4396 0.7266 0.4116 Normalized Paasche Quantity Index by Relative Population Size 0.2021 0.3675 0.5948 0.3688 0.2098 0.3863 0.5697 0.3769 0.2152 0.4265 0.5747 0.3828 Population Ratio (Comparison/GR) 0.4819 0.7101 0.7086 0.6935 0.4831 0.7116 0.7105 0.6935 0.4852 0.7141 0.7131 0.6938 Pp/Lp=Pq/Lq 0.9576 0.9667 0.7639 0.9263 0.9450 0.9469 0.8171 0.9291 0.9294 0.9703 0.7910 0.9298 r -0.0500 -0.0287 -0.0650 -0.0241 -0.0610 -0.0351 -0.1382 -0.0233 -0.0788 -0.0200 -0.1996 -0.0208 Vp 0.4819 0.3935 0.4639 0.5793 0.5146 0.4991 0.5016 0.5745 0.5338 0.4926 0.5467 0.6007 Vq 1.7619 2.9478 7.8313 5.2862 1.7511 3.0365 3.1691 5.3005 1.6799 3.0124 1.9152 5.6167 r.Vp.Vq -0.0424 -0.0333 -0.2361 -0.0737 -0.0550 -0.0531 -0.2196 -0.0709 -0.0706 -0.0297 -0.2090 -0.0702

PAGE 60

49 Table 1 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Germany 2000 2001 2002 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 78 77 73 83 76 73 73 83 74 73 70 80 Laspeyres Price Index (GR weighted) 0.9288 1.3144 1.2500 1.2063 0.8976 1.2551 1.2278 1.2151 0.9371 1.3122 1.2869 1.2709 Paasche Price Index (Own weighted) 0.8504 1.2718 0.9711 1.1059 0.8148 1.2152 0.9509 1.0969 0.8347 1.2548 0.9712 1.1326 Laspeyres Quantity Index 0.4811 0.6280 1.0248 0.6125 0.4781 0.6692 0.9748 0.6199 0.4778 0.7031 0.9085 0.5869 Paasche Quantity Index 0.4405 0.6077 0.7962 0.5615 0.4340 0.6479 0.7550 0.5596 0.4256 0.6723 0.6856 0.5230 Normalized Laspeyres Quantity Index by Relative Population Size 0.2345 0.4493 0.7334 0.4244 0.2353 0.4800 0.7009 0.4293 0.2374 0.5050 0.6557 0.4057 Normalized Paasche Quantity Index by Relative Population Size 0.2147 0.4348 0.5698 0.3891 0.2135 0.4647 0.5429 0.3875 0.2115 0.4829 0.4948 0.3616 Population Ratio (Comparison/GR) 0.4874 0.7155 0.7156 0.6929 0.4921 0.7172 0.7190 0.6925 0.4969 0.7183 0.7217 0.6913 Pp/Lp=Pq/Lq 0.9156 0.9676 0.7769 0.9168 0.9077 0.9682 0.7745 0.9027 0.8907 0.9563 0.7547 0.8912 r -0.0847 -0.0211 -0.1900 -0.0196 -0.1030 -0.0193 -0.1734 -0.0198 -0.1141 -0.0253 -0.1666 -0.0198 Vp 0.5637 0.5143 0.6045 0.6371 0.5175 0.5136 0.6334 0.6439 0.5571 0.5464 0.6779 0.6846 Vq 1.7677 2.9850 1.9418 6.6707 1.7311 3.2135 2.0532 7.6203 1.7194 3.1633 2.1716 8.0274 r.Vp.Vq -0.0844 -0.0324 -0.2231 -0.0832 -0.0923 -0.0318 -0.2255 -0.0973 -0.1093 -0.0437 -0.2453 -0.1088

PAGE 61

50 Table 1 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Germany 2003 Index Measures SPN UK FR ITY Number of ATC/Molecule Matching 72 73 70 79 Laspeyres Price Index (GR weighted) 1.0025 1.4142 1.3528 1.2845 Paasche Price Index (Own weighted) 0.9100 1.3936 1.0273 1.1281 Laspeyres Quantity Index 0.4439 0.6673 0.8402 0.5475 Paasche Quantity Index 0.4029 0.6576 0.6381 0.4808 Normalized Laspeyres Quantity Index by Relative Population Size 0.2241 0.4806 0.6093 0.3803 Normalized Paasche Quantity Index by Relative Population Size 0.2034 0.4736 0.4627 0.3339 Population Ratio (Comparison/GR) 0.5048 0.7201 0.7252 0.6945 Pp/Lp=Pq/Lq 0.9077 0.9854 0.7594 0.8782 r -0.0968 -0.0089 -0.1695 -0.0223 Vp 0.5155 0.5183 0.6190 0.6266 Vq 1.8495 3.1537 2.2930 8.7072 r.Vp.Vq -0.0923 -0.0146 -0.2406 -0.1218 The ratios of the Paasche to Laspeyres indexes, P/L, are uniformly less than one, except for Italy in 1994. This is consistent with a negative correla tion between price and quantity, i.e. a Gerschenkron effect, and can be interpreted as a demand-dominated drug market in which equilibrium quantity responds mo re to shifts in supply than in demand. It could also reflect a substitution effect in which consumers use relatively more of products that are relatively cheap. As in Danzons (2000) study of price differentials relative to the US, the magnitudes of the co rrelation between pri ce and quantity and coefficients of variations are small, showing that there are less variation in prices and quantities among the countries. Table A.7 and figures B.6 and B.7 report price and quantity indexes relative to Spain for bilaterally matched molecules. Th e Laspeyres price index shows that Spain has the lowest drug prices in the sample, with prices higher in the UK, Germany, France and Italy by 38.5%, 38.0%, 32.8% and 46.6% in 1994 and 44.9%, 9.9%, 23.9% and 30.5% in

PAGE 62

51 2003. The Paasche price index reports smaller but similar differentials for all countries except France. The quantity indexes show that per capita consumption is lower in Spain than in the other countries. On the other hand, P/L ratios are again less than one for all countries except France between 1994 and 2001. 5.1.2 Standard Unit Price Differences for Global Molecules The price indexes for global molecules re lative to Germany (table 2) generally show similar but some cases smaller price di fferences between countries than the indexes based on the larger bilaterally matched sample s (figure B.8 and B.9). The P/L ratios are again below one for all countries except Spain between 1994 and 1997. The smaller coefficients of variation for both quantity a nd price indicate that consumption and prices variation falls as drugs become more univers ally available. This could reflect the European Commissions efforts to coor dinate the EU pharmaceutical industry, particularly by increasing pa rallel trade and using the international price comparison regulatory method. Similar results are repor ted in table A.8 and figures B.10 and B.11 for global molecules relative to Spain.

PAGE 63

52 Table 2 Pharmaceutical Price and Quantity Indexes fo r Global Molecules, Relative to Germany 1994 1995 1996 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (GR weighted) 0.7716 1.0671 1.0809 0.9778 0.8257 1.1189 1.1765 0.9670 0.8751 1.1720 1.2267 1.0601 Paasche Price Index (Own weighted) 0.8103 1.0365 0.9550 1.0531 0.8614 1.0929 0.9980 0.9721 0.9008 1.1325 1.0164 1.0393 Laspeyres Quantity Index 0.4108 0.5281 1.0769 0.5780 0.4072 0.5354 1.0348 0.5820 0.3990 0.5605 0.9838 0.5625 Paasche Quantity Index 0.4314 0.5130 0.9516 0.6225 0.4247 0.5230 0.8778 0.5851 0.4107 0.5416 0.8151 0.5515 Normalized Laspeyres Quantity Index by Relative Population Size 0.1982 0.3752 0.7622 0.7360 0.1965 0.3805 0.7329 0.4057 0.1923 0.3980 0.6966 0.3908 Normalized Paasche Quantity Index by Relative Population Size 0.2082 0.3644 0.6734 0.4351 0.2049 0.3717 0.6217 0.4079 0.1979 0.3846 0.5772 0.3832 Population Ratio (Comparison/GR) 0.4825 0.7105 0.7077 0.6989 0.4825 0.7106 0.7083 0.6972 0.4819 0.7101 0.7081 0.6948 Pp/Lp=Pq/Lq 1.0503 0.9713 0.8836 1.0770 1.0432 0.9768 0.8483 1.0053 1.0293 0.9663 0.8285 0.9803 r 0.1188 -0.0675 -0.1973 0.1143 0.1000 -0.0632 -0.2544 0.0084 0.0584 -0.0903 -0.2677 -0.0303 Vp 0.3677 0.4060 0.3841 0.4830 0.3764 0.3864 0.3787 0.5218 0.4213 0.4014 0.4108 0.5548 Vq 1.1506 1.0480 1.5364 1.3955 1.1473 0.9500 1.5749 1.2042 1.1912 0.9297 1.5590 1.1710 r.Vp.Vq 0.0503 -0.0287 -0.1164 0.0770 0.0432 -0.0232 -0.1517 0.0053 0.0293 -0.0337 -0.1715 -0.0197

PAGE 64

53 Table 2 (Continued) Pharmaceutical Price and Quantity Indexes fo r Global Molecules, Relative to Germany 1997 1998 1999 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (GR weighted) 0.9468 1.2196 1.3003 1.1495 0.9795 1.2684 1.3120 1.1892 0.9940 1.2952 1.2882 1.2241 Paasche Price Index (Own weighted) 0.9482 1.1979 1.0230 1.0874 0.9732 1.2392 1.0325 1.1296 0.9708 1.2688 1.0209 1.1609 Laspeyres Quantity Index 0.4208 0.5998 1.0208 0.5518 0.4412 0.6587 1.0403 0.5557 0.4532 0.7033 0.9977 0.5713 Paasche Quantity Index 0.4215 0.5891 0.8030 0.5220 0.4383 0.6436 0.8187 0.5279 0.4427 0.6890 0.7907 0.5418 Normalized Laspeyres Quantity Index by Relative Population Size 0.2028 0.4259 0.7233 0.3827 0.2131 0.4688 0.7336 0.7834 0.2199 0.5022 0.7114 0.3964 Normalized Paasche Quantity Index by Relative Population Size 0.2031 0.4183 0.5690 0.3621 0.2117 0.4580 0.5817 0.3661 0.2148 0.4920 0.5638 0.3759 Population Ratio (Comparison/GR) 0.4819 0.7101 0.7086 0.6935 0.4831 0.7116 0.7105 0.6935 0.4852 0.7141 0.7131 0.6938 Pp/Lp=Pq/Lq 1.0015 0.9822 0.7867 0.9460 0.9936 0.9770 0.7870 0.9499 0.9767 0.9796 0.7925 0.9484 r 0.0026 -0.0463 -0.2795 -0.0772 -0.0104 -0.0570 -0.2449 -0.0724 -0.0377 -0.0494 -0.2146 -0.0756 Vp 0.4803 0.4185 0.4840 0.6064 0.5022 0.4318 0.5237 0.6106 0.5035 0.4393 0.5527 0.6322 Vq 1.2222 0.9187 1.5768 1.1533 1.2312 0.9349 1.6608 1.1339 1.2289 0.9399 1.7496 1.0797 r.Vp.Vq 0.0015 -0.0178 -0.2133 -0.0540 -0.0064 -0.0230 -0.2130 -0.0501 -0.0233 -0.0204 -0.2075 -0.0516

PAGE 65

54 Table 2 (Continued) Pharmaceutical Price and Quantity Indexes fo r Global Molecules, Relative to Germany 2000 2001 2002 Index Measures SPN UK FR ITY SPN UK FR ITY SPN UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (GR weighted) 1.0118 1.3495 1.3224 1.2915 0.9634 1.2798 1.3276 1.3107 1.0092 1.3928 1.4220 1.3807 Paasche Price Index (Own weighted) 0.9775 1.2897 1.0146 1.2258 0.9156 1.2172 0.9998 1.2214 0.9396 1.2847 1.0261 1.2842 Laspeyres Quantity Index 0.4521 0.7084 0.9971 0.5649 0.4487 0.7529 0.9691 0.5704 0.4392 0.7776 0.9238 0.5235 Paasche Quantity Index 0.4367 0.6770 0.7651 0.5362 0.4265 0.7161 0.7299 0.5315 0.4089 0.7172 0.6666 0.4869 Normalized Laspeyres Quantity Index by Relative Population Size 0.2204 0.5068 0.7135 0.3914 0.2208 0.5400 0.6968 0.3950 0.2182 0.5585 0.6667 0.3619 Normalized Paasche Quantity Index by Relative Population Size 0.2129 0.4844 0.5475 0.3715 0.2098 0.5136 0.5248 0.3681 0.2032 0.5152 0.4811 0.3366 Population Ratio (Comparison/GR) 0.4874 0.7155 0.7156 0.6929 0.4921 0.7172 0.7190 0.6925 0.4969 0.7183 0.7217 0.6913 Pp/Lp=Pq/Lq 0.9661 0.9557 0.7673 0.9491 0.9504 0.9511 0.7531 0.9319 0.9311 0.9224 0.7216 0.9301 r -0.0531 -0.0994 -0.2151 -0.0754 -0.0864 -0.1091 -0.2132 -0.1054 -0.1206 -0.1575 -0.2367 -0.1027 Vp 0.5000 0.4909 0.5905 0.6477 0.4424 0.5038 0.6057 0.6413 0.4294 0.5334 0.5937 0.6422 Vq 1.2765 0.9084 1.8321 1.0423 1.2970 0.8902 1.9113 1.0083 1.3308 0.9237 1.9813 1.0598 r.Vp.Vq -0.0339 -0.0443 -0.2327 -0.0509 -0.0496 -0.0489 -0.2469 -0.0681 -0.0689 -0.0776 -0.2784 -0.0699

PAGE 66

55 Table 2 (Continued) Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Germany 2003 Index Measures SPN UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 Laspeyres Price Index (GR weighted) 1.0980 1.5679 1.5275 1.4675 Paasche Price Index (Own weighted) 1.0590 1.5168 1.1217 1.3771 Laspeyres Quantity Index 0.3770 0.6946 0.8214 0.4576 Paasche Quantity Index 0.3636 0.6720 0.6031 0.4294 Normalized Laspeyres Quantity Index by Relative Population Size 0.1903 0.5002 0.5956 0.3178 Normalized Paasche Quantity Index by Relative Population Size 0.1835 0.4839 0.4374 0.2982 Population Ratio (Comparison/GR) 0.5048 0.7201 0.7252 0.6945 Pp/Lp=Pq/Lq 0.9645 0.9674 0.7343 0.9384 r -0.0649 -0.0676 -0.2373 -0.0959 Vp 0.3578 0.4876 0.4931 0.5735 Vq 1.5306 0.9891 2.2706 1.1200 r.Vp.Vq -0.0355 -0.0326 -0.2657 -0.0616 5.1.3 Country Price Differences for All and Gl obal Molecules, Relative to 1994 Tables A.10 and A.11 and figures B.12 and B.13 show the price and quantity indexes over the years for all bilaterally ma tched and global molecules in each country. For all bilateral matched molecules, prices consistently decreased in Germany and France starting in 1998, whereas th ey increased until around 2000 and then decreased in the other three countries. The results for global molecules are similar. 5.2 Quality Adjusted Standard Unit Price Differences This section shows the results of quasi-hedonic regressions that examine the contribution of various quality and market competition charac teristics to the large crosscountry dispersion of drug pri ces just documented. The re siduals of these regressions

PAGE 67

56 will then form the price variables in the pri ce convergence regressions to be discussed in the subsequent section. 5.2.1 Drug Quality and Market (Competition) Characteristics Table 3 shows descriptive statistics for quality and market characteristics in the bilaterally matched data on 3,790 molecule /country/year observa tions, while table 4 presents the same information for the 1,900 observations on globally diffused molecules. In the latter data, the SU price is lowe r while quantity and retail sales are higher. Table 3 Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Molecules Overall Variable Description N Overall Mean Overall SD Within SD Suthnds Standard units in thousands 3,790 73,189 125,996 41,562 Leuthnds Total standard unit retail sales in thousands in euro 3,790 18,343 33,034 13,665 Leusuprice Standard unit prices in euro 3,790 0.39 1.25 0.43 Quality Characteristics Strengthg Strength per gram 3,790 0.15 0.55 0.06 Molage Molecule age 3,790 19.62 12.15 2.87 Packsize Pack size 3,790 57.27 65.02 35.22 Formcode Form Code 3,790 7.67 11.99 2.59 Globpenet Global penetration 3,790 4.16 1.20 0.27 Market(Competition) Characteristics Gencompet Generic Competition 3,790 4.80 7.44 2.55 Thsubsmol Therapeutic Substitute Molecule 3,790 15.95 6.31 1.37 Thsubsolentlag Therapeutic Substitute Molecule Entry Lag 3,780 17.70 12.59 0.00

PAGE 68

57 Table 4 Pharmaceutical Drug Quality and Market (Competition) Characteristics for Global Molecules Overall Variable Description N Overall Mean Overall SD Within SD Suthnds Standard units in thousands 1,900 107,664 144,484 54,522 Leuthnds Total standard unit retail sales in thousands in euro 1,900 27,242 41,213 18,320 Leusuprice Standard unit prices in euro 1,900 0.27 0.22 0.04 Quality Characteristics Strengthg Strength per gram 1,900 0.18 0.58 0.09 Molage Molecule age 1,900 20.58 11.59 2.87 Packsize Pack size 1,900 58.08 74.38 41.44 Formcode Form Code 1,900 9.97 14.97 3.31 Globpenet Global penetration 1,900 5 0.00 0.00 Market(Competition) Characteristics Gencompet Generic Competition 1,900 6.46 9.36 3.15 Thsubsmol Therapeutic Substitute Molecule 1,900 14.73 5.47 1.28 Thsubsolentlag Therapeutic Substitute Molecule Entry Lag 1,900 13.53 9.99 0.00 In table A.11, the SU price is .58, .54, .30, .27 and .21 in Germany, the UK, Italy, France and Spain, respectively. Similarly, average SU retail sales are the greatest in Germany and the smallest in Spai n, while average total retail sales are the highest in France and the lowest in Spain. For globally diffused molecules, SU prices, in the order listed above, are .29, .32, .28, .27 and 0.21 (table A.12). The descriptive statistics are also listed by molecule, country and year, for all molecules in table A.13 and for global molecules in table A.14.

PAGE 69

585.2.2 Quality Adjusted Standard Unit Price Differences for All Molecules Table 5 shows the results of the quasi-he donic price regressions estimated by both fixed and random effect models, with Germa ny as the base country. Most of the quality and market variables have the expected sign and are statistically significant, but their effects are small in magnitude, indicating ove rall reimbursement and regulation effects. Standard unit price is increas ing in strength. The number of forms available is expected to be positively related in markets, if ra nge of formulation enha nces effectiveness, convenience and value. Additionally, intr oducing a new formation is a method of obtaining a price increase in c ountries that do not permit pric e increases for established products or when the product life cycle dec lines (Danzon and Chao 2000a). Here, form code is inversely related due to possible e xplanations of therapeu tic category-specific differences in medical norms and insurance. Further investigation is needed for this relationship in the regulated ma rkets. SU price is inversely related to molecule age, suggesting that newer molecules offer impr oved therapeutic quali ty, although molecule age may also reflect life-cycle regulatory effect s, but it is not signifi cant. Price decreases with pack size, consistent with economies of scale in packaging, a nd global penetration, which is a proxy for diffused therapeutic value. Generic competition lowers price as it is expected. Therapeutic substitute molecule is expected to be inversely related to price due to substitution effect (Danzon and Chao 2000b) but here again, it is di rectly related to the price. When a new molecule is introduced assuming better therapeutic treatment, a few good substitutes will be availa ble and the drugs with the ne w main ingredient will have high prices. But after time, competition elimin ates some of the substitutes while at the same time lowers prices. Further investig ation is needed for therapeutic substitute

PAGE 70

59 molecules. Lastly, therapeutic substitute mo lecule entry lag has a positive price effect but it is insignificant. All of the ATC3 du mmies are significant in the random effect model, implying significant differences in pr ices for different indications, presumably due to differences in therapeuti c value and insurance coverage. The Hausman test for fixed effect (FE) vs. random effect (RE) models tests the null hypothesis that the coefficients estimated by the random effects estimator, which is efficient but possibly inconsis tent, are the same as the ones estimated by the consistent fixed effects estimator. Therefore, a failure to reject the null hypothe sis means either that RE and FE estimates are sufficiently close so that it does not matter which is used, or the sampling variation is so large in the FE es timates that one cannot conclude practically significant differences are statistically signifi cant. A rejection using Hausman test is taken to mean that the key RE assumption, i.e. Cov(Xk,j,t, ck) = 0, is false, so that the FE estimates should be used. Here, the Hausman test fails to reject the null hypothesis, so the random effect estimator is presumed consis tent and hence used to interpret the price differentials. The main interest in thes e regressions is the coeffi cients of the country/year interactions, which trace out the pattern of price differences over time controlling for quality and market characteris tics and molecule identity. The individual country dummy coefficients give the 1994 pr ice differential between country and Germany, while the individual year dummy coefficients give th e price differential for Germany between each year and 1994. Table 6 and Figure 1 show the price differences in percentages, estimated with the random effect mode l, relative to Germany in 1994 (the omitted country/year

PAGE 71

60 combination), while table A.15 and figure B.14 do the same for prices relative to Spain in 1994. Table 5 Quasi-Hedonic Price Regression Results for All Molecules, Relative to Germany Fixed Effect Model Random Effect Model Dependent variable: logleusuprice Explanatory Variables Coefficient (Standard Error)a Coefficient (Standard Error)a Quality Characteristics STRENGTHG 0.0341*** (0.0185) 0.0313*** (0.0172) MOLAGE -0.0007 (0.0029) -0.0039 (0.0028) PACKSIZE -0.0020* (0.0001) -0.0021* (0.0001) FORMCODE -0.0046* (0.0009) -0.0049* (0.0009) GLOBPENET -0.0430*** (0.0236) -0.0627* (0.0203) Market (Competition) Characteristics GENCOMPET -0.0044* (0.0016) -0.0044* (0.0016) THSUBSMOL 0.0123* (0.0018) 0.0123* (0.0020) THSUBSMOLENTLAG 0.0006 (0.0017) Country dummies Yes Yes Time dummies Yes Yes Country/Time dummies Yes Yes ATC dummies No Yes N 3,790 3,780 R2 (Within) 0.2495 0.2490 Prob>F 0.0000 0.0000 a Standard errors are heteroskedasticity-robust. *, ** and *** reflect p<0.01, p<0.05 and p<0.10.

PAGE 72

61 Table 6 Quality Adjusted (by RE) Standard Unit Price Differentials for All Molecules, Relative to Germanya Year Germany France Italy United Kingdom Spain 1994 -31.91% -38.02% -12.12% -51.23% 1995 0.21% -29.25% -41.05% -10.47% -50.24% 1996 1.99% -29.16% -38.06% -10.84% -50.15% 1997 0.70% -27.71% -35.10% -10.57% -48.23% 1998 -0.11% -26.58% -32.51% -3.20% -47.35% 1999 -0.39% -25.65% -29.61% 1.96% -46.49% 2000 -0.78% -25.17% -27.84% -0.27% -46.65% 2001 -0.30% -25.76% -26.97% -1.98% -47.03% 2002 -4.83% -22.42% -22.52% 1.68% -44.60% 2003 -6.10% -22.43% -24.20% 2.57% -43.43% The main result is that price differences are still significant, but the percentage differences are consistent with the expectations that price differentials are decreasing over time. All the countries ha ve lower prices relative to Germany during the time period except the United Kingdom between 2002 and 2003. Prices in France, Italy and Spain move similarly, which is lik ely attributable to their pha rmaceutical industries having similar characteristics. UK prices are getting higher at small percen tages in 1999. It is observed that the price differences are decrea sing at increasing rate of possibly due to increased parallel imports and the Europ ean Commissions coordination efforts. a These are the coefficients of the country/time effects ( j,t) in the random effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1].

PAGE 73

62 Figure 1 Quality Adjusted (by RE) Standa rd Unit Price Differentials for All Molecules, Relative to Germany a -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 1994199519961997199819992000200120022003 Year% FR ITY UK SPN Price differentials estimated from the fi xed effect model are reported in tables A.16 and A.17 and figures B.15 and B.16. As expected, the results are quite similar. Additionally, the same models are estimated without year an d country indicators. The quality adjusted prices, i.e. the residuals, fr om these latter models are used to form the dependent variable in the pr ice convergence equations. 5.2.3 Quality Adjusted Standard Unit Pric e Differences for Global Molecules Table 7 shows the results from applying the same model to the globally diffused molecules. Coefficient signs are the same, a nd magnitudes remain similar. Molecule age variable is significant in thes e specifications, indicating that older molecules have lower prices. a Normalized to Germany.

PAGE 74

63 Table 7 Quasi-Hedonic Price Regression Results for Global Molecules, Relative to Germany Fixed Effect Model Random Effect Model Dependent variable: logleusuprice Explanatory Variables Coefficient (Standard Error)a Coefficient (Standard Error)a Quality Characteristics STRENGTHG -0.1975* (0.0534) -0.1735* (0.0465) MOLAGE -0.0104** (0.0044) -0.0158* (0.0033) PACKSIZE -0.0019* (0.0002) -0.0019* (0.0002) FORMCODE -0.0056* (0.0010) -0.0055* (0.0010) Market (Competition) Characteristics GENCOMPET -0.0028*** (0.0017) -0.0030*** (0.0017) THSUBSMOL 0.0094* (0.0022) 0.0106* (0.0025) THSUBSMOLENTLAG -0.0019 (0.0017) Country dummies Yes Yes Time dummies Yes Yes Country/Time dummies Yes Yes ATC dummies No Yes N 1,900 1,900 R2 (Within) 0.2593 0.2593 Prob>F 0.0000 0.0000 a Standard errors are heteroskedasticity-robust. *, ** and *** reflect p<0.01, p<0.05 and p<0.10.

PAGE 75

64 The estimated price differentials are reporte d in the table 8 and figure 2, relative to Germany, for global molecules. Price differe nces are significant and large relative to Germany, also much larger relative to Spain (tables A.18 and figure B.17). All countries price differences are decreasi ng at increasing rate s. In 2003, Spain has the highest price differences for all the years, representing the lo west priced country, re lative to Germany. Table 8 Quality Adjusted (by RE) Standard Unit Price Differentials for Global Molecules, Relative to Germany a Year Germany France Italy United Kingdom Spain 1994 -35.18% -40.37% -10.63% -50.84% 1995 1.26% -31.41% -43.44% -11.15% -49.78% 1996 2.87% -29.62% -39.71% -11.38% -48.88% 1997 0.60% -26.35% -35.67% -10.43% -45.12% 1998 0.52% -25.71% -32.66% -5.08% -43.25% 1999 1.79% -26.03% -29.49% -3.36% -41.80% 2000 1.01% -24.19% -26.46% -4.53% -40.75% 2001 1.69% -23.48% -24.38% -7.61% -40.76% 2002 -3.26% -18.58% -19.43% -4.36% -37.02% 2003 -6.32% -16.39% -18.32% -0.82% -34.03% Not surprising results are obtained when pr ice differentials are observed relative to Spain. Price difference decreases from 82% (1994) to 25% (2003) in the UK, where it decreases from 103% (1994) to 26% (2003) in Germany. Italy and France are % and 1% higher priced than Spain in 2003. Agai n, price differences decrease at increasing rates relative to Spain over the sample time period. a These are the coefficients of country/time effects ( j,t) in the random effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1].

PAGE 76

65 Figure 2 Quality Adjusted (by RE) Standard Unit Pr ice Differentials for Global Molecules, Relative to Germanya -60.00 -50.00 -40.00 -30.00 -20.00 -10.00 0.00 10.00 1994199519961997199819992000200120022003 Year% FR ITY UK SPN Fixed effect model price differences for global molecules are reported in tables A.19 and A. 20 and figure B.18, with simila r results. Moreover, the same models are estimated without year and c ountry indicators. The quality adjusted prices, i.e. the residuals, from these latter models are used to form the dependent variable in the price convergence equations for global molecules. 5.3 Price Convergence Results This section shows the price convergen ce results. The regression coefficients provide estimates of the speed of the convergen ce to the law of one price, which indicates how quickly deviations from long term price differentials are eliminated. a Normalized to Germany.

PAGE 77

66 Table 9 reports estimation results with Sp ain and Germany as base countries. The Hausman test rejects the null hypothesis of fi xed and random effects equivalence, thus fixed effects models are used to form th e dependent variables for the convergence equations. The results are robust for zero, one and two lags, in addition to Campbell and Perron (1992) top down approach, with the latter reported in the tables. Model 1 does not include molecule/country fixed effects and thus tests for convergence to the absolute version of the law of one price. The nu ll hypothesis is that the price differences converge to ward zero in the long run, i.e. < 0. The coefficient is negative in the specification where Germany is th e base country, with implied half-life of a shock, i.e. ln(2)/ln(1+ ), of 34.3 years. This half-lif e of 34.3 years is longer than the typical life cycle of a drug and provides of little, if any evidence of convergence. When Spain is the base country, the coefficient is positive, indi cating that there is no evidence of the absolute version of law of one pric e holding. The positive sign of the coefficient actually implies persistent price divergence. The results are mostly consistent with the international trade literature.

PAGE 78

67 Table 9 Results for Price Convergence Estimations for All Molecules (Adjusted by FE) Model 1 Model 2 Dependent Variable: Pi,k,t Base: SPN Base: GR Base: SPN Base: GR a 0.01 -0.02 -0.18 -0.17 Half-life of Shock (in years) 34.3 3.5 4.4 FRb 0.02 0.01 GRb 0.08 UKb 0.15 0.00 SPNb -0.08 ITYb 0.06 -0.08 Lags of Pi,k,t Yes(1)c Yes(1)c Yesd Yesd t-star P>t 8.89 1.000 -19.25 0.000 -8.39 0.000 -8.34 0.000 Molecule/Country Fixed Effects No No Yes Yes Time Trend No No No No N 2,940 3,210 2,940 3,210 Model 2 tests the relative version of th e law of one price and finds estimated ranges between .17 and .18. Based on the adju sted t statistics (t-s tar) values, the unit root hypothesis is rejected, signifying significant evidence of price convergence. The implied half-lives of shocks, according to these estimates, are between 3.5 and 4.4 years. These half-lives are longer than f ound in the recent international trade literature (Goldberg and Verboven 2005). a coefficients are estimated by the Levin et al. (2002) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each mo lecule/country by the Augmented Dickey Fuller regressions in StataTM 9.2 (dfuller) and then aver aged for each country. c The average number of lags for each molecule/country is 1. d The number of lags is determined by using the Campbell and Perron top-down approach.

PAGE 79

68 The i,k in model 2 captures molecule country fixed effects that account for nontime dependent, molecule specific price differe nces across countries. Such effects could include transportation costs, unobserved quality differences th at vary by destination, or markup differences. The presence of molecule /country fixed effects in the estimations implies testing the relative version of law of one price. The average molecule/country specific fixed effects are displayed in table 9. These large values of molecule/country fixed effects indicate market segmentation even if the relative version of law of one price holds in the data, consistent with other studie s in the literature. By dividing the fixed effects by long-term systematic price differences relative to Spain of 83% for the UK, 33% for Italy, 11% for France, and 44% for Germany are obtained. When Germany is the base country, long term pr ice differentials are 6% for France, 0% for the UK, and 47% for Spain and Italy. Similar convergence coefficients are obt ained for both the absolute and the relative version of law of one pr ice, using prices from random effect models (table A.21). Slightly smaller molecule/count ry price differentials emerge which are higher than Spain by 11% for France, 47% for Germany, 53% for the UK and 37% for Italy, and lower than Germany by 50% for Spain, 39% for Italy, 0% for France and the United Kingdom. In addition, all specifications are also estimated using model 3, which includes a time trend. The expected results, that the deviations from the long term differences are eliminated in approximately a year, are obtained. This is shorter than without the time trend in model 2, but the long term price differences remain about the same. In addition, the same hypotheses are tested for unadjusted prices. In this case, model 3, which includes a time trend, is al so reported in the tables for comparison

PAGE 80

69 purposes. Table A.22 reports the results for the three models. The results remain robust for the absolute version of the law of one price. In model 2, the estimated rates of convergence are very similar to the adjusted price estimates, but molecule/country fixed effects show very high price differences wh ich are expected due to not accounting for quality adjustments. The random effect quality adjusted price residuals for the globally diffused models are also used to test for price converg ence because the Hausman test fails to reject the null hypothesis. Table 10 shows that these results are similar to those from before. Model 1 rejects the hypothesis of a unit root where Germany is the base country and fails to reject the unit root where Spai n is the base. Model 2 shows that coefficients are negative, showing evidence of price convergence and implied half-lives of shocks of 2.9 and 4.6 years. The long term price diffe rentials are 36% (Germany), 186% (UK), 64% (Italy) and 0% (France) highe r than Spain and 24% (Spain) and 38% (Italy) less and 19% (UK) and 33% (France) higher than Germany. The results are similar when prices are formed using the residuals from the fixed ef fect model (table A.23). Lastly, the same three models are re-estimated for global mol ecules for unadjusted prices (table A.24). Again, all estimations remain the same but mo lecule/country fixed effects are larger than adjusted estimation results. In model 2, the long term price differen tials are much higher than the previous estimations.

PAGE 81

70 a coefficients are estimated by the Levin et al. (2002) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each mo lecule/country by the Augmented Dickey Fuller regressions in StataTM 9.2 (dfuller) and then aver aged for each country. c The average number of lags for each molecule/country is 1. d The number of lags is determined by using the Campbell and Perron top-down approach. Table 10 Results for Price Convergence Estimations for Global Molecules (Adjusted by RE) Model 1 Model 2 Dependent Variable: Pi,k,t Base: SPN Base: GR Base: SPN Base: GR a 0.01 -0.02 -0.14 -0.21 Half-life of Shock (in years) 34.3 4.6 2.9 FRb 0.00 0.07 GRb 0.05 UKb 0.26 0.04 SPNb -0.05 ITYb 0.09 -0.08 Lags of Pi,k,t Yes(1)c Yes(1)c Yesd Yesd t-star P>t 11.50 1.000 -17.08 0.000 -5.71 0.000 -12.40 0.000 Molecule/Country Fixed Effects No No Yes Yes Time Trend No No No No N 1,900 1,900 1,900 1,900

PAGE 82

71 Chapter 6 Conclusions This chapter summarizes the main findings regarding pharmaceutical prices in the European Union. Additionally, limitations and opportunities for fu ture research are discussed. 6.1 Main Findings This dissertation is the first attempt to invest igate the convergence of pharmaceutical prices in the EU. It uses annual panel data from 1994 on molecule level standard unit prices in the five largest pharmaceutical producing countries. The analysis has three main findings: 1. The first part of the analysis uses weighted price indexe s to compare price differences in the data for the selected nati ons. This approach is one of the suggested methods in the cross country price differen ces literature in the pharmaceutical industry (Danzon 2000). Using unadjusted per dose pr ices of bilaterally matching and global molecules, Laspeyres and Paasche indexes show that there are substantial diverse pharmaceutical price differences across EU markets. Bilaterally and using Germany as the base country, the results (tables 1 and fi gure B.4) indicate that Spain has consistent decreases in prices from -28% (1994) to 0.3% (2003) while the pric es increase from the beginning (1994) to the end (2003) of the time period by 7.5% to 41.4% in the UK, 15%

PAGE 83

72 to 35% in France, and -4.3% to 29% in Ital y by Laspeyrex index. Similar results are obtained when the indexes are calculated in th e global molecules (table 2 and figure B.8) such that price differences increase from -23% to 10% in Spain, 7% to 59% in the UK, 8% to 53% in France and -2% to 47% in Italy. When the base country is Spain (tables A.7 and A.8 and figures B.6 and B.10), the resu lts show that prices decrease in Germany, the UK and Italy but increase in France for a ll molecules; increase in the UK, France and Italy and decrease in Germany for global molecu les. Besides the choices of base country and index, these price differences depend on th e sample and method used. On their own, these price differences show price divergence across the countries and do not provide evidence of price convergence. 2. Even though the analysis employs molecule level standard unit prices, observed molecule characteristics vary across c ountries, particularly with regard to drug quality (form availability, pack sizes, strength levels) and market (competition) characteristics. Using quasi hedonic price regressions to control for this variation, price differentials are re-analyzed. The results show stronger evidence of decreasing price differences than those imputed from the inde x calculations. Price differences in all countries consistently decreas e from 1994 and 2003 regardless of the choice of the base country, method and the sample. Price diffe rentials decrease from 32% to 22% in France, 38% to 24% in Italy, 51% to 43% in Spain and -12% to 2.6% in the UK, relative to Germany for all molecules; there are smaller and again consistent decreases for global molecules in the time period of the study (tables 6 and 8, figures 1 and 2). Similar results are obtained relative to Spain (tables A.15 and A.18, figures B.14 and B.17). However,

PAGE 84

73 the price differences across countries still remain substantial even when the observed quality and market characteristics are controlled in the analysis. 3. Controlling for variation in quality a nd market characteristics across countries, both the absolute and relative versions of the law of one price are tested by employing panel data unit root tests. Price convergen ce is expected in integrated markets like the European Union, but factors like transportation costs, ta x differences and regulatory regimes that vary across countries, might produce fixed country-specific price differences. The relative version of the law of one price states that these price gaps tend to return to some long-run level over time, even if this level is not zero. Despite the fact that the data in this analysis control for observable quality and market characteristics, systematic price differences across countries could persist because of the nature of the pharmaceutical market and differences across countries with regard to demographics, culture and medicine consumption attitudes. For the absolute version, the half life of a shock is 34 years, indicating very slow convergence. The main interest in this study is not the absolu te version of the law of price because it is expected that cr oss country drug price differences cannot be completely eliminated. Results provide evidence in support of the relative version of the law of one price, which is a narrower defi nition of price convergence and is the key interest of the study. For the relative version, half lives ar e between 3 and 5 years. The estimated rates of convergen ce in the EU pharmaceutical industry are comparatively slower than the rates for analogous studies of different European industries, e.g. half lives of 1.3.6 years in the automobile industry. This could be expl ained by the different

PAGE 85

74 health care systems, pricing and reimburseme nt regulations by the national governments across countries. Long term price differentials imputed from the estimated molecule/country fixed effects show price differences across markets between 0% and 40% for countries other than the UK, providing evidence of market se gmentation. Possible sources of the larger long term price differences between the UK and other countries includ e the fact that the UK is not a member of the euro zone, transportation costs (e.g. the UK is not contiguous like the other na tions) and other different policies by the UK government. According to Kotzian (2004), the reason for thes e price differences co uld be the political arena in the EU: some governments grant hi gh prices for newly introduced products to encourage therapeutic innovations, others set a very low price due to health care budget concerns. This study contributes new results to the existing literature on European integration and pharmaceutical price convergence The results show that even though the pharmaceutical industry is one of the most heavily regulated markets with divergent methods (tables A.1-A.3) in the EU and ther e are no legislative actions by the European Commission for the pharmaceutical industry towa rd a single market, price differences across countries do converge over time, al beit relatively slowly, conditional on the molecule/country fixed effects. Although pri ce differences still exist, progress toward the single pharmaceutical market is evident. On e of the possible explanations of finding the evidence of price convergence could be paralle l import. However, this study does not test the direct impact of the parallel import on price conve rgence which also is not the scope of this dissertation. However, Gans landt and Maskus (2004) analyzed the direct

PAGE 86

75 impact of parallel imports on drug prices in Sweden, and concluded th at parallel imports represented a significant form of compe tition in markets and reduced manufacturing prices by 12%. Another possible explanati on could be internati onal reference pricing that create spillover of price levels from one country to another. The unique molecule level data set in th is dissertation includes branded drugs along with generics, licensed, OTCs and pa rallel imports, and allo ws controlling for quality and market characteristics. The empi rical results not only de monstrate consistent decrease in the prices overal l but also provide evidence of relative price convergence in the pharmaceutical industry. As it is emphasi zed by the European Commission that there is no reason to exempt pharmaceuticals from the single market ideal. Finally, this study attempts to be the firs t detailed empirical investigation of the drug price convergence in the EU. 6.2 Limitations This research has several limitations that should be noted. First, the data set analyzed in this study spans only ten years. Several other studies in the literature use periods of similar length, but as with a ll time series, a longer series would permit a stronger test of the price convergence hypothesis. It would also allo w for testing whether molecule/country effects are declining over time and if the speed of convergence has changed over time. Second, the data set is based on bilate rally and globally matched samples of molecule/ATC3 criteria that failed to ma tch nearly 50% of the total retail sales observations. Therefore, even the price indexe s may be unavoidably biased by selection.

PAGE 87

76 Third, quality and market characteris tics are based on the measurability of variables in the data set. A better market characteristic variable, such as an index variable compiling all the differe nt regulations and reimburseme nts in different countries, might be created in order to better capture the differences in prices. Fourth, since IMS fixed euro standard un it prices are used to flatten exchange rate fluctuations, the role of exchange rate changes must be investigated separately. Fifth, the data include all formations of the drugs including brand, OTC, parallel imports, etc. Because the structure of the data and variables, it is not possible to determine if the evidence of the price convergence is due to parallel importing or the alignment of government politics. Lastly, of the 25 members of the Europ ean Union only the largest five members are part of this study. 6.3 Future Research Future research should address several of the limitations mentioned above. The data should extend to more years, including ex change rates, and other quality and market characteristics to capture price differences In particular, further research should investigate whether molecule /country effects are declining over time along with the investigation of the sp eed of convergence. As a follow up, data from this disserta tion should employ a similar analysis but match data by molecule/ATC4 category, product level under the same ATC and international product name (IPN). IMS assigns drugs the same IPN that have two of the following: same chemical compositions, same brand name, or same corporation.

PAGE 88

77 The effect of price regulations (mostly demand-side regulations) on consumption patterns in relationship to the price convergence is an important topic for further research due to various cultural and consumpti on attitudes across countries. Even though the European Commission has never proposed any common policy for the EU pharmaceuticals, only recommendati ons, it should be investigated if the European Commission, the G10 Medicine Grou p or the establishment of the EMEA in 1995 has had any impact on price convergen ce. In addition, because of the standardization in packaging and labeling, it is expected that there will be an increase in parallel importing and therefore an impact on the price convergence. This could be another area of future research. Tables A.1-A.3 show all the different regulations across countries on both the demand and the supply side. It should be inve stigated if the penguin effect has resulted in any convergence of regulations in the EU pharmaceutical industry. It is expected that parallel imports will impact price differences. Ganslandt and Maskus (2004) developed a theoretical model and also empirically te sted whether parallel import reduced the price differences. As a follow up of this work, the theoretical work could be extended to the impact of the paralle l import on the relative version of the law of one price and half life of a shock in the regulated markets. Moreover, a detail investigation of the determinants of price convergence in the EU pharmaceutical industry is subject to future research. Finally, continued research in this area w ould give a better u nderstanding of the relationship between market integration and price convergence in the pharmaceutical

PAGE 89

78 industry in the EU. This dissert ation represents a preliminary investigation of this subject; however it does seem to be very open area for future research.

PAGE 90

79 References Germany Pharmaceutical Pricing and Reimbursement. LSE Survey on Pharmaceutical Pricing and Reimbursement Regulation in Europe Report by the London School of Economics and Social Science. Allen, C., Gasiorek, M., Smith, A., Fl am, H., and Sorensen, P.B. 1998. "The Competition Effects of Single Market in Europe." Economic Policy, 13:27, pp. 439-86. Allen, R.G.D. 1975. Index Numbers in the Theory and Practice: Aldine Publishing Company. Arfwedson, J. 2003. "Parallel Trade in Pharmaceuticals." CNE (Centre for the New Europe) Publications. Baltagi, B.H. 2005. Econometric Analysis of Panel Data.Third Edition: Wiley. Berndt, E.R. 2000. "Editorial: Intern ational Comparison of Pharmaceutical Prices? What Do We Know, and What Does It Mean?" Journal of Health Economics, 19, pp. 283-87. Berndt, E.R., Bir, A., Busch, S.H., Frank, R.G., and Normand, S.L.T. 2002. "The Medical Treatment of Depression, 19911996: Productive Inefficiency, Expected Outcome Variations, and Price Indexes." Journal of Health Economics, 21, pp. 373-96. Berndt, E.R., Cocburn, I.M., Griliches, Z., Keeler, T.E., and Baily, M.N. 1996. "Pharmaceutical Innovations and Market Dyna mics: Tracking Effects on Price Indexes for Antidepressant Drugs." Brookings Papers on Economic Activity. Microeconomics., 1996:1996, pp. 133-99. Berndt, E.R., Cutler, D.M., Frank, R. G., Griliches, Z., Newhouse, J.P., and Triplett, J.E. 1999. "Medical Care Prices and Output." Handbook of Health Economics, 1, pp. 119-77.

PAGE 91

80 Berndt, E.R., Griliches, Z., and Rosett, J.G. 1993. "Auditing the Producer Price Index: Micro Evidence From Prescrip tion Pharmaceutical Preparations." Journal of Business & Economic Statistics, 11:3, pp. 251-64. Blachier, C. and Kanavos, P. France Pharmaceutical Pricing and Reimbursement. LSE Survey on Pharmaceutical Pric ing and Reimburse ment Regulation in Europe Report by the London School of Economics and Social Science. Bornhorst, F. and Baum, C.F. 2001. "L evin-Lin-Chu Panel Unit Root Test." STATA. Burstall, M.L. 1991. "Europe After 1992: Implications for Pharmaceuticals." Health Affairs, 9, pp. 157-71. Camarero, M., Esteve, V., and Tamarit, C. 2000. "Price Convergence of Peripheral European Countries on the Way to the EMU: A Time Series Approach." Empirical Economics, 25, pp. 149-68. Capri, S. and Levaggi, R. 2005. Social Science Research Network: Working Paper Series. International Price Regulati ons in the Pharmaceutical Sector: A Common Model to Sharing Cost and Benefits. 2005. http://papers.ssrn.com Capri, S. and Levaggi, R. 2004. Social Science Research Network: Working Paper Series. Drug Pricing in a Regulated Market. 2004. http://papers.ssrn.com Caves, D.W., Christensen, L.R., and Diewert, W.E. 1982. "Multinational Comparisons of Output, Input, and Producti vity Using Superlative Index Numbers." The Economic Journal, 92:365, pp. 73-86. Caves, R.E., Whinston, M.D., Hurwitz, M.A., Pakes, A., and Temin, P. 1991. "Patent Expiration, Entry, and Compe tition in the U.S. Pharmaceutical." Brookings Papers on Economic Activity. Microeconomics, 1991:1991, pp. 1-66. Chaudhry, P., Dacin, P., and Peter, J. 1997. "Strategic Planning in the Pharmaceutical Industry in the European Union." European Management Journal, 15:6, pp. 686-97.

PAGE 92

81 Chaudhry, P., Dacin, P., and Peter, J. 1994. "The Pharmaceutical Industry and European Community Integration." European Management Journal, 12:4, pp. 442-53. Chaudhry, P. and Walsh, M.G. 1995. "Managing the Gray Market in the European Union: The Case of the Pharmaceutical Industry." Journal of International Marketing, 3:3, pp. 11-33. Communication Commission. 1998a. Commission Communication on the Single Market in Pharmaceuticals. COM(98)588 final. European Commission. Brussels Communication Commission. 1998b. Council Conclusions on Single Market in Pharmaceuticals. European Commission. Coopers, Lybrand N.V., UCL. 1996. Pri ce Convergence and Price Levels. 1996. http://ec.europa.eu/internal_market/studies/stud3.htm Danzon, P.M. 1997a. Pharmaceutical Price Regulation: National Policies versus Global Interests Washington, D.C.: The AEI Press. Danzon, P.M. 1997b. "Price Discrimination for Pharmaceuticals: Welfare Effects in the US and the EU." International Journal of the Economics of Business, 4:3, pp. 301(2). Danzon, P.M. 1998. "The Economics of Parallel Trade." Pharmacoeconomics, 13:3, pp. 293-304. Danzon, P.M. 1999. Price Comparisons for Pharmaceuticals: A review of U.S. and Cross-National Studies Washington, D.C.: The AEI Press. Danzon, P.M. and Chao, L.W. 2000a. "Does Regulation Drive Out Competition in Pharmaceutical Markets?" Journal of Law and Economics, 43:2, pp. 311-57. Danzon, P.M. and Chao, L.W. 2000b. "Cross-National Price Differences for Pharmaceuticals: How Large, and Why?" Journal of Health Economics, 19, pp. 159-95. Danzon, P.M. and Furukawa, M.F. 2003. "Prices and Availability of Pharmaceuticals: Evidence From Nine Countries." Health Affairs, Web Exclusive.

PAGE 93

82 Danzon, P.M. and Kim, J.D. 1998. "I nternational Price Comparisons for Pharmaceuticals." Pharmacoeconomics, 14:1, pp. 115-28. Darba, J. and Rovira, J. 1998. "Paral lel Imports of Pharmaceuticals in the European Union." Pharmacoeconomics, 14:1, pp. 129-36. Dickson, M. 1992. "The Pricing of Pharmaceuticals: An International Comparison." Clinical Therapeutics, 14:4, pp. 604-10. Dickson, M. and Jacobzone, S. 2003. "Pharmaceutical Use and Expenditure for Cardiovascular Disease and Stroke : A Study of 12 OECD Countries." OECD Health working Papers, DELSA/ELSA/WD/HEA (2003)1. Diewert, E. 2003. "Hedonic Regressions : A Consumer Theory Approach." Studies in Income and Wealth, 64. Diewert, E. 2005. "Adjacen t Period Dummy Variable Hedonic Regressions and Bilateral Index Number Theory." Discussion Papaer No: 05-11, Department of Economics, The University of British Columbia. DRI. 1996. "Subseries VImpact on Co mpetition and Scale Effects: Price Competition and Price Convergence." The Single Market Review Series. EFPIA. 1998. The Pharmaceutical Industry's Agenda Additional Statements from the Paris Round Table. European Commission. EFPIA. 2002. "2001-2002: The Year in Review." European Federation of Pharmaceutical Industries and Associations. EFPIA. 2005. "The Pharmaceutical Industry in Figures." European Federation of Pharmaceutical Industries and Associations, Key Data:2005 update, pp. 1-27. Ellison, S.F., Cockburn, I., Griliches, Z., and Hausman, J. 1997. "Characteristics of Demand for Pharmaceutical Products: An Examination of Four Cephalosporins." The RAND Journal of Economics, 28:3, pp. 426-46. EPHMRA. 2004. "Anatomical Clas sification Guidelines 2004."

PAGE 94

83 EphMrA/PBIRG, Classification Committee. 2005. "Who We Are What We Do." Ess, S.M., Schneeweiss, S., and Szucs, T. 2003. "European Healthcare Policies for Controlling Drug Expenditure." Pharmacoeconomics, 21:2, pp. 89-103. EUROPA. 2005. The EU At A GlanceThe History of the European Union. 2005. http://europa.eu.int July 2005. European Commission. 2004. European Economy: Convergence Report. Special Report No: 2 DirectorateGeneral fo r Economic and Financial Affairs. European Commission. 2000a. "Community Patent." Vol. COM(2000) 412 Final. European Commission. The Singl e Market in Pharmaceuticals. http://europa.eu.int/scad plus/leg/en/lvb/l21227.htm April 2005. European Commission. 2000b. Enterprise Directorate-General. Pharmaceuticals in the European Union. 2000b. http://pharmacos.eudra.org/F2/pha rmacos/docs/brochure/pharmaeu.pdf April 2005. European Commission. 2003. Communication from the Commission to the Council, the European Parliament, th e Economic and Social Committee and the Committee of the Regions: A Stronger Eur opean-based Pharmaceutical Industry for the Benefit of the PatientA Call for Action. COM(2003) 383 final. Brussels European Commission, Direct orate-General for Economic and Financial Affairs. 2002. European Economy European Integr ation and the Functioning of Product Markets. Special Report Number 2. Fan, C.S. and Wei, X. 2006. "Price Index Convergence in China." Lingnan University, Department of Economics. Feenstra, R.C. 1997. "Generics and New G oods in Pharmaceutical Price Indexes: Comment." The American Economic Review, 87:4, pp. 760-67. Flam, H. 1992. "Product Markets and 1992: Full Integration, Large Gains?" The Journal of Economic Perspectives, 6:4, pp. 7-30.

PAGE 95

84 Folland, S., Goodman, A.C., and Stano, M. 2004. The Economics of Health and Health Care.Fourth Edition: Pearson-Prentice Hall. Fontaine, P. 2003. Europe in 12 Lessons Luxembourg: Office for Official Publications of the European Communities. G10 Medicines, High Level Group on Innova tion and Provision of Medicines. 2002. Recommendations for Action. Gambardella, A., Orsenigo, L., and Pammolli, F. 2000. "Global Competitiveness in Pharmaceuticals: A European Perspectiv e," Report prepared for the Directorate General Enterprise of the Eu ropean Commission, pp. 1-100. Ganslandt, M. and Maskus, K. 2004. "P arallel Imports and the Pricing of Pharmaceutical Products: Evidence From the European Union." Journal of Health Economics, 23, pp. 1035-57. Garattini, L., Salvioni, F., Scopelliti, D ., and Garattini, S. 1994. "A Comparative Analysis of the Pharmaceutical Mark et in Four European Countries." Pharmacoeconomics, 6:5, pp. 417-23. Gaulier, G. and Haller, S. 2000. "The Convergence of Automob ile Prices in the European Union: An Empirical Analysis for the Period 1993-1999." CEPII, Working Paper, n-00:14, pp. 1-25. Gerschenkron, A. 1947. "The Soviet Indices of Industrial Production." The Review of Economic Statistics, 29:4, pp. 217-26. Gerschenkron, A. 1955. "Soviet Heavy Industry: A Dollar Index of Output, 1927/28-1937." The Review of Economics and Statistics, 37:2, pp. 120-30. Gil-Pareja, S. and Sosvilla-Rivero, S. 2004. "Export Market Integration in the European Union." Journal of Applied Economics, VII:2, pp. 271-301. Gilles, R. 1940. "International Comparison of Wholesale Prices." The Review of Economic Statistics, 22:3, pp. 150-56.

PAGE 96

85 Goldberg, P.K. and Verboven, F. 2001. "The Evolution of price Dispersion in the European Car Market." Review of Economic Studies:68, pp. 811-48. Goldberg, P.K. and Verboven, F. 2004. "Cross-Country Price Dispersion in the Euro Era: A Case Study of the European Car Market." Economic Policy, pp. 483-521. Goldberg, P.K. and Verboven, F. 2005. "Market Integration and Convergence to the Law of One Price: Evidence from the European Car Market." Journal of International Economics, 65, pp. 49-73. Griliches, Z. and Cockburn, I. 1994. "Generics and New Goods in Pharmaceuticals Price Indexes." The American Economic Review, 84:5, pp. 1213-32. Guillen, A.M. and Cabiedes, L. 2003. "Reforming Pharmaceutical Policies in the European Union: A "Penguin Effects"?" International Journal of Health Services, 33:1, pp. 1-28. Hausman, J. 1978. "Specification Tests in Econometrics." Econometrica, 46:6, pp. 1251-71. Hellerstein, J. 1998. "The Importance of the Physician in the Generic versus Trade-Name Prescription Decision." The RAND Journal of Economics, 29:1, pp. 108-36. Huttin, C. 1999. "Drug Price Divergence in Europe: Regulatory Aspects." Health Affairs, 18:3, pp. 245-49. IMS. 2002. "IMS Midas Customized Insi ghts: Data Elements, Measures and Statistics." IMS Health Incorporated or Its Affiliates. IMS. 2005a. "Guide to Pack Prices on IMS Midas Quantum." IMS. 2005b. IMS Midas Quantum Data Elem ents, Measures and Statistics. IMS, Health Data. 2006. "New Form Code/Counting Unit and Standard Unit Definitions."

PAGE 97

86 Jacobzone, S. 2000. "Pharmaceutical Policie s in OECD Countries: Reconciling Social and Industrial Goals." Labor Market ad Social Policy-Occasional Papers No: 40:DEELSA/ELSA/WD (2000)1, pp. 1-100. Jefferys, D. 1995. "The New Pharmaceutical Regulatory Procedures for Europe." TIPS, 16:Speacial Feature, pp. 226-31. Jommi, C. Italy Pharmaceutical Pricing and Reimbursement. LSE Survey on Pharmaceutical Pricing and Reimbursement Regulation in Europe Report by the London School of Economics and Social Science. Jonas, P. and Sardy, H. 1970. "The Gerschenkron Effect: A Re-Examination." The Review of Economics and Statistics, 52:1, pp. 82-86. Kanavos, P. 2000. "The Single Market for Pharmaceuticals in the European Union in Light of European Court of Justice Rulings." Pharmacoeconomics, 18:6, pp. 523-32. Kanavos, P. 2001. Communication of the Eu ropean Communities, Dg. Enterprise. Overview of Pharmaceutical Pricing and Reimbursement Regulation in Europe. 2001. http://www.pharmacos.eudra.orf/F3/g10/docs/synthesis.pdf April 2005. Kerem, K., Puss, T., and Viies, M. 2005. "Convergence of Health Care Expenditure in EU." The Global Economic and Research Conference, Istanbul. Kokoski, M.F. 1992. "New Research on In terarea Consumer Price Differences." Monthly Labor Review. Kotzian, P. 2004. "Pharmaceutical R& D Setting of Incomplete European Integration." International Journal of the Economics of Business, 11:2, pp. 177-95. Kravis, I.B. and Lipsey, R.E. 1969. "International Price Comparisons by Regression Methods." International Economic Review, 10:2, pp. 233-46. Kullman, D. United Kingdom-Pharmaceutical Pricing and Reimbursement Policies. LSE Survey on Pharmaceutical Prici ng and Reimbursement Regulation in Europe Report by the London School of Economics and Social Science.

PAGE 98

87 Levin, A. and Lin, C.F. 1992. "Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties." University of California, San Diego, Department of Economics, Discussion Paper 92-23. Levin, A. and Lin, C.F. 1993. "Unit R oot Tests in Panel Data: New Results." University of California, San Diego, Department of Economics:Discussion Paper 93-56. Levin, A., Lin, C.F., and Chu, C-S.J. 2002. "Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties." Journal of Econometrics, 108:2002, pp. 1-24. Liikanen, E. 2004. "Changes to the Europ ean Union Pharmaceutical Legislation: Implications of the "Pharma Re view" for Biotechnology Companies." Journal of Commercial Biotechnology, 11:1, pp. 38-43. Lopez-Casasnovas; Puig-Junoy, J. 2000. "R eview of the Literature on Reference Pricing." Health Policy, 54:2, pp. 87-123. Maddala, G.S. 1999. "On the Use of Pane l Data Methods with Cross-Country Data." Annales D'Economie Et De Statistique, 55-56. Maddala, G.S. and Wu, S. 1999. "A Comp arative Study of Unit Root Tests with Panel Data and a New Simple Test." Oxford Bulletin of Econom ics and Statistics, Special Issue. Maskus, K. 2001. Parallel Imports in Pharmaceuticals: Implications for Competition and Prices in Developing Countries. Final Report to World Intellectual Property Organization. Maskus, K. and Chen, Y. 2004. "Vertical Price Control and Parallel Imports: Theory and Evidence." Review of International Economics, 12:4, pp. 551-70. Mossialos, E., Mrazek, M., and Wally, T. eds. 2004. Regulating Pharmaceuticals in Europe: Striving for Efficiency, Equity and Quality: Open University Press. European Observatory on Health Systems and Policies Series. Norris, P. 1998. "The Impact of Eu ropean Harmonisation on Norwegian Drug Policy." Health Policy, 43, pp. 65-81.

PAGE 99

88 Noyce, P.R., Huttin, C., Atella, V., Brenner, G., Haaijer-Ruskamp, F.M., Hedwall, M., and Mechtler, R. 2000. "The Cost of Prescription Medicines to Patients." Health Policy, 52, pp. 129-45. OECD, Committee on Competition Law and Policy. 2000. "Competition and Regulation Issues in the Pharmaceutical Industry." Directorate for Financial, Fiscal and Enterprise Affairs: DAFFE/CLP(2000)29. OECD, Health Data. 2003. "A Compar ative Analysis of 30 Countries." Pammolli, F., Riccaboni, M., and Magazzini, L. 2004. European Competitiveness in Pharmaceuticals. Peck, M. 1989. "Industrial Organization and the Gains from Europe 1992." Brookings Papers on Economic Activity, 1989:2, pp. 277-99. Permanand, G. and Mossialos, E. 2004. "Theorising the Development of the European Union Framework for Pharmaceutical Regulation." LSE Health and Social Care Discussion Paper, Number 13. Pollard, S. 2002. CNE Publications. Saving the European Pharmaceutical Industry: Price Regulatio n and 'Recommendation VI' (A CNE White Paper). 2002. www.centrefortheneweurope.org Pollard, S. 2003. "Pharmaceutical R&D Jeopardized in the EU." Fraser Forum: 17-19. Ratfai, A. 2006. "How Fast Is Converge nce to the Law of One Price? Very." Economics Bulletin, 6:10, pp. 1-12. Read, I.M. 1998. Report on the Communication fr om the Commission on the Single Market in Pharmaceuticals. COM(98)0588-C4-0127/99. Committee on Economic and Monetary Affairs an d Industrial Policy. Redwood, H. 1994. "Public Policy Trends in Drug Pricing and Reimbursement in the European Community." Pharmacoeconomics, 6:1, pp. 1-10.

PAGE 100

89 Rogers, J.H. 2001. "Price Level Convergence, Relative Prices and Inflation in Europe." International Finance Discussion Pape rs #699, Board of Governors of the Federal Reserve System. Rogers, J.H., Hufbauer, G.C., and Wada E. 2001. "Price level Convergence and Inflation in Europe." Working Paper 01-1, Institut e for Interna tional Economics. Scherer, F.M. 1993. "Pricing, Profit s and Technological Progress in the Pharmaceutical Industry." The Journal of Economic Perspectives, 7:3, pp. 97-115. Scherer, F.M. 2000. "The Pharmaceutical Industry." Handbook of Health Economics, 1. Scut F.T.; Van Bergeijk, P.A.G. 1986. "International Price Dicrimination: The Pharmaceutical Industry." World Development, 14:9, pp. 1141-50. Seget, S. 2003. Pharmaceutical Pricing Strategies: Optimizing Returns Throughout R&D and Marketing. Reuters Business Insight Healthcare. Sirmans, G.S., Macpherson, D.A., and Zietz, E.N. 2005. "The Composition of Hedonic Pricing Models." Journal of Real Estate Literature, 13:1, pp. 3-43. Smith, D.L. and Wanke, J. 1993. "Comple ting the Single European Market: An Analysis of the Impact on the Member States." American Journal of Political Science, 37:2, pp. 529-54. Sosvilla-Rivero, S. and G il-Pareja, S. 2004. "Price C onvergence in the European Union." Applied Economics Letters:11, pp. 39-47. Towse, A. 1998. "The Pros and Cons of a Single "Euro-Price" for Drugs." Pharmacoeconomics, 13:3, pp. 271-76. Verboven, F. 1996. "International Price Discrimination in the European Car Market." RAND Journal of Economics, 27:2, pp. 240-68.

PAGE 101

90 Vogel, R. 2004. "Pharmaceutical Pricing, Price Controls, and Their Effects on Pharmaceutical Sales and Research and Development Expenditures in the European Union." Clinical Therapeutics, 26:8, pp. 1327-40. Wertheimer, P. 2003. "The Economics of Eu ropean Integration Lecture Notes." Normandy, France. WHO. 2002. "Model List of Essential Medi cines (12th List) Sorted According to ATC Classification." Wooldridge, J.M. 2001. Econometric Analysis of Cr oss Section and Panel Data: The MIT Press. Wooldridge, J.M. 2003. Introductory Econometrics: A Modern Approach.2: Thompson-South Western.

PAGE 102

91 Bibliography Wooldridge, Jeffrey M. 2002. Econometric Analysis of Cross Section and Panel Data. The MIT Press. Cambridge, MA.

PAGE 103

92Appendices

PAGE 104

93 Appendix A: Tables Table A.1 National Controls for Pharma ceuticals on the Supply-Side COUNTRY: NATIONAL HEALTH SYSTEMS: PRICING: REIMBURSEMENT: Germany GKV, statuary health insurance covers 88% of the population. Most of the remaining population had private insurance. Price freedom for new products a) Reference price for off-patent sector (products subjected to generic competition; reference price for identical molecule only) b) Drug budgets with caps re-introduced in 1999. c) Negative list d) Positive list UK The National Health Service since 1948 financed through central government. a) PPRS: Agreement with industry on profit control, renewed in 1999 for a five-year period b) Price cut, as part of PPRS, of 4.5% c) Free price modulation by 2001. a) Negative List b) Homogeneous budget given to PCGs c) Practice guidelines d) Guidance on cost-effectivene ss by NICE, influences prescribing France Universally covered (99% of the population) by statuary health insurance. a) Price fixing through negotiation (products medical value, prices of comparable medicines, volume sales and conditions used) b) Comparisons with other European Countries for innovative products c) Periodic price reductions for new and expensive products d) Price freedom has been introduced since 2003** a) Comite Economique du Medicament decides on reimbursable prices on advice from Transparency committee b) Positive List c) Medical References d) Targets for gate-keeping GP e) Pharmacoeconomic guidelines under development f) Prices of generics 30% lower than those of the original Italy SSN: National Health Service. Funds are supplemented by local taxes and health service charges. a) Average European Price (all EU countries) for old products and products registered with the nationa l procedure; AEP is calculated on ex-manufacturers price (excl. VAT) of top five selling equivalents, including generics. b) Price negotiation (contractual m odel) for new and innovative products (for drugs registered with EMEA or for those for which AEP cannot be calculated) c) Price freedom for non-reimbursable drugs d) Generics are priced at least 20% below the original e) Frequent use of price cuts/freezes a) Positive list b) Reference listing and same pri ces for same drugs principle for off-patent drugs c) Formal requirement for economic evaluation during price negotiations d) Guidelines and protocols de fined and managed at local level e) Official earmarked budget for innovative drugs introduced in 1998, representing 1% of national drug budget Spain The National statuary health insurance a) Price control trough negotiation on a cost-plus basis b) International price comparisons c) Price-volume agreement for expensive products a) Positive list b) Negative list c) Reference pricing for estimating maximum reimbursement for multi-source products Adapted from (Kanavos 2001) Compiled from (Seget 2003), (Blach ier and Kanavos), (Jommi), (Kul lman), (Mossialos et al. 2004), ** Added from (Mossialos et al. 2004)

PAGE 105

94 Table A.2 Summary of Approaches in the Regu lation of Pharmaceutical Prices by On-Patent and Off-Patent Drugs (2003) Countries: Market segment Free Pricing Direct Price Controls Use of international price comparisons Profit Controls Reference Pricing On-patent X X France Off-patent X On-patent X Germany On-patent X On-patent X X Italy Off-patent X On-patent X X Spain Off-patent X X On-patent X X UK Off-patent X Adapted from (Mossialos et al. 2004)

PAGE 106

95 Table A.3 Demand-Side Policies (Prescribing, Dispensing and Consumption) in the Member States Country: Positive List Negative List Budget Guidelines / Monitoring Generic Prescribing Substitution Incentives Co-payment France Yes No Yes Yes Yes (limitedgatekeepers) Yes Yes (gatekeepers) % Germany No (but planned) Yes Yes Yes Yes Yes Yes Flat Fee Italy Yes No Yes Yes No Yes No % + flat fee Spain Yes Yes No Yes Yes (limited) No No % up to a max per item UK No Yes Yes Yes Yes (limited) No Yes Flat Adapted from (Kanavos 2001)

PAGE 107

96 Table A.4 ATC Therapeutic Categories for Cardiovascular Disease ATC Code Category Name C1A Cardiac Glycosides and Combinations C2A Antihypertensives (of no n-herbal origin) Plain: It includes plain antihyper tensives and combinations other than those with diuretics, eg combinations of tw o synthetic antihypertensives or combinations of one synthetic antihypertensive with reserpine. C3A Diuretics: Combinations with potassium belong to C3A1, C3A2 or C3A3. C4A Cerebral and Peripheral Vasotherapeutics: This group includes all products (including citicoline) which are mainly recommend for cerebral vascular diseases or peripheral ci rculatory disorders excluding venous diseases. Combination products are only cl assified in this group if they do not belong to group C1-C3, C7-C11. C7A Beta-Blocking Agents, Plain: Includes, eg acebutolol, alprenolol, amosulalol, arotinol, atenolol, befunolo l, betaxolol, bevantolol, bisoprolol, bopindolol, bucumolol, bufetolol, bun itrolol, bupranolol, butofilolol, carazolol, carteolol, carvedilol, celiprolol, cloranolol, dilevalol, esmolol, indenolol, labetolol, levobunolol, me pindolol, metipranolol, metoprolol, nadolol, nifenalol, nipradilol, oxpreno lol, penbutolol, pindolol, practolol, propranolol, sotalol, tertanolol, tilisolol, timolol, toliprolol. C8A Calcium Antagonists, Plain C9A Ace Inhibitors, Plain : Angiotensin-Converting-Enzyme inhibitors. It includes eg alacepril, ben azepril, captopril, cilazepril, delapril, enalapril, fosinopril, imidapril, lisinopril, moexipri l, perindopril, quin april, ramipril, spirapril, temocapr il, trandolapril. C10A Cholesterol and Triglyceride Regulating Preparations: Includes all products regulating cholesterol and trig lycerides only. Combinations with products of group C4 shoul d be classified here. Adopted from (Jacobzone 2000)

PAGE 108

97 Table A.5 Largest Pharmaceutical Markets in the World, National Currency (million), growth: US$, NC 2000 2001 2002 2003 Rank 03 Country Mill NC +($) +NC Mill NC +($) +NC Mill NC +($) +NC Mill NC +($) +NC 1 USA 150,952 14 14 176,748 17 17 197,602 12 12 219,522 11 11 2 Japan 6,231,585 8 2 6,502,706 -7 4 6,603,811 -2 2 7,059,335 12 7 3 Germany 18,157 -8 6 19,921 7 10 21,515 14 8 24,631 30 14 4 France 18,111 -6 9 19,418 4 7 20,183 9 4 22,583 27 12 5 Italy 11,990 -1 15 13,441 9 12 14,136 11 5 15,592 25 10 6 United Kingdom 7,380 0 7 8,180 5 11 9,111 16 11 10,386 20 14 7 Spain 7,711 2 18 8,349 5 8 9,174 16 10 10,794 34 18 13 Australia 5,452 0 11 6,227 2 14 6,854 16 10 8,088 30 18 17 Belgium 2,722 -7 8 2,862 2 5 3,049 12 7 3,521 31 15 18 Poland 11,013 12 22 11,913 15 8 12,373 4 4 14,407 18 16 19 Greece 1,504 -1 19 1,805 15 20 2,281 33 26 2,898 44 27 20 Sweden 19,690 0 11 21,051 -5 7 22,737 15 8 24,711 22 9 21 Switzerland 2,971 -4 8 3,279 10 10 3,517 16 7 3,926 26 12 23 Austria 1,766 -9 6 1,864 3 6 2,038 15 9 2,284 27 12 24 Portugal 1,702 -6 8 1,848 5 9 1,998 14 8 2,183 24 9 29 Finland 1,071 -5 10 1,201 9 12 1,330 17 11 1,491 27 12 33 Denmark 7,105 -5 10 7,782 6 10 8,799 19 13 9,762 26 11 35 Norway 6,945 -3 9 7,742 9 11 8,912 30 15 9,384 16 5 36 Czeck Republic 28,254 0 11 29,896 7 6 32,763 27 10 38,138 29 16 Adapted from (Pammolli et al. 2004)

PAGE 109

98 Table A.6 Balanced Sample ATC/Molecule a nd Country Availability for 1994-2003 ATC Molecule Country Availability Global Molecule C7A Acebutolol FR,GR,ITY,SPN,UK X C10A Acipimox GR,UK C4A Alprostadil Alfadex GR C3A Amiloride FR, UK C8A Amlodipine FR,GR,ITY,SPN,UK X C7A Atenolol FR,GR,ITY,SPN,UK X C3A Azosemide GR C9A Benazepril FR,GR,ITY,SPN C4A Bencyclane GR C3A Bendroflumethiazide UK C10A Benfluorex FR,ITY,SPN C8A Bepridil FR C7A Betaxolol FR, GR, ITY, UK C10A Bezafibrate FR,GR,ITY,SPN,UK X C7A Bisoprolol FR,GR,ITY,SPN,UK X C4A Blood GR C4A Buflomedil FR,GR,ITY,SPN C3A Bumetanide FR, SPN,UK C2A Bunazosin GR C2A Cadralazine ITY C3A Canrenoic Acid FR,GR,ITY C9A Captopril FR,GR,ITY,SPN,UK X C7A Carteolol FR,GR,SPN C7A Carvedilol GR,ITY,SPN, UK C7A Celiprolol FR,GR, SPN, UK C3A Chlortalidone GR,ITY, SPN, UK C2A Cicletanine FR,GR C9A Cilazapril FR,GR,ITY,SPN,UK X C4A Cinnarizine GR,ITY,SPN,UK C10A Ciprofibrate FR,UK C4A Citicoline ITY,SPN C2A Clonidine FR,GR,ITY,SPN,UK X C3A Clopamide GR C10A Colestipol GR,SPN,UK C10A Colestyramine FR,GR,ITY,SPN,UK X C4A Cyclandelate FR,GR,ITY C1A Digitoxin GR,SPN C1A Digoxin FR,GR,ITY,SPN,UK X C2A Dihydralazine GR C4A Dihydroergocristine GR, ITY,SPN C4A Dihydroergotoxine FR,GR,ITY,SPN,UK X C8A Diltiazem FR,GR,ITY,SPN,UK X C2A Doxazosin GR,ITY,SPN, UK

PAGE 110

99Table A.6 (Continued) Balanced Sample ATC/Molecule a nd Country Availability for 1994-2003 ATC Molecule Country Availability (1993-2003 Period) Global Molecule C9A Enalapril FR,GR,ITY,SPN,UK X C7A Esmolol GR,SPN C10A Etofibrate GR C10A Etofylline Clofibrate GR C8A Felodipine FR,GR,ITY,SPN,UK X C10A Fenofibrate FR,GR,ITY,SPN,UK X C4A Flunarizine ITY C10A Fluvastatin GR,SPN C9A Fosinopril GR,ITY, SPN, UK C3A Furosemide FR,GR,ITY,SPN,UK X C8A Gallopamil GR,ITY C10A Gemfibrozil FR,GR,ITY,SPN,UK X C4A Ginkgo Biloba FR,GR,SPN C4A Glycosaminoglycan Polysulfate(S) ITY C2A Guanfacine FR C2A Hydralazine SPN,UK C3A Hydrochlorothiazide FR,GR,ITY,SPN,UK X C4A Ifenprodil GR C3A Indapamide FR,GR,ITY,SPN,UK X C2A Indoramin GR,UK C8A Isradipine FR,GR,ITY,UK C4A Kallidinogenase GR C2A Ketanserin ITY C7A Labetalol FR, ITY, SPN,UK C8A Lacidipine FR,ITY, SPN,UK C9A Lisinopril FR,GR,ITY,SPN,UK X C10A Lovastatin GR, SPN C7A Mepindolol GR C2A Methyldopa FR,GR,ITY,SPN,UK X C1A Metildigoxin GR, ITY,SPN C3A Metolazone GR, ITY,UK C7A Metoprolol FR,GR,ITY,SPN,UK X C2A Minoxidil FR,GR,ITY,SPN,UK X C4A Moxisylyte FR,UK C2A Moxonidine GR C7A Nadolol FR,GR,ITY,SPN,UK X C4A Naftidrofuryl FR,GR,ITY,SPN,UK X C8A Nicardipine FR,GR,ITY,SPN,UK X C4A Nicergoline FR, GR,ITY,SPN C8A Nifedipine FR,GR,ITY,SPN,UK X C8A Nilvadipine GR C4A Nimodipine GR,ITY,SPN,UK

PAGE 111

100Table A.6 (Continued) Final Balanced Sample ATC/Molecule and Country Availability for 1994-2003 ATC Molecule Country Availability (1993-2003 Period) Global Molecule C8A Nisoldipine GR,ITY,SPN C8A Nitrendipine FR,GR,ITY,SPN C7A Oxprenolol FR,GR, SPN,UK C10A Pantethine ITY,SPN C7A Penbutolol GR C4A Pentoxifylline FR,GR,ITY,SPN,UK X C9A Perindopril FR,GR,ITY,SPN,UK X C4A Phenoxybenzamine GR C4A Phentolamine UK C7A Pindolol FR,GR,ITY,UK C3A Piretanide FR,GR,ITY,SPN C4A Piribedil GR, ITY C10A Pravastatin FR,GR,ITY,SPN,UK X C2A Prazosin FR,GR,SPN,UK C10A Probucol SPN C7A Propranolol FR,GR,ITY,SPN,UK X C9A Quinapril FR,GR,ITY,SPN,UK X C9A Ramipril FR,GR,ITY,SPN,UK X C2A Rilmenidine FR C10A Simvastatin FR,GR,ITY,SPN,UK X C7A Sotalol FR,GR,ITY,SPN,UK X C3A Spironolactone FR,GR,ITY,SPN,UK X C7A Talinolol GR C2A Terazosin GR,ITY,UK C7A Tertatolol FR C7A Timolol FR,ITY,UK C3A Torasemide GR,ITY,UK C9A Trandolapril FR,GR,SPN,UK C3A Triamterene UK C2A Urapidil FR,GR,ITY C8A Verapamil FR,GR,ITY,SPN,UK X C4A Vincamine FR,GR,ITY,SPN C3A Xipamide FR,GR,ITY,SPN,UK X Summary of Data Set by ATC, Mole cule, Country and Global Molecules C1A=3 C2A=17 C3A=16 C4A=16 C7A=20 C8A=13 C9A=10 C10A=16 Total Molecules : 118 Country Totals : FR = 70 GR =94 ITY =75 SPN=69 UK =71 =379 Total Global Molecules: 38 C1A=1 C2A=3 C3A=5 C4A=3 C7A=7 C8A=6 C9A=7 C10A=6

PAGE 112

101 Table A.7 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Spain 1994 1995 1996 Index Measures GR FR UK ITY GR FR UK ITY GR FR UK ITY Number of ATC/Molecule Matching 81 63 71 82 78 62 67 78 78 65 68 79 Laspeyres Price Index (SPN weighted) 1.3851 1.3798 1.3284 1.4659 1.3104 1.3779 1.3303 1.3032 1.2570 1.3663 1.3253 1.3245 Paasche Price Index (Own weighted) 1.3797 1.2787 1.4740 1.3244 1.2941 1.2517 1.4795 1.1753 1.2323 1.2400 1.4615 1.2126 Laspeyres Quantity Index 2.3118 1.1846 2.5719 1.4422 2.3050 1.2043 2.3777 1.4065 2.3753 1.2605 2.2487 1.3634 Paasche Quantity Index 2.3029 1.0979 2.8538 1.3030 2.2764 1.0940 2.6444 1.2684 2.3287 1.1439 2.4797 1.2482 Normalized Laspeyres Quantity Index by Relative Population Size 4.7912 1.7442 3.7724 2.0888 4.7770 1.7737 3.4903 2.0323 4.9286 1.8571 3.3041 1.9655 Normalized Paasche Quantity Index by Relative Population Size 4.7727 1.6165 4.1859 1.8872 4.7178 1.6112 3.8817 1.8327 4.8319 1.6853 3.6435 1.7995 Population Ratio (Comparison/SPN) 2.0725 1.4724 1.4667 1.4484 2.0725 1.4728 1.4679 1.4449 2.0750 1.4733 1.4693 1.4417 Pp/Lp=Pq/Lq 0.9961 0.9268 1.1096 0.9035 0.9876 0.9084 1.1122 0.9018 0.9804 0.9075 1.1027 0.9155 r -0.0007 -0.1296 0.0502 -0.0492 -0.0003 -0.2261 0.1227 -0.1022 -0.0090 -0.2202 0.0473 -0.0879 Vp 0.6560 0.3607 0.4627 0.8024 0.6686 0.3707 0.4238 0.8001 0.6694 0.3709 0.4013 0.7553 Vq 8.0652 1.5659 4.7176 2.4459 68.146 1.0926 2.1567 1.2005 3.2668 1.1321 5.4115 1.2731 r.Vp.Vq -0.0039 -0.0732 0.1096 -0.0965 -0.0124 -0.0916 0.1122 -0.0982 -0.0196 -0.0925 0.1027 -0.0845

PAGE 113

102 Table A.7 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Spain 1997 1998 1999 Index Measures GR FR UK ITY GR FR UK ITY GR FR UK ITY Number of ATC/Molecule Matching 77 65 68 74 78 69 68 75 77 66 66 73 Laspeyres Price Index (SPN weighted) 1.1929 1.3553 1.3137 1.3165 1.1559 1.3937 1.2885 1.2978 1.1565 1.4089 1.2701 1.3178 Paasche Price Index (Own weighted) 1.1423 1.2391 1.4314 1.2074 1.0924 1.2606 1.3847 1.2032 1.0748 1.3154 1.3463 1.2239 Laspeyres Quantity Index 2.3843 1.2934 2.1498 1.2633 2.3024 1.3120 2.0025 1.2147 2.2541 1.3322 1.9368 1.2151 Paasche Quantity Index 2.2831 1.1825 2.3424 1.1586 2.1758 1.1867 2.1520 1.1261 2.0949 1.2438 2.0531 1.1285 Normalized Laspeyres Quantity Index by Relative Population Size 4.9471 1.9057 3.1610 1.8180 4.7662 1.9327 2.0366 1.7273 4.6459 1.9607 2.8465 1.7375 Normalized Paasche Quantity Index by Relative Population Size 4.7374 1.7423 3.4442 1.6673 4.5042 1.7481 3.1650 1.6167 4.3177 1.8306 3.0174 1.6137 Population Ratio (Comparison/SPN) 2.0749 1.4735 1.4703 1.4391 2.0701 1.4731 1.4707 1.4356 2.0611 1.4717 1.4697 1.4299 Pp/Lp=Pq/Lq 0.9576 0.9143 1.0896 0.9171 0.9450 0.9045 1.0746 0.9271 0.9294 0.9336 1.0600 0.9288 r -0.0222 -0.0933 0.0456 -0.1073 -0.0398 -0.1098 0.0302 -0.0207 -0.0413 -0.0991 0.0741 -0.0143 Vp 0.6367 0.3640 0.3832 0.6963 0.5776 0.3860 0.3487 0.6146 0.5245 0.3573 0.3166 0.5521 Vq 2.9994 2.5251 5.1234 1.1090 2.3900 2.2530 7.0822 5.7195 3.2576 1.8747 2.5596 8.9937 r.Vp.Vq -0.0424 -0.0857 0.0896 -0.0829 -0.0550 -0.0955 0.0746 -0.0729 -0.0706 -0.0664 0.0600 -0.0712

PAGE 114

103 Table A.7 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Spain 2000 2001 2002 Index Measures GR FR UK ITY GR FR UK ITY GR FR UK ITY Number of ATC/Molecule Matching 78 65 64 72 76 63 65 70 74 63 62 69 Laspeyres Price Index (SPN weighted) 1.1759 1.4349 1.2938 1.3770 1.2274 1.4384 1.2964 1.3824 1.1980 1.4492 1.2680 1.3870 Paasche Price Index (Own weighted) 1.0767 1.3567 1.3402 1.2773 1.1141 1.3513 1.3029 1.2768 1.0671 1.3633 1.2643 1.2938 Laspeyres Quantity Index 2.2703 1.3576 1.9679 1.2270 2.3042 1.4704 1.9692 1.2806 2.3495 1.5478 1.8747 1.2126 Paasche Quantity Index 2.0787 1.2836 2.0385 1.1382 2.0916 1.3814 1.9789 1.1827 2.0927 1.4560 1.8693 1.1312 Normalized Laspeyres Quantity Index by Relative Population Size 4.6576 1.9927 2.8891 1.7441 4.6828 2.1433 2.8773 1.8023 4.7283 2.2374 2.7230 1.6872 Normalized Paasche Quantity Index by Relative Population Size 4.2646 1.8840 2.9927 1.6179 4.2508 2.0136 2.8915 1.6645 4.2116 2.1047 2.7151 1.5738 Population Ratio (Comparison/SPN) 2.0515 1.4678 1.4681 1.4215 2.0323 1.4576 1.4612 1.4074 2.0125 1.4456 1.4525 1.3913 Pp/Lp=Pq/Lq 0.9156 0.9455 1.0359 0.9276 0.9077 0.9395 1.0050 0.9236 0.8907 0.9407 0.9971 0.9328 r -0.0539 -0.1530 0.0530 -0.1562 -0.0983 -0.1670 0.0061 -0.1786 -0.1061 -0.1698 -0.0034 -0.1665 Vp 0.5106 0.3418 0.2980 0.4875 0.4550 0.3450 0.3394 0.4594 0.4513 0.3367 0.3591 0.4283 Vq 3.0649 1.0428 2.2700 0.9506 2.0631 1.0508 2.3968 0.9319 2.2818 1.0373 2.3538 0.9425 r.Vp.Vq -0.0844 -0.0545 0.0359 -0.0724 -0.0923 -0.0605 0.0050 -0.0764 -0.1093 -0.0593 -0.0029 -0.0672

PAGE 115

104 Table A.7 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to Spain 2003 Index Measures GR FR UK ITY Number of ATC/Molecule Matching 72 61 61 71 Laspeyres Price Index (SPN weighted) 1.0989 1.4487 1.2392 1.3054 Paasche Price Index (Own weighted) 0.9975 1.3730 1.2339 1.1987 Laspeyres Quantity Index 2.4817 1.6173 1.8352 1.2398 Paasche Quantity Index 2.2528 1.5328 1.8275 1.1384 Normalized Laspeyres Quantity Index by Relative Population Size 4.9164 2.3072 2.6365 1.7057 Normalized Paasche Quantity Index by Relative Population Size 4.4628 2.1867 2.6254 1.5663 Population Ratio (Comparison/SPN) 1.9810 1.4266 1.4366 1.3758 Pp/Lp=Pq/Lq 0.9077 0.9478 0.9958 0.9183 r -0.0840 -0.1388 -0.0009 -0.1599 Vp 0.4386 0.3509 0.3353 0.4292 Vq 2.5050 1.0722 14.448 1.1907 r.Vp.Vq -0.0923 -0.0522 -0.0042 -0.0817

PAGE 116

105 Table A.8 Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Spain 1994 1995 1996 Index Measures GR UK FR ITY GR UK FR ITY GR UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (SPN weighted) 1.2340 1.3125 1.2547 1.2163 1.1609 1.3057 1.2838 1.0890 1.1102 1.3025 1.2958 1.1345 Paasche Price Index (Own weighted) 1.2961 1.2632 1.3138 1.2831 1.2111 1.2342 1.3294 1.1414 1.1427 1.2224 1.3255 1.1800 Laspeyres Quantity Index 2.3180 1.3020 2.3520 1.4253 2.3544 1.3519 2.2150 1.4134 2.4347 1.4448 2.0988 1.3784 Paasche Quantity Index 2.4345 1.2531 2.4627 1.5036 2.4561 1.2779 2.2937 1.4814 2.5060 1.3559 2.1468 1.4337 Normalized Laspeyres Quantity Index by Relative Population Size 4.8040 1.9171 3.4498 2.0643 4.8794 1.9911 3.2514 2.0422 5.0519 2.1286 3.0838 1.9871 Normalized Paasche Quantity Index by Relative Population Size 5.0455 1.8450 3.6121 2.1778 5.0902 1.8821 3.3670 2.1404 5.1998 1.9977 3.1543 2.0669 Population Ratio (Comparison/SPN) 2.0725 1.4724 1.4667 1.4484 2.0725 1.4728 1.4679 1.4449 2.0750 1.4733 1.4693 1.4417 Pp/Lp=Pq/Lq 1.0503 0.9624 1.0470 1.0550 1.0432 0.9453 1.0355 1.0481 1.0293 0.9385 1.0229 1.0402 r 0.0698 -0.1073 0.0976 0.2661 0.0615 -0.1625 0.0726 0.2386 0.0385 -0.1885 0.0424 0.1898 Vp 0.3816 0.3611 0.2755 0.3078 0.4081 0.3790 0.2725 0.2993 0.4400 0.3894 0.2915 0.2988 Vq 1.8860 0.9701 1.7490 0.6710 1.7209 0.8893 1.7967 0.6733 1.7273 0.8378 1.8489 0.7081 r.Vp.Vq 0.0503 -0.0376 0.0470 0.0550 0.0432 -0.0547 0.0355 0.0481 0.0293 -0.0615 0.0229 0.0402

PAGE 117

106 Table A.8 (Continued) Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Spain 1997 1998 1999 Index Measures GR UK FR ITY GR UK FR ITY GR UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (SPN weighted) 1.0546 1.2847 1.2949 1.1479 1.0275 1.3251 1.2857 1.1662 1.0300 1.3596 1.2707 1.2059 Paasche Price Index (Own weighted) 1.0562 1.2171 1.3064 1.1724 1.0209 1.2331 1.2730 1.1778 1.0060 1.2802 1.2474 1.2099 Laspeyres Quantity Index 2.3727 1.4794 2.0031 1.2827 2.2814 1.5418 1.9653 1.2414 2.2590 1.5841 1.8558 1.2458 Paasche Quantity Index 2.3762 1.4015 2.0208 1.3101 2.2667 1.4347 1.9459 1.2537 2.2064 1.4916 1.8218 1.2500 Normalized Laspeyres Quantity Index by Relative Population Size 4.9231 2.1799 2.9452 1.8459 4.7226 2.2713 2.8904 1.6909 4.6561 2.3314 2.7274 1.7814 Normalized Paasche Quantity Index by Relative Population Size 4.9305 2.0651 2.9713 1.8853 4.6922 2.1136 2.8618 1.7999 4.5476 2.1953 2.6775 1.7873 Population Ratio (Comparison/SPN) 2.0749 1.4735 1.4703 1.4391 2.0701 1.4731 1.4707 1.4356 2.0611 1.4717 1.4697 1.4299 Pp/Lp=Pq/Lq 1.0015 0.9473 1.0088 1.0213 0.9936 0.9306 0.9901 1.0099 0.9767 0.9416 0.9817 1.0033 r 0.0018 -0.1654 0.0152 0.0938 -0.0071 -0.1966 -0.0164 0.0434 -0.0247 -0.1685 -0.0303 0.0140 Vp 0.4710 0.3857 0.3009 0.2983 0.4592 0.4211 0.2875 0.2782 0.4347 0.3951 0.2702 0.2646 Vq 1.8240 0.8252 1.9343 0.7624 1.9879 0.8386 2.0952 0.8211 2.1714 0.8770 2.2367 0.9004 r.Vp.Vq 0.0015 -0.0527 0.0088 0.0213 -0.0064 -0.0694 -0.0099 0.0099 -0.0233 -0.0584 -0.0183 0.0033

PAGE 118

107 Table A.8 (Continued) Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Spain 2000 2001 2002 Index Measures GR UK FR ITY GR UK FR ITY GR UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 38 38 38 Laspeyres Price Index (SPN weighted) 1.0231 1.3745 1.2870 1.2717 1.0922 1.3833 1.3311 1.3591 1.0643 1.4404 1.3229 1.3533 Paasche Price Index (Own weighted) 0.9884 1.3077 1.2444 1.2639 1.0380 1.2974 1.2440 1.3193 0.9909 1.3461 1.2277 1.3147 Laspeyres Quantity Index 2.2897 1.5810 1.8399 1.2399 2.3449 1.7193 1.8959 1.2854 2.4456 1.7984 1.8710 1.2391 Paasche Quantity Index 2.2120 1.5041 1.7790 1.2323 2.2286 1.6126 1.7719 1.2477 2.2771 1.6806 1.7363 1.2038 Normalized Laspeyres Quantity Index by Relative Population Size 4.6974 2.3206 2.7011 1.7625 4.7655 2.5061 2.7703 1.8091 4.9217 2.5997 2.7176 1.7240 Normalized Paasche Quantity Index by Relative Population Size 4.5381 2.2078 2.6117 1.7517 4.5292 2.3505 2.5890 1.7560 4.5826 2.4295 2.5220 1.6748 Population Ratio (Comparison/SPN) 2.0515 1.4678 1.4681 1.4215 2.0323 1.4576 1.4612 1.4074 2.0125 1.4456 1.4525 1.3913 Pp/Lp=Pq/Lq 0.9661 0.9514 0.9669 0.9939 0.9504 0.9379 0.9346 0.9707 0.9311 0.9345 0.9280 0.9715 r -0.0337 -0.1378 -0.0520 -0.0264 -0.0552 -0.1665 -0.0847 -0.1164 -0.0715 -0.1786 -0.0811 -0.0964 Vp 0.4524 0.3770 0.2747 0.2510 0.3975 0.3795 0.3389 0.2787 0.4103 0.3652 0.3660 0.3137 Vq 2.2262 0.9358 2.3193 0.9243 2.2585 0.9828 2.2785 0.9036 2.3503 1.0042 2.4245 0.9442 r.Vp.Vq -0.0339 -0.0486 -0.0331 -0.0061 -0.0496 -0.0621 -0.0654 -0.0293 -0.0689 -0.0655 -0.0720 -0.0285

PAGE 119

108 Table A.8 (Continued) Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to Spain 2003 Index Measures GR UK FR ITY Number of ATC/Molecule Matching 38 38 38 38 Laspeyres Price Index (SPN weighted) 0.9443 1.4832 1.3183 1.3119 Paasche Price Index (Own weighted) 0.9107 1.3923 1.2291 1.2507 Laspeyres Quantity Index 2.7505 1.8956 1.8777 1.2621 Paasche Quantity Index 2.6528 1.7794 1.7506 1.2032 Normalized Laspeyres Quantity Index by Relative Population Size 5.4488 2.7043 2.6976 1.7364 Normalized Paasche Quantity Index by Relative Population Size 5.2552 2.5385 2.5150 1.6553 Population Ratio (Comparison/SPN) 1.9810 1.4266 1.4366 1.3758 Pp/Lp=Pq/Lq 0.9645 0.9387 0.9323 0.9533 r -0.0388 -0.1616 -0.0118 -0.1207 Vp 0.3891 0.3522 0.3369 0.3517 Vq 2.3560 1.0771 17.0891 1.1002 r.Vp.Vq -0.0355 -0.0613 -0.0677 -0.0467

PAGE 120

109 Table A.9 Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Germany Number of ATC/Molecule Matching 156 152 148 144 143 141 138 135 130 Laspeyres Price Index 0.9742 0.9582 0.9223 0.8931 0.8842 0.8694 0.8691 0.8358 0.7993 Laspeyres Quantity Index 1.0842 1.1919 1.2274 1.2385 1.2652 1.3397 1.4593 1.6497 1.8903 Paasche Price Index 0.9695 0.9406 0.8909 0.8518 0.8330 0.7997 0.7951 0.7405 0.6853 Paasche Quantity Index 1.0790 1.1700 1.1857 1.1812 1.1919 1.2322 1.3351 1.4615 1.6207 United Kingdom Number of ATC/Molecule Matching 93 93 92 92 90 87 82 81 81 Laspeyres Price Index 1.0072 1.0075 1.0047 1.0217 1.0819 1.1046 1.0319 1.0225 1.0333 Laspeyres Quantity Index 1.1164 1.2987 1.4457 1.5747 1.7187 1.8432 2.1377 2.4391 2.6650 Paasche Price Index 1.0071 1.0070 1.0097 1.0268 1.0695 1.0749 1.0293 1.0357 1.0485 Paasche Quantity Index 1.1162 1.2981 1.4529 1.5826 1.6990 1.7935 2.1324 2.4705 2.7041 France Number of ATC/Molecule Matching 101 99 97 95 92 89 88 85 84 Laspeyres Price Index 1.0221 1.0346 1.0348 1.0306 1.0235 1.0049 0.9910 0.9711 0.9485 Laspeyres Quantity Index 1.0154 1.0206 1.0437 1.0357 1.0197 1.0495 1.0909 1.1269 1.1445 Paasche Price Index 1.0194 1.0248 1.0168 1.0029 0.9849 0.9653 0.9571 0.9405 0.9338 Paasche Quantity Index 1.0126 1.0109 1.0256 1.0078 0.9813 1.0082 1.0536 1.0913 1.1268 Italy Number of ATC/Molecule Matching 116 111 104 101 99 97 95 91 88 Laspeyres Price Index 0.9015 0.9497 0.9686 0.9861 1.0162 1.0378 1.0479 1.0274 0.9574 Laspeyres Quantity Index 1.0667 1.1160 1.1118 1.1172 1.1658 1.2233 1.3335 1.3837 1.4648 Paasche Price Index 0.8997 0.9422 0.9511 0.9626 0.9874 1.0067 1.0162 1.0031 0.9387 Paasche Quantity Index 1.0646 1.1072 1.0917 1.0905 1.1326 1.1867 1.2931 1.3510 1.4362

PAGE 121

110 Table A.9 (Continued) Pharmaceutical Price and Quantity Indexes for All Molecules, Relative to 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Spain Number of ATC/Molecule Matching 100 98 95 94 93 88 86 84 81 Laspeyres Price Index 1.0161 1.0328 1.0516 1.0627 1.0619 1.0385 1.0023 0.9886 0.9924 Laspeyres Quantity Index 1.0950 1.1744 1.2461 1.2762 1.3187 1.3815 1.4652 1.5697 1.6227 Paasche Price Index 1.0145 1.0258 1.0345 1.0342 1.0719 0.9986 0.9632 0.9481 0.9407 Paasche Quantity Index 1.0933 1.1665 1.2258 1.2420 1.3312 1.3284 1.4080 1.5053 1.5383

PAGE 122

111 Table A.10 Pharmaceutical Price and Quantity Indexes for Global Molecules, Relative to 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Number of ATC/Molecule Matching 38 38 38 38 38 38 38 38 38 Germany Laspeyres Price Index 0.9609 0.9345 0.8837 0.8549 0.8454 0.8167 0.8084 0.7622 0.7020 Laspeyres Quantity Index 1.0920 1.2154 1.2801 1.2837 1.3281 1.4367 1.6112 1.8680 2.1973 Paasche Price Index 0.9538 0.9104 0.8468 0.8173 0.8070 0.7696 0.7618 0.6967 0.6300 Paasche Quantity Index 1.0840 1.1840 1.2267 1.2272 1.2677 1.3539 1.5184 1.7076 1.9721 United Kingdom Laspeyres Price Index 1.0064 1.0075 1.0033 1.0160 1.0995 1.0713 0.9986 0.9876 0.9980 Laspeyres Quantity Index 1.1067 1.2744 1.4106 1.5281 1.6626 1.7569 2.0333 2.3407 2.5819 Paasche Price Index 1.0061 1.0070 1.0086 1.0247 1.0932 1.0504 1.0068 1.0146 1.0319 Paasche Quantity Index 1.1064 1.2738 1.4181 1.5412 1.6531 1.7226 2.0501 2.4048 2.6696 France Laspeyres Price Index 1.0282 1.0453 1.0501 1.0455 1.0360 1.0242 1.0208 1.0068 0.9892 Laspeyres Quantity Index 1.0199 1.0388 1.0645 1.0748 1.0550 1.0980 1.1625 1.2273 1.2693 Paasche Price Index 1.0255 1.0358 1.0338 1.0204 1.0061 0.9905 0.9910 0.9773 0.9769 Paasche Quantity Index 1.0172 1.0294 1.0481 1.0490 1.0245 1.0619 1.1286 1.1913 1.2536 Italy Laspeyres Price Index 0.9095 0.9602 0.9814 1.0025 1.0380 1.0614 1.0724 1.0427 0.9749 Laspeyres Quantity Index 1.0676 1.1188 1.1151 1.1149 1.1671 1.2357 1.3554 1.4095 1.4925 Paasche Price Index 0.9068 0.9500 0.9580 0.9713 0.9979 1.0179 1.0326 1.0197 0.9601 Paasche Quantity Index 1.0643 1.1069 1.0885 1.0802 1.1220 1.1851 1.3051 1.3783 1.4699 Spain Laspeyres Price Index 1.0186 1.0372 1.0574 1.0715 1.0792 1.0578 1.0114 0.9969 0.9928 Laspeyres Quantity Index 1.0795 1.1618 1.2523 1.3038 1.3664 1.4528 1.5588 1.6939 1.7708 Paasche Price Index 1.0167 1.0287 1.0377 1.0389 1.0369 1.0104 0.9683 0.9525 0.9376 Paasche Quantity Index 1.0774 1.1523 1.2289 1.2641 1.3129 1.3877 1.4924 1.6185 1.6724

PAGE 123

112 Table A.11 Pharmaceutical Drug Quality and Market (Competiti on) Characteristics for All Molecules by Country Country Germany France United Kingdom Variable N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD Suthnds 940 100,061187,30067,855 700 97,023 104,881 26,537 710 67,201 126,89242,091 Leuthnds 940 21,105 33,634 13,395 700 26,278 36,961 12,509 710 17,958 42,364 21,945 Leusuprice 940 0.58 1.79 0.13 700 0.27 0.22 0.02 710 0.54 1.93 0.98 Quality Characteristics Strengthg 940 0.15 0.55 0.03 700 0.133 0.48 0.01 710 0.12 0.47 0.13 Molage 940 20.77 12.65 2.87 700 18.9 10.2 2.87 710 21.23 15.28 2.87 Packsize 940 88.42 27.74 9.25 700 32.26 9.57 3.54 710 88.39 130.71 80.58 Formcode 940 16.66 19.1 4.32 700 3.27 2.98 1.18 710 8.56 9.57 2.72 Globpenet 940 3.79 1.42 0.32 700 4.24 1.15 0.24 710 4.30 1.10 0.29 Market(Competition) Characteristics Gencompet 940 9.86 12.03 4.23 700 2.80 3.16 1.84 710 2.42 2.81 0.99 Thsubsmol 940 21.05 7.64 1.50 700 13.94 4.34 1.16 710 13.51 3.82 1.33 Thsubsolentlag 930 20.46 15.45 0.00 700 15.5 9.88 0.00 710 14.97 10.69 0.00

PAGE 124

113 Table 11 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics by Country Country Italy Spain Variable N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD Suthnds 750 50,636 72,026 19,584 690 43,078 55,333 16,430 Leuthnds 750 15,267 26,607 8,800 690 10,272 17,515 6,619 Leusuprice 750 0.30 0.30 0.05 690 0.21 0.17 0.02 Quality Characteristics Strengthg 750 0.14 0.44 0.05 690 0.21 0.76 0.01 Molage 750 18.97 11.04 2.87 690 17.85 10.33 2.87 Packsize 750 27.89 9.93 2.01 690 40.13 15.09 2.02 Formcode 750 3.49 3.16 1.26 690 3.51 3.58 0.93 Globpenet 750 4.18 1.15 0.18 690 4.43 0.91 0.28 Market(Competition) Characteristics Gencompet 750 3.46 3.78 1.28 690 3.82 5.06 2.29 Thsubsmol 750 15.40 4.72 1.52 690 14.16 5.64 1.21 Thsubsolentlag 750 19.21 12.65 0.00 690 17.39 11.45 0.00

PAGE 125

114 Table A.12 Pharmaceutical Drug Quality and Market (Competiti on) Characteristics for Global Molecules by Country Country Germany France United Kingdom Variable N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD Suthnds 380 176,624 232,607103,012380123,465105,591 30,273380100,552 139,457 48,511 Leuthnds 380 35,317 42,138 19,520 38035,337 44,564 16,44438029,019 53,787 28,876 Leusuprice 380 0.29 0.22 0.05 3800.27 0.19 0.02 3800.32 0.25 0.05 Quality Characteristics Strengthg 380 0.13 0.32 0.05 3800.19 0.64 0.00 3800.18 0.64 0.17 Molage 380 21.66 11.81 2.88 38020.13 10.97 2.88 38021.13 12.28 2.88 Packsize 380 88.12 15.69 2.23 38032.84 8.65 1.95 380102.05 149.72 92.66 Formcode 380 25.79 24.26 6.14 3804.20 3.55 1.53 38011.97 11.54 3.43 Globpenet 380 5.00 0.00 0.00 3805.00 0.00 0.00 3805.00 0.00 0.00 Market(Competition) Characteristics Gencompet 380 16.81 14.90 5.77 3803.71 3.81 2.25 3803.19 3.59 1.28 Thsubsmol 380 19.16 6.85 1.44 38013.63 4.49 1.11 38012.91 3.71 1.24 Thsubsolentlag 380 14.08 11.57 0.00 38011.95 8.63 0.00 38012.97 8.90 0.00

PAGE 126

115 Table A.12 (Continued) Pharmaceutical Drug Quality and Mark et (Competition) Characteristics for Global Molecules by Country Country Italy Spain Variable N Overall Mean Overall SD Within SD N Overall Mean Overall SD Within SD Suthnds 380 78,600 87,262 24,399 380 59,079 63,228 20,453 Leuthnds 380 22,462 32,826 11,203 380 14,078 21,154 8,419 Leusuprice 380 0.28 0.24 0.05 380 0.21 0.17 0.02 Quality Characteristics Strengthg 380 0.19 0.60 0.07 380 0.20 0.64 0.01 Molage 380 20.32 11.47 2.88 380 19.68 11.30 2.88 Packsize 380 27.04 9.77 2.58 380 40.33 11.26 1.74 Formcode 380 3.77 2.96 1.36 380 4.11 4.37 1.13 Globpenet 380 5.00 0.00 0.00 380 5.00 0.00 0.00 Market(Competition) Characteristics Gencompet 380 3.95 3.94 1.57 380 4.63 6.21 2.71 Thsubsmol 380 14.48 3.74 1.43 380 13.44 5.40 1.14 Thsubsolentlag 380 14.61 10.09 0.00 380 14.05 10.31 0.00

PAGE 127

116 Table A.13 Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: Germany Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 91,341 94,200 97,565 96,550 94,591 94,139 97,388 102,283 110,876 121,675 St.Dev. 197,718 192,491 189,559 179,558 171,411 166,979 171,646 179,957 199,644 223,877 Suthnds N 94 94 94 94 94 94 94 94 94 94 Mean 0.5818 0.5809 0.5805 0.5719 0.5678 0.5680 0.5805 0.5963 0.5888 0.5838 St.Dev. 1.7368 1.7368 1.7254 1.7007 1.7019 1.7015 1.8159 1.9337 1.9505 1.9648 Leusuprice N 94 94 94 94 94 94 94 94 94 94 Mean 18,819 19,838 21,193 20,701 19,989 20,007 20,372 22,102 23,299 24,727 St.Dev. 30,386 29,658 30,389 29,043 28, 792 29,638 31,483 36,453 41,151 45,850 Leuthnds N 94 94 94 94 94 94 94 94 94 94 Quality Characteristics Mean 0.1418 0.1439 0.1456 0.1474 0.1488 0.1506 0.1520 0.1533 0.1545 0.1562 St.Dev. 0.5390 0.5417 0.5441 0.5473 0.5503 0.5549 0.5587 0.5626 0.5667 0.5723 Strengthg N 94 94 94 94 94 94 94 94 94 94 Mean 16.2660 17.2660 18.2660 19.2660 20.2660 21.2660 22.2660 23.2660 24.2660 25.2660 St.Dev. 12.3812 12.3812 12.3812 12.3812 12.3812 12.3812 12.3812 12.3812 12.3812 12.3812 Molage N 94 94 94 94 94 94 94 94 94 94 Mean 88.6616 86.9820 86.9964 87.6721 88.8467 88.9582 88.7587 88.8862 89.1031 89.3121 St.Dev. 42.4639 33.0368 30.5640 27.2622 25.0869 23.8694 22.9830 22.4613 21.9649 21.8510 Packsize N 94 94 94 94 94 94 94 94 94 94 Mean 16.5000 16.0426 16.2553 16.2447 16.4894 16.7340 16.7128 17.2340 17.3085 17.0745 St.Dev. 19.028 18.5507 19.1928 18.5829 18.7093 18.8965 18.6689 19.5879 20.4135 20.1015 Formcode N 94 94 94 94 94 94 94 94 94 94 Mean 3.8192 3.8192 3.8617 3.8404 3.8085 3.7872 3.7660 3.7553 3.7340 3.7234 St.Dev. 1.3675 1.3675 1.3804 1.4165 1.4239 1.4208 1.4400 1.4568 1.4678 1.5127 Globpenet N 94 94 94 94 94 94 94 94 94 94

PAGE 128

117 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition ) Characteristics for Molecules by Country by Year Country: Germany Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 6.8618 7.5852 8.2128 8.7660 9.9681 10.3936 11.0745 11.3936 11.7660 12.5851 St.Dev. 9.0739 9.8771 10.4572 11.4351 12.3432 12.4721 12.6807 12.7290 13.50548 13.9588 Gencompet N 94 94 94 94 94 94 94 94 94 94 Mean 21.7979 21.4894 21.3936 21.6170 21.2447 21.2234 20.8404 20.7660 20.6915 19.4575 St.Dev. 8.8956 8.7838 8.8076 7.6074 7.2047 6.9207 6.6661 6.9645 7.5020 6.6812 Thsubsmol N 94 94 94 94 94 94 94 94 94 94 Mean 20.4624 20.4624 20.4624 20.4624 20.4624 20.4624 20.4624 20.4624 20.4624 20.4624 St.Dev. 15.5295 15.5295 15.5295 15.5295 15.5295 15.5295 15.5295 15.5295 15.5295 15.5295 Thsubsmolentrylag N 93 93 93 93 93 93 93 93 93 93

PAGE 129

118 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: France Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 101,660 100,599 97,956 98,079 95,477 93,620 94,809 96,060 96,284 95,787 St.Dev. 105,062 102,405 98,983 101,014 101,152 102,217 105,804 109,388 113,001 115,067 Suthnds N 70 70 70 70 70 70 70 70 70 70 Mean 0.2666 0.2706 0.2740 0.2741 0.2735 0.2742 0.2727 0.2715 0.2691 0.2657 St.Dev. 0.2267 0.2245 0.2250 0.2248 0.2240 0.2242 0.2249 0.2255 0.2231 0.2259 Leusuprice N 70 70 70 70 70 70 70 70 70 70 Mean 25,201 26,121 26,439 26,840 26,327 25,483 25,684 26,524 26,957 27,201 St.Dev. 28,465 29,733 31,465 33,227 34, 938 34,002 36,017 40,896 45,616 51,180 Leuthnds N 70 70 70 70 70 70 70 70 70 70 Quality Characteristics Mean 0.1324 0.1326 0.1329 0.1330 0.1333 0.1332 0.1335 0.1336 0.1337 0.1324 St.Dev. 0.4812 0.4813 0.4814 0.4814 0.4814 0.4813 0.4814 0.4813 0.4814 0.4809 Strengthg N 70 70 70 70 70 70 70 70 70 70 Mean 14.4000 15.4000 16.4000 17.4000 18.4000 19.4000 20.4000 21.4000 22.4000 23.4000 St.Dev. 9.8487 9.8487 9.8487 9.8487 9.8487 9.8487 9.8487 9.8487 9.8487 9.8487 Molage N 70 70 70 70 70 70 70 70 70 70 Mean 31.8639 31.8353 31.9500 32.0411 32.1769 32.0760 32.3648 32.6652 32.7235 32.9488 St.Dev. 9.1641 8.9247 9.0061 9.1370 9.0886 9.5272 9.8104 10.1974 10.4168 10.7817 Packsize N 70 70 70 70 70 70 70 70 70 70 Mean 2.7714 2.6857 2.8714 3.1143 3.3000 3.5286 3.7143 3.6286 3.5429 3.5857 St.Dev. 2.1814 2.0039 2.3277 2.7586 2.9750 3.4963 3.7500 3.5105 3.0960 3.1091 Formcode N 70 70 70 70 70 70 70 70 70 70 Mean 4.3000 4.2857 4.3000 4.2857 4.2857 4.2429 4.2286 4.2000 4.1571 4.1571 St.Dev. 1.0948 1.1183 1.1209 1.1183 1.1183 1.1602 1.1443 1.1869 1.2469 1.2469 Globpenet N 70 70 70 70 70 70 70 70 70 70

PAGE 130

119 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: France Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 1.9143 1.8714 2.0857 2.3143 2.6429 2.8143 3.1143 3.3429 3.8000 4.1286 St.Dev. 1.4915 1.39275 1.7590 2.1839 2.7295 2.9747 3.4538 3.7836 4.4055 4.6685 Gencompet N 70 70 70 70 70 70 70 70 70 70 Mean 14.8143 14.0857 14.3143 14.3000 14.2571 13.9857 13.6000 13.6143 13.2714 13.1286 St.Dev. 5.2483 4.6054 4.2683 4.5248 4.4516 4.5028 4.3218 3.9647 3.6153 3.6828 Thsubsmol N 70 70 70 70 70 70 70 70 70 70 Mean 15.5 15.5 15.5 15.5 15.5 15.5 15.5 15.5 15.5 15.5 St.Dev. 9.9488 9.9488 9.9488 9.9488 9.9488 9.9488 9.9488 9.9488 9.9488 9.9488 Thsubsmolentrylag N 70 70 70 70 70 70 70 70 70 70

PAGE 131

120 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: Italy Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 46,682 48,395 50,129 49,193 48,513 49,447 50,817 53,763 53,815 55,612 St.Dev. 66,011 66,638 69,011 67,751 67, 963 70,521 73,589 78,333 79,016 83,040 Suthnds N 75 75 75 75 75 75 75 75 75 75 Mean 0.2894 0.2646 0.2800 0.2891 0.2952 0.3042 0.3100 0.3125 0.3126 0.2984 St.Dev. 0.2945 0.2611 0.2823 0.3007 0.3024 0.3087 0.3159 0.3179 0.3178 0.3102 Leusuprice N 75 75 75 75 75 75 75 75 75 75 Mean 13,075 12,542 13,775 13,861 14,094 15,090 16,159 17,795 18,233 18,049 St.Dev. 20,816 19,397 21,791 22,168 22, 830 25,104 27,843 31,874 34,237 35,113 Leuthnds N 75 75 75 75 75 75 75 75 75 75 Quality Characteristics Mean 0.1322 0.1329 0.1334 0.1346 0.1476 0.1483 0.1484 0.1481 0.1485 0.1486 St.Dev. 0.3965 0.3987 0.3876 0.3784 0.4751 0.4799 0.4798 0.4797 0.4798 0.4798 Strengthg N 75 75 75 75 75 75 75 75 75 75 Mean 14.4667 15.4667 16.4667 17.4667 18.4667 19.4667 20.4667 21.4667 22.4667 23.4667 St.Dev. 10.7242 10.7242 10.7242 10.7242 10.7242 10.7242 10.7242 10.7242 10.7242 10.7242 Molage N 75 75 75 75 75 75 75 75 75 75 Mean 27.9435 27.9282 27.9028 27.7540 27.7421 27.8486 27.9149 28.0924 27.9160 27.8151 St.Dev. 9.9150 9.8735 9.9394 10.0877 10.0829 9.9887 9.9177 10.0440 10.0108 10.0791 Packsize N 75 75 75 75 75 75 75 75 75 75 Mean 3.4933 3.4533 3.4400 3.4133 3.3333 3.3733 3.5067 3.6000 3.61330 3.6800 St.Dev. 3.4656 3.2768 3.0369 2.9459 2.7427 2.7202 3.0330 3.4011 3.4363 3.5723 Formcode N 75 75 75 75 75 75 75 75 75 75 Mean 4.1867 4.1867 4.21337 4.2000 4.2133 4.1867 4.1600 4.1600 4.1467 4.1467 St.Dev. 1.1472 1.1472 1.1306 1.1508 1.1542 1.1589 1.1745 1.1745 1.1705 1.1705 Globpenet N 75 75 75 75 75 75 75 75 75 75

PAGE 132

121 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: Italy Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market(Competition) Characteristics Mean 3.2400 3.3200 3.3333 3.2533 3.2267 3.2533 3.4933 3.6400 3.85330 4.000 St.Dev. 3.2708 3.3978 3.4222 3.4371 3.3715 3.4410 3.8882 4.2542 4.4591 4.6702 Gencompet N 75 75 75 75 75 75 75 75 75 75 Mean 16.5467 16.4000 15.9733 15.3467 15.4800 15.4533 15.3867 15.0133 14.3333 14.0933 St.Dev. 5.8663 5.7657 4.8601 4.2569 4.4032 4.4700 4.6438 4.0520 4.1275 3.9837 Thsubsmol N 75 75 75 75 75 75 75 75 75 75 Mean 19.2133 19.2133 19.2133 19.2133 19.2133 19.2133 19.2133 19.2133 19.2133 19.2133 St.Dev. 12.7272 12.7272 12.7272 12.7272 12.7272 12.7272 12.7272 12.7272 12.7272 12.7272 Thsubsmolentrylag N 75 75 75 75 75 75 75 75 75 75

PAGE 133

122 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: Spain Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 38,772 40,293 41,373 42,429 42,390 42,742 43,766 45,188 46,565 47,256 St.Dev. 51,009 51,949 52,499 53,422 53, 697 54,523 56,228 58,283 60,739 62,791 Suthnds N 69 69 69 69 69 69 69 69 69 69 Mean 0.2072 0.2084 0.2099 0.2120 0.2135 0.2134 0.2087 0.2069 0.2053 0.2057 St.Dev. 0.1698 0.1692 0.1692 0.1688 0.1676 0.1669 0.1641 0.1658 0.1661 0.1690 Leusuprice N 69 69 69 69 69 69 69 69 69 69 Mean 7,822 8,691 9,436 10,112 10,368 10,641 10,847 11,098 11,704 12,004 St.Dev. 12,392 13,607 14,684 15,781 16, 501 17,375 18,189 19,187 21,906 23,155 Leuthnds N 69 69 69 69 69 69 69 69 69 69 Quality Characteristics Mean 0.2072 0.2075 0.2082 0.2099 0.2105 0.2109 0.2112 0.2115 0.2111 0.2122 St.Dev. 0.7619 0.7620 0.7622 0.7629 0.7631 0.7633 0.7635 0.7636 0.7637 0.7638 Strengthg N 69 69 69 69 69 69 69 69 69 69 Mean 13.3478 14.3478 15.3478 16.3478 17.3478 18.3478 19.3478 20.3478 21.3478 22.3478 St.Dev. 9.9836 9.9836 9.9836 9.9836 9.9836 9.9836 9.9836 9.9836 9.9836 9.9836 Molage N 69 69 69 69 69 69 69 69 69 69 Mean 40.6834 40.6355 40.6007 40.1174 39.9884 39.8821 39.8462 39.8963 39.8558 39.7986 St.Dev. 15.1235 15.1269 15.3012 15.1153 15.0211 15.1568 15.1621 15.2691 15.2716 15.2803 Packsize N 69 69 69 69 69 69 69 69 69 69 Mean 3.3478 3.3768 3.4058 3.4493 3.4203 3.4783 3.6087 3.6957 3.6522 3.6232 St.Dev. 3.1381 3.3082 3.3752 3.5253 3.4145 3.4878 3.9266 3.9790 3.9064 3.8810 Formcode N 69 69 69 69 69 69 69 69 69 69 Mean 4.4203 4.4493 4.4928 4.4783 4.4493 4.4203 4.4203 4.4058 4.3768 4.3768 St.Dev. 0.8644 0.8666 0.8335 0.8679 0.9161 0.9299 0.9300 0.9750 0.9717 0.9717 Globpenet N 69 69 69 69 69 69 69 69 69 69

PAGE 134

123 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: Spain Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 3.0580 3.1594 3.1739 3.3043 3.4493 3.6087 3.9565 4.3913 4.9420 5.1304 St.Dev. 3.1243 3.3678 3.3779 3.6312 3.9090 4.4465 5.2087 6.2006 7.0521 7.6233 Gencompet N 69 69 69 69 69 69 69 69 69 69 Mean 15.0435 14.3333 14.4348 14.1884 14.8696 14.4348 13.9855 13.7971 13.5507 12.9565 St.Dev. 6.3765 5.8251 6.1822 5.9267 6.0729 5.8648 5.6944 5.0631 4.9809 4.0960 Thsubsmol N 69 69 69 69 69 69 69 69 69 69 Mean 17.3913 17.3913 17.3913 17.3913 17.3913 17.3913 17.3913 17.3913 17.3913 17.3913 St.Dev. 11.5214 11.5214 11.5214 11.5214 11.5214 11.5214 11.5214 11.5214 11.5214 11.5214 Thsubsmolentrylag N 69 69 69 69 69 69 69 69 69 69

PAGE 135

124 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: United Kingdom Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 50,900 53,770 58,921 61,324 64,127 67,759 68,702 76,668 82,708 87,136 St.Dev. 100,046 102,315 109,760 112,259 116,617 124,097 125,245 142,848 154,858 166,074 Suthnds N 71 71 71 71 71 71 71 71 71 71 Mean 0.4178 0.5119 0.8223 0.5085 0.5925 0.4795 0.4738 0.5508 0.5590 0.5078 St.Dev. .9723 1.724 4.3158 1.6306 2.1351 0.9227 0.9210 1.5446 1.5928 1.1460 Leusuprice N 71 71 71 71 71 71 71 71 71 71 Mean 9,984 11,258 13,124 14,676 16,274 18,514 19,968 22,168 25,455 28,164 St.Dev. 21,531 23,474 26,867 30,259 33, 912 37,465 40,047 47,892 60,288 71,467 Leuthnds N 71 71 71 71 71 71 71 71 71 71 Quality Characteristics Mean 0.1605 0.1125 0.1139 0.1132 0.1131 0.1129 0.1132 0.1114 0.1086 0.1079 St.Dev. 0.6221 0.4768 0.4768 .4760 0.4704 0.4702 0.4673 0.4449 0.4192 0.4115 Strengthg N 71 71 71 71 71 71 71 71 71 71 Mean 16.7324 17.7324 18.7324 19.7324 20.7324 21.7324 22.7324 23.7324 24.7324 25.7324 St.Dev. 15.1072 15.1072 15.1072 15.1072 15.1072 15.1072 15.1072 15.1072 15.1072 15.1072 Molage N 71 71 71 71 71 71 71 71 71 71 Mean 124.6009 125.0463 122.1820 109.1184 102.8499 81.8099 58.1690 54.9476 53.4879 51.6353 St.Dev. 176.9906 179.8889 180.3949 155.8690 153.4981 106.6297 53.1618 46.4842 43.1163 39.3169 Packsize N 71 71 71 71 71 71 71 71 71 71 Mean 7.2676 7.4366 7.7887 8.3521 8.8451 9.3662 9.4085 9.2394 8.9155 9.0141 St.Dev. 7.7367 7.9977 8.5522 9.5620 10.0933 11.1153 10.5919 10.2378 9.8282 9.6651 Formcode N 71 71 71 71 71 71 71 71 71 71 Mean 4.3239 4.3239 4.3521 4.3380 4.3380 4.3099 4.2817 4.2535 4.2254 4.2254 St.Dev. 0.9968 1.0250 1.0433 1.0683 1.0683 1.1031 1.1487 1.2038 1.1976 1.1976 Globpenet N 71 71 71 71 71 71 71 71 71 71

PAGE 136

125 Table A.13 (Continued) Pharmaceutical Drug Quality and Market (Competition) Characteristics for All Mole cules by Country by Year Country: United Kingdom Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market(Competition) Characteristics Mean 1.9155 2.0423 2.1127 2.3099 2.3944 2.4366 2.5211 2.7887 2.7887 2.8451 St.Dev. 1.8107 2.0663 2.3758 2.8009 3.1052 2.9215 2.9074 3.4555 3.1934 3.0503 Gencompet N 71 71 71 71 71 71 71 71 71 71 Mean 13.3802 13.3803 13.7324 14.0282 14.1690 14.0282 13.5775 12.9155 12.7465 13.0986 St.Dev. 4.5556 4.5556 4.3588 4.2392 4.0602 3.6877 3.2145 3.1018 2.9506 2.8593 Thsubsmol N 71 71 71 71 71 71 71 71 71 71 Mean 14.9718 14.9718 14.9718 14.9718 14.9718 14.9718 14.9718 14.9718 14.9718 14.9718 St.Dev. 10.7597 10.7597 10.7597 10.7597 10.7597 10.7597 10.7597 10.7597 10.7597 10.7597 Thsubsmolentrylag N 71 71 71 71 71 71 71 71 71 71

PAGE 137

126 Table A.14 Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Germany Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 154,306 160,153 167,148 168,119 162,890 164,158 172,419 183,792 204,108 229,144 St.Dev. 245,310 238,622 233,069 224,878 207,050 205,637 211,184 220,526 252,621 291,580 Suthnds N 38 38 38 38 38 38 38 38 38 38 Mean 0.3080 0.3032 0.3023 0.2981 0.2940 0.2944 0.2913 0.2893 0.2745 0.2519 St.Dev. 0.1985 0.1949 0.1965 0.2061 0.2141 0.2209 0.2333 0.2408 0.2461 0.2188 Leusuprice N 38 38 38 38 38 38 38 38 38 38 Mean 31,065 32,358 34,381 33,675 32,621 33,295 34,350 37,988 40,430 43,005 St.Dev. 36,526 35,633 36,125 35,027 34, 550 36,720 39,544 46,870 53,988 60,798 Leuthnds N 38 38 38 38 38 38 38 38 38 38 Quality Characteristics Mean 0.1168 0.1215 0.1245 0.1282 0.1315 0.1357 0.1391 0.1422 0.1454 0.1494 St.Dev. 0.2544 0.2683 0.2797 0.2949 0.3087 0.3288 0.3448 0.3605 0.3767 0.3974 Strengthg N 38 38 38 38 38 38 38 38 38 38 Mean 17.1579 18.1579 19.1579 20.1579 21.1579 22.1579 23.1579 24.1579 25.1579 26.1579 St.Dev. 11.5981 11.5981 11.5981 11.5981 11.5981 11.5981 11.5981 11.5981 11.5981 11.5981 Molage N 38 38 38 38 38 38 38 38 38 38 Mean 85.5251 86.0092 86.2022 87.3814 88.9049 89.0981 88.9529 89.1485 89.7112 90.3038 St.Dev. 15.9256 15.7505 15.7250 15.7112 15.7485 15.7806 15.8148 15.8579 15.8178 15.8275 Packsize N 38 38 38 38 38 38 38 38 38 38 Mean 24.9737 24.6316 25.2105 24.9211 25.0789 26.0263 26.1053 27.1579 27.1842 26.5789 St.Dev. 24.2359 23.7889 24.8891 23.8864 24.2814 24.3160 23.8234 24.8624 25.9791 25.2315 Formcode N 38 38 38 38 38 38 38 38 38 38 Mean 5 5 5 5 5 5 5 5 5 5 St.Dev. 0 0 0 0 0 0 0 0 0 0 Globpenet N 38 38 38 38 38 38 38 38 38 38

PAGE 138

127 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Germany Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market(Competition) Characteristics Mean 11.3158 13.0000 14.210 15.2368 16.9474 17.6842 18.8684 19.3947 20.1053 21.3684 St.Dev. 11.7338 12.7513 13.2056 14.5592 15.6377 15.5148 15.2938 15.1239 16.2776 16.5667 Gencompet N 38 38 38 38 38 38 38 38 38 38 Mean 19.3421 19.0263 19.1053 19.8947 19.6316 19.6842 19.2632 19.1316 18.8158 17.7368 St.Dev. 8.1548 7.9270 8.1067 7.2328 6.8396 6.4352 6.0390 6.1035 6.2379 5.3407 Thsubsmol N 38 38 38 38 38 38 38 38 38 38 Mean 14.0789 14.0789 14.0789 14.0789 14.0789 14.0789 14.0789 14.0789 14.0789 14.0789 St.Dev. 11.7137 11.7137 11.7137 11.7137 11.7137 11.7137 11.7137 11.7137 11.7137 11.7137 Thsubsmolentrylag N 38 38 38 38 38 38 38 38 38 38

PAGE 139

128 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: France Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 127,328 126,144 123,101 123,032 120,912 118,308 120,800 123,526 125,324 126,174 St.Dev. 109,921 106,117 101,089 102,219 102,264 100,450 104,124 108,503 114,037 118,120 Suthnds N 38 38 38 38 38 38 38 38 38 38 Mean 0.2636 0.2701 0.2752 0.2761 0.2735 0.2712 0.2701 0.2701 0.2677 0.2644 St.Dev. 0.2004 0.1962 0.1966 0.1967 0.1961 0.1966 0.1978 0.1987 0.1928 0.1983 Leusuprice N 38 38 38 38 38 38 38 38 38 38 Mean 31,951 33,417 34,378 35,163 35,040 33,914 34,752 36,809 38,322 39,620 St.Dev. 31,438 33,114 36,154 38,829 41, 850 40,525 43,508 50,576 57,310 65,306 Leuthnds N 38 38 38 38 38 38 38 38 38 38 Quality (Market) Characteristics Mean 0.1926 0.1931 0.1937 0.1939 0.1937 0.1936 0.1939 0.1940 0.1940 0.1940 St.Dev. 0.6459 0.6459 0.6459 0.6459 0.6459 0.6459 0.6460 0.6459 0.6460 0.6460 Strengthg N 38 38 38 38 38 38 38 38 38 38 Mean 15.6316 16.6316 17.6316 18.6316 19.6316 20.6316 21.6316 22.6316 23.6316 24.6316 St.Dev. 10.7186 10.7186 10.7186 10.7186 10.7186 10.7186 10.7186 10.7186 10.7186 10.7186 Molage N 38 38 38 38 38 38 38 38 38 38 Mean 32.7990 32.6109 32.4992 32.4883 32.6507 32.7460 33.0276 33.2668 33.1910 33.1579 St.Dev. 9.3614 9.2018 9.1312 9.0082 8.8119 8.7658 8.3847 8.3167 8.2380 8.2098 Packsize N 38 38 38 38 38 38 38 38 38 38 Mean 3.3684 3.1842 3.5263 3.9737 4.2368 4.5789 4.9474 4.8421 4.6316 4.6842 St.Dev. 2.5300 2.2763 2.7187 3.2507 3.5446 4.2339 4.5259 4.2140 3.6423 3.6621 Formcode N 38 38 38 38 38 38 38 38 38 38 Mean 5 5 5 5 5 5 5 5 5 5 St.Dev. 0 0 0 0 0 0 0 0 0 0 Globpenet N 38 38 38 38 38 38 38 38 38 38

PAGE 140

129 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: France Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 2.2105 2.1579 2.55269 3.0000 3.6053 3.8947 4.2895 4.6579 5.1579 5.5263 St.Dev. 1.7267 1.6362 2.1270 2.6610 3.3574 3.6449 4.1323 4.5398 5.1282 5.4263 Gencompet N 38 38 38 38 38 38 38 38 38 38 Mean 14.2632 13.6579 13.8158 13.7632 13.8684 13.8158 13.4474 13.4737 13.1579 13.0789 St.Dev. 5.5395 4.8339 4.5432 4.7558 4.6625 4.6374 4.4339 4.0653 3.7742 3.8088 Thsubsmol N 38 38 38 38 38 38 38 38 38 38 Mean 11.9474 11.9474 11.9474 11.9474 11.9474 11.9474 11.9474 11.9474 11.9474 11.9474 St.Dev. 8.7300 8.7300 8.7300 8.7300 8.7300 8.7300 8.7300 8.7300 8.7300 8.7300 Thsubsmolentrylag N 38 38 38 38 38 38 38 38 38 38

PAGE 141

130 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Italy Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 72,514 75,073 77,957 76,583 75,321 76,959 79,191 83,609 83,220 85,570 St.Dev. 81,844 81,868 84,434 82,710 83, 105 86,145 89,584 95,191 95,360 99,768 Suthnds N 38 38 38 38 38 38 38 38 38 38 Mean 0.2676 0.2428 0.2581 0.2675 0.2729 0.2837 0.2916 0.2977 0.2959 0.2832 St.Dev. 0.2371 0.1978 0.2140 0.2314 0.2320 0.2424 0.2567 0.2682 0.2742 0.2676 Leusuprice N 38 38 38 38 38 38 38 38 38 38 Mean 18,911 18,307 20,100 20,208 20,478 22,082 23,787 26,467 27,179 27,099 St.Dev. 25,141 23,657 26,719 27,278 28, 070 30,883 34,262 39,621 42,924 44,171 Leuthnds N 38 38 38 38 38 38 38 38 38 38 Quality Characteristics Mean 0.1756 0.1770 0.1778 0.1796 0.2046 0.2061 0.2058 0.2057 0.2059 0.2060 St.Dev. 0.5371 0.5400 0.5234 0.5091 0.6489 0.6556 0.6553 0.6550 0.6550 0.6549 Strengthg N 38 38 38 38 38 38 38 38 38 38 Mean 15.8158 16.8158 17.8158 18.8158 19.8158 20.8158 21.8158 22.8158 23.8158 24.8158 St.Dev. 11.2391 11.2391 11.2391 11.2391 11.2391 11.2391 11.2391 11.2391 11.2391 11.2391 Molage N 38 38 38 38 38 38 38 38 38 38 Mean 27.3968 27.3100 27.2803 26.9495 26.7784 26.7462 26.8004 27.1335 27.0267 26.9813 St.Dev. 9.5066 9.4423 9.5875 9.9184 10.0963 10.0612 9.8308 10.0760 10.1157 10.1591 Packsize N 38 38 38 38 38 38 38 38 38 38 Mean 3.3158 3.3158 3.4474 3.5000 3.5263 3.6842 3.9474 4.1579 4.2895 4.5526 St.Dev. 2.1699 2.1699 2.2263 2.3795 2.4576 2.6108 3.0839 3.6057 3.9105 4.1700 Formcode N 38 38 38 38 38 38 38 38 38 38 Mean 5 5 5 5 5 5 5 5 5 5 St.Dev. 0 0 0 0 0 0 0 0 0 0 Globpenet N 38 38 38 38 38 38 38 38 38 38

PAGE 142

131 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Italy Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 3.2895 3.3684 3.5263 3.5000 3.5263 3.6842 4.1053 4.4474 4.8947 5.1579 St.Dev. 2.5352 2.8037 3.0733 3.2529 3.3590 3.4882 4.2477 4.8418 5.2029 5.3753 Gencompet N 38 38 38 38 38 38 38 38 38 38 Mean 15.0000 14.9211 14.7632 14.5000 14.7105 14.6842 14.6579 14.5000 13.7105 13.3947 St.Dev. 4.8990 4.6230 3.8304 3.2777 3.4634 3.5421 3.6854 3.3430 3.3281 3.1153 Thsubsmol N 38 38 38 38 38 38 38 38 38 38 Mean 14.6053 14.6053 14.6053 14.6053 14.6053 14.6053 14.6053 14.6053 14.6053 14.6053 St.Dev. 10.2074 10.2074 10.2074 10.2074 10.2074 10.2074 10.2074 10.2074 10.2074 10.2074 Thsubsmolentrylag N 38 38 38 38 38 38 38 38 38 38

PAGE 143

132 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Spain Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 51,162 53,068 54,806 56,982 57,673 58,805 60,943 63,471 66,094 67,786 St.Dev. 59,583 59,411 58,957 59,923 60, 683 62,113 64,462 67,196 70,371 73,228 Suthnds N 38 38 38 38 38 38 38 38 38 38 Mean 0.2076 0.2088 0.2104 0.2136 0.2165 0.2187 0.2157 0.2135 0.2116 0.2103 St.Dev. 0.1714 0.1689 0.1679 0.1670 0.1660 0.1677 0.1680 0.1713 0.1708 0.1702 Leusuprice N 38 38 38 38 38 38 38 38 38 38 Mean 10,340 11,348 12,358 13,437 14,006 14,650 15,179 15,607 16,683 17,168 St.Dev. 14,540 15,768 17,035 18,530 19, 658 21,002 22,179 23,549 27,237 28,779 Leuthnds N 38 38 38 38 38 38 38 38 38 38 Quality Characteristics Mean 0.1940 0.1946 0.1961 0.1991 0.1999 0.2005 0.2010 0.2012 0.2013 0.2017 St.Dev. 0.6483 0.6485 0.6489 0.6506 0.6510 0.6514 0.6518 0.6521 0.6523 0.6527 Strengthg N 38 38 38 38 38 38 38 38 38 38 Mean 15.1842 16.1842 17.1842 18.1842 19.1842 20.1842 21.1842 22.1842 23.1842 24.1842 St.Dev. 11.0572 11.0572 11.0572 11.0572 11.0572 11.0572 11.0572 11.0572 11.0572 11.0572 Molage N 38 38 38 38 38 38 38 38 38 38 Mean 40.9292 40.7530 40.6194 40.3590 40.1945 40.0689 40.0768 40.1406 40.1108 40.0166 St.Dev. 11.9422 11.8425 11.6951 11.3057 11.2812 11.2371 11.1980 11.1785 11.1456 11.0794 Packsize N 38 38 38 38 38 38 38 38 38 38 Mean 3.6842 3.7632 3.8684 4.0000 3.9211 4.0526 4.2632 4.5000 4.5263 4.5526 St.Dev. 3.6101 3.8934 4.0281 4.2490 4.1680 4.2487 4.8862 4.9688 4.9415 4.9193 Formcode N 38 38 38 38 38 38 38 38 38 38 Mean 5 5 5 5 5 5 5 5 5 5 St.Dev. 0 0 0 0 0 0 0 0 0 0 Globpenet N 38 38 38 38 38 38 38 38 38 38

PAGE 144

133 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: Spain Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 3.4737 3.6579 3.6579 3.8947 4.0789 4.4211 4.8947 5.4211 6.2632 6.5789 St.Dev. 3.7034 4.1085 4.1347 4.4829 4.8120 5.5343 6.5669 7.6463 9.2871 8.5793 Gencompet N 38 38 38 38 38 38 38 38 38 38 Mean 14.0526 13.5000 13.6053 13.4211 14.0789 13.7895 13.3158 13.1579 13.0000 12.5263 St.Dev. 6.0402 5.5641 5.9436 5.7452 5.8558 5.7664 5.5513 4.8576 4.8267 3.9094 Thsubsmol N 38 38 38 38 38 38 38 38 38 38 Mean 14.0526 14.0526 14.0526 14.0526 14.0526 14.0526 14.0526 14.0526 14.0526 14.0526 St.Dev. 10.4389 10.4389 10.4389 10.4389 10.4389 10.4389 10.4389 10.4389 10.4389 10.4389 Thsubsmolentrylag N 38 38 38 38 38 38 38 38 38 38

PAGE 145

134 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: United Kingdom Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Mean 80,298 83,879 90,308 93,006 96,537 100,756 100,849 111,846 120,814 127,227 St.Dev. 122,753 123,225 128,652 128,780 131,682 137,061 134,530 151,547 162,743 173,136 Suthnds N 38 38 38 38 38 38 38 38 38 38 Mean 0.2965 0.3014 0.3060 0.3080 0.3224 0.3308 0.3266 0.3234 0.3238 0.3252 St.Dev. 0.2224 0.2281 0.2361 0.2446 0.2601 0.2657 0.2571 0.2680 0.2760 0.2764 Leusuprice N 38 38 38 38 38 38 38 38 38 38 Mean 17,005 18,936 21,825 24,195 26,628 29,711 31,382 34,813 40,388 45,309 St.Dev. 27,598 30,002 34,274 38,623 43, 236 47,376 49,965 59,890 76,642 91,943 Leuthnds N 38 38 38 38 38 38 38 38 38 38 Quality Characteristics Mean 0.2665 0.1759 0.1767 0.1771 0.1766 0.1770 0.1773 0.1731 0.1676 0.1665 St.Dev. 0.8392 0.6464 0.6464 0.6451 0.6373 0.6369 0.6329 0.6020 0.5666 0.5559 Strengthg N 38 38 38 38 38 38 38 38 38 38 Mean 16.6316 17.6316 18.6316 19.6316 20.6316 21.6316 22.6316 23.6316 24.6316 25.6316 St.Dev. 12.0818 12.0818 12.0818 12.0818 12.0818 12.0818 12.0818 12.0818 12.0818 12.0818 Molage N 38 38 38 38 38 38 38 38 38 38 Mean 145.2151 148.0258 146.2829 124.6675 119.4850 93.2642 66.3229 61.2544 58.9442 57.0293 St.Dev. 202.9629 209.2266 208.3658 173.3241 172.0150 123.1002 65.1457 57.0984 52.4440 46.5931 Packsize N 38 38 38 38 38 38 38 38 38 38 Mean 9.9211 10.1579 10.7105 11.6316 12.2368 13.2368 13.2632 13.1053 12.6316 12.8421 St.Dev. 9.2543 9.6493 10.4103 11.7161 12.3626 13.6232 12.8771 12.3284 11.7068 11.4880 Formcode N 38 38 38 38 38 38 38 38 38 38 Mean 5 5 5 5 5 5 5 5 5 5 St.Dev. 0 0 0 0 0 0 0 0 0 0 Globpenet N 38 38 38 38 38 38 38 38 38 38

PAGE 146

135 Table A.14 (Continued) Pharmaceutical Drug Quality and Market (Competition) Ch aracteristics for Global Molecules by Country by Year Country: United Kingdom Year Variable Measure 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Market (Competition) Characteristics Mean 2.2895 2.5789 2.6579 3.0789 3.2105 3.2895 3.4474 3.8684 3.7632 3.7632 St.Dev. 2.3005 2.6573 3.0868 3.6271 4.0414 3.7626 3.7032 4.4001 4.0366 3.7592 Gencompet N 38 38 38 38 38 38 38 38 38 38 Mean 12.5263 12.5263 12.9474 13.3421 13.5263 13.4474 13.1053 12.5000 12.3684 12.7895 St.Dev. 4.4826 4.4826 4.2359 4.1151 3.9232 3.6145 3.1860 3.0202 2.8798 2.8867 Thsubsmol N 38 38 38 38 38 38 38 38 38 38 Mean 12.9737 12.9737 12.9737 12.9737 12.9737 12.9737 12.9737 12.9737 12.9737 12.9737 St.Dev. 9.0060 9.0060 9.0060 9.0060 9.0060 9.0060 9.0060 9.0060 9.0060 9.0060 Thsubsmolentrylag N 38 38 38 38 38 38 38 38 38 38

PAGE 147

136 Table A.15 Quality Adjusted (by RE) Standa rd Unit Price Differentials for All Molecules, Relative to Spaina Year Spain France Italy United Kingdom Germany 1994 39.63% 27.09% 80.21% 105.06% 1995 2.25% 42.17% 18.46% 79.93% 100.96% 1996 4.25% 42.12% 24.26% 78.87% 100.62% 1997 6.91% 39.62% 25.36% 72.73% 93.15% 1998 7.84% 39.45% 28.19% 83.87% 89.94% 1999 9.30% 38.94% 31.54% 90.54% 86.88% 2000 8.55% 40.26% 35.24% 86.94% 87.43% 2001 8.30% 40.15% 37.86% 85.04% 88.77% 2002 8.11% 40.03% 39.86% 83.54% 80.51% 2003 8.93% 37.12% 33.99% 81.30% 76.76% Table A.16 Quality Adjusted (by FE) Standa rd Unit Price Differentials for All Molecules, Relative to Germanyb Year Germany France Italy United Kingdom Spain 1994 30.43% 36.76% 11.62% 50.02% 1995 0.02% 27.80% 39.93% 10.07% 49.10% 1996 1.44% 27.70% 36.90% 10.47% 49.04% 1997 0.01% 26.34% 33.97% 10.24% 47.14% 1998 0.92% 25.35% 31.51% 3.03% 46.34% 1999 1.39% 24.45% 28.63% 2.20% 45.50% 2000 1.99% 24.02% 26.88% 0.08% 45.72% 2001 1.81% 24.61% 26.02% 1.59% 46.11% 2002 6.42% 21.28% 21.62% 1.99% 43.71% 2003 7.81% 21.43% 23.46% 2.72% 42.61% a These are the coefficients of country/time effects ( j,t) in the random effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1]. b These are the coefficients of country/time effects ( j,t) in the fixed effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1].

PAGE 148

137 Table A.17 Quality Adjusted (by FE) Standa rd Unit Price Differentials for All Molecules, Relative to Spaina Year Spain France Italy United Kingdom Germany 1994 39.19% 26.53% 76.83% 100.07% 1995 1.86% 41.85% 18.01% 76.67% 96.46% 1996 3.41% 41.88% 23.83% 75.70% 96.24% 1997 5.74% 39.36% 24.92% 69.80% 89.19% 1998 6.38% 39.10% 27.63% 80.70% 86.35% 1999 7.52% 38.63% 30.96% 87.51% 83.48% 2000 6.43% 39.98% 34.72% 84.39% 84.24% 2001 5.86% 39.91% 37.28% 82.62% 85.57% 2002 5.39% 39.83% 39.23% 81.17% 77.64% 2003 5.85% 36.92% 33.38% 78.98% 74.25% Table A.18 Quality Adjusted (by RE) Standa rd Unit Price Differentials for Global Molecules, Relative to Spainb Year Spain France Italy United Kingdom Germany 1994 31.87% 21.32% 81.81% 103.44% 1995 3.45% 36.58% 12.63% 76.92% 99.13% 1996 6.98% 37.68% 17.94% 73.35% 95.62% 1997 12.30% 34.21% 17.22% 63.23% 82.23% 1998 16.06% 30.90% 18.65% 67.24% 76.20% 1999 20.52% 27.09% 21.15% 66.06% 71.83% 2000 21.76% 27.95% 24.11% 61.13% 68.77% 2001 22.55% 29.17% 27.65% 55.96% 68.81% 2002 23.94% 29.29% 27.94% 51.86% 58.79% 2003 25.72% 26.75% 23.81% 50.34% 51.59% a These are the coefficients of country/time effects ( j,t) in the fixed effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1]. b These are the coefficients of country/time effects ( j,t) in the random effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1].

PAGE 149

138 Table A.19 Quality Adjusted (by FE) Standa rd Unit Price Differentials for Global Molecules, Relative to Germanya Year Germany France Italy United Kingdom Spain 1994 -34.7% -40.2% -10.7% -50.6% 1995 0.7% -30.9% -43.3% -11.4% -49.5% 1996 1.7% -29.1% -39.6% -11.5% -48.6% 1997 -1.0% -25.8% -35.6% -10.6% -44.9% 1998 -1.7% -25.1% -32.5% -5.2% -42.9% 1999 -0.9% -25.5% -29.3% -3.4% -41.5% 2000 -2.3% -23.6% -26.2% -4.5% -40.4% 2001 -2.2% -22.9% -24.2% -7.7% -40.4% 2002 -7.5% -18.0% -19.2% -4.4% -36.7% 2003 -11.0% -15.7% -18.0% -0.7% -33.6% Table A.20 Quality Adjusted (by FE) Standa rd Unit Price Differentials for Global Molecules, Relative to Spaina Year Spain France Italy United Kingdom Germany 1994 32.16% 20.88% 80.66% 102.29% 1995 2.83% 36.87% 12.29% 75.52% 98.02% 1996 5.78% 37.97% 17.55% 72.05% 94.50% 1997 10.43% 34.51% 16.82% 62.17% 81.36% 1998 13.60% 31.09% 18.25% 66.09% 75.14% 1999 17.28% 27.32% 20.78% 65.04% 70.88% 2000 17.78% 28.20% 23.81% 60.23% 67.84% 2001 17.88% 29.44% 27.34% 55.01% 67.91% 2002 18.54% 29.55% 27.54% 50.94% 57.93% 2003 19.52% 27.06% 23.45% 49.60% 50.64% a These are the coefficients of country/time effects ( j,t) in the fixed effect quasi-hedonic regression model. Percentages are calculated as 100[Exp( )-1].

PAGE 150

139 Table A.21 Results for Price Convergence Estimations for All Molecules (Adjusted by RE) Model 1 Model 2 Dependent Variable: Pi,k,t Base: SPN Base: GR Base: SPN Base: GR a 0.01 -0.02 -0.19 -0.18 Half-life of Shock (in years) 34.3 3.3 3.5 FRb 0.02 0.00 GRb 0.09 UKb 0.10 0.00 SPNb -0.09 ITYb 0.07 -0.07 Lags of Pi,k,t Yes(1)c Yes(1)c Yesd Yesd t-star P>t 8.80 1.000 -19.89 0.000 -8.92 0.000 -8.16 0.000 Molecule/Country Fixed Effects No No Yes Yes Time Trend No No No No N 2,940 3,210 2,940 3,210 a coefficients are estimated by the Levin et al. (1992) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each molecu le/country by the Augmented Dickey Fuller unit root test module in StataTM 9.2 (dfuller) and then aver aged for each country. c The average number of lags for each molecule/country is 1. d The number of lags is determined by using the Campbell and Perron top-down approach.

PAGE 151

140 Table A.22 Results for Price Convergence Estimati ons for All Molecules (Unadjusted) Model 1 Model 2 Model 3 Dependent Variable: Pi,k,t Base: SPN Base: GR No Base Base: SPN Base: GR No Base Base: SPN Base: GR No Base a -0.00 -0.01 0.00 -0.33 -0. 14 -0.30 -0.77 -0.68 -0.65 FRb 0.16 -0.02 -0.66 0.16 -0.02 -0.90 GRb 0.02 -0.32 0.02 -0.44 UKb -0.17 0.02 -0.25 -0.17 -0.17 -0.81 SPNb -0.19 -1.08 -0.79 -1.50 ITYb 0.06 -0.00 -0.35 0.06 -0.30 -0.94 Lags of Pi,k,t Yes(1)c Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) t-star p>t -14.19 0.000 -8.52 0.000 0.18 0.575 -30.00 0.000 -15.82 0.000 -30.78 0.000 -94.37 0.000 -27.88 0.000 -41.37 0.000 Molecule/Country Fixed Effects No No No Yes Yes Yes Yes Yes Yes Time Dummies No No No No No No Yes Yes Yes N 2940 3210 3790 2940 3210 3790 2940 3210 3790 a coefficients are estimated by Levin et al. (2002) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each molecule/country by Augmented Dickey Fuller regressions (dfuller) in StataTM 9.2 and then averaged for each country. c The average number of lags for each molecule/country is 1.

PAGE 152

141 Table A.23 Results for Price Convergence Estimations for Global Molecules (Adjusted by FE) Model 1 Model 2 Dependent Variable: Pi,k,t Base: SPN Base: GR Base: SPN Base: GR a 0.02 -0.02 -0.14 -0.21 Half-life of Shock (in years) 34.3 4.6 2.9 FRb 0.00 0.07 GRb 0.08 UKb 0.19 0.06 SPNb -0.08 ITYb 0.07 -0.09 Lags of Pi,k,t Yes(1)c Yes(1)c Yesd Yesd t-star P>t 11.96 1.000 -12.69 0.000 -5.95 0.000 -12.69 0.000 Molecule/Country Fixed Effects No No Yes Yes Time Trend No No No No N 1,900 1,900 1,900 1,900 a coefficients are estimated by the Levin et al. (2002) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each mo lecule/country by the Augmented Dickey Fuller regressions in StataTM 9.2 (dfuller) and then aver aged for each country. c The average number of lags for each molecule/country is 1. d The number of lags is determined by using the Campbell and Perron top-down approach.

PAGE 153

142 Table A.24 Results for Price Convergence Estimation s for Global Molecules (Unadjusted) Model 1 Model 2 Model 3 Dependent Variable: Pi,k,t Base: SPN Base: GR No Base Base: SPN Base: GR No Base Base: SPN Base: GR No Base a 0.01 -0.02 0.00 -0.14 -0.08 -0.19 -0.64 -0.65 -0.62 FRb 0.05 0.14 -0.17 0.29 0.04 -0.02 GRb 0.20 0.12 0.74 0.60 UKb 0.08 0.04 -0.03 0.15 -0.20 -0.24 SPNb -0.20 -0.19 -0.74 -0.29 ITYb 0.08 0.02 -0.09 0.17 -0.27 -0.38 Lags of Pi,k,t Yes(1)c Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) Yes(1) t-star P>t 10.78 1.000 -18.85 0.000 0.33 0.630 -7.99 0.000 -6.16 0.000 -13.13 0.000 -25.76 0.000 -21.94 0.000 -27.72 0.000 Molecule/Country Fixed Effects No No No Yes Yes Yes Yes Yes Yes Time Dummies No No No No No No Yes Yes Yes N 1900 1900 1900 1900 1900 1900 1900 1900 1900 a coefficients are estimated by Levin et al. (2002) panel unit root test module in StataTM 9.2 (levinlin). b Country fixed effects are estimated for each molecule/country by Augmented Dickey Fuller regressions (dfuller) in StataTM 9.2 and then averaged for each country c The number of lags for each molecule/country is 1.

PAGE 154

143 Appendix B: Figures Figure B.1 Summary of EU Pharmaceutical Background Free movement of goods by Treaty of Rome Permits arbitrage of these price differences Market integration (SMP and EMU) Process Invites Parallel Imports Export low prices from low to other potentially high priced countries Price Convergence? Single Pharmaceutical Market Regulation based on international p rice com p arisons Difference in national health systems and pricing and reimbursement regulations EU Pharmaceutical Pricing a nd Reimbursement Regulations Leads price differences

PAGE 155

144 Figure B.2 Pharmaceutical Production in the European Union (In million dollars at exchange rate) 0 5000 10000 15000 20000 25000 30000 199019911992199319941995199619971998199920002001 Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom Source: OECD Health Data (2003)

PAGE 156

145 Figure B.3 Total Pharmaceutical Sales in the European Union (In million dollars at exchange rate) 0 5000 10000 15000 20000 25000 30000 350001990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002YearsMillion Dollars Denmark Finland France Germany Greece Italy Netherlands Portugal Sweden United Kingdom Source: OECD Health Data (2003)

PAGE 157

146 Figure B.4 Bilateral Price Differences for All Molecules Between 1994-2003 by Laspeyres Index, Relative to Germany -40.00% -30.00% -20.00% -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 1994199519961997199819992000200120022003 SPN UK FR ITY Figure B.5 Bilateral Price Differences for All Molecules Between 1994-2003 by Paasche Index, Relative to Germany -40.00% -30.00% -20.00% -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 1994199519961997199819992000200120022003 SPN UK FR ITY

PAGE 158

147 Figure B.6 Bilateral Price Differences for All Molecules Between 1994-2003 by Laspeyres Index, Relative to Spain 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% 40.00% 45.00% 50.00% 1994199519961997199819992000200120022003 GR UK FR ITY Figure B.7 Bilateral Price Differences for All Molecules Between 1994-2003 by Paasche Index, Relative to Spain -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 1994199519961997199819992000200120022003 GR UK FR ITY

PAGE 159

148 Figure B.8 Bilateral Price Differences for Global Molecules Between 1994-2003 by Laspeyres Index, Relative to Germany -30.00% -20.00% -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% 1994199519961997199819992000200120022003 SPN UK FR ITY Figure B.9 Bilateral Price Differences for Global Molecules Between 1994-2003 by Paasche Index, Relative to Germany -30.00% -20.00% -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 1994199519961997199819992000200120022003 SPN UK FR ITY

PAGE 160

149 Figure B.10 Bilateral Price Differences for Global Molecules Between 1994-2003 by Laspeyres Index, Relative to Spain -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 1994199519961997199819992000200120022003 GR UK FR ITY Figure B.11 Bilateral Price Differences for Global Molecules Between 1994-2003 by Paasche Index, Relative to Spain -20.00% -10.00% 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 1994199519961997199819992000200120022003 GR UK FR ITY

PAGE 161

150 Figure B.12 Country Price Changes for All Molecules by Laspeyres Index, Relative to 1994 Germany-25.00% -20.00% -15.00% -10.00% -5.00% 0.00% 199519961997199819992000200120022003 United Kingdom 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 199519961997199819992000200120022003 France-6.00% -5.00% -4.00% -3.00% -2.00% -1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 199519961997199819992000200120022003 Italy -12.00% -10.00% -8.00% -6.00% -4.00% -2.00% 0.00% 2.00% 4.00% 6.00% 199519961997199819992000200120022003 Spain -2.00% -1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 199519961997199819992000200120022003

PAGE 162

151 Figure B.13 Country Price Changes for Global Molecules by Laspeyres Index, Relative to 1994 Italy-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 199519961997199819992000200120022003 Year Germany-0.35 -0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0 199519961997199819992000200120022003 Year Spain-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 199519961997199819992000200120022003 Year United Kingdom-0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 199519961997199819992000200120022003 Year France-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 199519961997199819992000200120022003 Year

PAGE 163

152 Figure B.14 Quality Adjusted (by RE) Standa rd Unit Price Differentials for All Molecules, Relative to Spain 0.0 20.0 40.0 60.0 80.0 100.0 120.0 1994199519961997199819992000200120022003 Year% FR ITY UK GR Figure B.15 Quality Adjusted (by FE) Standa rd Unit Price Differentials for All Molecules, Relative to Germany -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 1994199519961997199819992000200120022003 Year% FR ITY UK SPN

PAGE 164

153 Figure B.16 Quality Adjusted (by FE) Standa rd Unit Price Differentials for All Molecules, Relative to Spain 0.0 20.0 40.0 60.0 80.0 100.0 120.0 1994199519961997199819992000200120022003 Year% FR ITY UK GR Figure B.17 Quality Adjusted (by RE) Standa rd Unit Price Differentials for Global Molecules, Relative to Spain -20.00 0.00 20.00 40.00 60.00 80.00 100.00 120.00 1994199519961997199819992000200120022003 Year% FR ITY UK GR

PAGE 165

154 Figure B.18 Quality Adjusted (by FE) Standa rd Unit Price Differentials for Global Molecules, Relative to Germany -60.00 -50.00 -40.00 -30.00 -20.00 -10.00 0.00 10.00 20.00 1994199519961997199819992000200120022003 Year% FR ITY UK SPN Figure B.19 Quality Adjusted (by FE) Standa rd Unit Price Differentials for Global Molecules, Relative to Spain 0.00 20.00 40.00 60.00 80.00 100.00 120.00 1994199519961997199819992000200120022003 Year% FR ITY UK GR

PAGE 166

About the Author Aysegul Timur obtained her undergraduate degree in Business Administration in 1993 and her MBA in 1997 from the University of Istanbul. After holding positions in the information technology industry as a total qual ity coordinator and a corporate trainer in Turkey, she moved to the USA in 1998. While teaching at International College as a full time professor in the Business Administration Department, Mrs. Timur entere d the Ph.D. program in the Department of Economics at the University of South Florida in 2002. Her major area of research is in Health and International Economics.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001926313
003 fts
005 20080207091811.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080207s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001911
035
(OCoLC)191864149
040
FHM
c FHM
049
FHMM
090
HF71 (ONLINE)
1 100
Timur, Aysegul.
4 245
The single market and pharmaceutical industry in the European Union :
b is there any evidence of price convergence?
h [electronic resource] /
by Aysegul Timur.
260
[Tampa, Fla] :
University of South Florida,
2006.
3 520
ABSTRACT: During the last two decades, the European Union (EU) has experienced closer market integration through the removal of trade barriers, the establishment of a single market, and the reduction of exchange rate volatility. In addition, there have been several structural reforms in product markets designed to increase competition, monitor cross-country price differences and increase transparency. One anticipated effect of market integration is price convergence, because of the reduced potential for price discrimination across the EU. This dissertation explores market integration and price convergence in the European pharmaceutical market, which is the fifth largest industry in the EU. Since 1985, many EU directives have been adopted to achieve a single EU-wide pharmaceutical market, with the aim of enhancing the quality of life for European citizens and the European pharmaceutical industry's competitiveness and research and development capability. Using annual 1994--2003 data from five EU countries on prices of drugs used to treat cardiovascular disease, this dissertation explains how the integration process has affected cross-country drug price dispersion in the EU. The results show strong evidence of price convergence in the pharmaceutical market, with long term price differences arising from country fixed effects.
502
Dissertation (Ph.D.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 154 pages.
Includes vita.
590
Adviser: Gabriel Picone, Ph.D.
653
Market integration.
European pharmaceutical market.
Unit root.
Hedonic regression.
Laspeyres.
Paasche.
0 690
Dissertations, Academic
z USF
x Business Administration
Doctoral.
773
t USF Electronic Theses and Dissertations.
856
u http://digital.lib.usf.edu/?e14.1911