NNNNNH lNN h This is to certify that the dissertation entitled LICENSING STRATEGIES AND PERFORMANCE: AN EMPIRICAL ANALYSIS IN THE GIDBAL PHARMACEUTICAL INDUSTRY presented by Kathleen R. Whitney has been accepted towards fulfillment of the requirements for Ph. D. , Business degree in C/Nw wa/ Major professor S. Tamer Cavusgil, Ph. DU November 1 , 1991 Date MSUisa nAfflratm 11'de onu/Eq arulOppotn irylnst (in! 0-12771 gmgm ‘ m M PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. ll DATE DUE DATE DUE { DATE DUE h MSU Is An Affirmative Action/Equal Opportunity Institution ammut LICENSING STRATEGIES AND PBRFORNANCB: AN EMPIRICAL ANALYSIS IN THE GLOBAL PHARMACEUTICAL INDUSTRY BY Kathleen R. Whitney A DISSERTATION Submitted to Michigan State University in partial fulfillment for the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation Administration 1991 ABSTRACT LICENSING STRATEGIES AND PERFORMANCE: AN EMPIRICAL ANALYSIS IN THE GLOBAL PHARMACEUTICAL INDUSTRY BY Kathleen R. Whitney Statement of the Problem. Pharmaceutical firms have a complex problem marketing diverse products globally. They possess little theory to guide them and there is limited empirical research relating specific strategies to sales and market share performance. Further, strategies are constrained by governmental regulations. Finally, licensing seems to be used extensively without an adequate understanding of its impact upon performance. Under these conditions firms may rely upon past product and market strategies as decision criteria, possibly excluding more effective strategies. An important opportunity exists to empirically test global marketing strategy as it relates to diverse products, worldwide markets, use of strategic alliances, and regulations. Ineoretical Eramewgrk. These concerns are addressed by incorporating these issues into a theoretical framework: (1) since global product and market diversity is a problem, the number of products and markets a firm manages may impact performance; (2) licensing strategies may impact performance; and (3) global strategies may be affected by regulatory differences. o o o . Selected pharmaceutical product classes are studied. The pharmaceutical industry is chosen because of its extensive global experience, costly technology, and stringent governmental regulation. Secondary data are used to compile 504 observations by firm, therapeutic class, and country, for a period of dramatic industry change, 1982- 1987. Approximately 76% of world pharmaceutical sales is represented by the markets studied: the United States, Japan, West Germany, Italy, France, and the United Kingdom. Multiple regression analysis with time-lagged variables are used to test relationships between strategy and performance, past and current strategies, and regulation and product introductions. Analysis is conducted at the firm, therapuetic class, and country levels. uajgr Eindings. There are four major findings: (1) Sales and market share performance is enhanced by the number of product introductions and by constraining product acquisitions to a few product classes; (2) there is a critical difference between the market leader and other firms with regard to the relationship between product introductions and performance: (3) market representation does not vary with performance; and (4) product introduction behavior does not always relate to regulatory conditions as expected. Copyright by KATHLEEN ROGERS WHITNEY 1991 DEDICATION For David ACKNOWLEDGMENTS I would like to gratefully acknowledge the individuals. who have played a significant role in accomplishing this research. Each committee member made a unique and vital contribution. S. Tamer Cavusgil directed me to pursue my research in an industry in which I have personal strengths. Robert W. Nason persisted in drawing the best he could from me. David J. Closs provided guidance in methodology and organization. Mary Jane Sheffet gave generously of her time, reading every word of each and every draft. As with any project of this size, considerable resources were needed. John Hornick provided access to the Upjohn Business Library and assisted in obtaining the gag; ge Hagn data base. The Michigan State University CIBER grant also contributed to data acquisition costs. Finally, my family provided me with moral support and endured persistent benign neglect. I extend my special thanks and appreciation to each of them: David, my husband: Kendra, my daughter: and Ellen; my mother. vi Chapter II. III. TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . LIST OF FIGURES O O O O O O O O O O O O 0 INTRODUCTION . . . . . . . . . Statement of the Problem . . . Research Scope and Objectives Limitations of the Study . . . Potential Contributions . . . Organization of the Study . . SURVEY OF THE LITERATURE . . . Introduction . . . . . . . . Global Strategy Literature . Strategic Alliances . . . Product Acquisition and Market Access Inseparability of Product Acquisition and Market Access . . . . Product Acquisition . . . Market Access . . . . . . Strategy and Performance . . Product Portfolio Analysis . . Pharmaceutical Global Product Portfolio Conclusion . . . . . . . . . . . . . . . PHARMACEUTICAL INDUSTRY BACKGROUND . . . . Introduction . . . . . . . . . . . . . . Study Position within the Pharmaceutical Industry . An Historical Perspective Nature of the Industry . . Environmental Constraints Research and Development . Regulatory Issues . . . . National Health Insurance National Formulary Systems R&D Pricing Incentives . . Acceptance of Nondomestic Clinical Testing . . . . Patent Protection . . . . Compulsory Out-Licensing Generic Substitution . . Regulatory Issues and Strategy Foeeeeee g... eemeeeeee :3 vii Page xii 62 65 65 70 71 72 73 73 74 75 75 75 77 77 Chapter IV. Page In the United States . . . . . . . . . . . . 78 Japan . . . . . . . . . . . . . . . . . . . 83 European Community . . . . . . . . . . . . . 84 Competitive Environment . . . . . . . . . . 85 Patent Protection . . . . . . . . . . . . 86 The Regulatory Life-Cycle . . . . . . . . . 87 Analysis of Industry Strategic Elements . . 90 Place . . . . . . . . . . . . . . . . . . 90 Price . . . . . . . . . . . . . . . . . . 91 Promotion . . . . . . . . . . . . . . . . 91 Product . . . . . . . . . . . . . . . . . 93 Conclusion . . . . . . . . . . . . . . . . . 94 RESEARCH DESIGN AND METHODOLOGY . . . . . . . 96 Introduction and Design Objectives . . . . . 96 Theoretical Framework . . . . . . . . . . . 97 Hypotheses . . . . . . . . . . . . . . . . 100 Past Position and Strategy . . . . . . . 104 Strategy and Performance Goals . . . . . 105 Regulation and Market Decisions . . . . . 109 Data Collection . . . . . . . . . . . . . . 110 Operationalization of the Variables. . . . . 116 Paul de Haen Variables Used . . . . . . . 117 Pharmaceutical Line-of—Business Data . . 118 The Regulatory Environment . . . . . . . 122 Method of Analysis . . . . . . . . . . . . . 124 ANALYSIS AND RESULTS . . . . . . . . . . . . . 127 Introduction . . . . . . . . . . . . . . 127 Testing of Hypotheses . . . . . . . . . . . 127 Hypothesis--Family One . . . . . . . . . . . 129 Managerial Use of Past Product Acquisition Information . . . . . . . . . . 134 Managerial Implications-- Product Acquisition Strategies. . . . . . . 141 Managerial Implications-- Market Access . . . . . . . . . . . . . . . 143 Summary--Hypothesis Family One . . . . . . . 144 Hypothesis-~Family Two . . . . . . . . . . . 146 Linear Model Predicting Sales . . . . . . . 148 Managerial Implications-- Linear Model Predicting Sales . . . . . . . 155 Quadratic Regression Model Predicting Market Share . . . . . . . . . . . . . . 158 Summary of Strategy Models . . . . . . . . . 164 Concurrent Strategies . . . . . . . . . . . 167 Hypothesis--Three . . . . . . . . . . . . . 169 Summary . . . . . . . . . . . . . . . . . . 182 Chapter VI. CONCLUSIONS AND IMPLICATIONS . . . Introduction . . . . . . . . . . Research Implications . . . . Past and Current Strategies . Strategy and Performance Goals . . Regulation and Product Introductions A Revised Global Strategy Framework for Pharmaceuticals . . . . . . . Managerial Implications . . . . . . Past and Current Strategies . . . Strategy and Performance Goals . Revised Pharmaceutical Global Product Portfolio . . . . . Regulation and Market Decisions Public Policy Implications . . . . Conclusion . . . . . . . . . . . . APPENDIX A: Literature Review Matrix . . . . . . . APPENDIX B: Summary of Variables . . . . . . . . . APPENDIX C: Definition of Terms . . . . . . . . . . APPENDIX D: Summary of Results--Strategy Performance Model . . . . . . . . . . BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . GENERAL REFERENCES . . . . . . . . . . . . . . . . . ix Page 187 187 188 188 193 194 195 197 197 199 203 206 207 213 215 220 227 227 230 241 Table 1. 10. 11. 12. 13. 14. 15. 16. LIST OP TABLES Page Estimated Value of Leading Pharmaceutical Markets-~1987-1988 . . . . . . . . . . . . . . . 52 Major Drug Product Classifications . . . . . . . . 52 Major Causes of Health-Related Mortality in Countries of Interest . . . . . . . . . . . . 53 Synthesis of Global Literature The Strategy-Performance Relationship . . . . . . 57 Major Ethical Drug Manufacturers--1988 . . . . . . 67 Top Thirty Selling Drugs Worldwide-- 1988 and 1989 O O O O O O O O O O O O O O O O O O 68 Number of Drug Representing 50% or Greater of Company Sales--l985 . . . . . . . . . . . . . 69 Number of New Chemical Entities Introduced, by Drug Category--1987 and 1988 . . . . . . . . . 74 Summary of Country Regulations . . . . . . . . . . 76 Regulatory Life-Cycle of Pharmaceutical Products . . . . . . . . . . . . . . . . . . . . 88 Summary of Strategy and Performance Variables . . 101 Tests of Hypotheses . . . . . . . . . . . . . . . 102 Sample Single Chemical Entities by Therapeutic Class . . . . . . . . . . . . . . . 113 Pharmaceutical Line-of-Business Sample by Domestic Country . . . . . . . . . . . . . . 115 Summary of Acronyms and Variables . . . . . . . . 128 Summary of Results--Hypotheses 1A and 13 Past Product Development and In-Licensing Strategies . . . . . . . . . . . . . . . . . . . 130 Table Page 17. Correlations of Past In-House Product Development and Past In-Licensing with Current Product Development within Product Classes . . . . . . . . . . . . . . . . 135 18. Correlations of Past In-House Product Development and Past In-Licensing with Top Drugs within Product Classes . . . . . . . . . . 138 19. Hypotheses 1C and 1D: Summary of Results . . . . 141 20. Relationships of Pairs of Past and Current Statistics . . . . . . . . . . . . . . . 145 21. Hypothesis 2A: Regression Model for Strategy and Sales . . . . . . . . . . . . . . . 149 22. Hypothesis 28: Regression Model for Strategy and Market Share . . . . . . . . . . . 160 23. Correlation Matrix-~Hypothesis Family Two . . . . 168 24. Stepwise Regression Analyses Predicting Product Introductions by Regulations Affecting the Global Pharmaceutical Industry . . . . . . . 171 25. Stepwise Regression Analyses Predicting Product Introductions by Firm Home Country . . . 173 26. Stepwise Regression Analysis Predicting Product Introductions by Country of Origin . . . 175 27. Product Introductions into Regulatory conditions O O O O O O O O O O O O O O O O O O O 177 28. Stepwise Regression Analyses Predicting Product Introductions into Country Markets by Country of Origin . . . . . . . . . . . . . 178 29. Frequencies of Product Introductions into Country Markets by Country of Origin . . . . . . 180 30. Summary of Study Findings . . . . . . . . . . . . 184 31. Findings in Contrast to Previous Literature . . . 189 A-l. Literature Review Matrix . . . . . . . . . . . . . 215 0-1. Summary of Results--Strategy Performance Model . . 229 xi LIST OF PIGURES Figure Page 1. The Boston Consulting Group Matrix . . . . . . . 42 2. Line-of-Business Ranked First--1987 . . . . . . . 48 3. Line-of-Business Ranked Fifteenth--1987 . . . . . 49 4. Global Regulatory Environmental Framework Pharmaceuticals . . . . . . . . . . . . . . . . 61 5. Global Strategy Framework for Pharmaceuticals . . . . . . . . . . . . . . . . 98 6. Strategic Choices . . . . . . . . . . . . . . . . 131 7. Linear Regression Model Predicting Sales . . . . 150 8. Quadratic Regression Model Predicting Market Share . . . . . . . . . . . . . . . . . . 161 9. Revised Global Strategy Framework for Pharmaceuticals . . . . . . . . . . . . . . . . 196 10. Revised Pharmaceutical Global Product Portfolio . . . . . . . . . . . . . . . . . . . 205 CHAPTER I INTRODUCTION Statement of the Problem Global strategic marketing requires a complex set of decisions concerning products, place, promotion, and pricing. It is more complex than domestic strategy because there is greater variation on all four elements. However, recent literature Ohmae 1986a and b; Porter 1990) reflects an emphasis on the product and place variables. Specif- ically, the method of acquiring new products (by internal product development or strategic alliances such as cross- 1icensing) has been of interest to researchers. The place element has also been addressed in global strategy because firms face a similar strategic decision (use of internal or external resources) to gain access to various geographic markets. Therefore, the primary concern of this research involves product acquisition and market access strategies. The dramatic increase in international availability of transportation and communication technologies and the rapid rate of technological change has invalidated classic theory which supported gradual dissemination of new products into iJTternational markets (Ohmae 1986a and b). Although it is 2 clear that gradual dissemination of product innovations no longer works (Ayal 1981; Porter 1990), alternative global strategy concepts have not yet been empirically tested. In light of the decreasing life-span of products, new marketable products have become more important to fill established channels. Foxall (1983), Meyer and Roberts (1988), Ohmae (1989), and Porter (1990) have emphasized continuous innovation in the form of internal product development, as an avenue to success. Porter stressed that development efforts should be aimed at updating current products, which implies focus on product categories rather than focus on a particular technology. Meyer and Roberts' (1988) findings empirically supported the importance of what they termed an innovation focus which stresses relatedness of technology rather than product categories in formulating strategy. On the other hand, Foxall (1983) also encouraged a constrained search for new products. New products do not necessarily supersede older products and they can be de- signed for existing or new functional markets. Two notable theorists have addressed innovation or internal product development (Porter 1990) and global dis- tribution or market access (Ohmae 1986a and b). Ohmae's "Triad World View" (1986a and b) states that to be success- ful, firms must simultaneously begin marketing new products in the three major developed countries: Japan, the United States, and Western Europe. Their market importance stems 3 from their large size with a combined population of approximately 630 million people. Because consumers of many countries demand similar product characteristics for many products, there has been a tendency not to establish separate operations in each coun- try. Even though in-country presence may not be necessary for purposes of adapting a product to special needs, it is recommended by Thorelli (1990) for developing and maintain- ing access to that geographic market. Ohmae noted that it is vital for firms to "become an insider in each of the triad regions (Ohmae 1986a, p. 18)." At the same time Thorelli (1990) and Yip (1990) emphasized the need for depth of commitment to country markets. Thorelli used the notion of networking which he defines as the process of estab- lishing and maintaining contacts within countries. He emphasized the need to become a good corporate citizen to firmly establish an in-country network. He stressed the need to establish buyer linkages which include information, technology, and social and financial aspects. The network concept is aimed at what Yip (1989) described as committed involvement which also involves establishing in-country linkages with a long term perspective. Committed involvement requires management to take a long-term view of in-country investments. If operations and a long-term view of investment are important, then firms face the problem of simultaneously managing continuous product acquisition and developing and 4 maintaining access to many country markets. The central managerial problem is how to direct both efforts most effec- tively for marketing success. These two tasks should not be separated from one another. Product acquisition must be directed towards identified needs and wants of consumers within or across country markets. Without an identified market segment and targeted geographic markets, products are less likely to be commercializable. Conversely, distribu- tion channels for targeted markets cannot be efficiently developed and maintained without regard to product charac- teristics. For example, pre- and post-sales support for main-frame computer systems is quite different from support for tailored clothing. In effect, the planning of product acquisition and market access is inseparable because each is dependent upon the other. The current environment strength- ens the bond between product acquisition and market access strategies. This will be addressed more fully in Chapter II. The complexity of the problem does not end with the need to simultaneously manage product acquisition and market access. Managers must plan to fill the strategic gaps between current capabilities and those which are still needed for success. Ohmae describes the problem in the following statement (1989, p. 143), "With enough time, money, and luck, you can do everything yourself. But who has enough?” In a global environment in which the cost of distribution and research and development (R&D) costs are 5 very high, the appropriate strategy requires entente--the striking of an alliance between firms (Ohmae 1989). Manage- ment decisions need to include consideration of firm strengths and weaknesses and those of available external resources. Robinson (1978) called this the make or buy decision. An appropriate combination of internal and external product acquisition and market access strategies must be selected to fill existing gaps. But first, these strategic gaps must be identified. Capon and Glaser (1987) succinctly state the following problems concerning acquiring and marketing new products: - evaluation of a wide range of options for develop- ing and acquiring technology, as well as for mar- keting technologies in the firm's inventory, [and] - utilization of the technology portfolio as an organizing tool that can help the firm both to evaluate the current technological portfolio and to plan the optimal decisions for future technolo- gy scenarios (p. 12). Limited time and money in a rapidly evolving global market lead strategists to consider obtaining specific functions externally through alliances with other firms. This decision requires evaluating potential partner strengths against strategic gaps. The above statement of Capon and Glaser (1987) specifies acquisition of technology rather than products. However, technology must be packaged as a product at some level for a specific target market. Therefore, development of marketable products rather than technology is the appropriate subject of analysis for marketers . 6 The second portion of the statement calls for use of a technology portfolio to identify strategic technology gaps. A useful portfolio would include both product and market elements to assist a firm in identifying specific strategic gaps, because attention to the marketability of technology is appropriate for marketers. In summary, product acquisition and market access strategies appear to be elements which are critical to success in the global environment. Firms face parallel decisions for managing these elements more frequently in today's global environment. Not only must each element be addressed individually and in tandem, but each possesses two dimensions, depth and breadth. That is, depth of commitment to a single product or single market and breadth of commit- ment across several products or several markets. Once strategic elements and their dimensions have been identi- fied, appropriate internal commitments and external alliances must be planned. Though not fully understood, product acquisition and market access, their dimensions of breadth and depth, and use of strategic alliances have all been conceptually related to success. . In Chapter II, the product acquisition and market access elements are more fully addressed and a portfolio is defined which is capable of considering them in tandem. The empirical portion of this dissertation, Chapters IV, V, and VI, relates these elements to desired performance outcomes. 7 The following section formalizes the empirical objectives of the study. Research Scope and Objectives In this section, the scope of the study is first defined, then three study objectives are presented. The scope of this research can be described as global because it addresses the strategy of firms which must operate across many national boundaries. One way of determining whether or not a firm is global is to examine the forces which shape its strategy. These forces have been called globalization drivers (Yip working paper) and are addressed in Chapter II. Because the same or similar environmental forces generally affect all the firms in an industry, one method to control for differences in globalization drivers is to study a Sing 1e industry . In this dissertation, innovative ethical pharmaceutical manufacturers are studied. The industry is experiencing a high cost of R&D, a lengthening R&D process, and a rapid rate of technological change. All three of these environ- mental forces add to the pressure to rapidly commercialize p1“><311cts which contribute to the drive for globalization. In fact the industry has decades of global marketing experi- ence and is well suited to a study of global strategy. Constraining the scope of the study to a single indus- try sample also provides relative homogeneity with respect to Study of product acquisition and 1regulatory structure . 8 market access is especially interesting in pharmaceuticals because promotion and pricing are effectively constrained by regulation. The pharmaceutical industry is also attractive for a global study because a major portion of its global market is represented by six countries which have similar product needs. The study countries which represent approximately 75% of the world pharmaceutical market include the United States, Japan, West Germany, Italy, France, and the United Kingdom. Many firms engage in other non-pharmaceutical busi- nesses. This study addresses only their pharmaceutical lines-of-business which may be comprised of the entire firm, a Single strategic business unit, or more than one strategic bus iness unit. There are three research objectives. The first two research objectives examine strategic behavior in a teC-Zl'xhologyobased, highly regulated, and global industry. P15'<>4 7 Source: TeII 1565, p. :53 antibiotic market was controlled by the same firms, 40% was accounted for by only three firms. Thus, niche markets can be considerably more concentrated and lucrative (Yashikawa 1989). The total number of pharmaceutical manufacturers is difficult to assess because of the relatively small market share held by all but the top 100-200 firms. However, innovative firms are highly visible because of their high sales volume and top-performing drugs. In 1974 in the United States the Pharmaceutical Manufacturers Association (PMA) had 135 members and an estimated 650 firms were not members. At that time PMA members accounted for 95% of industry volume (Silverman and Lee 1974). Current lists name 119 member firms (PMA 1990, p. 28). Japan currently has over 1,300 pharmaceutical firms, most of them are newer and smaller than in the United States and the European Community (EC) firms. The total number of EC firms is in excess of 2,700 (Burstall 1985). Of this large group of firms, only 70 about 190 meet the definition of an innovative ethical drug manufacturer used in this study. nv ro e tal C t t The channel through which pharmaceutical products reach the marketplace is anything but straightforward. The end- consumer generally does not select the product and, more often than not, does not fully pay for it. The industry has directed the majority of its marketing efforts toward the physician who fills the role of learned intermediary or surrogate buyer in writing prescriptions. The regulatory environment surrounding the industry has focused upon health and safety, antitrust, and patent protection issues. The Federal Drug Administration (FDA) in the United States and its counterparts throughout the world require extensive drug testing for efficacy and safety. This process involves a minimum of two to five years pre- marketing commitment for each developed drug in each clinical market. The industry has become accustomed to simultaneously managing parallel testing and administrative processes to meet a multitude of regulations. Industry-wide positive economies of scale are primarily driven by two forces: the high cost and uncertainty of research and development and the regulatory environment. The highly regulated nature.of the industry has created the need for a large amount of administrative support to manage the 71 plethora of documentation required to bring even a single product to market. Civil law in the United States and many European coun- tries also plays a large role in strategic decisions for these firms. Product liability is a major industry concern. Many countries have rapidly put significant legislative reforms into place in both Europe and the United States in the wake of the thalidomide tragedies. The sensitive nature of pharmaceuticals carries heavy legal and moral responsi- bilities. Perhaps due in part to this sensitivity, the industry has been described as highly conservative in lobby- ing efforts that address safety issues. so c v o e The productivity of research and development is uncer- tain. In Japan approximately 7,000 compounds are investi- gated to develop one which is pharmaceutically promising (Yashikawa 1989). Estimates in the European Community (EC) range as high as 10,000 to one. This hit ratio requires a large investment in R&D which is difficult for small scale firms to finance. The cost of developing a single new medicine is estimated to be $231 million U. S. dollars in 1990 (PMA 1990, p. 5). R&D expenditures are reported to be 16.8% of sales in 1990, up from 11.7% in 1980. In absolute terms, current growth of R&D expenditures is at a rate of 12.3% annually. 72 New Chemical Entities (NCEs) and New Drug Applications (NDAs) are indicators of industry research and development productivity. Between 1961 and 1980, 710 NCEs were registered by the EC firms, 353 by U.S. firms, and 155 by Japanese firms. The origin of major drugs in the EC market in 1982 included 175 from EC firms, 55 from U.S. firms, and 17 from Japanese firms. Sources for this data (Burstall 1985, p. 119) comment that Japanese firms have been conspic- uous by their absence from the EC market. Between 1961 and 1980, 353 new drugs were introduced in the United States and only 153 in Japan. One explanation for these differences is the relative immaturity of Japanese manufacturers compared to United States and EC firms. Japanese firms began mar- keting a total of 61 NCEs in 1987 and 53 in 1988. e a s s The world's largest pharmaceutical markets are the United States, Japan, and Western Europe (Sg;ip_Annnnl Bgyigy_1ggg), acCounting for approximately three-fourths of global volume (see Table 1). The domestic legal environment in these markets has a profound effect upon pharmaceutical manufacturers. The U.S. market has an interesting legal history which has shaped the structure of consumer drug procurement and the pharmaceutical industry. A brief presentation of this historical perspective is presented later in this chapter. In Japan, a close second in size, legislation has been 73 highly protective of domestic pharmaceutical firms, though Market-Oriented-Sector-Selective (MOSS) negotiations have increased access to this market. Reimbursement in Japan is strongly affected by its National Health Insurance plan. The Western European market is still somewhat diverse in its regulation but continues to converge under the influence of the EC. Most EC countries also have national health insur- ance plans which affect purchasing patterns and pricing. A country-specific listing of relevant regulation is summa- rized in Table 8. National Health Insurance Plans With the exception of the United States, all countries represented in Table 8 provide nation-wide health care coverage for their citizens. National health plans create two countervailing forces which act upon the pharmaceutical industry. First, the positive impact is removal of the ability-to-pay consideration at the individual level. The negative force stems from the vested interest of governments to constrain drug pricing. National Formulary Systems National formularies appear in the form of positive and negative lists of marketed drugs. Positive lists specify drugs that can be marketed: negative lists specify only those which cannot be marketed. Some countries require a positive listing for any drug to be marketed. Others re- quire positive listing only for those drugs which are 74 TABLE 8 NUMBER OF NEW CHEMICAL ENTITIES INTRODUCED, BY DRUG CATEGORY--1987 and 1988 1987 1988 Cardiovasculars 12 20 Anti-Infectives 14 11 Psychotropics 5 Castro-intestinal 5 3 Antihistamines 2 Anticancers 5 2 Other} 15 10 reimbursed by national health plans. In the United States some private payors use insurance coverage plan-specific formulary systems. Negative lists specify drugs for which there is no national health plan reimbursement. Sometimes this occurs in the form of categories of drugs such as contraceptives and diet aids. Ran Pricing Incentives Pricing incentives are built into some national health plan drug reimbursement schemes. Differential pricing is intended to encourage research and development activity of pharmaceutical firms. Incentives are based upon relative availability of similar drug therapies and/or marketing firm's investment in R&D. For instance, in Japan a drug entering the market will obtain higher reimbursement if it is not viewed as an alternative therapy to a drug already marketed in Japan. 75 Acceptance of Nondomestic Clinical Testing All countries represented in Table 9 require pre— marketing clinical testing. Some accept clinical tests per- formed in other countries to satisfy product registration requirements. Non-acceptance of testing performed in other countries acts as a protectionist barrier against non- domestic pharmaceutical firms. Patent Protection With the exception of Italy, the countries of interest have consistently provided some form of patent protection for pharmaceutical firms. Firms have reportedly avoided entry into Italy to avoid the risk of losing control over their intellectual property. Without patent protection drugs can be reverse-engineered, manufactured, and marketed by competing firms. Sole marketing rights, however, are not guaranteed by patents alone. Compulsory Out-Licensing In some countries, such as Great Britain, firms holding legal patent rights in a country may be forced to out- license production and marketing rights to competitors. The rationale for compulsory out-licensing is to maximize availability of drugs to consumers. The practice also has 76 TABLE 9 SUMMARY OF COUNTRY REGULATIONS France Regulation National we Health Plan National Formulary NO yes R&D Pricing Incentives Acceptance of Nondomestic Clinical Testing Patent Protection since 19 yes but Compulsory [an SC Out- licensing since Generic mm Substitution Best Great Italy Japan Germany Britain since 1979 yes yes yes N0 yes yes yes NO NO IIO yes NO NO yes yes since 1978 yes yes yes yes but yes but seldom seldom N0 yes often IIO NO wnbv u) since 1989 USA NO NO NO YES NO since 1984 Corlpilei from various sources. 77 the effect of price competition early in the commercial life of a product. Generic Substitution Generic substitution is the practice of dispensing a non-branded drug in the place of a branded drug. In most countries generic substitution is possible only if specifi- cally endorsed by the prescribing physician. The United States, Great Britain, and West Germany are exceptions. In the United States, generic substitution for certain drugs is required for reimbursement by the Federal Medicare system. Many state Medicaid policies have emulated this practice. Great Britain is tracking physician utilization of branded products versus unbranded products to encourage generic substitution. The program is likely doomed to failure because physicians have consistently resisted loss of auton— omy in prescribing treatments and reporting diagnoses. The new policy instituted by Germany in late 1989 has a stronger substitution language. Pharmacists must dispense generics unless specifically requested by the physician to provide branded drugs. 0 an S te Industry participants operate within this highly regu- lated environment and necessarily manage market access relative to regulatory issues. The following hypothesis relates regulatory status to strategy: 78 E-3 The number of products a business introduces into a country market is a function of regulations con- cerning compulsory out-licensing, generic substi- tution, national formularies, national health plans, acceptance of nondomestic clinical testing, and pricing incentives. Note that patent protection is excluded from the hypothesis because there is no variation between study countries during the 1982-1987 time period. In thg United States The character of the drug delivery system has evolved almost accidentally as a result of interactions among a number of legislative changes (Temin 1980). The first federal law was the Pure Food Act of 1906 which was amended by the Federal Food, Drug and Cosmetic Act in 1938. The 1906 legislation was the result of public concern over food safety. The major provisions of the legislation were safety oriented and specifically addressed the issue of product adulteration and misbranding (Gibson 1976). In fact, drug consumption at this time was largely of patent medicines. Any drug could be obtained directly from the pharmacist. Physicians' prescriptions were largely used to provide a remedy recipe for the compounding of drugs. Major struc- tural revisions and a new focus on drugs came with the 1938 Act which addressed concerns about the overall safety of chemical substances. This legislation accomplished four 79 objectives: (1) instituted a requirement for New Drug Applications (NDAs), (2) stipulated that drug instructions be provided for consumers, (3) required that drugs be labelled with appropriate warnings (with some exceptions), and (4) outlawed unsafe drug products. This bill did establish procedures for monitoring new drugs marketed: however, the FDA had no process for enforcement of new drug approval. ManufaCturers were free to act unless the FDA interfered. The volume of prescribed drugs was still less than a quarter of the current total and the FDA's original intent was safer self-medication not its elimination. The FDA evidently had a change of heart between passage of the Kefauver-Harris amendments and instituting regula- tions. Temin (1980) hypothesized that the dramatic surge in the physiologic power of pharmaceuticals was the reason. Both therapeutic effectiveness and seriousness of adverse side effects from new drug therapies increased dramatically. Knowledge of clinical usefulness of penicillin increased, spurred its use, and the search for additional powerful drugs increased. Wide-spread use of prescriptions was an effect of regulations put into effect in the two years following the 1938 Act rather than a result of the act itself. These regulations required prescriptions for dispensing a new class of drugs which, according to the FDA, could be dangerous in the hands of the consumer. These newer more potent drugs are repackaged before dispensing to the consumer and therefore exempt from labeling 80 requirements. Consequently the U.S. ethical or prescription drug distribution channel requires a learned intermediary to choose individual drug therapy. Since NDAs were not subject to a critical review process until the 1962 Drug Amendments, this new class of drugs was in effect defined by the drug companies under exceptions to the labelling regulation. The only controls for specific drugs remained those listed for FDA supervision under the Pure Food and Drug Act of 1906, and the Harrison Anti-Narcotics Act of 1914 which is en- forced by the Justice Department. It is also notable that little or no evidence of either American Medical Association or drug manufacturer lobbying accompanied either the 1906 or 1938 Acts (Temin 1980). Based on this, Temin hypothesized that the emergence of prescription medicines was an artifact of regulation, a function of the exceptions to labelling clause of the 1938 legislation. Penicillin was not patentable because it was a known substance prior to establishing its enormous clinical value and it was initially produced by many firms. Streptomycin, on the other hand, was purposefully extracted for clinical use. Merck succeeded in patenting the process of extracting Streptomycin. Though a landmark for the industry, Strepto- mycin did not change the way drugs were marketed. Merck licensed its production process to other manufacturers on an unrestricted basis. Senator Kefauver introduced legislation designed to foster competition among drug companies and increase FDA 81 safety surveillance over drug manufacturing and new drugs. The proposed legislation would have made out-licensing of drug patents compulsory. However, the Kefauver-Harris bill, which was passed in late 1962, did not include this provision. This legislation did make three significant contributions to law. It (1) changed the standards for approving a new drug, (2) required FDA npppgynl of NDA, and (3) gave the FDA jurisdiction over new drug testing. The change in new drug approval standards went from "effica- cious" to "effective." The FDA then commissioned the Drug Efficacy Study from the National Research Council of the Academy of Sciences. The three-year study began in 1966 and its findings were based on the expert opinion of panels of clinicians. The process used to establish product efficacy was the forerunner of FDA product approval for NDAs and over-the-counter drugs. Thus, the worldwide pharmaceutical industry became subject to lengthy testing procedures in its largest market. The 1962 Drug Amendments gave a decided short-term advantage to firms with existing, patented, approved drugs and to firms who had established testing facilities. The overall effect of the new legislation was to reduce the availability of new drugs by lengthening the approval process thereby reducing the effective life of patents due to time spent in clinical testing. In April of 1979 the Carter Administration asked Con- gress to overhaul the nation's pharmaceutical regulations. 82 The overall aim was to speed marketing approval of new drugs and encourage the sale of relatively inexpensive drugs. In May the Pharmaceutical Manufacturing Association (PMA) was denied the right to block federal promotion of generics. Medicare Title XVII was ultimately passed and forced the industry toward a stronger focus on product innovation. New products were needed from a strategic standpoint to bolster shrinking profits. For Medicare recipients to receive reimbursement they must accept generic substitution for all but a specific few branded pharmaceuticals. Medicaid Title XIX adopted similar policies which applied to outpatient drug coverage. Patented drugs receive favorable pricing since no generic substitutes are available. General business legislation also impacted the industry during the 1970s. The IRS tightened loopholes in foreign corporate tax shelters. This specifically hurt those firms with highly profitable subsidiaries in Puerto Rico. Tort law made a mark on the industry during this dec- ade. The Supreme Court remanded the Sindeli vs. Abbopp Lgpgnnpppig§_§p_nip (1980) case back to California courts for trial. Due to the time lag between use of the drug and its deleterious side effects, it was impossible to trace the manufacturer of the actual product. When the manufacturer is unknown this ruling effectively allocated blame to all product manufacturers according to market share (Sheffet 1983). The case established a new doctrine of causation in 83 product liability, which is called market share liability. However, this has been accepted in a only few states. £5232 Japan has been characterized as highly protective of its domestic pharmaceutical industry. The most striking restriction has been the Foreign Exchange Control Law. This law imposed strong controls over market access which stimu- lated the MOSS negotiations. Foreign companies were re- quired to obtain the sponsorship of a Japanese firm in order to apply for government approval (shonin) and a license (kyoka). This requirement was eliminated in the mid-70s opening doors to foreign competition (Yoshikawa 1989). A current constraint to pharmaceutical trade in Japan is its strong national formulary system. Drugs cannot be marketed unless they are listed under the formulary. Addi- tionally, Japan's National Health Insurance (NHI) covers virtually the entire population when it sets prices. The NHI out 1980 pricing 44% by 1986. Industry analysts have noted that new products receive more favorable reimburse- ment. Currently as the world's second largest pharmaceuti- cal market, Japan's favorable treatment of new drugs adds urgency to the global search for new drugs and places spe- cial emphasis on early market entry because the opportunity for favorable pricing can be lost to competing products which enter the market first. 84 W The twelve members of the European Community are striv- ing to eliminate trade barriers that restrict the flow of products, services, capital, and people. These nations retain their independent sovereignty but have agreed to harmonize numerous trade laws to make them competitive in the global market. The pharmaceutical industry is watching the development of an EC with mixed feelings. Easier move- ment of capital and goods, and streamlined policies, will lessen the administrative burden associated with the product approval process. The EC White Paper Agreement calls for unity in product technical and safety standards. The savings to the pharmaceutical industry worldwide is expected to be considerable. On the other hand, the industry fears the spread of policies which reduce allowable reimbursement costs for drug products. An example of one such policy has already been mentioned, the newly enforced West Germany generic substi- tution requirement. The EC must also focus on making social benefits similar among the member nations to prevent workers from crossing borders to obtain better benefits. Countries of the community have long participated in socialized medi- cine. Free movement of workers among the nations could greatly amplify costs to some firms and governments if benefits are not equal among the member nations. The EC acknowledges that in the United States and Canada new drug testing and safety regulations are more 85 stringent than the community's (Burstall 1985). Even with standardized regulation within the EC, firms in member states must continue to meet U.S. standards to access the 29% of the world market controlled by those standards. W In the current regulatory environment, product intro- duction is possible only for the largest firms with pockets deep enough to finance not only research and development but also product introduction. The NDA process is such that considerable investment is required beyond the discovery period for clinical testing and, once approved, large scale promotion. Promotion is required to foster rapid product adoption, maximize product life-span, increase overall profitability, and sustain a first mover advantage in market share. R&D and promotion are very expensive in the pharma- ceutical industry and, because of the large necessary in- vestment, affordable by only the large firms. Competitive market entry and growth of smaller firms are thus preempted in the highly profitable therapeutic classes. However, thousands of smaller firms continue to address geographical- ly smaller and less lucrative market niches. Reimbursement regulations strongly influence the abil- ity of successful firms to compete with an innovative strat- egy. These regulations may reduce the role of pricing when the market might support higher pricing. Assuming substan- tial investment is required for R&D and advertising, an 86 innovative strategy cannot be sustained by smaller firms. Again, the cycle is reinforced in favor of larger firms. As a result, few newcomers have risen to the top 20 global industry players in the last twenty years. Patent Protection The length of time required to clinically test a new drug is justification enough for the industry to be highly protective of its products. However, the R&D process is also expensive and lengthy, adding considerably to the time required to pggin marketing a new drug. Additionally, product protection and differences in health and safety regulations among countries require a large amount of admin- istrative effort and expense long before a compound's market potential can be tapped. Even with the most aggressively managed patent policies, companies are not able to effect complete protection. The problem is exacerbated in coun- tries which either lack patent laws or refuse to enforce them. Examples include India and some Latin American coun- tries. Differences have also been significant in the more developed nations. For instance, Italy offered no patent protection whatever between 1939 and 1978. Historically, only Belgium and Panama offer protection as comprehensive as in the United States (Silverman and Lee 1974). The EC Patent Convention offers 20 years protection from date of filing. Many but not all European countries have adhered to the rule. The effective patent life, post- 87 product approval, ranges from eight to twelve years (Key Wise 1988. p- 74) in Europe- e e f - Faro (1990) used a life-cycle model to study regulation governing the introduction and diffusion of hip joint re- placement prostheses. Medical products such as equipment, prosthetics, and pharmaceuticals have similar regulatory environments associated with product introduction, market- ing, and post-marketing stages. The model is adapted here as a framework for discussion of the pharmaceutical industry (see Table 10). During the product development stage regulation ad- dresses protection of intellectual property and development of "orphaned drugs." Orphaned drugs are those which have a socially beneficial use but are abandoned due to excessive cost of further development or product launching. These regulations provide subsidization for the development of drugs which are not commercially viable. Because of their low commercial value these drugs are not of interest in this study. The length of the market approval process is determined by the regulation of clinical trial testing for safety and efficacy and the product registration mandated by law. The approval process varies in length among countries as a function of clinical standards and whether foreign clinical trials are accepted. As already noted, the United States 88 TABLE 10 REGULATORY LIFE-CYCLE OP PHARMACEUTICAL PRODUCTS Life-Cycle Stage Type of Regulation Development Intellectual Property Orphan Drug Laws Market Approval Clinical Trials for Safety and Efficacy Product Registration Manufacture Good Manufacturing Practices Good Laboratory Practices Packaging and Labelling Purchase Prescribing and Dispensing Promotion Reimbursement Government Health Plans Generic Substitution Formulary Price Listing Post-Marketing Surveillance Adverse Drug Reactions Drug Recall Procedures Adapted from Faro (1990). has the most stringent standards for product acceptance. Manufacturing regulations are generally uniform among the developed countries. Much variation exists among devel- oping and third world countries: however, the developed country-markets and regulations are not of interest to this study. Regulations which govern access to ethical pharma- ceuticals also tend to be practiced uniformly in these markets. Some differences in how particular drugs are classified occur with some drugs such as older antibiotics available over-the-counter (rather than by prescription) in some Western European countries. However, most of these drugs are accessed under the recommendation of a physician, 89 if not by prescription. These differences are at the prod- uct level rather than the firm or country level and are therefore not of interest in this study. Restrictions governing promotional practices also influence purchasing behavior. Again, the largest markets tend to be similar in their promotional regulation and practices. Currently promotional efforts are directed almost exclusively at the physician. Some changes are in process, however. In the United States, some ethical drug advertising now targets the end consumer. To protect the consumer, post-marketing surveillance is centered around monitoring adverse drug reactions and drug recall procedures. These generally do not affect marketers except through the promotion variable. However, the United Kingdom is attempting to institute a computerized drug utilization system to monitor appropriate utilization of drugs. Physicians will be required to report each prescrip- tion and the condition for which it is prescribed. It is important to note that the discretionary power exercised by the physician can circumvent attempts to regulate the use of pharmaceutical products. Off-label pre- scribing is the practice of prescribing drugs for uses which have no official clinical testing documentation. Physicians very frequently prescribe off-label and communicate efficacy and adverse side effects through medical journals. It is common practice for physicians to enter secondary or inaccu- rate diagnoses to obtain reimbursement for procedures. The 90 practice can easily be extended to drug therapies to protect the physician's latitude in drug utilization. The intent of the 0.x. system is to justify prescription utilization and reduce off-label prescribing. While the number of prescrip- tions can be monitored if tied to reimbursement procedures, it is unlikely that off-label prescribing can be avoided as long as physicians can exercise latitude in assigning diagnoses. We Marketing strategy can be defined as a set of princi- ples used to adjust the firm's marketing mix to environmen- tal changes. The strategy employed by current industry leaders can be described as emergent which Hofer and Schendel (1978) describe as reactive rather than proactive. Industry channel structure has been hypothesized to be an unintended result of legislative interactions (Temin 1979). Industry leaders were not active in lobbying for industry interests until the 1962 Drug Amendments (Temin 1979). Marketing elements which are controllable by the firm, the classic four Ps, (place, price, promotion and product) are highly constrained in the pharmaceutical industry. Place Place can be evaluated in terms of where the consumer obtains pharmaceuticals, and the territory of the firm. The distribution systems for the pharmaceutical industry are well developed. The vast majority of ethical products are 91 distributed to retailers or physicians via wholesalers in each of the three major markets. Legislative or regulatory changes would be required in the largest world markets to change the point of sale for drugs. Industry rules of thumb dictate simultaneous product introduction in the three largest country markets: Japan, the United States, and West Germany (Hornick 1990). Firms deem it especially favorable to gain rapid access to the Japanese market to ensure favorable pricing prior to intro- duction of a competing product. Market selection and entry strategy are appropriate to the industry and this study. Price Price has been largely mitigated by national health plans in Japan and Western Europe and health maintenance organizations and maximum allowable costs in the United States. Price competition is not viable because it is highly controlled by outside agencies. The one possible exception is a new drug which has no substitute. Pharmaceutical pricing is frequently controlled by country-specific legislation. In this study pricing is viewed as a structural element of country markets and exam- ined in its relationship to market selection and market entry decisions. Promotion Promotion is the most highly controllable variable available to the pharmaceutical manufacturer. Promotion 92 takes many forms including selling, advertising, direct mail, and detail men (sales personnel). A 1970 study by the FDA revealed that physicians rely heavily upon detail men, advertising in medical journals, and direct mail to provide them with up-to-date drug information. Detail men are considered essential to the success of a product. A typical U.K. pharmaceutical company will call on about 17,000 of the 24,500 general practitioners at least once a year. Of these about 6,000 high frequency prescrib- ers are visited three to four times annually (Slatter 1977). The Pharmaceutical Manufacturers Association (PMA) estimates 25,000 U.S. detail representatives were employed at a cost of over $25,000 per employee (Silverman and Lee 1974). Current salary costs can be conservatively estimated at $40,000 per employee. Cooperative agreements between firms with noncompeting products have been made to access domestic detail sales force. For example, Glaxo rented F. Hoffman- LaRoche & Co.'s team of 800. Study of the effectiveness of promotion is difficult because current figures on the promotional aspects of phar- maceutical marketing are very difficult to obtain. Given the high level of sensitivity toward profits obtained in the industry, these costs are generally obscured at the product and line-of-business level in most public documents. Industry promotional practices also tend to be obscured when ail‘lillyzed at the firm level. The study of promotion is more meaningful at the product level since these strategies are 93 product-specific. For instance, in the year a product is launched promotional costs may easily exceed sales. Blitz- ing efforts later in the product life-cycle may be in response to the entry of competing products such as with non-steroidal anti—inflammatory agents and beta blockers. On the other hand, some products such as the anti-ulcer drug Prozac have few competitors, enjoy a large degree of press attention, and require little promotional spending. Product Product has been the area of highest differentiation between firms. Marketing a single successful product has been considered a worldwide prerequisite to the success of a pharmaceutical firm (Slatter 1977). The gestation period for developing a new drug is upwards of ten years. In-house costs prior to the 1962 Drug Amendments were estimated at $500,000 per NDA, costs immediately post-1962 were estimated to be over $7.5 million per NDA (Scherer 1975). Current costs for development of a single commercializable product are calculated to be $231 million (DiMasi et al 1990, p. 3). Joint ventures have also been in evidence at least as early the 19708. Unfriendly takeovers, however, have not been considered a risk until recently as illustrated in this quote from Inc Wnii Signet gnurnni: ... Hoffmann-LaRoche & Co.'s $4.2 billion tender offer for Sterling Drug Inc. marks the end of an era in which an unspoken gentleman's agreement' among major drug firms has all but prevented hos- tile takeover in the industry. (January 4, 1988, P- 4) 94 The number of licensing agreements, mergers, and joint ventures in the news jumped from six in 1979 to twenty in 1988 (Wall spzegt Journai Indiges 1979 and 1988). Product extensions are also highly useful in differen- tiating pharmaceutical products. As many as 200 dosages, forms, and packagings may be used for a single product in the international market. Threats to competitive product advantage include super— seding technology in the form of newer, better drugs: cures for conditions and diseases; change in public attitudes (e.g., towards usage of tranquilizers): changes in medical practice: and expiration of patent rights. The market life of a patent is considerably shortened by the testing and approval processes, however for many drugs their effective patent life is determined by these other threats. There- fore, the competitive advantage of a highly differentiated product must be maximized quickly. Promotion and market access provide this key. Management of the product variable relates largely to the R&D process. Qonclusion The current state of the pharmaceutical industry re- quires a high degree of focus on product acquisition and market access strategies. The rapid pace and high cost of R&D are industry drivers towards globalization (Yip 1991). The legislative environment brings multiple forces to bear upon the industry. These forces simultaneously increase the 95 cost of commercializing drug products while decreasing product pricing. Clearly the challenge of the future is to effectively manage resources to maintain firm performance in both the short and long runs. The methodology for testing the hypothesis formulated in this chapter is presented in Chapter IV. A discussion of the results follows in Chapter V. Then implications for further research are presented in the final chapter. CHAPTER IV RESEARCH DESIGN AND METHODOLOGY introduction and Design ijectives The empirical phase of the study draws upon relevant literature and industry experience to address the research objectives stated in Chapter I. These objectives have been developed and restated as testable hypotheses in Chapters II and III. This chapter addresses study design and method- ology used to empirically test these hypotheses. First a theoretical framework is presented and explained to illustrate hypothesized relationships. For reference purposes hypotheses are restated in families according to study objectives. Variables are then related to the hypotheses with statistical tests. This is followed by data collection methodology and a profile of the study sample. Next, variables are operationalized using secondary data. Finally, the statistical techniques for testing study hypotheses are discussed. This study is designed to explore the relationships between (1) past and current product acquisition and market access strategies, (2) product acquisition, market access, licensing strategies and line-of—business performance, and 96 97 (3) specific regulations and product introduction into country markets. The study also considers the element of time and is designed for a global context. Theopgtical Fpamework The theoretical framework for this study is drawn from the previous discussion of relevant literature and the industry overview. The study specifically addresses global product acquisition and market access in the form of licensing versus use of internal resources. Figure 5 illustrates the relationships predicted in the first two families of hypotheses which relate past to current strategies, and strategy to performance. The last hypothesis relates specific government regulations to pharmaceutical product introduction into country markets. This third set of relationships is illustrated in Figure 5. Since these regulations are fully discussed in Chapter II, only a brief review is provided in the hypothesis section of this chapter. This section will present a full discussion of the framework for the first two families of hypotheses. The number of products previously developed, in- licensed, introduced, and out-licensed are shown as contributing to a business's current product line and access to markets. These measures of past strategy are hypothesized to be related to current internal product development and market introduction strategies. Product acquisition variables are measured within a therapeutic g§m0§m>wEGm._. Emozoo H>9.— § 3980.750 moo—69E. Encoun— wmmog Ems... 832.. 888...... Santa 8:8... zO.....m500< hogan—Omn— >Omk0mh<¢hw 5(a— 99 class and market access variables are measured within a country market. Theoretically, the greater the number of past product developments and in-licensed products, the greater the efforts towards current product developments and introductions within a therapeutic class. Likewise, the greater the number of past in-house product introduction and out-licensed products, the greater the efforts toward current product introductions within a country market. Following this reasoning a step further, the greater the number of total products developed, the greater the number of top-performing drug products. Under current strategy, the framework combines product acquisition and market access variables across therapeutic classes and country markets respectively, to relate them to performance at the level of the pharmaceutical business. This variation is preserved across therapeutic classes and country markets in the framework under diversity because the variation of global products and markets is hypothesized to affect business performance. ProdUct line diversity is functionally defined according to therapeutic class and represented by the number of classes which the business markets. Likewise, diversity of internal product develop- ment is defined as the number of classes in which the business has developed products. The notion of breadth of top-performing drugs is similarly captured. Variation in market access is a more complex measure because the size of the potential markets is also predicted to impact 100 performance. Therefore firm market representation is more useful as a percentage of world volume rather than the number of country markets accessed. Product line breadth, development breadth, and market representation are considered as they relate to performance. Licensing strategies are also hypothesized to have some relationship to general performance. Finally, industry criteria, top- performing drugs and the number of classes of top drugs, are related to sales performance and market share. All current strategy and diversity variables are related to performance at the business level. The next section provides a discussion of the nature of these relationships and sets forth three families of hypotheses which are designed to test them. Hypgpneses Hypotheses developed in Chapters II and III are restated in this section and grouped as they relate to research objectives. A full key to the variables used in this study is provided in Appendix B. An explanation of how the variables are operationalized is provided in a sub- sequent section of this chapter. The reader is also provided a summary of strategy and performance variables in Table 11 and a summary of hypotheses tested in Table 12. 101 TABLE 11 SUMMARY OF STRATEGY AND PERFORMANCE VARIABLES VARIABLES W Products developed Products introduced Wales In-licensed products Out-licensed products 'v s't Breadth Market representation Development breadth DESCRIPTION The number of new products a firm develops The number of new products a firm introduces The number of products a firm licenses fron another firm The number of products a firm licenses pg another firm The number of therapeutic classes in which a firm markets products The percentage of the world market in which a firm has an in- country marketing presence The number of therapeutic classes in which a firm develops products Ingnsppy 2gpfopnnnce Criteria Top-performing drugs Top-performer breadth The number of drugs a firm markets which rank in the top thirty sales performing drugs The number of therapeutic classes in which a firm markets top- ‘ performing drugs fienepni ngfopnance Criteria Sales Market share 1 The sales volume of a pharmaceutical line-of-business in U.S. dollars A surrogate measure'--the percentage of total sales of a constant pool of pharmaceutical firms Note: A surrogate market share measure is used because global pharmaceutical volt-e is calculated only with proprietary data which has not been made available for this study. 102 TABLE 12 TESTS OF HYPOTHESES Independent Dependent Variable(s) Variable Test Ens; and Cunpent gppgpegy 3 2 3 A1x1 + 12x2 91 MR 2 A1x1 + Azx2 P2 MR P1 cor A1x32 across product class A1x‘ P3 .cor. w1th1n countries 0 t te a d e liyz + RQ'ZI’EHYZ + Byr24-Egy2 + Byr S MR 11 2 , 3 4 4 5 see 1 + B7y7 + Bay8 + 39y9 across years and countries 13,1712 + Bzyzz + 133y32 + 3,31,} + 1353752 + 36y: 32 MR + 37y7 + Bay8 + 39y9 across years and countries ato Structure an Market Selection C1z1 + C222 + C3z3 + Cyz,’ + C52S + C‘sz,5 I MR across countries 103 TABLE 12 (cont'd) 511.12.!ariablea East )9 = products developed in-house within a class )9 = products in—licensed within a class 1% = breadth of products developed )9 = total products marketed to a country y1 = products developed in-house y2 = products introduced y3 = in-licensed products y‘ = out-licensed products y5 = breadth of in-house product development y6 = product line breadth y7 = top-performing drugs Ya = top-performer breadth y9 = market representation z1== compulsory out-licensing 22 = generic substitution z, a national formulary z‘ = national health plan 25" acceptance of non-domestic clinical testing 26 = R&D pricing incentives Current Wi'n Cs P1 = products developed P2 = top- performing drugs Within n quntrx P3 = products introduced By Firm S1 = Sales 52 = Market Share CO“ I = Products introduced within a country 104 Past Position and Strategy The first study objective and family of hypotheses predicts a positive relationship between historic product and market position and subsequent product innovations and introductions. The past variables are measured by year using the number of products which a firm developed within a product class or introduced within a country market during the previous nine years. Current year variables are similarly measured for the observed year only. E-la Product developments in a therapeutic class is a function of the number of products a business prev iously developed and the number of products previously in-licensed. E-lb The number of top sales producing drugs the busi- ness markets in a therapeutic class is a function of the number of products a business previously developed and in-licensed. The expected relationship of products developed to current product development and top sales is expected to be positive. The relationship of in-licensed products to subsequent product development is expected to be an inverted U-shaped curve. This is because in-licensing is a form of commitment to a product class but does not build internal resources. Firms which do in-license are building channel resources to effectively market the drugs. Because this variable is calculated upon a fixed nine-year period the downturn of the inverted U may not be present in the data 105 even if it does exist when not constrained to nine years. Therefore, both linear and curvilinear relationships are tested. H-lc The number of current products developed by a busi ness is a function of the number of therapeutic classes in which a business previously developed products. Based upon Foxall's notion of constrained search for new products, the relationship between breadth of product development and current products developed is expected to be negative. E-ld The number of current product introductions into that country is a function of the number of products previously introduced to a country market. The relationships expressed in this hypothesis are based upon Yip's notion of committed involvement (1989), Thorelli's networking concept (1990), and Cavusgil's (1980) progressive internationalization of firms. Previous product introduction and out-licensed products are expected to be positively related to subsequent product introductions in a country. Strategy and Performance Goals The second family of hypotheses is designed to test the relationships between product acquisition and market access strategies and performance. These strategies include 106 licensing, the variety, and diversity of products and country markets, and top sales performing products. Since a firm can acquire products by internally developing or licensing them to market, the number obtained in each manner is used. Likewise, a firm can market its own products or license them to other firms to gain access to markets. The variety of products, product line breadth, is measured by the number of therapeutic classes in which a firm markets products. The variety of products developed is the number of therapeutic classes in which a firm develops products. To capture market representation, the percent of world market represented by each country in which a firm does business is summed. All of these strategy variables capture nine years of historic data plus observed year data. The number of top sales performing drugs and breadth of product development, i.e., the number of therapeutic classes, are measured across the case year. This is to avoid multiple counts of the same drug product and to provide an accurate representation of the current competi- tive environment. H-Za Pharmaceutical business sales revenue is a func- tion of the number of products developed, the numb er of product introductions into country markets, the number of in—licensed products, the number of out-licensed products, breadth of product 107 development, product line breadth, the number and breadth of top drugs, and market representation. H-zb Pharmaceutical business market share is a function of the number of products developed, the number of product introductions into country markets, the number of in-licensed products, the number of out- licensed products, breadth of product development, product line breadth, the number and breadth of top drugs, and market representation. The relationship of the number of products developed and the number of products introduced to sales and market share is expected to form an inverted U-shaped curve. This is because experience with products and markets should lead to better performance up to a point. Beyond that hypotheti- cal point, efforts may be spread too thinly to be effective. This is consistent with Foxall's (1983) notion of con- strained search for new products and markets. For the same reasons breadth of product development is expected to have an inverted U-shaped relationship to performance measures. However, the nature of strategy studies tends toward a bias of successful firms, consequently the downswing in the inverted U may not be visible in the data. Therefore, these three variables are tested for both a linear and a curvi- linear relationship to sales and market share. However, market representation, as the percent of world .maziket in which a firm does business, should have a linear and positive relationship to performance measures. This 108 assumes that the industry must operate in a global environment to sUrvive. This is consistent with industry expert opinion as expressed in Chapter II. It is also consistent with the review of the industry and its globalization drivers which was presented in Chapter III. The relationship of in-licensed and out-licensed products to sales and market share is also expected to be an inverted U-shaped curve. This relationship is a proposed synthesis of two opposing viewpoints. The upswing in the curve is based upon Ohmae's (1989) hypothesis which expresses a positive relationship between use of strategic alliances and performance. The downswing in the curve is based upon Levitt's (1990) hypothesis that long-term performance will suffer with use of strategic alliances. Both of these variables may not demonstrate the downswing in the inverted U because they are constrained to a nine year period. Therefore these two variables are tested for both a linear and curvilinear relationship to market share and sales. The relationship between top-performing drugs and breadth of top performers is expected to be linearly positive. There are two reasons for expecting this relationship. First, top-performing drugs and top-performer breadth are in themselves a subset of current performance arui would be expected to remain positively related to sales and market share at any level. Second, industry experts Predict this relationship. 109 Regulation and Market Decisions The relationship between government regulation and pharmaceutical line—of—business strategy is examined with this hypothesis. Choice of markets and timing of entry into country markets is thought to be affected by the regulatory environment. The specific regulations analyzed are (l) compulsory out-licensing which requires firms to license products to competitors to gain market access, (2) generic substitution which permits nonbranded drugs to be replaced by branded drugs, (3) national formularies which require every drug be approved and listed by the regulatory agency to be marketed, (4) national health plans which often restrict reimbursement for pharmaceuticals, acceptance of nondomestic clinical testing which can reduce drug approval time, and (5) pricing incentives which reward innovating firms with higher drug reimbursement. Patent protection is excluded from the hypothesis because there is no variation in the study sample. The following hypothesis addresses the relationship between country regulatory structure and a business introducing a pharmaceutical product into a country: H-3 The number of pharmaceutical products a business introduces into a country market is a function of regulations concerning compulsory out-licensing, generic substitution, national formularies, 110 national health plans, acceptance of nondomestic clinical testing, and pricing incentives. The relationship between each regulation and product introduction into a country market can be predicted based upon constraints to competition and barriers to entry. Compulsory out-licensing, generic substitution, national health plans and national formularies are anti-competitive forces. These are expected to be somewhat negatively related to timing of market entry. Acceptance of non- domestic clinical testing and R&D pricing incentives are pro-competitive forces which should positively affect product introductions. Data Collection This section describes the data collection technique and profiles the study sample. Secondary data sources are presented for all data elements. Then selection and relevant characteristics of the study sample are discussed. The basic unit of analysis is the pharmaceutical line- of-business accompanied by measures of product acquisition and market access strategies and performance. Data has been gathered for the time period 1982-1987 because it was a period of dramatic industry change. This change was characterized by impending patent expiration of many top- Performing drugs, an increase in licensing activity, and the beginning of industry consolidation. To calculate some 111 variables information from nine additional years (1973-1981) is required. Multiple sources were used to obtain the data required to test hypotheses. However, for all variables, the same source was used to measure all observations. Product acquisition and market access data are measured from two proprietary data bases. The Drug Product Indices are maintained by Paul De Haen International, Inc. (De Haen 1989 a and b). This data is a collection of worldwide product information and is updated annually. Data is maintained on marketed drugs in eight countries. On-line data searches and professional medical and pharmaceutical journals are routinely accessed by Paul de Haen International, Inc. for current information. The data bases are widely accepted by the pharmaceutical industry (Schifrin 1983). The Paul de Haen International data spans the years 1970-1988 and the U.S. data spans 1950-1988. Each data base consists of nine elements: non-proprietary drug name, trade name, therapeutic class, manufacturer, country manufacturer, originating firm, originating country, date of beginning marketing in-country, and country market. The international database focuses upon the Western European and Japan markets with over 3,000 entries and the United States database with over 5,000 entries. Each observation in the data base represents the initial marketing of a product in a country. One entry is made for each drug, or single chemical entity (SCE) , in each 112 country. Only one entry is made at the time marketing begins regardless of the number of product forms, dosages, or trade names eventually marketed. Information is recorded by SCE, which is the unique name associated with all product forms, dosages, and trade names. The International and United States Paul de Haen data bases have been merged for the purposes of this study. The product and market sample from which variables were calcu- lated includes all single chemical entities represented which are marketed in the six major markets and belong to one of the nine major therapeutic classes presented in Table 13. These therapeutic classes were selected to correspond to the major causes of mortality in the study countries (see Table 3). These causes were very similar among countries and provide a degree of sample homogeneity with regard to country markets. A longitudinal data base has been constructed by pharmaceutical line-of-business and year. Variables which measure product development and introductions are calculated from these merged de Haen data bases using a series of crosstabs. 3,395 single chemical entities were used to calculate these variables for each pharmaceutical line-of- business. Selection of pharmaceutical lines-of—business was based uPorn sales performance data which was obtained from 5.9.21.9 M (1983-1988) . By virtue of inclusion in the W, these firms rank in the top 100-150 113 TABLE 13 SAMPLE SINGLE CHEMICAL ENTITIES BY THERAPEUTIC CLASS Number of Therapeutic Class Single Chemical Entities Antihistamines and 71 Antiallergy Anti-infectives 949 Antineoplastics 192 Cardiovascular Agents 471 Central Nervous System 674 Eye, Ear, Nose, and Throat 70 Gastrointestinal 184 Hormones and substitutes 411 Skin and mucous membrane 373 Total Sample 3,395 114 sales producers. However, the number of businesses which compete globally is small. Therefore, with reasonable assurance, small firm bias is avoided within the set of global competitors. Table 14 provides a breakdown of sample businesses by country. Country markets were selected to represent the majority of worldwide pharmaceutical sales volume. France, Germany, Great Britain, Italy, Japan, and the United States represent approximately 66% of the global pharmaceutical market (Sgpip Xenppngk_iggg, p. 21). These countries also represent a homogeneous market group in terms of economic development and therapeutic needs. Other country markets represent two percent or less of the world market and are therefore less likely to affect the global strategies examined here. Country information concerning regulatory status has been gleaned from a combination of secondary sources including interviews with industry executives, legal journals, on-line databases, and reference books. Table 9 provides a summary of regulations by country. The original sample includes 101 top-performing global pharmaceutical firms during the six year period 1982-1987. 0f the potential 606 observations by firm (101 firms times six years) 504 observations were constructed using data at the line-of—business level. The range of sales performance of sample business is $750 million to $ 16,340 million in 1982, and $1,520’million to $26,800 million in 1987. 115 TABLE 14 PHARMACEUTICAL LINE-OP-BUSINESS SAMPLE BY DOMESTIC COUNTRY L Number of Country Pharmaceutical Lines-of—Business Japan 37 United States 26 West Germany 9 United Kingdom 8 France 6 Italy 5 Switzerland 4 Sweden 2 Belgium 1 Denmark 1 Finland 1 Netherlands 1 Total Sample Size 101 116 e a iona io o h v e The longitudinal nature of this study is captured in two ways. First, the data base covers a six-year time period and is constructed by year. Second, although observations are pooled across years, use of moving counts in calculating past variables (nine years) incorporates information and current year variables into a pooled cross- sectional analysis. These variables capture the number of products a firm markets by therapeutic classes and by country during their peak commercial period as defined by average effective patent life. Variables measured include products developed, breadth of product development, products introduced, in-licensed products, out-licensed products, product line breadth, market representation, top-performing drugs, and top- performer breadth calculated for the time period 1982-1987 from the proprietary data base. Information regarding the number of top sales-producing drugs by therapeutic class and sales is also compiled from Sg;ip_Lengne_Inplg§ (1983-1988). Since actual world volume is proprietary information, when available, a surrogate measure of market share must be used. These are be calculated for each company by year using a sum of actual annual sales from all businesses which consistently appear in the six study years. Longitudinal studies are subject to changes in measure- ment over time. An obvious but controllable measurement change in this study is the value of financial statistics. 117 These values are corrected for inflation using Gross National Product deflators (U.S. Bureau of the Census 1983 and 1990) prior to analysis. All strategy and performance measures relate directly to the pharmaceutical line-of-business. Variable measures were derived by an appropriate count of single chemical entities (SCEs) from the Paul de Haen data base. The Paul de Haen variables are defined in the following section followed by Operationalization of the strategy, performance, and regulatory variables. Paul de Haen Variables Used Singig Qnenical Entity (SCE) The non-proprietary drug name uniquely associated with a chemical compound regardless of product form, dosage, or brand name. T u 'c 58 THERA The therapeutic class of the single chemical entity as recorded by the American Hospital Formulary System. This study uses the two digit codes associated with the thera- peutic classes in Table 2. MAIKEIIDQ_Ei£m_iMAEEEIEBI The manufacturer and marketer of a single chemical entity for a country market. QQHDLIY_M§I£§§_iQQflNIBXI The country in which this single chemical entity is marketed under the brand name specified. 118 W The year in which this single chemical entity was introduced in this country market. r n nu tur 0V R The firm which developed this single chemical entity in-house. Pharmaceutical Line-of-Business Data al 'n -o -B s' 58 Designates the line(s)-of—business of a firm which are engaged in ethical drug manufacturing. The same codes are used for INNOVATR and MARKETER in the merged Paul de Haen data base. W The year for which case data is recorded. -_.._ : D- - OO‘d w' in . he_2oe ' .ss 'RI_V Innovative efforts are measured by a count of SCEs, within a therapeutic class, which are self-innovated and marketed by a business. The historic variable is a moving count of SCEs introduced during the nine previous years. The nine year time period corresponds to the average effective patent life of pharmaceutical products during the study time frame. The therapeutic classes examined in the study are summarized in Table 2. Count {SCEs within THERAPY for nine previous years (past )} and {SCEs within THERAPY for YEAR (current)} where INNOVATOR = PLB and PLB = MARKETER}. 119 7. a ' .._ : I- - ace: 3, . ,1e 39‘- '_ _.s—- _-'u This variable is the sum of current and past PRODDEV across therapeutic classes for each business and represents ten years of information. d v 1 m nt EVC S This variable is the number of classes in which a business has developed product during the current year or previous nine years. ,- ' 61-1! '_o- s W_tgin . - .oe fc _.ss I L In-licensing arrangements are measured by a count of new SCEs introduced. The past variable is a moving count of SCEs introduced during the nine previous years. The nine year time period corresponds to the average effective patent life of pharmaceutical products. The current variable is a count of SCEs introduced in the case year. Count {SCEs for nine previous years within THERAPY (past)} and (SCEs for YEAR within THERAPY (current)} where MARKETER = PLB AND MARKETER not = INNOVATR}. In this study, INLIC is used only as an intermediate variable to calculate PRODNO. o - e se Produ s N The total number of drugs a firm has acquired by in- licensing is the sum of current and past INLIC. 0 -, u..." 0 _!1-°S u: k “0. t”. -. 1‘ -Oetl' as: The number of drugs a business markets within a therapeutic class is the sum of TPRDEV and TINLIC. 120 W Breadth of product line is the sum of the number of therapeutic classes marketed in the current year or previous nine years. 02 1 .._ 'us ' a! . 0-1 e ,._ .- '1, R0 This variable is a count of the number of SCEs a business introduces in a country. The past variable is a moving count of the SCEs introduced in the previous nine years. The current variable is a count of self-innovated and marketed product introductions within a country. This is a measure of a business's commitment to a country market. To calculate: Count {SCEs where YRINTRO = YEAR - 1 to YEAR - 9 (past)} and {SCEs where YRINTRO = YEAR within a country (current)} and MARKETER = case PLB. t d tio cros C u e LIEBINTBQI This variable is the sum of current and past (nine previous years) product introductions across country markets for each business. - O I Products innovated but not marketed by a business are out-licensed to access markets. The past variable is a moving count of SCEs innovated by a business and licensed to another business for marketing in a specific country during the previous nine years. The current variable is a similar count for the case year. To calculate: Count (SCE where YEAR = YRINTRO and PLB = INNOVATR and INNOVATR not = 121 MARKETR. In this study, OUTLIC is used only as intermediate variable to calculate MKTNO. Topgi gnp-iieensing (TOUTLIC) This variable is the sum of current and past (nine previous years) products out-licensed across country markets. WI The number of drugs marketed within a country is the sum of TOTINTRO and TOTOUTLIC for that country. ankeping Beppesentapion ncnoss Counppy Markeps (METREP) Marketing representation is measured as the percent of world market accessed by a business. This variable is calculated by multiplying the percent weight for each country market in which the firm markets a therapeutic class times the number of products marketed in that class, and summing across countries. This product is then divided by the total number of drugs marketed in that country. The highest value for this study variable is 66, which means that the business markets in sixty-six percent of the world ethical pharmaceutical market. Values used to weight country markets are provided in Table 1. To calculate: E (MKTNO + TPRDEV) x country market weight. S ss OP G Market-winning innovations for the current year are represented in this variable. Single chemical entities which ranked in the top thirty global sales performers during the year are counted. 122 - ucce s c s R This variable captures top product performer breadth. The information is recorded as the number of therapeutic classes in which the pharmaceutical line-of—business has a top-performing drug during the year. This variable provides a measure of the pharmaceutical line-of-business against industry criterion one which is discussed on page 49. e- - us' a 8 Revenue of each pharmaceutical line-of—business is recorded in current year dollars and deflated using Gross National Product deflators (U.S. Bureau of the Census 1983 and 1990) prior to analysis across time periods. faint ‘_ a "O "_ 1“: 9°. ‘ 9'. '- .!_-i_s._ The size of the world pharmaceutical market is not easily determined nor readily available. Therefore a surrogate market share measure is computed relative to the sales volume of the entities which consistently appear in every study year. Summed sales of these businesses is used as the denominator for the entire study period. To compute: SALES + (E SALES of the firms which are represented in each study year} The Regulatory Environment Hypothesis three examines a business's product intro- duction decisions relative to country regulatory structure. Information was obtained from industry literature and regulatory agencies to dummy variable regulatory issues over 123 the study time frame. These variables are coded as 1 if there is a formalized regulation and 0 if there is no formalized regulation. W The study is conducted across six countries. This categorical variable represents one of the following: France, Great Britain, Italy, Japan, the United States, or West Germany. anpnlsgpy Out-Licensing (COMPLLQ) COMPLIC records the presence or absence (1,0) of mandatory out-licensing within a country during the year. nc u a in Ge ric S s itu ion EN R C GENERIC records the presence or absence (1,0) of a regulation which encourages generic substitution within a country during the year. Fo u a S s e 0 FORM records the presence or absence (1,0) of a national formulary system within a country during the year. WW). PLAN records the presence or absence (1,0) of a nationally funded health care plan within a country during the year. 0-1 e. lo ‘9 -1c- 0 o '10u‘S 7c C 51' a_ '1 s CLINIC records the presence or absence (1,0) of a country's acceptance of other non-domestic clinical trials during the year. 124 'c 'ves N N INCENT records the presence or absence (1,0) of pricing policies which provide an incentive for R&D investment. In summary, these variables are calculated from reliable secondary sources to test research hypotheses stated earlier. All strategy and performance measures represent actual recorded behaviors and/or performance. The regulatory measures are dummy variables which correspond to the formal regulatory environment of each country by year during the study period. Met 0 o a s Because this study is limited to specific relation- ships, a full path analytic technique is not employed. Rather, a combination of univariate, multivariate, and modified time series techniques are used. For each hypothesis, observations are pooled across years for empirical testing. However, though the statistical tests themselves use pooled data, multiple regression, and in some instances simple correlations, the time series nature of the study is captured in Operationalization of the variables. Variables are calculated in a modified time series fashion to calculate moving sums incorporating nine prior years of information. _ In the first family of hypotheses, both univariate and multivariate analysis is indicated to examine the relation- ship between past and current strategies. Each observation 125 by firm and year provides observations in nine therapeutic classes and six country markets. Therefore hypotheses 1a, lb and 1c are tested using 4,536 observations and hypothesis 1d is tested using 3,024 observations. A bare bones meta- analysis is used to test and correct for sampling error prior to pooling (Hunter and Schmidt 1990, p.100). Hypotheses 1a and 1b are then analyzed using multiple regression analysis. A quadratic polynomial is used for in- licensed products because the relationship is hypothesized to have a curvilinear monotonic relationship to further product development and top drugs. Hypothesis family 2 is multivariate and uses quadratic polynomials for products developed, breadth of product development, products introduced, in-licensed products, out- licensed products, and product line breadth because they are all hypothesized to have a curvilinear monotonic to both performance measures. Hypothesis 3 is analyzed using multiple regression analysis. All multiple regression analyses are first tested using backwards step-wise multiple regression analysis. To ensure homogeneity of error distribution among independent variables, a second simultaneous regression equation is then run with only those variables found to be significant in the step-wise regression (Cohen and Cohen 1983). Standardized regression coefficients are reported and used for comparison purposes among the predictor variables. 126 In summary, this chapter has presented a theoretical framework for testing study hypotheses which relate global product acquisition and market access strategies to performance. The objective of the study design and method- ology is to measure and test these relationships over time. Chapter V presents an analysis and discussion of results by hypotheses. Then, study implications for global strategy literature, marketing managers, the pharmaceutical industry, and policy-makers are explored in Chapter VI. Recommendations for further research conclude the chapter. CHAPTER V ANALYSIS AND RESULTS W This chapter presents the analysis and the empirical results associated with each study hypothesis as presented in Chapters II through IV. First, the analysis and results are presented in three groups by families of hypotheses. Then, these results are related to the study objectives as stated in Chapter I. The reader is referred to Table 15 for a summary of the acronyms and variables which are used during the following discussion. Tesping of gypotnpsgs The first family of hypotheses is concerned with the relationships of past product acquisition and market access strategies on the one hand: and on the other hand current strategy, firm sales, and product success. The second family of hypotheses is concerned with the relationships of firm sales and product success to product acquisition, market access, and licensing. The third hypothesis family relates product introductions to regulatory issues which impact the pharmaceutical industry. 127 128 TABLE 15 SUMMARY OF ACRONYMS AND VARIABLES ACRONYMS VARIABLE DESCRIPTION lPRDEV Products previously developed in-house within a class CPRDEV Products currently developed in-house within a class HIILIC Products in-licensed within a class UEVCLASS Breadth of products developed in-house HPRIITRU Products previously introduced into a country CPRTITIO Products currently introduced into a country HPIUDUCT Products introduced into or out-licensed to reach a country market TPRDEV Total products developed in-house across classes TPRIITIU Total products introduced across countries TIILIC Total products in-licensed jggn another firm TOUTLIC Total products out-licensed in another firm SREADTH Product line breadth across classes IKTREP Market representation TUPDRUG The number of top-performing drugs by firm TOPTHEI The number of therapeutic classes in which a firm has top-performing drugs ASALES Firms sales were totaled across sales in all countries as reported in U.S. dollars and then adjusted by annual Gross National Product Deflators for observed years SHARE A surrogate measure of firm market share in observed year. CUIPUL Compulsory out-licensing regulation GENERIC Generic substitution regulation FUII National formulary system PLAN National health plan CLIIIC Acceptance of non-domestic clinical testing *mA sUrrogate market share measure 1s used because global R80 oricing incentives regulated market size is availEEIe_DhTy_thqugh proprietary sources which were not made available for this research. 129 Hypothegis--znniiy One E-la In-house product developments in a therapeutic class is a function of the number of products a business previously developed and the number of products previously in-licensed. The sample is composed of 4,536 observations, with 504 observations in each of nine selected therapeutic classes (see Table 13). Using a backwards stepwise regression analysis, each of the two independent variables, past in- house product development (HPRDEV) and past in-licensing (HINLIC), was regressed on the dependent variable, current product developments (CPRDEV). The variable, products previously developed in-house (HPRDEV), is hypothesized to be positively related to current product development (CPRDEV). The number of products previously in-licensed (HINLIC) is hypothesized to positively affect the subsequent effectiveness of in-house product development up to a point and then negatively beyond that point. Therefore, HINLIC was used as a quadratic polynomial. HINLIC was also modelled as a monomial since the U-shaped relationship may not be apparent because the element of time was constrained to nine years in this model. See Table 16 for a summary of results of both models. Figure 6 also provides a graphic representation of the results for the entire first family of hypotheses. The quadratic regression equation first loaded both variables resulting in an adjusted R2 of .1170 but was 130 TABLE 16 SUMMARY OF RESULTS--HYPOTHESIS 1A AND B PAST IN-HOUSE PRODUCT DEVELOPMENT AND IN-LICENSING STRATEGIES STANDARD BETA ERROR OF T - ADJUSTED WEIGHTS BETA STATISTIC R2 H-la: Dependent Variable - Current products developed Linea; egnation HPRDEV .341 .016 21.8418 .1301 HINLIC .040 .016 2.5678 a 'c ion HPRDEV .342 .034 10.021. .1156 HINLIC2 n.s. H-lb: Dependent Variable - Top-performing drugs Linear egnation HPRDEV n.s. - - .0008 HINLIC .032 .015 2.1408 uad 'on HPRDEV n.s. - - .0000 HINLIC2 n.s. - - a - denotes significance at the p < .001 level 131 PAST STRATEGY CURRENT STRATEGY PRODUCT ACQUISITION .116 PRODUCT ACQUISITION + Products Developed —> + Products Developed + ln-lloensed Products + In-llcensed Products BREADTH 5 \ +BreadthofProducts Developed (sq) Top Performing Drugs MARKET ACCESS MARKET ACCESS .064 + Products Introduced + Products Introduced # + Out-licensed Products + Out-licensed Products KEY: CONCEPTS Variables Vduesreponedare Wquuares. Non-reportedvdueswereuatlstlcalyormenagerlsly non-slums. FIGURE 6 STRATEGIC CHOICES 132 subsequently reduced to .1156 when past in-licensing was dropped from the equation using a backwards stepwise elimination method (probability of F to remove = .05). The linear regression equation provided a slightly higher explanatory power using both variables for an adjusted Rzrmf .1305 with an F statistic of 340.12 (significant F = .000). Two observations are relevant. First, the Beta weights on the linear equation suggests that the past in-house product development variable (HPRDEV) provides approximately ten times the explanatory power of the in-licensing variable (HINLIC). Second, the sign of the Beta weight for past in- licensing is also positive suggesting that some direct relationship between past in-licensing and current in-house product development exists. The positive sign could also be consistent with the left side of an inverted U-shaped relationship as hypothesized. This suggests a longer time period be modelled in subsequent studies if the quadratic curve is to be identified. Generally accepted practice is to utilize lower order polynomials over higher order polynomials (Cohen and Cohen 1983) for the virtues of simplicity, flexibility, and general descriptive accuracy. This is particularly applicable when two or more variables are used as predictors. Given this practice and its higher R2, the linear equation is the preferred model. This simpler 133 explanation implies that the in-licensing strategies have some, but relatively minor, affect upon subsequent product development success. It bears repeating that these past strategy variables reflect a fixed, and perhaps relatively short, nine-year time period. If this time period were lengthened or manipulated in the model, a stronger relationship might be found. Though the linear model is the generally accepted best fit, the discrepancy in findings relative to the quadratic model bears some attention because this study is concerned with the relationship between in-licensing and future product development success. The hypothesized relationship reconciles Porter's (1990) and Ohmae's (1990) theories concerning strategic alliances and predicts an inverted U- shaped curve. The effect size, or R2, associated with both models is relatively small (.130 and .116). However, strategy is a highly complex phenomenon which involves many variables both known and unknown. An effect size of .01 is considered interesting in other highly complex fielded research, psychology for example (Hunter and Schmidt 1991). There- fore, given the complexity of strategy, the significance of the findings, and attribution of the effect to a single variable, an effect size of .13 is of sufficient interest to researchers and managers to merit further investigation of the relationship. 134 Correlations of past in-house product development to current product development and past in-licensing to current product development by therapeutic class were subjected to a bare bones meta-analysis to detect differences in distri- bution among classes. Correlations of both variables to current product developments are shown in Table 17. The analysis showed that 5.38% of the past product development and current product development variables among classes is attributable to sampling error. Likewise, analysis of the correlation of past in-licensing and current in-house product development showed that approximately 16.55% of the variance among classes is attributable to sampling error. The homogeneity chi-square in the first meta-analysis (167.22) and the second analysis (54.38) both show that significant (p < .01) heterogeneity is present in the samples. This suggests that an unknown moderator variable may account for these differences. Hypothesis 1a is supported but warrants additional research. ganagepial Use 0; Past product Acgnisition Information Managers can use information regarding competitors past product acquisition behavior to predict product classes in which competitors are likely to introduce new products. The relationship between past and current strategic behavior was found to be significant but small. However, in-house product development intuitively and statistically represents 135 TABLE 17 CORRELATIONS OP PAST IN-HOUSE PRODUCT DEVELOPMENT AND PAST IN-LICENSING WITH CURRENT PRODUCT DEVELOPMENT WITHIN PRODUCT CLASSES PAST PRODUCT PAST PRODUCT CLASS DEVELOPMENT IN-LI CEN S ING Antihistamines -.0206 .0105 Anti-infectives .4335” .3174” Anti-cancer .4750b .2232b Cardiovascular .3053b .0688 Central nervous .4339” .2038” system Eye, ear, nose, .0127 -.0134 and throat Gastro- .2355b .0412 intestinal Hormones .1131a .0000 Skin and mucous .3370” .1089' membrane O ' fiteO SIMITICODCO It p ‘ .3! b - denotes significance at p < .01 a greater commitment to a product market than does in- licensing. There are two managerial ramifications of this difference in level of commitment to product classes. First, expectations for further product introductions into a particular category should be stronger when products are developed in-house versus externally. This means that research and development activity would be a better indicator of future product introductions than past in- licensing. Second, these findings imply there has been a 136 relatively small risk of future competition from R&D of firms who in-license products in a particular class. Caution should be exercised in generalizing these findings to other industries. In particular industries in which the R&D cycle is short may have intense competition using both in-licensing and in-house product development interchangeably. However, it is likely that these findings can be tentatively generalized to other industries with costly and lengthy R&D cycles. n-lb The number of top sales producing drugs the busi— ness markets in a therapeutic class is a function of the number of products a business previously developed and in-licensed. As in hypothesis 1a, the sample is composed of 4,536 observations by firm and by therapeutic product class, pooled across classes based upon homogeneous findings across classes as determined by meta-analysis. Using a backwards stepwise regression analysis, each of the two independent variables, past in-house product development (HPRDEV) and past in—licensing (HINLIC), was regressed on the independent variable, top-performing drug (TOPDRUG). A summary of model statistics is provided in Table 16. The linear model was significant with an adjusted R2== .0008 and an F = 4.578 (significant F = .032). Past in- house product development was not incorporated into the model. The quadratic model did not load past in-licensing 137 or reach significance with an adjusted R2== .0000. Though statistically significant, the effect size of the linear model is so small as to be of little predictive value. Two explanations can be offered for the small effect size in the significant linear model. The first is the low number of top-performing drugs relative to the total number of products developed in-house, and therefore few top- performing drugs are available for in-licensing by other firms. Second, the short nine-year time span associated with the in-licensing and product development variables may obscure stronger relationships. It is interesting to note the significant relationship, though very small, is between in-licensing and top- performing drugs rather than internally developed products and top-performing drugs. However, the low predictive value of this relationship renders additional analysis trivial in a now univariate model. A meta-analysis was done using the correlations of past in-house product development and past in-licensing with top drugs (see Table 18). The analysis revealed heterogeneity in the correlations between past in- house product development and top drug with a non- significant homogeneity Chi-square of 58.09. This indicates that a moderator variable is operating between past in-house drug development and the incidence of top drugs. However, homogeneity is present with the correlations between past in-licensing and top drugs with a homogeneity Chi-square of 15.49 (p < .05). This meta-analysis indicates 138 TABLE 18 CORRELATIONS OP PAST IN-EOUSE PRODUCT DEVELOPMENT AND PAST IN-LICENSING 'ITH TOP DRUGS UITEIN PRODUCT CLASSES = - - PAST IN-HOUSE PAST PRODUCT CLASS PRODUCT IN-LICENSING DEVELOPMENT Antihistamines .0100 -.0127 Anti-infectives .1315b .0032 Anti-cancer .1155b .0020 Cardiovascular .0015 -.0387 Central nervous -.0068 -.0031 system Eye, ear, nose, .0292 .0395 and throat Gastro- -.0272 -.0197 intestinal Hormones .0392 -.0222 Skin and mucous .0121 .3525b membrane a - denotes significance at 05' P‘- b - denotes significance at p < .01 139 that the small effect size between in-licensing and top drugs is not due to sampling error and holds across therapeutic classes. Findings relevant to both independent analyses direct further research to seek additional variables to account for success in producing top-performing drugs. However, hypothesis 1b is partially, but nominally, supported. Combining the findings of hypotheses 1a and 1b, it is possible to summarize that past in-house product development plays a role in predicting the dependent variable, sales, in some but not all therapeutic classes. However, past in- house product development may or may not play a role in predicting success of producing top-performing drugs. Further research within product classes, or group classes by other relevant criteria, is needed to determine the true relationship. Past in-licensing plays a small role in predicting sales performance and development of top-performing drugs. This role is consistent in relationship to producing top- performing drugs but varies due to an unknown moderator variable in relationship to sales. H-lc The number of current products developed by a business is a function of the number of thera- peutic classes in which a business previously developed products. A Spearman Product Moment Correlation was calculated between the independent variable, breadth of product 140 developments (DEVCLASS), and the dependent variable, current product developments (CPRDEV). The sample is composed of a total of 504 observations with the line-of-business the unit of analysis. A correlation of .1753 (significant at p < .01), with an.13 = .031 was observed. This effect size, though significant, is small. The relationship is positive implying that the broader the product developments across classes, the higher the number of top-performing drugs. This is inconsistent with Foxall's (1983) theory of constrained search across product classes. However, a correlation of .1704, r2 = .029 (significant at p < .01), was also found between DEVCLASS2 and current product developments. A summary of these models is provided in Table 19. While the linear relationship is the preferred model because of its simplicity, the quadratic relationship cannot be ignored. This is particularly true because the sign of the Beta weight for the quadratic is positive, which is opposite to the relationship hypothesized. It is possible that the relationship with the quadratic polynomial might be stronger if the element of time is modelled as a variable. Alternatively, observation of any existing quadratic relationship might be curtailed by a limitation in firm resources which prevents an over-broad commitment to product development. Since the relationships tested in hypotheses 1a and 1b showed significant heterogeneity across thera- peutic classes, different functions might also be found for 141 this hypothesis if the data were split using one or more moderator variables as described in the discussion of the first two hypotheses. Hypothesis 1c is partially supported. TABLE 19 HYPOTEESES 1C AND 1D: SUMMARY OF RESULTS HYPOTHESIS CORRELATION :2 SIGNIFICANCE l-c Dependent Variable--Current Products Developed Linear DEVCLASS .1753 .032 (.01 Quadratic DEVCLASS .1704 .029 <.01 1-d Dependent Variable--Current Products Introduced Linear HPRDEV .2535 .064 <.01 r 1 Im licat o s-- ct c s n t ate 193 Though these findings are inconclusive relative to the question of breadth of search for new products, the relationship between product acquisition strategy and successful product development is of vital importance. Strategic research on an industry scale is difficult because unsuccessful firm strategies remain unstudied due to firm failures and management turnovers. This is an area in which managerial interviews and support of case studies have the 142 potential to contribute significantly to distinguishing successful from unsuccessful strategies. If the hypothesized relationship does not hold as these tentative findings suggest, managers should not constrain their search to a few product classes or categories. They imply quite the contrary, that a much broader search for new products yields better results than a moderately broad search. H-ld The number of current product introductions into that country is a function of the number of products previously introduced to a country market. The sample is composed of a total of 3,024 observations by country and by firm. 504 observations from each of the six study countries are included. A Spearman Product Moment Correlation was performed with the independent variable, past products introduced (HPRODUCT), and the dependent variable, current product introductions (CPRINTRO). A significant correlation of .2535 (significant at p < .01), with an r2 = .064 was determined. This effect size, though significant, is somewhat small. Here again, given the large number of variables which affect market access strategy, the effect size is an interesting basis for further research. The relationship is linear and positive. That is, the higher the number of products previously marketed, the higher the likelihood of current product introductions into a country. Correlations of past product introductions to 143 current product introductions across countries were subjected to a bare bones meta-analysis to detect differ- ences in distribution among countries. The analysis showed that 67.67% of the variance in correlation of products previously marketed and current product introductions among countries is attributable to sampling error. The homogeneity chi-square in this meta-analysis (8.87) is not significant (p < .20) indicating that this relationship holds true across countries. Therefore hypothesis 1d is supported. a a m c s-- r t Although these findings statistically support the notion that existing market relationships are used for subsequent product introductions, from a managerial perspective the correlation is smaller than would be expected for maximizing these relationships. The pharmaceutical industry engages in licensing quite heavily, which provides opportunities to fill existing channels. Given the high cost of developing and maintaining nondomestic market relationships, firms would be well advised to seek additional products to market with their own. This study reveals that these firms utilize in- licensing almost twice as much as other product acquisition strategies. Previous research cautions managers that these products should be chosen to use the same pre- and post- sales support system as their current product line. 144 Marketing in-licensed products also provides a means for preparing a marketing system for future product introductions. ggmgg;1--Hypothegis Family One Although all four of the hypotheses in this family are statistically supported, the actual affect sizes are quite small. It seems prudent to further assess the relationships of pairs of past and current strategies. For this analysis only firms which engage in the current strategy examined were used. Table 20 sets forth Phi statistics (Pearson Chi-square probabilities) for each pair of strategies studied in this family of hypotheses. These statistics provide a useful historic description of firms which engage in certain strategies. The findings are consistent with the multivariate analysis done earlier. However, a closer examination of significant pairs sheds some additional light on licensing motivations. First, as is logical, firms which engaged in product development in the past currently develop products in-house. Second, firms which currently market top-performing drugs have acquired in-licensing agreements in the past. Third, firms currently introducing products tend to have introduced products in the past. Further, the relationship of current introductions is more significant with past out-licensing of the firm products than with past self-introductions. However, the relationship between past and current 145 TABLE 20 RELATIONSHIPS OP PAIRS OP PAST AND CURRENT STRATEGIES Relates to Past Current Phi1 Hypothesis Strategy Strategy 1a HPRDEV CPRDEV .3382 HINLIC CPRDEV .180 1b HPRDEV TOPDRUG .105 HINLIC TOPDRUG .5812 lc DEVCLASS CPRDEV .277 1d apnooucmf CPRINTRO .7222 HOUTLIC CPRINTRO .2643 HPRINTRO CPRINTRO .180 ‘"iSEEE‘ Statistics inc - only true engaging in t e current §EF3E85§"éxanu 1 - Pearson Chi-square probability 2 - Significant a P g .01 3 - Significant a P 5 .05 l. - apawucr is the sun of 800111: and HPRINTRO introductions is stronger with than without past self- introductions. The first significant relationship is readily evident. That is, firms with R&D capacity tend to continue to engage in R&D activities and strategies. The second relationship, between past in-licensing and marketing top-performing drugs, suggests that firms with in- house R&D are more likely to out-license a product than to introduce it themselves. This is important, especially in light of the findings of hypotheses 2a and 2b, which are presented in the next section. The models which follow show a strong relationship between product introductions and performance. It can be extrapolated that to maximize performance, firms which engage in R&D must also develop their channels of distribution. 146 Finally, firms currently introducing products have a strong history of out-licensing. That is, these firms are either learning that they need to introduce products themselves 9; they cannot find firms which will introduce their products into key markets. I gypothesis--zamily ng H-Za: Pharmaceutical business sales revenue is a function of the number of products developed, the breadth of products developed, the number of in- licensed products, product line breadth, the number of product introductions into country markets, the number of out-licensed products, market representation, and the number and breadth of top drugs. The sample for the second hypothesis family is composed of 504 observations by line-of—business over the study time period (1982-1987). Hypothesized relationships included potential polynomials for six of the independent variables: products developed (TPRDEV), products introduced (TPRINTRO), in-licensed products (TINLIC), out-licensed products (TOUTLIC), product line breadth (BREADTH), and breadth of products developed (DEVCLASS). Two explanations are possible to account for a potentially better model using one or more monomials versus the hypothesized polynomial variables. The first explanation involves successful firm bias. The sample may exclude firms which may represent the 147 downswing of hypothesized U-shaped relationships. The second explanation involves measurement constraint which holds the element of time to a fixed period of nine years in calculation of variables. For this reason, a modelling approach was used to determine the best fit of combinations of polynomials as both monomial and polynomial variables. This analysis resulted in a possible sixty-four combinations of variables for each of the two models. In each analysis all nine independent variables are regressed on the dependent variable using a backwards stepwise method. Appendix D reports the adjusted R2 for each model. The independent variables include products developed (TPRDEV), products introduced (TPRINTRO), in- licensed products (TINLIC), out-licensed products (TOUTLIC), breadth of product line (BREADTH), breadth of products developed (DEVCLASS), breadth of top-performing drugs (TOPTHER), the number of top-performing drugs (TOPDRUG), and market representation (MKTREP). Hypothesis 2a models sales, adjusted for inflation, (ASALES) as the dependent variable and hypothesis 2b models a surrogate measure of market share (SHARE) as the dependent variable. The best fit model for each hypothesis was subsequently analyzed with a simultaneous model utilizing only the variables which loaded at p < .05 in the original stepwise equations. This method provides a more normal distribution of error over the variables included in the model and avoids 148 over or understatement of Beta weights as a result of order of entry. 0 c s The model predicting sales tested hypothesis 2a and utilized seven of the nine independent variables to obtain an R2== .589. A summary of this model is provided in Table 21. A graphic representation of both the linear and quadratic model solutions are also provided in Figure 7. The linear model was preferred over the quadratic model because of its slightly higher predictive value and its simplicity. However, it should be acknowledged that many quadratic bivariate relationships are valid, although they did not contribute to the model in the presence of other elements. The following discussion first addresses the variables excluded from the sales model followed by a discussion of the model itself. Relevant bivariate relationships are address along with discussion of independent variables of the model. The model predicting is addressed in the next section of this chapter. Market representation (Y§)'was not included in the final model predicting sales because it did not add significantly to the equation with the highest predictive value. In preliminary univariate analysis, market representation showed significant correlations with top- performing drugs (r = .34, p < .01), out-licensed drugs (r = .27, p < .01), and in-licensed drugs (r= .09, p < .05) 149 TABLE 21 HYPOTHESIS 2A: REGRESSION MODEL FOR STRATEGY AND SALES Quadratic Solugion Ingepaadant Variablas Dependent Variable Y12 Y2 Y Y Ysz Y 2 Y7 Y1 Y9 ASALES :22 .585 Bertaa n.s. -.330 n.s. .860 n.s. n.s. .603 n.s. ms 1 -256 6&fl Lax sig 1 .013 .000 .000 Liaaar Solution Independent Variables Dependent “Mb“ Y, Y: Y! itE Ya Ya Y, Yg Y9 ASALES R2= . 594 Beta' -.184c .151c .102 .234C n.s. .107 n.s. .578c n.s 1 -2.916 2.598 2.716 5.696 2.416 18.700 Sig 1 .006 .010 .007 .000 .016 .000 a - adjusted R2 b - standardized Beta Heights c - denotes significance at the p s .001 level Y1 8 number of products developed in-house Y2 8 nulber of products introduced Y3 8 umber of products in-licensed Y4 8 nmber of products out-licensed Y5 8 breadth of product developments 76 = prochct line breadth Y7 = tuber of top-performing drugs Y3 8 breadth of top-performing drugs Y9 8 market representation market share 150 CURRENT STRATEGY DIVERSITY PERFORMANCE PROWCT AMUISITION BREADTH + meq) 4' WUMMDQ) *1 MOI .151 63$) MARKETAOCESS + Producfflntrodlmd + 0111mm 1.1me KEY N.S.-DDTW STANDARDIZED BETAWElGl-ITS W V.“ nuns-.6“ (.605) «mmmmm FIGURE 7 LINEAR REGRESSION MODEL PREDICTING SALES 151 but did not have a significant correlation with sales or market share. There may be a significant correlation if differences in the drug products which are represented in the various markets is also taken into account. One such difference among these products is age. Forty percent of the observations showed a zero market representa- tion as calculated for this variable. This means that forty percent of the observations represent a year between 1982- 1987 in which a firm did not introduce or in-license a product for marketing to any of the six study countries. Since the historic variables were also accounted for in this variable, it also means that these same firms also did not introduce or in-license products in any of the nine years prior to the observation years. The sample firms represent the top 100-150 sales performing firms in the study years. There are three interpretations which might follow from this observation. First, it is possible that many of the top 100-150 firms primarily market outside of the top six country markets. If this is the case, these firms are giving priority to smaller markets, which is intuitively unlikely. Second, many of those firms that do market within the six countries have not introduced products which are listed in the Paul de Haen data base for the ten-year period observed. Given the thorough and continuous updating procedures used by Paul de Haen International, Inc. and the industry confidence in the databases, exclusion of significant 152 products or a significant number of products is unlikely. Finally, regardless of age of their products these firms can and do produce sizable sales revenues. Therefore it is more probable that the average age of top sales-performing products is high and that these products had already been introduced into the top six markets before or at the beginning of the study time span. Breadth of products (Y3) developed also did not load in the model. This is probably a function of low variation in this variable. Approximately 77% of the sample had a product line breadth of three or fewer product classes. However, it is consistent with the notion of building complementary product lines within markets. Therefore subsequent studies should address product breadth within markets as well as across markets. The variables which were used to predict sales are discussed in order of relative importance as demonstrated by their Beta weights in the model. Breadth of top-performing drugs (Y,) has the highest Beta weight (.574) representing a positive linear relationship with sales. This is consistent with the hypothesized relationship and industry criteria as derived from interviews with industry executives. Simply stated, as the number of classes in which a firm has top- performing drugs increases, firm sales increase. Given the skewed distribution in this variable--on1y Merck has top performers in three or more classes--a second model was run eliminating Merck as an outlier. 153 With Merck eliminated, the R2 dropped slightly (from .589 to .585) and top-performing drugs replaced top- performer breadth in the equation (Beta = .404). These findings are consistent with accepted industry success criteria which tie survival to the number and breadth of top performers. The number of products out-licensed (Y,) loaded into the model with the second highest Beta weight (.234). In the model excluding Merck, this Beta weight ranked first (.880) indicating that aggressive out-licensing of products is an important part in the sales strategy of most firms. This describes a positive linear relationship between the number of out-licensed products and sales. The hypothesized relationship is an inverted U-shape which is not inconsist- ent with this finding. However, if the downswing in the hypothesized relationship is to be found, subsequent studies should expand the number of years represented in the sample. This can be accomplished by increasing the number of observed years or the number of years' data used in the calculation of this historic variable. These findings are consistent with Ohmae's (1989) pro-strategic alliance views. The third largest Beta weight (-.184) describes a negative linear relationship between in-house product development (Y3) and sales. This is inconsistent with Porter's hypothesis supporting in-house development strategies over external acquisition of products. The simple correlation (.244) between the quadratic variable and 154 sales defines a U-shaped curve, suggesting a learning curve and/or economies of scale may be present in the sales success of products developed in-house. The next variable of importance, products introduced (Yé), carries a Beta weight of .151, which describes a positive linear relationship with sales. Products intro- duced is the only variable which dramatically changes by remaining significant but reversing its sign (Beta = -.330) with Merck removed from the sample. Though Merck is an outlier in the sample, its success is envied by the industry and its strategy is of particular interest. The reversal of the sign of this Beta weight illustrates the importance of not only introducing new products, but of introducing good products with effective strategy since, for the model without Merck, the impact of product introductions is negative in the presence of other strategies. The hypothesized relationship is an inverted U-shaped curve. As previously discussed with out-licensing, a negative linear finding is not inconsistent with the hypothesis. Subsequent research should address longer time periods if this relationship is to be found. Constrained firm resources may also prevent manifestation of the downswing in this curve. That is, because product introduction is quite costly, only a limited number of eligible products may be introduced. Product line breadth (Y6,) loaded into the equation with the fifth highest Beta weight (.107), describes a 155 positive linear relationship to sales. This relationship is not contrary to the hypothesized relationship which pre- dicted an inverted U-shape. The sixth most important variable, products in-licensed (Y3), carries a Beta weight of (.102) in this model. The relationship described with sales is positive and linear. As with out-licensing and product introductions, the hypothesized relationship is an inverted U-shape. The argument regarding the time period captured in this model is also relevant to this variable. That is, findings are not inconsistent with the inverted U-shaped relationship which may be found in a model with a longer time frame. Managarial Implications--Linear Model Predicting Salaa This section summarizes the sales model with a review of its major components. These three components are product acquisition, market access, and industry portfolio criteria. First, the two variables relevant to product acquisition, product developments and in-licensed products, were significant predictors in the model. Internal product development shows a much stronger (and negative) relation- ship than did the external strategy, in-licensing. This is inconsistent with speculation from industry executives, which predicts a lesser sales effort for in-licensed products. The reduced effort is thought to be a result of lack of identity with the product, lower sales commissions, lower profit margins, or a combined effect of all three. 156 Both product acquisition variables became non-significant when Merck was eliminated from the sample. The variance became more associated with products introduced and out- licensed products. This is interesting because the shift from strength in in-licensing to strength in out-licensing suggests that Merck's strategies for product introduction are superior to those of the rest of the industry. Furthermore, firms are cognizant of their lack of the necessary ingredients for some successful product launches as is evidenced by their extensive reliance upon out-licensing to produce sales. Managers and researchers should specifically address product introduction strategies to enhance sales performance. Executives should be particularly cautious against over-reliance upon a narrow product line. This need is illustrated by the difference in contribution of measures of product line breadth between the Merck and no-Merck models. This might be accomplished by building in—country sales forces who can market an increased number of in- licensed products as well as reduce reliance upon out- licensing. Both products developed and in-licensed products were hypothesized to have an inverted U-shaped relationship with sales. The number of products developed was demonstrated to have the expected curvilinear relationship. The expected relationship of in-licensed products to sales was neither proven or disproved. 157 Second, the variables relevant to market access, product introductions, and out-licensed products, were also significant predictors. In this case the external strategy, out-licensing, had a stronger relationship with sales than did the internal strategy of product introductions except when outliers were removed. These two market access variables were also predicted to have inverted U-shaped relationships with sales. In both cases only a positive linear relationship was shown. There is still the potential for this relationship over a longer time frame. The model failed to provide sufficient evidence to disprove either Ohmae's (1989) hypothesis favoring use of strategic alliances or Porter's (1990) hypothesis against use of strategic alliances. However, there is support for the benefits of using both in-licensing and out-licensing at least in the short run. In fact, out-licensing was found to have a very strong relationship with sales, much greater than does product introductions, which has also been shown to be negative. In the case of product acquisition, product development was shown to be more closely associated with sales than was in-licensing (although negatively) in the short run. The third component of this model is to examine the relationship of industry-defined criteria to sales. The variables which represent these criteria in the model are market representation, product line breadth, top-performing 158 drugs, and breadth of top-performing drugs. Market representation failed to load on the model but may be an artifact of the nine-year time frame used for the study. Findings relative to product line breadth supported the expected relationship. Not surprisingly, top-performing drugs also demonstrated a highly positive relationship with sales either directly or through top-performer breadth. The most predictive variable for the entire sample was breadth of top-performing drugs, though out-licensing was far more predictive in the absence of Merck. The second model addresses the same independent variables to predict market share. The empirical results associated with the model are presented in the following section followed by a comparison of the two models. Quadratic gegreasion Modal Predicting Market agate H-zb: Pharmaceutical business market share is a function of the number of products developed, the breadth of products developed, the number of in-licensed products, product line breadth, the number of product introductions into country markets, the number of out-licensed products, market representation, and the number and breadth of top drugs. The model predicting market share utilized five of the nine independent variables to obtain an R2== .683. As with the model predicting sales, this model was calibrated a 159 second time with Merck removed from the sample. This resulted in a somewhat lower R? (.612). A summary of both the linear a quadratic solutions to the general model is provided in Table 22. The quadratic model was chosen because of its higher predictive power. A graphic representation of the results of the preferred model is provided in Figure 8. The figure also includes the Beta weights for the solution without Merck noted in parentheses. The following discussion first addresses variables which were excluded from the model, followed by a discussion of the model itself. Three variables were excluded from the model predicting market share: breadth of top-performing drugs, market representation, and in-licensed products. Three variables (top-performing drugs, breadth of top-performing drugs, and market representation) are industry success criteria. One excluded variable, market representation, was also excluded from the model predicting sales. Two of the excluded variables (top-performing drugs and breadth of top- performing drugs) appear to be predictors of sales and market share either directly or indirectly. Market representation and market share did not correlate significantly (r = .050). The variables which were used to predict market share are discussed in order of relative importance as demon- strated by their Beta weights in the model. As expected, 160 TABLE 22 HYPOTHESIS 2B: REGRESSION MODEL FOR STRATEGY AND MARKET SHARE W25 0 e V b e Dependent Variable Y12 Y2 Y YSZ Y; Y7 Y1 Y9 SHARE 112 . .683” set.” -.266” .227” n.s. .237” -.126” .299” .586” n.s. n.s. 1 -6.095 3.253 5.165 -1.903 6.961 16.269 sm 1 .001 .003 .000 .058 .000 .000 Linaar Solution Inganangent Variablas Dependent Variable Y1 Y 494i Y5 Y9 Y7 Yfir Y9 SHARE R2 = .633a Betab -.230” .151” n.s. .2182” n.s. .197” n.s. .665” -.230” 1 -3.557 2.720 5.130 6.813 21.868 -3.557 $16 1 .001 .078 .000 .000 .000 .001 a - adjusted 112 b - standardized Beta Heights c - denotes significance at the p s .001 level Y1 8 nmber of products developed in-house Y2 8 nulber of promcts introduced Y3 8 nulber of products in-licensed Y‘ 8 nwber of products out-licensed Y5 8 breadth of product developments Y6 8 product line breadth Y7 8 number of top-performing drugs '8 8 breadth of top-performing drugs Y9 8 market representation 161 CURRENT STRATEGY DIVERSITY PERFORMANCE PRODUCT ACQUISITION BREADTH + ProductsDeveloped(sq) +ProductUneBroadIh(84) + ln-Iloensed Drugs Top-PerforIner Breadth Top-Pedonnlng Drugs 227 (270) MARKET ACCESS N. s. + Products Introduced + Out-licensed Products .237 (M MOI 8mm KEY N. s. - N01 SIGNIFDANT CONCEPTS Vlldllas STANDARDIZED BETA WEIGHTS n w . .533 (.312) (Values mm m m M FIGURE 8 QUADRATIC REGRESSION MODEL PREDICTING MARKET SHARE 162 top-performing drugs were the highest predictor of market share with a Beta weight of .584. Breadth of product line (Y3) has the second highest Beta weight (.299), indicating an U-shaped relationship with sales. This relationship is opposite to the inverse U-shape hypothesized. As in the model predicting sales, two explanations for this phenomena are offered. First, as firms begin to offer a greater number of products, they experience a learning curve with respect to managing multiple product categories. As they offer an even greater number of products this experience accumulates and is reflected in higher sales and subsequently higher market share. The second explanation relates to specialization. Initial product line expansion across therapeutic classes may not carry sufficient critical mass to warrant specialized marketing and management. As critical mass is achieved, specialization becomes more feasible. Higher sales and market share then result from specialized marketing and management skills. The number of products developed (Y5) loaded into the model with the third highest Beta weight (-.244). This describes an inverted U-shape relationship as hypothesized. This finding is consistent with Foxall's (1983) notion of constrained search resulting in higher in-house development productivity, leading to greater market share. By relating product development to market share, other factors such as promotion, market access, competition and other 163 idiosyncrasies of management are also implied in the equation. It is very informative that even with several of these (e.g., market access, and licensing strategies) included in this model, this inverted U-shaped relationship is third in importance. The fourth largest Beta weight (.237) describes a positive linear relationship between the number of out- licensed products (Yg) and market share. The hypothesized relationship is an inverted U-shape, which is not inconsistent with this finding. Again, if the downswing in the hypothesized relationship is to be found, subsequent studies should expand the number of years represented in the sample. These findings are consistent with Ohmae's (1989) pro-strategic alliance views. The fifth variable in the model, products introduced (Yé), carries a Beta weight of .227, which describes a linear relationship with market share. This is not incon- sistent with the hypothesized relationship, an inverted U-shaped curve. As discussed with product line breadth in the sales model, this finding can be explained by applying the principles of specialization, positive economies of scale, and the learning curve. As firms introduce a greater number of products they experience reduced effec- tiveness as they begin a learning curve with respect to the product introduction process. As they introduce more products their effectiveness increases with experience and is reflected in higher market share. With respect to 164 specialization, initial growth in product introductions may not carry sufficient critical mass to warrant specialized marketing and management. As critical mass is achieved, specialization becomes more feasible. Higher market share then results from specialized in-country marketing and management skills. The sixth and final variable in the model, breadth of product developments (Y3), with a negative Beta weight (-.126) has a negative relationship with market share. The hypothesized relationship is an inverted U-shape. That is, as the number of classes in which products are developed increases, the firm's market share increases. The findings are contrary to the hypothesized relationship. Summary of Strategy Models To summarize, a review of model elements relative to product acquisition, market access, and industry criteria is presented. This summary is accompanied by a comparison of the sales and market share models developed in testing this second family of hypotheses. The market access variables which denote internal strategy (product introductions) and external strategy (out- licensed products) were both significant in the market share model. As in the sales model, out-licensed products loaded more heavily than product introductions, indicating advantages to out-licensing strategies at least in the short run. 165 Both of the product acquisition variables which represent internal strategy (product developments) and external strategy (in-licensed products) did not load into the market share model as they did in the sales model. In- licensed products, while correlating significantly with market share, did not add significantly to the model. Product developments demonstrated the hypothesized inverted U-shaped relationship with market share, as with sales. Breadth of product development loaded only on the market share models with an inverted U-shaped curve. Breadth of product line also displayed a similar relationship with market share, as with sales. In both the sales and market share models the relationship was found to be a U-shaped relationship, which might be explained by the learning curve, positive economies of scale, and/or specialization. Findings of these models can also be summarized as they relate to use of external versus internal resources. The models showed similar rankings within product acquisition and market access. That is, product development was consistently more important than in-licensed products and out-licensing was consistently more important than product introductions. In an overall sense with respect to external and internal strategy, out-licensing ranked first relative to sales, and product development ranked first relative to market share. Product introductions and in-licensing retained third and fourth strategy ranks respectively across the two models. Taken as a whole, these findings seem to 166 indicate that internal product acquisition and external market access are the stronger strategies in a sales producing and a competitive sense. These relationships may shift if studied over a longer period, but do have managerial implications which are discussed in Chapter VI. Major differences in the importance of the industry success criteria were found between the sales and market share models. Both the sales and market share models did not incorporate the market representation variable. An explanation for this lack of importance may be an artifact of the time span studied. That is, the high age of top- performing drugs may have reduced the variation in the market representation variable within the observed sample. Top-performing drugs showed a surprisingly negative linear relationship to sales and are observed to be relatively unimportant in the market share model. Breadth of top-performing drugs was the best predictor in the sales model, yet did not add significantly to the market share model. This may be the case since a sufficiently broad product line is necessary to appropriately develop and maintain distribution channels in several country markets. A well developed distribution network is a necessary precursor to exploiting sales potential of top drug products. 167 goncurrent atrategles Correlations between pairs of licensing strategies is also interesting (see Table 23). The single most frequently used strategy in the sample was product out-licensing (401), followed by the number of products developed in-house (299). These two strategies are complementary and logically occur together the most frequently and logically have the highest correlation (.655). The third most frequently occurring strategy is in-licensing (139). The least frequently used strategy is product introduction (89). In-licensing and product introductions are also logically highly correlated (.521) . These sample frequencies, when compared to Merck's strategies, are consistent with the earlier observation that Merck seems to manage product introductions better than other firms. In fact, other firms as a whole are not attempting introductions as a major strategy. These correlations raise the question of multi- collinearity in the predictive models. The effects of collinearity as a potential problem in marketing research has been recently examined by Mason and Perreault (1991). Using exhaustive Monte Carlo experiments they concluded that the effects of collinearity are over-represented in the literature, they found the danger of Type II errors are more problematic than Type I. "Any connination of small sample size, low overall model fit, gr extreme intercorrelation of predictors precludes confidence in inference (p 277)." 168 TABLE 23 CORRELATION MATRIX--HYPOTHESIS FAMILY TWO I= =fi 11 v12 12 122 13 L 15 I 11 1.000 112 .8696* 1.0000 l v; .8656* . 7706* 1 . 0000 112 .7189- .8625* .8722* 1 .0000 v3 .5626* .3951* .5207* .6200* 1 .0000 II 1,, .6550* .5003* .6328* 5239* .5615* 1.0000 15 .7369* .6305* .5827* .3578* .5071* .6377* 1.0000 I 1‘ .6851* .3918* .5973* .3853* .6099* .6323* .8910* 1,,” .7602* .6863* .6836* .6766* .6697* .6531* .8266* 17 .3088* .1661* .2782* .1603* .3096* .2686* .3269* 1 __!a .3616* .1628* .3119* .1585* 2953* .2603* .3639* 13 -.o186 -.0361 .0816 .0892! .101“ .3656* .1608* I sues .6262* .2602* .6372* .2867* .6638* .6829* .6639* I 116m .6006* .2050* .6063* .2652* .6098* .6360* .6333* I m J - ”a 16 32 11 m 19 sues same I 1 1.0000 1 2 .9336* 1.0000 17 .3556* 3355* 1.0000 _!a .3667* .3696* .9816* 1.0000 19 .2128* .1259* -.0279 -.0357 1.0000 sues .6926* .5044" .6880* .6923* .7068 1 .0000 MT .6903* .6958* . 7300* . 7662* .0556 .9703* 1 .0000 £181.11; - Sigm cant gfifim # - Significant 5 LE .05 Y1 8 rulber of products developed in-house Y2 8 numer of products introduced Y; 8 nuber of promote in-licensed Y). 8 nulber of products out-licensed Y5 8 breadth of product developments '6 8 product line breadth Y7 8 nulber of top-performing drugs '8 8 breadth of top-performing drugs 169 to their Monte Carlo runs (n s 300) and the overall fit of both models, it can be reasonably concluded that collinearity is not a problem. Hypothesis--Three H-3: The number of products a business introduces into a country market is a function of regulations concerning compulsory out-licensing, generic substitution, national formularies, national health plans, acceptance of nondomestic clinical testing, and pricing incentives. The original sample includes 101 top-performing global pharmaceutical firms during the period 1982-1987. 0f the potential 606 observations (101 firms by six years), 504 were constructed at the line-of—business level. This resulted in 3,024 observations by line-of—business, by year, and by country of product introduction. A stepwise backward analysis was performed regressing six of the dummy regulation variables on the number of product introductions. The regulation variables were lagged by three years (1979-1984) to allow for the average length of time between application and approval of drugs for marketing. Variables found significant in the stepwise regression were rerun using a simultaneous method to distribute any sampling error equally. Since there was no variation in patent protection among the sample countries 170 during the study period, that factor was not included in the study. ' The regressions found a relationship for only two of the independent variables: acceptance of nondomestic clinical testing and national health plans (R2 .006 [p < .0011). Though statistically significant, these findings do not provide meaningful policy implications from the hypothesis as stated. To restate, these findings an imply that firms choose to enter markets regardless of generic substitution plans, national formulary systems, compulsory out-licensing, and pricing incentives. A summary of standardized Beta weights for the analysis is presented in Table 24. A second multiple regression was run with compulsory out-licensing recoded "0" for France, Italy, and Japan because they reportedly seldom enforce this regulation. Four regulatory variables were retained in this equation-- generic substitution, national formularies, national health plans, and compulsory out-1icensing--which still yields an I? = .006 (p < .001). The explained variance remains unchanged from the first regression, yet the four variables which loaded significantly did change. A third regression analysis was done to determine whether country market had an impact on the number of product introductions. This regression was also significant (p < .001); however, it provided only R2 = .004. 171 TABLE 24 STEPWISE REGRESSION ANALYSES PREDICTING PRODUCT INTRODUCTIONS BY REGULATIONS AFFECTING THE GLOBAL PHARMACEUTICAL INDUSTRY Ingependent yarianles Dependent GENERIC NAT IGIAL NAT I ML ACCEPTS R810 CGIPULSORY Variable SUSSTI - FMWLARY HEALTH NGIDUIESTIC PRICING WT' TUT ICI PLAN CLINICAL INCENTIVES LICENSING TESTING Fiat was CPRINTRO 222.006” n.s. n.s. - .063 - .052 n.s. n.s. Betab -2.232 -2.682 1 .026 .007 Sig T §22209 Model cpnxu1ao 82-.006” b - .041 - .072 - .055 n.s. n.s. .054 Beta -1.798 -2.836 -2.837 1.863 1 .072 .005 .005 .063 Sig T a - adjusted 117- b - standardized beta weights c - denotes significance at the p s .001 level 172 First, it is notable that the regulations showing any relationship are anti-competitive in nature. Except for compulsory out-licensing and product introductions, each of these slight relationships was negative, as expected. In a purely competitive environment, initial use of such constraints would be likely to deter market introduction. The actual environment, however, has a long history of regulation, and this may have upset balance of the market. The country sample in this study represents considerable purchasing power for pharmaceuticals (approximately three- fourths of the world market). In the last two or three decades, buyer power has exerted itself to such an extent that pharmaceutical firms cannot avoid these regulations and must comply to access the largest markets. This is consistent with the definition of an industry with a high degree of globalization drive, that is, subject to forces which encourage the firm to operate across many national boundaries (Yip 1991). It is also interesting to examine differences among firms' strategy relative to regulation. Table 25 reports these same regulations regressed, in six stepwise equations, upon product introductions grouped by firm's home country. Adjusted st range from .030 for national health plans to .136 for generic substitution. These are all considerably more predictive than the general regulatory equation with a predictive power of .006. 13713 EPIKIELJB 1215 STEPWISE REGRESSION ANALYSES PREDICTING PRODUCT INTRODUCTIONS BY REGULATIONS AND COUNTRY OF ORIGIN Qanandent Variables FRENCH ITALIAN GERMAN ENGLISH USA JAPANESE FIRMS FIRMS FIRMS FIRMS FIRMS FIRMS Independent Variables - Standardized Beta Heights 1 511101101 Adjusted 112 = .136 -.141 -.130 n.s. n.s. n.s. -.372 mm]. gamma: Adjusted n2 = .110 -.098 .194 n.s. .095 n.s. -.216 EAIIQMAL u LT p A11 Adjusted 112 -.o3o 1 s1 0 lNlCAL TESTINQ 2 n.s. .114 .117 .084 n.s. -.213 Adjusted R 8.094 £52. PRjgjgg luceu1lg§s Adjusted 112 2.030 £9!EQL§QBI. 9!I:Li§§!§i!§ Adjusted 112 2.051 Note: Beta weights are significant 0 p s .001 level. 174 French, Italian, and Japanese firms all behave somewhat adversely relative to generic substitution (R3== .136). Japanese product introduction behavior is over twice as strong (-.372 Beta weight) as French and Italian firms. All significant relationships were negative, which is consistent with the study hypothesis. The second strongest relationship is with national formularies (R”== .110). As with generic substitution, French and Japanese firms had negative Beta weights (-.098 and -.216 respectively). However, Italian and English firms showed positive Beta weights (.194 and .095 respectively). The study hypothesis is partially supported with regard to national formularies. That is, at least for French and Japanese firms, national formularies are negatively related to product introduction strategies. Acceptance of nondomestic clinical testing was hypothesized to have a positive relationship with product introduction behavior. In relationships which were significant, three of the four were positive. Italian, German, and English firms all introduced more products into countries which accepted nondomestic clinical testing. Japan, on the other hand, had a Beta weight of close to double the other three but with a negative valence. This can be rather easily explained by a general strategy of Japanese firms, which has kept them operating primarily domestically until very recently (Yashikawa 1989). 175 Compulsory out-licensing appears to affect product introductions by some of the study country firms (R2== .051). As hypothesized, the predicted relationships were negative with French and Japanese firms. English firms did, however, introduce products significantly more often under compulsory out-licensing. This is easily explained because their domestic market was the only country in which the regulation has been consistently enforced. National health plans were expected to have a negative relationship with product introductions. This is not the case. German firms appear to introduce more products into countries with than those without national health plans. The United States is the only study country not to have such a plan. Since German firms are not introducing as many products into the United States or Japan (see Table 26), this could be a function of a Europe based strategy rather than aversion to countries without national health plans. TABLE 26 STEPWISE REGRESSION ANALYSIS PREDICTING PRODUCT INTRODUCTIONS BY COUNTRY OF ORIGIN Ingenendent Variables FRENCH ITALIAN GERMAN ENGLISH USA FIRMS JAPANESE FIRMS FIRMS FIRMS FIRMS FIRMS Dependent Variable - Total Firm Product Introductions Adjusted 112 a .082 Beta Heights ufiiifly 176 Descriptive product introduction statistics by regulatory condition are provided in Table 27. Table 28 sets forth a breakdown of product introductions into country markets by country of origin. R&D pricing incentives appeared to be related to strategies of Japanese and German firms, both countries' firms had positive Beta weights. The stronger relationship with Japanese firms (.186) can be explained by its strong domestic marketing strategy and the Japanese RSD incentive price program. German firm strategy has less than half the strength relationship (.075). Though this relationship is significantly positive, with an R2 of only .030, existing R&D pricing incentives cannot be considered a success in attracting pharmaceutical product introductions into country markets. Finally, the potential for sales of drug products over- the-counter (0TC)--that is, without a physician's prescription--might also affect product introduction regression analysis was run using three conditions as dummy variables: CNS drugs introduced into the United States, all categories of drugs introduced into France, and all others. The study sample of 504 firm-level observations over six years yielded 27,218 observations when split by country of introduction and nine therapeutic classes. As expected, 177 TABLE 27 PRODUCT INTRODUCTIONS INTO REGULATORY CONDITIONS Generic let i onal National Acceptance R80 Coapul - Subet i tu- Formu- Health of Pricing sory tion lary Plans Nondomestic Incentives Out- Systems Clinical licensing Testing m R la 1 n Effec Products 129 138 44 132 104 77 Introduced (75%) (80%) (25%) (76%) (60%) (45%) (beervations by 72 111 33 105 88 49 fire and year Ragulation Not in Effgt Predicts 44 35 129 41 69 96 Introchced (25%) (20%) (75%) (24%) (40%) (55%) (beervations by 28 32 110 38 45 84 firm and year Total Predicts 173 173 173 173 173 173 Introduced (100%) (100%) (100%) (100%) (100%) (100%) _hi-_squarU_” 35.26 , 50.99 __3526__ 88.5 _ 88.5 178 TABLE 28 STEPWISE REGRESSION ANALYSES PREDICTING PRODUCT INTRODUCTIONS INTO COUNTRY MARKETS BY COUNTRY OF ORIGIN Ingangnggnt yariables - Prggggt Introductions intg countrigg; FRANCE ITALY GERMANY ENGLAND USA JAPAN Adjusted R2 . .037 .136 .112 .051 .116 .107 ————————_—_—_—.—————== Dependent Variables] Country of Introduction - Standardized Beta Heights FRENCH FIRMS n.s. -.094 n.s. -.083 -.164 n.s. ITALIAN FIRMS n.s. .294 n.s. n.s. -.122 n.s. GERMAN FIRMS .104 n.s. .247 n.s. -.114 n.s. ENGLISH FIRMS n.s. n.s. n.s. .106 n.s. n.s. USA FIRMS n.s. -.101 n.s. n.s. n.s. n.s. JAPANESE FIRMS -.130 -.208 -.182 -.184 -.351 .330 Note: Beta weights are significant 0 p 5 .531 level. behavior. The study countries have very little variation in regulations which govern introduction of new chemical entities as OTC products. England, West Germany, Japan, and Italy all require that new products be classified as prescription-only for the first five years after introduction. France is slightly more lenient in requiring prescription-only status for a minimum of only two years. The United States has the most flexible categorization laws on the books: however, in effect these laws are as or more restrictive than five years. In addition, central nervous system drugs and other drugs with a high potential for consumer abuse may never be reclassified to OTC status. Nevertheless, a multiple even with the statistical power of this enormous sample size, neither product introductions into France or CNS 179 introductions into the United States loaded into the equation. Though a multivariate model is unable to substantially predict product introduction strategies, a comparison of introduction frequencies with respect to presence or absence of each regulation does provide some insights. The differences between the number of product introductions with and without each regulation in effect and without were significant (Chi-square significant 0 p < .01). Policy makers should note three specific results. First, the strong positive relationships between product introductions and generic substitution and national formularies is contrary to the negative impact hypothesized. Second, the other anti-competitive regulations, national health plans, compulsory out-licensing, and non-acceptance of foreign clinical trials do relate negatively to product introductions as hypothesized. Third, when controlling for domestic activity, fewer product introductions were made into countries with R80 incentives. Table 29 provides frequencies of introductions into country markets by country of origin. It is interesting to note that the largest block of introductions are generally into domestic markets. However, firms from three countries of origin did introduce a larger percentage of products into foreign markets: German (60%), US (61%), and English (82%). 180 TABLE 29 FREQUENCIES OF PRODUCT INTRODUCTIONS INTO COUNTRY MARKETS BY COUNTRY OF ORIGIN FRANCE ITALY GERMANY ENGLAND USA JAPAN Total Number into country 24 13 33 22 44 39 FRENCH FIRMS 2 - 1 - - - foreign - 1 (33%) total - 3 ITALIAN FIRMS - - - - - - foreign - 0 total - 0 GERMAN FIRMS 5 5 12 4 3 1 foreign - 18 (60%) total - 30 ENGLISH FIRMS 3 2 8 6 14 1 foreign - 28 (82%) total - 34 USA FIRMS 11 5 12 11 27 4 foreign - 43 (61%) total - 70 JAPANESE FIRMS 1 1 - 1 - 33 foreign - 3 (8%) total - 36 181 A larger sample or different research design probably would not yield a stronger effect for several reasons: (1) the statistical power of the research is high; (2) a portion (76 percent) of the world market is represented the sample; (3) The sample size is large; (4) six countries have been sampled: (5) nine therapeutic classes are included: (6) the secondary databases are widely accepted; (7) the sample period is long (1982-1987); and (8) an exaggeration effect is created by the dummy coding of regulation variables. For all these reasons, if an effect does exist, it should have been apparent in a study such as this one. The fact that the study hypothesis generally was not supported by these findings suggests interesting implications for macromarketers and public policy makers. This lack of the expected relationships between regulation and behavior in the pharmaceutical industry in these countries supports the notion that regulators have been able to exercise a free hand over the industry with respect to anti-competitive legislation without impacting availability of drugs. It is still possible, however, that if this free-hand approach continues, especially in terms of increased compulsory out-licensing, it may affect future product introductions. The current state of regulation and enforcement has not impacted strategic product introduction decisions. This may be because they are based upon contribution margins, not sunk R&D costs. Compulsory out- licensing, however, may have the potential to impact future 182 product introductions by reducing profit potential for the industry and influencing R&D investment decisions. EEEEEEI This chapter has provided a discussion of the analysis and results of this dissertation. In summary, the three objectives of this study have been fulfilled by empirical testing of three families of hypotheses. The first hypothesis, to relate past product acquisition and market access strategies to current strategy, was partially supported. The second was to examine the relationship between product acquisition and market access strategies and performance. Specifically, the effect of licensing strategies versus internal strategies were related to sales and market share and produced two models with reasonably good predictive values (R2== .602 an .632 respectively). The third, to examine the relationship between regulatory issues and product introduction, was not supported. A summary of study findings is presented in Table 30. The first objective was met with an empirical analysis of four hypotheses. Hypothesis 1a was partially supported with a positive linear relationship between past in-house product development and current product development within a therapeutic class. Test of hypothesis 1b provided nominal support in a statistical sense but provides little support for the relationship between past in-licensing and top- performing drugs. Hypothesis 1b showed no relationship 183 between past in-house product development and top- performers. The third hypothesis in this first family (1c) was empirically tested to study the relationship between past breadth of product development and current product development success. The relationship was expected to be an inverted U-shaped curve and was found to have a positive linear relationship, partially supporting this hypothesis. The last hypothesis of this family (1d) showed a significant positive relationship between past and current product introductions to a country. Hypothesis 1d was fully supported. The second objective was met with two models predicting performance. Performance measures used are sales and market share. The predictors used in both models are product acquisition, market access, licensing strategies, and industry success criteria. Internal product acquisition strategy, product development, was found to be more important in predicting sales and market share than in- licensed products. At the same time, external market access strategy (out-licensing) was found to be more predictive than product introductions. 184 TABLE 30 SUMMARY OF STUDY FINDINGS HYPOTHESIS SUPPORT KEY FINDINGS 1-a R2 8 *130b Past in-house product development and past in- licensing have positive linear relationships with current in-house product development. 1-b r2 8 .001c Past in-licensing has a very small, positive linear I relationship to incidence of top drugs. 1-c r2 8 .0328 The breadth of products developed has a positive linear relationship to the number of products currently developed. A positive quadratic relationship is also significant. 1-d r2 8 .0648 The number of products previously introduced into a country is positively related to the number of current product introductions. 2-a R2 8 .594 In the multivariate model, sales is a significant linear function of total products developed (negative), breadth of product development (negative) and product line breadth, products introduced, products in-licensed, products out- licensed, top drugs (negative), and breadth of top drugs. 2-b R2 8 .683 In the multivariate model, market share is a significant quadratic function of total products developed (negative), breadth of product development, and breadth of product line, alga a linear function of out-licensed products and products introduced. 3 R2 8 .004c Regulatory issues were found to be non-significant in relationship to product introductions. Compulsory out-licensing was the exception with a slight (Beta 8 .004) relationship. when compulsory out-licensing was recorded to reflect enforcement the same R was obtained using three variables: generic substitution, national health plans, and compulsory out-licensing. a 8 significant at p < .01 b 8 significant at p < .001 c 8 statistically significant but not managerially significant 185 Breadth of product development showed a consistent relationship with sales and market share. Its inverted U- shape relationship is consistent with Foxall's (1983) notion of success with constrained search for new products. Product line breadth in both models and product introduc- tions in the market share model showed U-shaped relation- ships which suggest presence of learning curve effects, positive economies of scale, and/or specialized management and marketing. Industry success criteria did not predict sales and market share in the same manner. Market representation failed to be predictive in either model. Breadth of top- performing drugs was very important in the sales model but did not add significantly to the market share model. The third objective was met with an empirical test of hypothesis 3, which related regulatory issues to product introductions in a country. Compulsory out-licensing was the only regulatory issue found to have a significant overall relationship with product introductions. However, when product introductions were analyzed by firm home country it was found that some countries' firms introduce fewer products into countries with generic substitution (France, Italy, and Japan) and countries with compulsory out-licensing (France and Japan). In general France introduced fewer products into countries with anti- competitive regulations. Italy introduced somewhat fewer into countries with generic substitution and Japan 186 introduced fewer to all non-domestic countries than at home. This hypothesis was partially supported in a statistical sense but provided nominal explanation for product introduction behavior. Chapter VI addresses the managerial, public policy, and methodological implications, as well as limitations of the study, and recommendations for further research. CHAPTER VI CONCLUSIONS AND IMPLICATIONS Introductlon International business literature has been missing a balanced and empirical approach to the study of licensing strategies and its relationship to performance. The relationships conceptualized by Foxall (1983), Meyer and Roberts (1988), Ohmae (1989), and Porter (1990) were synthesized and adopted to develop models for such an empirical analysis. The models incorporated internal strategies of product development and introduction and external strategies of in-licensing to obtain products and out-licensing to access markets. First, past in-house product acquisition and market access behaviors were related to current strategies. Then, past and current strategies where related to two performance measures, sales and market share. Finally, since international business must access numerous country markets, product introduction strategies were examined as a function of relevant regulatory issues. This study has provided an empirical analysis of the strategy-performance relationships reviewed above. Based upon the findings presented in Chapter V, this chapter 187 188 presents a revised global strategy framework, a revised global pharmaceutical product portfolio, research and public managerial implications of the study followed by public policy implications and recommendations. The conclusion of this chapter summarizes study limitations and contributions. as arch m one This section will relate each of these families of hypotheses to the literature. A summary of key study findings as they relate to the literature can be found in Table 31. Past and Current strategies The four hypotheses in the first family were all supported to varying degrees. The first, which relates past in-house product development and past in-licensing to current product development, was fully supported. The stronger relationship is between past and current product development. The weaker relationship is between past in- licensing and current product development. Both relation— ships are positively linear, logically indicating that increased product development success evolves from an increase in interest in the product class both by develop- ment activity and, to a lesser extent, indirectly by in- licensing. This is very similar to Meyer and Roberts' (1988) notion of technology focus, using a product category rather than a type of technology. It is also consistent with the portion of Foxall's theory (1983), which positively 189 TABLE 31 FINDINGS IN CONTRAST TO PREVIOUS LITERATURE PROPOSITION AUTHOR THIS STUDY (year) Products are currently selected for country markets Anderson and consistent without regard to product complementarity. Coughlan (1983) A constrained search for innovation is positively Foxall (1983) consistent related to performance. Technology focus is important relative to Meyer and Roberts consistent performance. (1988) Strategic alliances enhance firm performance. Ohmae (1989) consistent Strategic alliances lead to mediocrity. Porter (1990) consistent (4-6 years) Networking, or strategic cooperation, contributes to Thorelli (1990) consistent achievement of firm goals. Strategic alliances lead to mediocrity. Porter (1990) inconsistent (short-run) 190 relates search within a limited number of categories to success. This study shows in-licensing to have a negative impact on market share during the later portion of the study period. These results may also be consistent with Porter's (1990) theory which associates long term success with internal rather than external product acquisition strategies. The second hypothesis is partially, but nominally, supported with a linear relationship between past in- licensing and producing top-performing drugs. These results did not show a significant relationship between product development and producing top-performing drugs. This is inconsistent with Foxall (1983), Meyer and Roberts (1988), and Porter (1990). However, the product categories samples have been determined to have heterogeneous relationships in the models for both hypotheses 1a and lb. Differences in the sample across therapeutic classes suggests that a moderator variable is at work between product development and in-licensing, and the two dependent variables (current product developments and top-performing drugs). Further studies should address this issue. The moderator(s) may be related to stage of the life-cycle associated with the therapeutic class, presence or absence of top-performing drugs within the therapeutic class and/or complementarity across product classes. Other variables may be less easily quantified: e.g., competition within the class, existence of related technologies within the firm, 191 strategic partnerships, and corporate mission statements. Further research is also needed both in technology-intensive and less technology-intensive industries. Similar findings might be expected in high-technology industries. However, industries which do not require an extensive financial and time commitment for product development may differ signifi- cantly. The relationship between in-licensing and product development and top-performing products is likely to be much stronger in low-end technologies because these types of products can be more readily copied or improved upon. These first two hypotheses (la and lb) both predicted a negative quadratic (inverted U-shaped) relationship between past in-licensing and both dependent variables. Findings of this study should not be construed to be inconsistent with these hypotheses. This is because the past variables were calculated with a time constraint which may limit variation in this relationship to less than what is needed to determine the relationship over the long term. Further studies should direct efforts towards studying a lag period in excess of nine years to detect the downswing in the hypothesized relationship. The third hypothesis in this family relates past breadth of product development to current product develop- ment and was partially supported. The findings showed a small but significant linear relationship as well as a slightly smaller but significant positive quadratic relationship. As with the first two hypotheses, further 192 studies should investigate this relationship over longer time periods and split the data using one or more moderator variable. The fourth and final hypothesis in this family relates past products marketed to current product introductions. This hypothesis was fully supported across country markets. However, variance between therapeutic classes within country markets was not examined and may provide additional insights. It also would be valuable and interesting to marketers to study this relationship in other industries and other countries. There are two reasons to explore these relationships in other industries. The first is to test whether a similar relationship exists in industries with similar globalization drivers. That is, are other high-technology, long term R&D, and rapidly changing industries behaving in the same manner? Second, is there a difference in behavior when the globali- zation drivers affecting an industry are different? For example, a globalization driver operating on the information industry is favorable logistics: for instance, electronic data transmission capability. Would the behavior of firms in the information industry differ significantly from the behavior of pharmaceutical firms? Also in terms of globalization drivers, it would be interesting to study behavior of firms with lesser globalization drivers. Multiple studies might contribute to a methodology for 193 measuring globalization drive and determine thresholds which stimulate firms to attempt access to global markets. The relationship between products marketed in the past and current products introduced should also be studied in countries other than the largest markets available to the firm or industry. Country markets may be attractive to firms and industries for reasons other than size. Proximity to manufacturing, home office, or raw materials may mitigate sheer size and warrant inclusion and/or control in further studies. Competitive strength and strategic alliances may also play a part in market access strategies. Strategy and Performance Goals The results of this study indicate need for further research relative to the strategy performance models developed for the pharmaceutical industry. First, the time frame should be expanded to include information beyond the ten-year scope of the strategy variables used in this study. Second, differences between product categories should be addressed in subsequent models. These differences may be reflected in product complementarity, relative product importance, product quality, and stage in product life- cycle. Third, additional predictive information should be sought. For example, differences in managerial focus and marketing programs may provide important information. A more sophisticated model for the pharmaceutical industry would incorporate moderator variables as suggested 194 in this study and identified in other research. A longitudinal path-analytic or structural equation model would be well suited to this purpose. The findings of this study are not generalizable without further research in both similar and dissimilar industries. The single industry design of this study offers a measure of control over regulation and globalization drivers. Subsequent studies should also address industry specific variables which maximize the opportunity to identify moderator variables which operate between strategy and performance. It is suggested that future samples address variation in technology, regulatory influence, and globalization drivers to further illuminate their role in the strategy-performance relationship. Regulation and Product Introductions As with the research implications for the strategy- performance relationship, future research relating regula- tion to market entry decisions should address other levels of technology and various globalization drivers. Research implications specific to the pharmaceutical industry relate to compulsory out-licensing and R&D incentives. First, compulsory out-licensing has the potential to impact a large portion of global pharmaceutical volume with a single stroke through EC regulation. An understanding of the nature and extent of negative externalities which may 195 accrue to other countries the long-term ramifications on the industry are worthy and complex research goals. Second, signs of industry consolidations may be reflective of extraordinary and growing pressure on the industry. The nature, the extent, and the reach of such pressure needs to be determined to protect social welfare interests. A Revised Global Strategy Framework for Pharmaceuticals The global strategy framework initially proposed in Chapter II has been revised to reflect the findings of this dissertation and is presented in Figure 9. Market represen- tation has been eliminated from the framework since it was not shown to significantly impact sales and market share. Another interesting facet of the revised model is presence of three quadratic variables: products developed, product line breadth, and breadth of products developed. All three variables have been shown to have an inverted U-shaped relationship with the performance measures. It is possible that they are also similarly moderated by other variables. Subsequent studies will require specific attention to the ceiling effect present in each of these quadratic variables so that separate path analytic models can be studied for each condition of the variable: that is, at a low value during the upswing, at higher values accompanied by the downswing, and especially at the mid- range as associated with acceptable and desirable w.._<0_5m0<2m<1m $0.... ¥m0>>mEOm._.mm o mmDOE 197 performance outcomes. Empirical studies can assist researchers in identifying firms and groups of firms for case level investigation for each of these quadratic variables. The next section discusses managerial implications and presents a revised pharmaceutical global product portfolio based upon the study findings. Managerial Imnllcations Past and Current Strategies There are several managerial implications of the first family of hypotheses. First, the relationship between past and current strategies has competitive implications. Second, the relationship between past breadth of product development and current product development has R&D manage- ment and in-licensing strategy implications. Past product development and in-licensing predict only about thirteen percent of the variance in current product development. Although this study does not bring to light alternative predictive sources, it does provide a relative weight for managerial knowledge of past strategy informa- tion. Observing competitors' product acquisition strategies is not sufficient to predict their future product develop- ment strategy. In the pharmaceutical industry, and possibly other industries with costly and lengthy R&D, product development appears to be considerably more predictive of sales and market share than in-licensing. This seems to be 198 the case over a time period approximately equal to the effective patent life in the pharmaceutical industry. There are also differences in this relationship between thera- peutic classes which may be attributable to other, as yet unidentified, variables such as contribution margins and availability of product substitutes. The relationship of past in-licensing to current strategies is also of interest to firms planning internal versus external market access strategies. That is, choosing to introduce a product directly or out-license it to gain market access. The small degree of the relationship between in-licensed products and future developments within that product class should provide reasonable assurance that out- licensing to a firm will not in itself result in subsequent product competition in the short term. However, this relationship has only been tested over a nine-year period and shows some indication of strengthening slightly beyond that time frame. The nature of this relationship may be different in industries with shorter product development cycles. 6 Unlike product acquisition strategies, a strong relationship between past and current market access strategies does not exist. This is consistent with Anderson and Coughlan's (1987) notion of pyramiding products within established international channels without regard to product complementarity. The relationship is significant, but very small. Therefore, in scanning the competition, very little 199 value can be placed on predicting where a particular firm will market products based upon past market access strategy. This has reverse implications which lead to questions regarding the reason for lack of a stronger relationship. One reason may be under-utilization of established global channels. That is, existing distribution channels should support a greater number of products to maximize the benefits of maintaining these channel relationships. Managers should evaluate country channels with an eye to maximizing their potential. Complementary products should be sought (through internal development or in-licensing) to more efficiently utilize these channels. Strategy and Performance Goals The strategy-performance model developed and revised in this study suggest three sets of managerial implications. The first set is concerned with product acquisition strategies; the second is concerned with market access strategy: and the third is concerned with industry success criteria. The last set is used to revise the pharmaceutical global product portfolio initially presented in Chapter III. Internal product development appears to have a stronger relationship with subsequent product development, sales performance, and market share than in-licensing does with product development. In-licensing does contribute to sales and market share. However, firms with a diverse, but not over-broad, base of product development have a competitive 200 edge over firms who do not. Managers may wish to plan product acquisition strategies to augment rather than replace product development for a few product categories. These findings also suggest that multiple-product category firms perform better than single-category firms in terms of both sales and market share. The relationship holds whether products are in-licensed or developed in-house. However, products developed in-house are considerably more predictive of sales and market share. The major implication is that product diversification is highly desirable whether by internal or external means. External market access through out-licensing appears to have a stronger relationship with sales and market share than direct product introduction. This relationship is strong in the short-run for pharmaceuticals and likely to hold true in other highly technical or highly regulated industries. Managers should out-license products where and when necessary to reach key markets. This is applicable when goals are short term sales oriented and when market share is the goal. 4 Finding and maintaining the appropriate degree of product diversification involves selecting a few strong categories for product development and in-licensing to fill out the existing product line. The number of product categories should be limited because performance increases with the number of categories up to a point and then drops off. The number of products actually in-licensed shows a 201 positive relationship. Together these findings suggest that a critical mass is necessary to enable specialization to occur. To this end, firms should in-license products to fill out their product line, thereby justifying speciali- zation. This also implies that appropriate selection of in- licensed products which augment the current product line is more important than the number of in-licensed products. Market access by out-licensing seems to be consistently associated with improved performance in the pharmaceutical industry. On the other hand, direct product introductions seem to behave in a manner similar to product line breadth, first decreasing and then increasing relative to perform- ance. Here again, it appears that a critical mass is preferable. This may be due to the opportunity to specialize. Market access by either out-licensing or product introduction is important in predicting sales and market share. However, the percent of market representation was not predictive, which is consistent with a mature industry. The most interesting and valuable implication of this study is that there are strong differences, possibly managerial, in nay products are introduced. In the majority of firms, product introductions have a distinctly negative relationship with performance. The industry leader, Merck, has an equally strong pggitiya relationship between product introductions and performance. Additionally, firm reports refute the myth of aging drugs pulling down performance. 202 Much to the contrary, Merck is enjoying a considerable profit contribution from older products (Mayer 1990). As a whole, this information paints a picture of superior channel and licensing management by Merck. In summary, managers must strive for selective diversification into each of the major country markets. Channels in each country should be filled with complementary products. These products may be developed internally or externally. Firms with inadequate resources to support the learning curve associated with managing a critical mass of multiple product categories should not pursue a diversifi- cation strategy. Also, firms which choose to diversify should plan a sufficiently broad product line to maximize channel efficiency. Pharmaceutical industry criteria which were used to develop the product portfolio in Chapter II were found to be more useful in predicting sales versus market share. Top- performer breadth was not predictive of market share, probably because only one firm had top performers in three or more product categories. Top performers remained predictive however, and top-performer breadth can be considered as a benchmark indicating success in both market access and product development. Managers should strive for a balanced line of products which maximize utilization of market channels. Top-performing drugs alone do not produce high level performance. They also require well developed specialized market channels. Continuous, well managed 203 product development and specialized marketing channels are required for consistently high firm performance. The next section presents a revision of the global product portfolio to reflect the managerial implications of this study. Revised Pharmaceutical Global Product Portfolio As originally developed, the global product portfolio incorporated two industry decision rules. The first rule suggested the necessity of a firm marketing three top- performing drugs. The second rule suggested that a firm must market its products in the three key markets: Japan, the United States, and West Germany. The results of this research have shown that both top-performing drugs and market representation are directly or indirectly predictive of sales or market share in the pharmaceutical industry. The portfolio model has been revised to incorporate additional study findings directly associated with high sales performance and market share in the industry (see Figure 10). First, top-performing drug sales were replaced in the left bar with sales accumulated by several high- performing drug products. This enables a firm to evaluate the breadth of its product line both within a therapeutic class and, by comparing drug sales bars, across therapeutic classes. The portfolio also encourages strategists to evaluate current utilization of their distribution system by therapeutic specialty. Use of the portfolio to design 204 product lines, rather than simply track top performers, is useful to managers. The right bar continues to indicate market presence in the six largest pharmaceutical markets. Access to these markets should reflect a concerted effort to manage channels necessary to market both existing products and those under development. An imbalance between market representation and marketable products in a therapeutic class will point management towards an appropriate strategic remedy. It is _ important to evaluate each product line relative to all markets. The portfolio can also be used to evaluate use of marketing channels within country markets. This could be done by using separate portfolios for each country market with a product sales bar to record sales of individual products. This would to enable managers to visualize both product line depth and breadth within the country market. Insufficient market representation associated with large sales may indicate lost sales volume opportunity in ready markets for these proven performers. Insufficient product sales may indicate a need to enlarge a product line with in-licensing, train detail personnel in marketing the therapeutic class, or a need to redirect firm resources to other therapeutic classes. It is especially important when top-performing drugs are in the R&D pipeline. Good product line and channel 205 r w E 8 (s uouuw) 89188 Omo O_._OuE.mO¢ #030011 17.5040 1_I F 206 management also assure the firm of a functioning system when top performers are added to the product line. Regulation and Market Decisions This study also relates product introductions to regulatory issues which affect pharmaceutical product introductions into six country markets: France, Great Britain, Italy, Japan, the United States, and West Germany. While there is variation in regulatory issues across countries, product introductions appeared not to vary widely with regulations. The significant relationship is with compulsory out-licensing and is surprisingly small. The implication to managers (and policy makers) in the pharmaceutical industry is that there appears to be profit potential regardless of regulatory condition. However, since Great Britain and Italy are the only countries to consistently enforce compulsory out-licensing regulation the situation has the potential for considerable change. If all the countries with existing regulations (France, Italy, and Japan) begin to enforce out-licensing, the portion of the global market under this anti-competitive regulation would increase from 4% to 41%. If the EC adopts a similar regulation, over 26% of the world market would be affected by a single regulatory change. With this competitive pressure the profit potential for the industry would inevitably decrease, at least in the short run, 207 because contribution margins may decrease substantially. However, as long as there are contribution margins available and significant barriers do not arise from regulation and competition, existing products are likely to be introduced. In the long run, lower pricing induced by compulsory out- licensing will render R&D investments more risky and less rewarding. Managers should direct research efforts towards understanding the short- and long-run ramifications of compulsory out-licensing. The following section discusses the public policy ramifications of these findings. Public Policy lmnllcatigna There are also implications for public policy makers in Western Europe, Japan, and the United States. The first implication concerns the choice of short- and long-term strategies. The second is more specific to regulations affecting the pharmaceutical industry, yet has more general implications as well. The goal of policy in regulating the ethical drug industry has been to provide safe, effective, accessible products to the public. To this end, numerous special regulations have been imposed upon the industry. These may have the effect of reducing profit potential for the industry and increasing the risk of subsequent RSD invest- ment. Despite the current high degree of regulation, the industry appears relatively unaffected in its product introduction behavior. The only relationships, though 208 slight, are with anticompetitive regulations. This suggests either that the industry was able to meet product introduc- tion criteria during the period studied or that it has no choice about which markets to access. Findings support the notion that policy makers in the context of this study have had a free hand in the industry because high contribution margins induce product introductions even under the most stringent of regulatory environments. As discussed in the managerial implications section, this situation could change dramatically with increased enforcement of compulsory out-licensing. Current firm strategies would no longer yield the same ROI. This might not affect product introduction strategy for up to fourteen years (the average product development time frame). This is because firms will continue to seek returns on sunk R&D costs. However, future product introduction strategies may include a reduction in R&D investment (and new product development) to reflect lower industry ROI. Policy makers thus should carefully consider the potential impact of compulsory out-licensing before enforcement. As much as 41 percent of the global market would be affected if existing regulations were enforced. Countries which do not require out-licensing may experience higher pharmaceutical pricing as firms continue to maintain profit margins. If the percentage of the market enforcing this regulation changes substantially the social impact in non-enforcing countries may be as dramatic as in those which do enforce 209 compulsory out-licensing. The long-term effects might be to stifle R&D thereby reducing the availability of new and improved drug products. Another regulatory issue, incentive pricing for firms engaging in R&D, was also studied. In the context of other regulations affecting the industry, no significant connec- tion to product introduction was found, with the exception of German and Japanese firms. This suggests either that the industry does not significantly alter strategic behavior in response to this incentive, or that present R&D investment is so far removed from the incentive as to leave current behavior unaffected. The strongest relationships were with firms whose home country has pricing incentives. Thus, this incentive appears to be an inadequate motivator for foreign industry R&D and requires rethinking if future R80 is to be induced on a global scale. When examined without other regulations and controlled for domestic activity, signifi- cantly fayar product introductions were made into countries with existing R&D incentives. The study sample represents the largest country markets (76 percent of global volume) but does not address the smaller markets. Smaller country markets, e.g. those of South America and Africa, tend to have few, if any, regula- tory restrictions on the industry. These countries may not have the resources, the sophistication, or the desire to regulate the industry. Furthermore, if these smaller markets were regulated, they may detract product entries 210 because the cost of satisfying regulatory requirements may exceed potential profit margins. As a result, pharma- ceutical products would not be available to consumers in such countries. The single industry nature of this study has provided an opportunity to closely examine the regulatory issues which impact product introduction. The industry has been characterized as subject to numerous globalization drivers. These include high R&D costs, a rapid rate of technological change, and an increasing rate of technological change. By definition, these factors force an industry to seek global markets in order to survive. Therefore, firms must and do comply with whatever regulations are imposed. One consequence of noncompliance in an‘industry with high globalization drivers is attrition due to insufficient market access. Attrition can also result from insufficient product development due to inadequate or poorly managed resources. Evidence of this has been noted recently with respect to aging patents (SmithKline's Tagamet) and the increasing incidence of mergers. Three mergers between major players have occurred in the last three years: (1) SmithKline and Beecham, (2) F. Hoffman-LaRoche & Co., and (3) Bristol-Myers Squibb. This suggests that high costs of R&D, including those imposed by regulatory bodies, may actually have had a negative impact upon long-term availability of drugs. Policy makers need to consider the long-term implications of loss of competition in the 211 pharmaceutical and other industries in which R&D commitment is of vital public concern. To restate the potential problem, anti-competitive regulation may have little effect on industries with a high globalization drive, at least in the short run. However, if enforcement of compulsory out-licensing increases, fewer new products may eventually reach the market as the industry attempts to maintain profits. The net long-term affect upon industry RSD may also be negative, resulting in a negative social impact. This is a fertile area for further research. A second point revolves around the long-run effect of anti-competitive regulation, particularly for industries with an emphasis on R&D. The high cost of continuous R&D is one of the more compelling globalization drivers because it must be covered by profits from one or more market. If ROI is insufficient, the long-run effect is a reduction in R&D. If firms nnat seek critical markets to survive, continued product introductions in these markets do not necessarily indicate a healthy industry. On the contrary, continued marketing in countries which require out-licensing may significantly affect future industry health because competitive presSures would reduce profitability and decrease capital available for R&D. To monitor unproductive or counter-productive regulatory effects upon the industry, academicians and policy makers may wish to develop measures for determining cross-country effects of compulsory out- licensing on pricing and profit contributions. Furthermore, 212 the current incentive programs should not be construed as global stimulators for future R&D. Instead, they appear to operate more in the domestic sphere. The danger here is the potential for policy makers to over-rely on these incentives to outweigh the effect of progressive enforcement of compulsory out-licensing. In summary, the lack of relationship between the number of product introductions and these regulatory issues suggests that either (a) the industry is obtaining suffi- cient profits to support any regulatory demands thus far imposed: gr (b) the industry health may be failing if these firms have no choice but to introduce products into anti- competitive regulatory environments to achieve contribution margin goals. Given the number of major mergers which have occurred in the past three years relative to the lack of previous mergers, the second explanation may be the more plausible one. The current industry situation may be attributable to the stiff regulatory environment andzor to lack of management expertise. If the explanation is either the regulatory environment or lack of management expertise, the problem is widespread. The implication to policy makers in the six study countries is that during the study period pharmaceutical product availablity was not affected at least in the short- run. However, the long-run effects are not yet fully understood. This research is consistent with the notion that the long-term effects of stiff regulatory intervention 213 may eventually decrease availability of pharmaceutical products as a result of reduced industry-wide ROI and R&D investment. This decrease would be considered a negative social impact. Additionally, findings of this study also indicate that regulation is probably not the primary determinant of product introduction behavior. Qongluglon This dissertation is limited in terms of its single industry nature, relatively short time frame, and the secondary nature of the data. Numerous unresolved issues remain for further research. These issues concern product acquisition strategy differences between product classes, the effects of licensing in the long run, differences in the strategy-performance relationship across industries, and potential ramifications of compulsory out—licensing on the pharmaceutical industry. This study has, however, contributed to the academic literature statistically powerful, longitudinal, empirical tests of the relationships between (a) past and current product acquisition and market access strategies, (b) licensing strategies and firm performance, and (c) regu- lation and global strategy of pharmaceutical firms. Existing literature has also been strengthened by the findings of this dissertation. First, Foxall's (1983) hypothesis concerning the positive relationship between a constrained search for new products and firm success was 214 supported in the pharmaceutical industry. Meyer and Roberts' (1988) technology focus was adapted to focus on product function and was supported. Porter's (1990) hypothesis concerning positive performance outcomes of internal innovation were supported. Study results also indicated past strategic behavior consistent with Anderson and Coughlan's (1987) findings with respect to a disregard for product complementarity in utilization of international channels. Furthermore, these results showed superior sales and market share performance in firms which have accomplished product complementarity within their international pharmaceutical channels. Contributions to the pharmaceutical industry include an empirical test of its commonly used success criteria and a synthesis of these criteria into an industry-specific product portfolio. The study empirically showed a relation- ship between these criteria and market share and sales. The global product portfolio has been revised for to increase its managerial value by incorporating additional predictive criteria for the industry. It is hoped that academicians will find this dissertation useful as a basis for further research in global licensing strategies, and that pharmaceutical executives will find these results insightful. APPENDICES APPENDIX A LITERATURE REVIEW MATRIX TABLE A-l LITERATURE REVIEW MATRIX INDUSTRY MARKETING STRATEGY ECONOMICS INT'L BUSINESS STRUCTURE Burstall 1985 Aaronson 1986 Brads 1980 Brastow 1988 Comanor 1986 .1387" 4 Schendel Caves 1982 Cocks 1973 Egan, Niginbotham Carlson 1983 Schnee 1979 C Heston 1982 Nutt Mokwa 8 Shapiro 1985 Comanor 1986 Faro 1990 Hu 1988 Glennie 1971 Gabowski Vernon & Thomas 1978 Goedde 1983 Schnee 1979 Gaggowski 2 Vernon Sherer 1985 Hartley, Levers & Schwartzman 1976 Maynard 1986 Slatter 1977 Neiduk 1982 Silvermen & Lee Kirim 1985 1974 Moran 1986 Temin 1980 Scherer 1985 Scott 8 Reekie 1985 Tesler, Best, E an 8 Niginbotham 1 75 Vernon 1971 CONDUCT Koenig & Lublin Aaronson 1986 boggles 8 Craig General 1989 19 Strategy Davidson 1985 Schnee 1979 Porter 1980, 1986 . . C 1990 Hilliams 1984 Yoshikawa 1989 215 216 TABLE A-l (cont'd) INDUSTRY MARKETING STRATEGY EMICS INT'L BUSINESS Product Beckhaus 1983 Carlson 1983 Grabowski 1968 Innovation DiMasi et al 1990 Cool 8 Schendel Hartley, Levers & 1988 Maynard 1986 Gorecki 1986 Foxall 1983 Park 8 Smith 1990 Porter 1990 Tushmen 8 Anderson 1986 Hu 1988 Product Bitondo 8 Frohman Harrell 8 Kiefer Portfoliot 1981 1981 Bennet 8 Hind 8 Douglas Cunningham 1985 1981 C 8 Glazer 13E?“ Carlson 1983 Cardoza 8 Smith 1983 Cardozo 8 Hind 1985 Cummi s 8 Dale 1981 ng y Devi 8 Stewart 1988"".y Guiltinen 8 Donnelly 1983 Hambrick, MacMillan 8 Day 1982 Hambrick I MacMillan 1982 MacMillan Hambrick A Day 1982 Mgzgr & Roberts Sherden 1983 Hind & Mahajan 1981 217 TABLE A—l (cont'd) INDUSTRY MARKETING STRATEGY ECONOMICS INT'L BUSINESS m Place/ Cohen 1985 Ayal C Ziff 1978 Market Entr y Kaminiski & Rink Anderson 8 1984 Gatignon 1986 Lilien 1985 Anderson & Coughlan 1987 Buzzel 1965 Cavusgil 1980 8 1990 Cavusgil C Nevin 1981 Devlin C Bleekley 1988 Ghauri 1986 Goodnow 1980 Goodnow C Hansz 1972 Hamel 8 Prahalad 1985 Harrigan 1987 Johanson 8 Hiedersheim-Paul 1975 Kale 8 Sudharshan 1987 Ohmae 1986 a l b, 1989, 1990 Porter 1986 Reid 1980 Robinson 1978 Simmonds 1985 Thorelli 1990 Yip 1989 PERFORMANCE Curtis 1985 Cool I Schendel Buzzell & Gale Firm 1988 1987 Miller 1988 N088 1989 Puetz 1987 Social Fisk 1982 Bodenheimer 1984 Hells 1982 Nason 1986 Brade 1980 Nason et el 1986 Bradfield 1989 Edner 1986 Leidlaw 1981 Pelzman 1982 Price 1988 Squire 1980 APPENDIX B SUMMARY OF VARIABLES Paul de Haen Variables Used Singla Qnenical Entity (SCE) The non-proprietary drug name uniquely associated with a chemical compound regardless of product form, dosage, or brand name. Ingrapeutic Class (THERAPY) The therapeutic class of the single chemical entity as recorded by the American Hospital Formulary System. This study uses the two digit codes associated with the thera- peutic classes in Table 2. t' irm The manufacturer and marketer of a single chemical entity for a country market. u ket TR The country in which this single chemical entity is marketed under the brand name specified. Yaar lntrodugad in Qonntry (XEAB) The year in which this single chemical entity was introduced in this country market. 218 219 'n a u a u e O The firm which developed this single chemical entity in-house. Pharmaceutical Line-of-Business Data Enarnagentigal Line-or-Business (PLB) Designates the line(s)-of-business of a firm which are engaged in ethical drug manufacturing. The same codes are use for INNOVATR and MARKETER in the merged Paul de Haen data base. MAB). The year for which case data is recorded. r u s eve o ed wi h' T e a u 'c C ass P V Innovative efforts are measured by a count of SCEs, within a therapeutic class, which are self-innovated and marketed by a business. The historic variable is a moving count of SCEs introduced during the nine previous years. The nine year time period corresponds to the average effective patent life of pharmaceutical products during the study time frame. The therapeutic classes examined in the study are summarized in Table 2. Count {SCEs within THERAPY for nine previous years (historic)) and (SCEs within THERAPY for YEAR (current)} where INNOVATOR = PLB and PLB = MARKETER}. j..- s 1- e ..e. . . : 71- .s- 'c .: -- “:18. II This variable is the sum of current and historic PRODDEV across therapeutic classes for each business and represents ten years of information. W This variable is the number of classes in which a business has developed product during the current year or previous nine years. 1- 81“! '_o-- t~ W' ! 1 1 1‘ .e-_ ' .;s L In-licensing arrangements are measured by a count of new SCEs introduced. The historic variable is a moving count of SCEs introduced during the nine previous years. The nine year time period corresponds to the average effective patent life of pharmaceutical products. The current variable is a count of SCEs introduced in the case year. Count {SCEs for nine previous years within THERAPY (historic)} and {SCEs for YEAR within THERAPY (current)} where MARKETER = PLB AND MARKETER not = INNOVATR}. In this study, INLIC is used only as an intermediate variable to calculate PRODNO. I9IS1_EnEhsI_2I_In:li§§n§s§_EIQQHQI§_iIIELIQl The total number of drugs a firm has acquired by in- licensing is the sum of current and historic INLIC. 9 1,, \-J.L.°‘, . .1°‘ 9'. 1‘ 3° W ,1 l 8. ,9‘,3.°e_ .C :7: The number of drugs a business markets within a therapeutic class is the sum of TPRDEV and TINLIC. 221 -mg- . 15373 1 wt' 1 ' 0!- : e e Dev- 0... D V a; This variable is a count of the number of therapeutic classes in which the PRODEV variable is a non-zero value. W Breadth of product line is the sum of the number of therapeutic classes marketed in the current year or previous nine years. --_ t oduc onsw _ .. .-ltfi .. - '1 1111110 This variable is a count of the number of SCEs a business introduces in a country. The historic variable is a moving count of the SCEs introduced in the previous nine years. The current variable is a count of self-innovated and marketed product introductions within a country. This is a measure of a business's commitment to a country market. To calculate: Count (SCEs where YRINTRO 8 YEAR - 1 to YEAR - 9 (historic)) and {SCEs where YRINTRO 8 YEAR within a country (current)} and MARKETER = case PLB. . . ° oo c _n .. _o ~ . ass . 1 e .a .- : 1128111189). This variable is the sum of current and historic (nine previous years) product intrudoctions across country markets for each business. MW). Products innovated but not marketed by a business are out-licensed to access markets. The historic variable is a moving count of SCEs innovated by a business and licensed to another business for marketing in a specific country during 222 the previous nine years. The current variable is a similar count for the case year. To calculate: Count {SCE where YEAR = YRINTRO and PLB = INNOVATR and INNOVATR not = MARKETR. In this study, OUTLIC is used only as intermediate variable to calculate MKTNO. W This variable is the sum of current and historic (nine previous years) products out-licensed across country markets. W The number cf drugs marketed within a country is the sum of TOTINTRO and TOTOUTLIC for that country. do. .- 1* {'9 -;-, ._ *1 a *;: --1 e 9' .- : 91191- ' Marketing representation is measured as the percent of world market accessed by a business. This variable is calculated by multiplying the percent weight for each country market in which the firm markets a therapeutic class times the number of products marketed in that class, and summing across countries. This product is then divided by the total number of drugs marketed in that country. The highest value for this study variable is 66 which means that the business markets in sixty-six percent of the world ethical pharmaceutical market. Values used to weight country markets are provided in Table 1. To calculate: E (MKTNO + TPRDEV) x country market weight. 223 i ce 5 P Market-winning innovations for the current year are represented in this variable. Single chemical entities which ranked in the top thirty global sales performers during the year are counted. ; -.. , . ,1 ,-~-__ -; _ * ..- ~ 103,7_; This variable captures top product performer breadth. The information is recorded as the number of therapeutic classes in which the pharmaceutical line-of—business has a top-performing drug during the year. This variable provides a measure of the pharmaceutical line-of—business against industry criterion one which is discussed on page 49. WW Revenue of each pharmaceutical line-of—business is recorded in current year dollars and deflated using Gross National Product deflators (U.S. Bureau of the Census 1983 and 1990) prior to analysis across time periods. 'Iesua -_ ' ., _, --- -: ~ll‘S? 9:..‘ 1a - .317 The size of the world pharmaceutical market is not easily determined nor readily available. Therefore a surrogate market share measure is computed relative to the sales volume of the entities which consistently appear in every study year. Summed sales of these businesses is used as the denominator for the entire study period. To compute: SALES + (E SALES of the firms which are represented in each study year). 224 The Regulatory Environment W The study is conducted across six countries. This categori- cal variable represents one of the following: France, Great Britain, Italy, Japan, the United States, or West Germany. anpnlsory Out-Licensing (QOMELIQ) COMPLIC records the presence or absence (1,0) of mandatory out-licensing within a country during the year. ;-* e '1 nc-u ao' . ‘1‘ 7 “ss - ., GENERIC records the presence or absence (1,0) of a regulation which encourages generic substitution within a country during the year. WWW FORM records the presence or absence (1,0) of a national formulary system within a country during the year. 'o Hea PLAN records the presence or absence (1,0) of a nationally funded health care plan within a country during the year. . I e. 'cceotance o_ 01".”‘5 _ _'ni . __'. s CLINIC records the presence or absence (1,0) of a country's acceptance of other non-domestic clinical trials during the year. c nce tives INC N INCENT records the presence or absence (1,0) of pricing policies which provide an incentive for R&D investment. APPENDIX C DEFINITION OF TERMS In-licensing - is the practice of obtaining contractual relationship to market a drug innovated by another firm. Industry criteria - are rules of thumb for gauging potential for marketing success. These originate from the expertise of industry managers. Innovating firms - are firms where marketable products are researched and developed. Out-licensing - is granting contractual permission for another firm to market products which were innovated within the granting firm. Performance - is measured using annual sales in current year U.S. dollars, normalized to 1985 dollars prior to analysis. Pharmaceutical line-of—business - is the group or groups within a firm which are engaged in the pharmaceutical line of business. This may be an entire company or one or more strategic business units. Product portfolio - is a tool or model for analyzing an array of products with which a firm works. 225 226 Single chemical entity - is the non-proprietary name of a unique chemical compound which may be marketed for clinical use under one or more trade names by one or more manufacturers. Top performing drugs - are year-specific and are those thirty drugs with the highest reported revenue in a given year. This closely approximates 1990 sales levels of above $300 million. SUMMARY OF RESULTS--STRATEGY PERFORMANCE MODEL SUMMARY OF RESULTS-- Independent Variables TABLE D-1 APPENDIX D STRATEGY PERFORMANCE MODEL Adjusted R2 sq sq sq sq sq sq Y1 YR .ASAJJMS SHMJUB - - - .59446 .63236 - - - .59297 .62455 — - - .58136 .61675 - - - .57917 .61198 - - - .56841 .60456 - - - .59270 .62964 sq - - .59233 .61133 - - - .58611 .62111 - - - .57526 .61031 - - - .57895 .62114 - - - .57895 .62455 sq - - .59851 .62276 - - - .58561 .61832 - - - .57844 .61330 - - - .59683 .63096 sq - - .59683 .63096 - - - .56634 .60317 - - - .59683 .63096 sq - - .59810 .61696 - - - .59002 .62964 sq - - .58652 .60615 sq - - .57884 .59903 - - - .57982 .61280 227 Yl Ya YI YI Yi Yfi Yz Y. 3 ASALES SHARE - sq sq sq - - - - .55502 .59349 - - sq sq sq - - - .57691 .61951 - - - sq sq sq - - .58873 .61424 sq sq - sq - - - - .56955 .60613 sq sq - - sq - - - .59553 .63236 sq sq - - - sq - - .59382 .61288 sq - sq sq - - - - .57171 .60534 sq - - sq sq - - - .58440 .62379 sq - - - sq sq - - .58526 .68299 - - sq sq - sq - - .57164 .59214 - sq sq - - sq - - .58190 .60277 sq - sq - sq - - - .57719 .59733 sq - sq - - sq - - .59134 .61262 sq - - sq - sq - - .58598 .60707 - sq - sq sq - - - .57568 .61922 - sq - - sq sq - - .59446 .62072 - sq - sq - sq - - .57719 .59733 - - sq - sq sq - - .59542 .62276 sq sq sq sq - - - - .55780 .59555 - sq sq sq sq - - - .57345 .61744 - - sq sq sq sq - - .58511 .61424 sq - sq sq sq - - — .58014 .62218 sq - - sq sq sq - - .58769 .61665 sq - sq sq - sq - - .58014 .62218 sq - sq - sq sq - - .59643 .62443 - sq - sq sq sq - - .58761 .61181 - sq sq - sq sq - - .59115 .62072 - sq sq sq - sq - - .56701 .58798 sq sq - - sq sq - - .58166 .63236 sq sq sq - - sq - - .58663 .60623 sq sq sq - sq - - - .59297 .63236 sq sq sq sq sq - - - .57953 .62228 sq sq sq sq - sq - - .57015 .59041 228 TABLE D-i (cont'd) Independent Variables Adjusted R? {229 TABLB 0-1 (cont'd) nan-unuuul u nun- Independent Variables Adjusted R? Y1 Y2 Y3 YI Yi Yi Y; Y. Ya ASALES SHARE sq sq sq - sq sq - .59823 .62540 sq sq - sq sq sq - .59117 .61741 - sq sq sq sq sq - .58314 .61181 sq sq sq sq sq sq - .58775 .61665 sq sq - sq - sq - .58063 .60021 sq - sq sq sq sq - .58769 .61665 - - s - s s - .59542 .62276 sq enotes var a e 3 square . denotes variable is not squared. BIBLIOGRAPHY BIBLIOGRAPHY Abernathy and Utterback (1982), "Pattern os Industrial Innovation," from M. T. Tushman and W. L. Moore, , Marshfield, HA: Pittman Publishing, 97-108. Allen. Bruce T- (1983). nanagerial_329n2migs. New York: Harper and Row. Lu‘ 1 y~-9. ogu. 59 a - P1 - n.0qu: .0; (1990), American Society of Hospital Pharmacists. Anderson, Eric and Anne T. Coughlan (1987), "International Market Entry and Expansion via Independent or Integrated Channels of Distribution," Qanznal 9f HQIKQLIDQ. 51 (January) 71-82- Anderson, Eric and Hubert Gatignon (1986), "Modes of Foreign Entry: A Transaction Cost Analysis and Propositions,” lQurnal_2i_InLernatignal_§u§iness_§tudie§. (Fall) 1-26- Ayal, Igal (1981), "International Product Lifecycle: A Reassessment and Product Implications," 1an:na1_a£ Marketing. 91‘5- Backhaus, Jurgen (1983), ”Competition, Innovation and Regulation in the Pharmaceutical Industry," Managanial and_negi§i_n_figgngmigs 4 (June) 107- 21. Bartlett, Christopher A. and Sumatra Ghoshal (1987), "Managing Across Borders: A New Strategic Require- ment," Sloan nanagemang Reyiew, Summer, 7-17. Bennett, Rex 0. and Lawrence F. Cunningham (1985), "Determining Profitable Products/Market Segments: An Expanded Portfolio Analysis Approach," 19n:na1_af Retail Banking, 7 (Fall) 53-62. Bitondo, Domenic and Alan Frohman (1981), "Linking Techno- logical and Business Planning," Raaaangn_nanagamant, 24 (November) 19-23. 230 231 Brada, Josef C. (1980), "Industry Structure and East-West Technology Transfer: A Case Study of the Pharmaceutical Industry." Assegiatign_fgr_22m2aratixe E29n2m12_§tudie§_§ulletin. 22. 1 (Spring) 31-59- Brand, Horst (1974), "Productivity in the Pharmaceutical Industry." Monthlx_Labgr_Bexieu. 97 (March) 9-14- Buc, Nancy L. (1982), ”Current Regulatory Issues in Marketing Prescription Drugs, Comparative Claims and Advertising to Consumers,” £22d_Drug_£2§netig_Lan Igurnal. 37. 402-8- Burstall, M. L. (1985), ' ' Inanagny, Luxembourg: Commission of the European Communities: Washington, D. C.: European Community Information Services. Buzzell, Robert D. (1968), "Can You Standardize Inter- national Marketing?" nargard_fiu§ine§§_3exieu. 102-113- Buzzell, Robert D. and Bradley T. Gale (1987), Iha_£1u§ 2rin2i2lesi__Linking_§trateg¥__2_£erfgrmance New York: The Free Press. Caqlarcan. et a1: (1979). The_2harmageutigal_1ndu§tr1;_ 0 On a ‘ 09u1_1‘ 0!. 0 ‘ 1n?! "0L,: ,01, New York: John Wiley & Sons Inc. Capon, Noel and Rashi Glazer (1987), "Marketing and Technology: A Strategic Coalignment," Journal of Marketing. 51 (JUIY) 1'14- Cardoza, Richard N. and David K. Smith, Jr. (1983), "Applying Financial Portfolio Theory to Product Portfolio Decisions: An Empirical Study," lannnal_a§ Marketing. 47 (Spring) 110-119- Carlson, Rodger D. (1983), "Planning Product Portfolio Strategies," MiQ:§QB§D.BB§1D§§§:EQQIB§1r 3 (January) 3-7. Cavusgil, S. Tamer (1980), "On the Internationalization Process of Firms," En;nnaan_3aaaangn, 8, 6 (November) 273-81. (1990), "A Market Oriented Clustering of Countries." Internati2nal_narketing_§trategie§. 3rd edition, Hans B. Thorelli and S. Tamer Cavusgil, editors, Toronto: Pergamon Press. 232 and John R. Nevin (1981), ”State-of—the-Art in International Marketing," in Bayiag_nf_uazkaning, B. Enis and K. Roering, eds., Chicago: American marketing Association, 195-209. Cocks, Douglas L. and John R. Virts (1974), "Pricing Behavior of the Ethical Pharmaceutical Industry,” Journal_gf_nusiness. 47 (July) 349-62- Cohen, Jacob and Patricia Cohen (1983), Annliadnunltinlg finianaaa, 2nd edition, Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Comanor, William S. (1979), "Competition in the Pharmaceutical Industry," 63-68, in Iaanaa_1n Bharmageutigal_zgongmiga. Robert I. Chien. ed- Lexington, Mass., and Toronto: Heath, Lexington Books. Cook, Thomas D. and Donald T. Campbell (1979), Qnaai; Sanginga, Boston: Houghton Mifflin Company. Cool, Karel and Dan Schendel (1988), "Performance Differences Among Strategic Group Members," Stranagig narketing_uanagement_lournal 9 207- 23 Cox, Douglas L. and John R. Virts (1975 unpublished), ”Market Definition in the Ethical Pharmaceutical Industry," unpublished Lilly Staff Paper. Cummings, William Theodore and James M. Daley (1981), "Strategic Management: A Marketing Segmental Approach," Baalnaaa, 31 (November/December) 10-15. Curtis, Donald A. (1985), "The Modern American Accounting System - Does It Really Measure Corporate Performance?" EEi__The_nagazine_for.£inangial_flxegutixes 1 (JanuarY/ February) 58- 62. De_naenLs_Drug_Br2du2t_1ndex_:_lnternatignal (1989). Volume II, Englewood, CO: Paul de Haen International, Inc. e ' - (1989), Volume II, Englewood, CO: Paul de Haen International, Inc. (1985), New York: United Nations, 151-156, and 761-846. Demographig_xearbggk_§erie§ (1986). New York: United Nations, 404-423. 233 ' (1987), New York: United Nations, 63-68, 404-423, and 486-487. Deutsch, Claudia H. (1989), "The Brouhaha Over Drug Ads," The_New_Ygrk_Time§. (May 14). 8-14- Devinney, Timothy M. and David W. Stewart (1988), "Rethink- ing the Product Portfolio: A Generalized Investment Model.” Managemen;_fisien2e. 34 (September) 1080-1095- Devlin, Godfrey and Mark Bleakley (1988), "Strategic Alliance-~Guidelines for Success,” Lgng_3anga_£1anning, 21, 5, 18-23. DiMasi, Joseph A. (1990), "Remarks,” Press Briefing-- Washington, D. C., (April 19). Douglas, Susan P. and Yoram Wind (1987), "The Myth of Globalization.” C9lumbia_1surnal_of_flgrld_nusiness. 22, 4 (Winter), 19-29. Faro, Laurie M. C., R. Huiskes, and W.B. van der Mijn (1990), ”Regulatory Scenario's for Medical Devices," a presentation to the E29d1_Drugi_and_Law_In§titute uedigal_nexige_Law_flork§hon. Brussels. June 20- Fisherow, W. Benjamin (1987), "The Shape of Prescription Drug Advertising: A Survey of Promotional Techniques and Regulatory Trends," E22d_Drug_§2smetig_Law_lournal. 42, 213-36. Foxall, Gordon (1983), "Limited Strategic Innovation: Coping with the Complexity of New Product Development," ' (UK), 1, 28-39. Ghauri, Perve N. (1986), "Guidelines for International Business Negotiations." International_narketing_3exiew. 9, 2 (Autumn) 72-82. Gibson. J- Tyrone (1976). uedigation_Law_and_Behaxior. New York: Wiley and Sons. Glaxo Annual Report 1989. Goedde, Alan G. (1982), ”International Diversification and US Regulatory Controls in the Pharmaceutical Industry," QuarteIlx_Beyiew.2f_E22n2miQ§_and_Busine§§. 22 (Spring) 101-10. 234 Goodnow, James D. and James E. Hansz (1972), "Environmental Determinants of Overseas Market Entry Strategies," I9urnal_2f_Internstignal_nusiness_Studies. 3 (Spring) 33-50. Grabowski, Henry G., John M. Vernon and Lacy Glenn Thomas (1978), "Estimating the Effects of Regulation on Innovation: An International Comparative Analysis of the Pharmaceutical Industry," Jgnrnal_9£_Lgn_§n§ Reopenigs. 21 (April) 133-63- Grabowski, Henry G. and John M. Vernon (1979), "Substitution Laws and Innovation in the Pharmaceutical Industry," .Law_and_§2ntemporarx_2roplems. 43 (Winter-Spring) 43-66. Grabowski, Henry G. (1968), "The Determinants of Industrial Research and Development: A Study of the Chemical, Drug, and Petroleum Industries," Janrnal_gf_£gli§igal EQQDQEXI 76: 292-306° Guiltinan, Joseph P. and James H. Donnelly (1983), "The Use of Product Portfolio Analysis in Bank Marketing Planning.” Isurnal_9f_Betail_Banking. 5 (Spring) 15-24- Hair, Joseph F., Rolph E. Anderson, and Ronald L. Tatham (1987). Eultixariate.22ta.bnal¥sis. 2nd Edition. New York: Macmillan Publishing Company. Hambrick, Donald C. and Ian C. MacMillan (1982), "The Product Portfolio and Man's Best Friend," Qalijgnnia Management_3exiew. 25 (Fall) 84-95- , Ian C. MacMillan and Diana L. Day (1982), "Strategic Attributes and Performance in the BCG Matrix - A PIMS-Based Analysis of Industrial Product Businesses." a2adepx_2f_nanagement_lournal. 25 (September) 510-531. Hamel, Gary and C. K. Prahalad (1985), "Do You Really Have a Global Strategy," rv s ev' (July/August) 139- 148. I Harrell, Gilbert D. and Richard O. Kiefer (1981), "Multinational Strategic Market Portfolios," finainaaa Toples. (Winter)- Harrigan, Kathryn (1987), "Strategic Alliances," innnnia Isurnal_gf__2rld_fiu§iness. (Summer) 67-9- Hofer. 0- & 0- Schendel (1978). Strategx_£ormulationi_ Analytigal_ggnsepts. st. Paul. MN: West. 235 Hornick, John (1990), "Presentation to MTA 862,” Michigan State University (January). Hunter, John E. and Frank L. Schmidt (1990), nannaga_afi - s O ' , Newberry Park, CA: Sage Publications. Jain. Subhash C- (1984). Internati2nal_narketing_uanagement. Boston: Kent Publishing Company. (1989), "Standardization of International Marketing Strategy: Some Research Hypotheses," Journal_gf_uarketing. 53 (Winter). 70-9- Johanson, Johny and Weidersheim-Paul (1975), "The Inter- nationalization of the Firm--Four Swedish Cases," The Jgnrnal of Managanant Spanies, 12 (October) 306-7. Kale, Sudhir and D. Sudharshan (1987), "A Strategic Approach to International Segmentation," Innannaniana1_nankaning Raglan, (Summer) 60-70. Kashani, Kamram (1986), ”Beware the Pitfalls of Global Marketing." narxard_nu&ine§§_zexeiw. (September/ October), 91-8. Kirim, Arman S. (1985), "Reconsidering Patents and Economic Development: A Case Study of the Turkish Pharma- ceutical ludustrY." EQ£l§_De¥§122m§nt. 13. 2 (February) 219-36. Koenig, Richard and Joann S. Lublin (1989), "Global Industry Appears to Be Headed For Big Consolidation," Hal] Street_lournal. April 13. Kogut, Bruce (1985), "Designing Global Strategies from Operational Flexibility." Sloan_nanagement_3exiew. (Fall), 27-38. Kotabe, Misaaki and Glenn 8. Omura (1989), "Sourcing Strategies of European and Japanese Multinationals: A Comparison." Journal_2f_International_hu§ine§s §£BQ1§§. 20, 1, (Spring), 113-130. Laidlaw, W. Angus (1981), Social gosn-Benefit Analyais and the_Appraisal_2flT2urism_Eroiest§_in_a_nexeloping Ialang_fignnany, a M.A. Thesis, Canada: Carlton University. Lall, Sanjaya (1978), "Price Competition and the International Pharmaceutical Industry," Oxforg Bulletin 9: Econgnics and Statisgics, 1 (February) 9-21. 236 Levitt, Theodore (1983), "The Globalization of Markets," Harxard_nnain§§§_3exiew. (Nay/June) 92-102- (1960). ”Marketing Myopia." Harxard_nusiness ngiag, (July/ August) 45-56. Liander, Bertil, Vern Terpestra, M. Y. Yoshino and Aziz sherbini (1967). Q2mparati2e_analxsi§_f2r_1nternatignal Manganing, Boston, MA: Allyn Bacon. Mason, Charlotte H. and William D. Perrault, Jr. (1991), ”Collinearity, Power, and Interpretation Multiple Regression Analysis." 12urnal_gf_narketing_Besearch. 28 (August) 268-280. Masson, Alison and Paul H. Rubin (1986), "Plugs for Drugs," Regulation. (October) 37-53. Mayer. N- 0- (1990). HercK_and_Q2mpanxi_Inci_§2mpanx_Beport. (October 26), New York: Sanford C. Beinstein & Co., Inc. Maynard, Alan and Keith Hartley (1984), "The Regulation of the Pharmaceutical Industry," in Enannageugical om' ° e s t 05' s , Bjorn Lindgren, ed. Malmo, Sweden: Liber, 122-37. McCarthy, E. Jerome and William Perrault 91989), Baaig nangaging, Homewood, Illinois: Richard D. Irwin, Inc. McDonald, Malcolm H. B. (1984), "Applying Portfolio Management to Retailing: Some New Insights," uankaning Intelligense_§_£lanning (UK). 2. 23-35- Meyer, Marc H.and Edward B. Roberts (1988), "Focusing Product Technology for Corporate Growth," Siaan Management_3exiew. 29 (Summer) 7-16- Moran, Patrick J. (1986), "U.S.A.: Recent Developments Relating to the Taxation of International Technology Transfers in the Pharmaceutical Industry," Bulletin_for Internati2nal.£iscal_ngcumentation. (April/May) 210-14- Morris, Louis A. and Lloyd G. Millstein (1984), "Drug Advertising to Consumers: Effects of Formats for Magazine and Television Advertisements,” Fang_pzng QQ§m§21Q_L§!, 39: 497-503- Novitch, Mark (1984), ”Direct-to-Consumer Advertising of Prescription Drugs." E22d_Drug_and_Qosmetic_Law Journal. 39. 306-11- 237 Ohmae, Kenichi (1986a), "Becoming A Triad Power: The New Global Corporation." Internatignal_narke§ing_3sxiew. (Autumn), 7-20. (1989), "Planting for a Global Harvest," Hazgard , (July-August) 136-145. (1986b), "The Triad World View,” Ina_lgnrna1_gfi Busine§§_§trategx. 8-19- Penrose. Edith (1968). The9rY_9f_;he_§rowth_2f_the_£irn. Oxford: Basil Blackwell. Perri, Matthew and Arthur A. Nelson, Jr. (1987), "An Exploratory Analysis of Consumer Recognition of Direct- to-Consumer Advertising of Prescription Medications," I2urnal_2f_nealth_gare_narketing. 7. 1 (March) 9-17- Pharmaceutical Manufacturer's Association (1990), 1229 Annnal Reponn, Washington D. C. Porter, Michael E. (1986), "Changing Patterns of Inter- national Competition." california_uanagement_aexiew. 28, 2 (Winter) 9-40. (1990), "The Competitive Advantage of Nations," , (March/April) 73-93. Prahalad, C. K. and Yves L. D02 (1987), Ina_uni;inaninnai 0 ° 1 an ' , New York: Free Press. Reekie, W. Duncan (1979), "Pricing in the Pharmaceutical Industry in the United Kingdom and the United States," 101-04. in Issues_in_Pharmaceutical_ficgnomics. Robert I. Chien, ed. Lexington, Mass., and Toronto: Heath, Lexington Books. and D. E. Allen (1985), "Generic Substitution in the UK Pharmaceutical Industry: A Markovian Analysis," Eanagerial_and_neci§ion_zconomics. 6 (June) 93-101. Robinson, Richard D. (1978), "Some New Competitive Factors in International Marketing." adxance§_in_1nternati2nal Marketing. 1. 1-20- Roth, Kendal, David Schweiger and Allen J. Morrison (1989), "An Empirical Study of Porter's International Strategy Types," presented at the Annual Conference of the Academy of Management. 238 "Rx for Growth, Global Drug Industry Appears to be Headed for Big Consolidation” (1989). fla11_§treet_lgnrnal. (April 13). Scherer, F. M. (1985), ”Post-patent Barriers to Entry in the Pharmaceutical Industry." 19nrnal_9f_flealth_ficonomics (OECD), 4 (March) 83-87. Schifrin, Leonard G. (1982), "Lessons from the Drug Lag: A Retrospective Analysis of the 1962 Drug Regulations," Har1ard_l9urna1_9f_Law_and_£ublig_£olicx. 5.91-129. Scott, D. R. and W. D. Reekie (1985), "Competition in Atomistic and Ologopsonistic Markets: The South African Pharmaceutical Industry," finn;n_Afnigan Journa1_2f_Ecgn2migs. 53 1 (March) 39-54- s (1983), Richmond, England: PJB Publications Ltd. (1984), Richmond, England: PJB Publications Ltd. (1985), Richmond, England: PJB Publications Ltd. (1986), Richmond, England: PJB Publications Ltd. (1987), Richmond, England: PJB Publications Ltd. (1988), Richmond, England: PJB Publications Ltd. Scrip_xearbgok_1229 (1990). Richmond. England: PJB Publications Ltd. Sethi, S. Prakash (1971), "Comparative Cluster Analysis for World Markets." l2urna1_of_narketing_Re§earch. (August) 348-54. Sheffet, Mary Jane (1983), "Market Share Liability: A New Doctrine of Causation in Product Liability, aanxnai_af Manganing, 47 (Winter) 35-43. and Stephen Kopp (1990), "Advertising Prescrip- tion Drugs to the Public: Headache or Relief?" lannnai of Puniic Enligy and Markening, 19, 42-61. Sherden, William A. (1983), "Strategic Planning: Do the Concepts Apply?" Egan's Baviaw (Bronzgaaualgy), 84 (October) 28-34. Shumpeter. J- A- (1942). 9apita1i§m1_§gcialism_aad_nem92: zany, New York: Harper & Rowe. 239 Silverman, Milton and Philip R. Lee (1974), Eill§1_EI2Ii£§1 ana_£aliniaa, Berkley and Los Angeles, CA: University of California Press, Ltd. Slatter. stewart st. P- (1977). c9npetition.and_narketing Strategies_in_tha.2harmaceutica1_1ndu§trx. London: Croom Helm. Teff, Harvey (1985), ”Drug Approval in England and the United states.” The_American.19nrna1.2f.§9mparatixe .Lag, 33, 567-610. Temin, Peter (1980), ' , Cambridge, MA: Harvard University Press. (1979), "Technology, Regulation, and Market Structure in the Modern Pharmaceutical Industry,” gall Journa1_gf_Ec9n2mics. 2 (Autumn) 427-46- The Boston Consulting Group (1973), "The Experience Curve Reviewed, IV, " Earnfolio. Thorelli, Hans B. (1990), "Networks: The Gay Nineties in International Marketing," in anxanagiaa, S. T. Cavusgil and H. B. Thorelli, eds. "Performance on Audits: The MNC through the Glasses of the LDC," in finxanagiaa, S. T. Cavusgil and H. B. Thorelli, eds. 0- 8 Bureau of the Census (1990). Statistical_5bstract_gf the_flni_ed_§tate§1__1_29. The National Data Book, U. S. Department of Commerce, Bureau of the Census, 480. (1983), ' a st t o e figaaaaa__12§z;§1, The National Data Book, U. S. Department of Commerce, Bureau of the Census, 454. Vernon, John M. and Peter Gusen (1974), "Technical Change in Firm Size: The Pharmaceutical Industry," Bagiag_a§ Economic§_and_§tatistics. 56 (August) 294-301- Eall_§treet_lournal_lndex (1979). Barron's IndeX. New York: Dow Jones and Company. Na11_§treet_lgurnal_lndex (1988). Barron's IndeX. New York: Dow Jones and Company. Williamson. Oliver (1975). Market§_and_nierarchies. New York: Free Press, 176-233. 240 Wind, Yoram and Susan P. Douglas (1981), "International Portfolio Analysis and Strategy," Janrnal_af MW. 69-82. and Vijay Mahajan (1981), ”Designing Product and Business Portfolios." W. 59 (January/February) 155-165. Yoshikawa, Aki (1989), "The Other Drug War: U. S.-Japan Trade in Pharmaceuticals,” Qalifarnia_flanag§m§n§ Bayian, (Winter) 76-90. Yip, George S. (1989), "Global Strategy...In a World of Nations?” Wigs. (Fall). 29-41- (working paper), "An Exploratory Test of Globalization Framework," Working Paper No. 90-01, Marketing Science Institute (September 24). Referenced with permission. GENERAL REFERENCES csusnar nsrsnswcss Anonymous (1988), ”The Year in Rxs--Flat But Hardly Dull," Drcc_Icpice. 132. 5 (April 4) 24-33- Baldo, Anthony (1988), "At the Ebb,” £inanaial_flazlg, 157, 19 (September 6) 20-1. Bezold. 0-. editor (1983). 2hermececcice1e_in_the_Yeer_zccce_ The_chencinc_ccntexc_fcr_nrcc_3_e_n. Alexandria. VA: Institute for Alternative Futures. Bradfield, R. E. (1989),"Social Cost-Benefit Analysis in Project Evaluation: A Practical Framework for Decision Making," St Ec ' s Econometrics. 13 (August) 25-35- Campbell, Nigel C. G. (1988), "Market-Share Patterns and Market Leadership in Japan."Internecicne1_Sccciee_cf nenecement_§_crcenizecicn. 18 (Spring) 50- 66- Cardozo, Richard N. and Jerry Wind (1985), "Risk Return Approach to Product Portfolio Strategy," Lang_Ranga Planning (UK), 18 (April) 77-85. Carpenter, Gregory S. (1987), "Modeling Competitive Marketing Strategies: The Impact of Marketing-Mix Relationships and Industry Structure," Manketing Science. 6 (Spring) 208-21- Caves, Richard (1982), : , 5th edition, Englewood Cliffs, NJ: Prentice-Hall Inc. Clifford, Mark and Nayan Chanda (1988), "Foreign Pharmaceutical Companies Accuse South Korean Rivals of Stealing Ideas: Patent Drugs Abuse," £a:_£aa§azn Eccncmic_3exiew. 140. 26 (June 30) 48-9- Coe, Barbara J. (1990), "Strategy in Retreat: Pricing Drops Out." 1cnrne1_cf_Bce1necc_§_1nccecrie1_nerkecing. 5 (Winter/Spring) 5-25. 241 242 Commanor, William S. (1986), "The Political Economy of the Pharmaceutical Industry." lccrne1_cf_Eccncmic Ligaxagnna, xxxv 1173-1217. Day, George S. and David B. Montgomery (1983), "Diagnosing the Experience curve." ccnrnal_ef_nerketing. 47 (Spring) 44-58. Dickson, Peter R. and James L. Ginter (1987), "Market Segmentation, Product Differentiation, and Marketing Strategy.” Iccrna1_cf_uerkecinc. 51 (April) 1-10- Douglas, Susan P. and C. Samuel Craig (1989), "Evolution of Global Marketing Strategy: Scale, Scope and Synergy," cc1cmbie_1curne1_ef_flcrlc_nueineee. 24 (Fall) 47-59- Egan, John W., Harlow N. Higinbotham, and J. Fred Weston (1982). Eccncmice_cf_the.2hermececcice1_lnccecrx. New York: Praeger Publishers. Erickson, Tamara J., John F. Magee, Philip A. Roussel, and Kamal N. Saad (1990), "Managing Technology as a Business Strategy," Sloan Managanena gaviaw, 31 (Spring) 73-78. Friedman, Hershey H. and Joshua Krausz (1986), "A Portfolio Theory Approach to Solve the Product Elimination Problem." nic:Ac1entic_cccrne1_cf_hueineec. 24 (Summer) 43-48. Gorecki, Paul K. (1986), "The Importance of Being First: The Case of Prescription Drugs in Canada,"1n§an; neticne1_Iccrne1_cf.1nccetrie1_crcenizecicn. 4. 371-95. Hannon, Kerry (1988), ”Girdling the Globe," Ragnaa, 142, 14 (December 26) 56. Hartley, K., R. J. Lavers and A. K. Maynard (1986), "Regula- tion and Development Times in the U.K. Pharmaceutical Industry." SccctichcIccrne1_cf_£clicice1_zccncmx. (November) 355-69. Hartley, Keith and Alan Maynard (1982), "The Regulation of the UK Pharmaceutical Industry: A Cost-Benefit Analysis.” Menace:iei.enc_cecicicn_nccncmice (UK). 3 (September) 122- -30. Heiduk, Gunter (1982), "Multinationalization in the Pharmaceutical Industry as a Response to National Health Policies." ueaeceriei.end_necieicn_Eccncmice (OECD), 3 (December) 194-204. 243 Heins, John (1986), "Take Two of These and Call Tokyo in the Morning,” Eaxnaa, 137, 6 (March 24) 178-80. Helms, Robert B. (1975), American Enterprise Institute for Public Policy. I Hofer, Charles W. (1975), "Toward a Contingency Theory of Business Strategy.” Acecemx_cf_Nenecemenc_cccrnal. 18 (December) 784-810. Inaba, Yu (1988), "Achieving 100 Billion Yen Sales Target,” Tckxc_Bueineee_chex. (May) 54-5- (1988), "No Magic Pill for Pharmaceutical Makers." Tck¥c_nceineee_chex. (June) 34-8- Industry Analysis Division, United States Department of Commerce, International Trade Association (1985), . sus‘ : i=:‘=:u‘! 0 _n‘ _ “oxen: Industry- Jain, Subhash C. and Girish Punj (1987), "Developing Marketing Strategy: A Framework," Mankaaing Intelligence_i_£1enning (UK). 5. 34-39- JameS. B- G- (1977). The_£cccre_cf_che_uclcineticnel 2hermecectice1_1ncuecrx_cc_1222 , London: Associated Business Programmes. Johanson, Johny and Hans B. Thorelli (1985), "International Product Positioning." Icurne1_cf_Interneticne1_Bccineee Studies. (Fall) 57-75. Kaitin, Kenneth, Nancy Mattison, Frances Northington, and Louis Lasagna (1989), "The Drug Lag: An Update of the New Drug Introductions in the United Kingdom, 1977 through 1987," nice Pharma 0 o a d e a e ' , 46, 2, 121-138. Lammey, Glenn D. (1987), "New Product Portfolio Power," ' , 72 (October) 64-70. Larreche, Jean-Claude (1980), "The International Product/ Market PortfoliO." in Interneticne1_Herketinc_§tretecY (revised edition), Hans Thorelli and Helmut Becker, eds. Pergamon Policy Studies on Business, New York: Oxford: Sydney and Paris: Pergamon, 295-305. Leffler, Keith B. (1981), "Persuasion or Information? The Economics of Prescription Drug Advertising," Qanrnal_af Lec_enc_Eccacmice. 24. 45-74. 244 Lusch, Robert F.and Gene R. Laczniak (1989), "Macroenviron- mental Forces, Marketing Strategy and Business Performance: A Futures Research Approach," Jan;nal_a£ the_Acecem¥_cf_uerkecinc_§cience. 17 (Fall) 283-295- Lynn, Robert A. (1987), ”Anticipating Competitive Reaction: Marketing Strategy in the 19803,” lanrnal_g£_§2n§nnez Marketing. 4 (Winter) 5'12- MacMillian, Ian C., Donald C. Hambrick and Diana L. Day (1982), ”The Product Portfolio and Profitability - A PIMS-Based Analysis of Industrial-Product Businesses.” Acecemx_cf_nenecement_lcnrnel. 25 (December) 733-755- Mattison, Nancy, Eilleen Thomas, A. Gene Trimble, and William W. Wardell (1984), ”The Development of Self- Originated Drugs by Swiss Pharmaceutical Firms, 1960- 1980." Becc1eccrx_Tcxicelccx_enc_£hermecclccx. 4. 157- 73. Miller, Danny and Peter H. Friesen (1986), "Porter's (1980) Generic Strategies and Performance: An Empirical Examination with American Data. Part I: Testing Porter." greenizecicn_§ccciee. 7. 37-55- Mintzberg, Henry and Alexandra McHugh (1985), "Strategy Formation in an Adhocracy." Administrecixe_§cience Qnaxnagly, 30 (June) 160- 197. , David Brinberg, Ron Klimberg, Carole Rivera, and Loyd G. Millstein (1986), ”Miscomprehension Rates for Prescription Drug Advertisements," SHIIEDL_I§§B§§_QDQ Research_in_acxertieinc. 2. 1 and 2. 93-117- Noss, Robert P. (1989), "Baseline Analysis: An Analytical Tool for Product Profitability," Bank_narka;ing, 21 (February) 30-33. Onkvisit, Sak and John J. Shaw (1989), ”The International Dimension of Branding: Strategic Considerations and Decisions." Incernecicne1_uerketinc_3exien. 22-34. Park, C. When and Daniel C. Smith (1990), "Product Class Competitors as Sources of Innovative Marketing Strate- gies." Iccrne1_cf_ccnecmer_uerkecing. 7 (Spring) 27-38- Parkinson, Gerald (1988), "Syntex Prepares for the Future," chemicel_fleek. 143. 3 (July 20) 40-1- Peltzman, Sam (1985), "The Health Effects of Mandatory Prescriptions." lccrne1_cf_Lew_enc_Eccncmice. 207-238. 245 Quelch, John A. and Edward J. Hoff (1986), "Customizing Global Marketing." Harxard_nneineee_3exiew. (Nay/June) 68. Rheinstein, Peter (1982), "A Head Start, a Broader Audience, and an Emphasis on Difference: The New Frontiers of Prescription Drug Promotion.” Eeec_nrnc_ceenetic_iew ignrnfilr 37' 330-59 Ronstadt, Robert and R. J. Kramer (1982), ”Getting the Most Out of Innovation Abroad.” Harxerc_nceineee_nexiew. (March/April) 94-9. Sadik, Abdul-Karim T. (1978), "A Note on Some Practical Limitations of Social Cost-Benefit Analysis Measures," World Development, 6 (February) 221-25. Scherden, William A. (1983), ”Strategic Planning: Do the Concepts Apply?" fiaa;ia_3ayiag, (October) 28-34, 113. Schnee, Jerome E. (1979), "Regulation and Innovation: U.S. Pharmaceutical Industry.” ce1ifernie_nenecement_8exiew. 22 (Fall) 23-32. Schwartzman, David (1975), , Washington D. C.: American Enterprise Institute. (1976), "Pharmaceutical R&D Expenditures and Rates of Return," in , Robert Helms, ed. Washington, D. C.: American Enter- prise Institute for Public Policy Research, 63-113. Sarin (1989), ”AFR data and Pharmaceutical Price Regulation Scheme (PPRS) queries in UK,” 1400 (April 5) 5. Sarin (1989), "The British Institute of Regulatory Affairs (BIRA's) diploma in regulatory affairs," 1422 (June 21) 9. fianip (1989), ”Drug regulation/research/risk reviewed," 1404 (April 19) 8. again (1989), "East-West perspectives on drug regulation," 1423 (June 23) 7. again (1989), European Offices for Regulatory Affairs Professionals (RAPS) & Drug Information Association (DIA)," 1435 (August 4) 1. $2212 (1989), "France and European Economic Community orphan drug regulations," 1397 (March 24) 6. 246 fiarip (1989), ”FRG biotech regulatory problems," 1391 (March 3) 3. fiarin (1989), "International Conference of Drug Regulatory Authorities (ICDRA) in Paris in October,” 1403 (April 14) 24. §arin_(1989), "International regulatory activity," 1385 (February 10) 21. Sarip (1989), Italian Society for Regulatory Affairs (SIAR) course on registration applications in Italy," 1466 (November 22) 3. Sarip (1989), "Italian view of EEC pharmaceutical regulation,” 1454 (October 11) 3. Scrip (1989), "Japanese regulations not subject to 301 action," 1422 (June 21) 22. garin (1989), "MEGRA (the Central European Association for Regulatory Affairs) - EEC (European Economic Community)/EFTA (European Free Trade Association) pharmaceutical interface," 1471 (December 8) 4. §arin (1989), "Ministry of Health and Welfare (MHW) tightens radiopharmaceutical regulations," 1408 (May 3) 26. §££iP (1989), "New Swedish regulatory authority by 1990," 1422 (June 21) 6. Sarip (1989), "New Swiss IKS (drug regulatory authority) commission," 1427 (July 7) 2. fiarin (1989), "Orphan regulations on US Food and Drug Administration (FDA) agenda," 1465 (November 17) 17. aarin (1989), "Swedish-Australian regulatory link," 1468 (November 29) 22. garip (1989), "Swedish regulatory productivity," 1385 (February 10) 7. garip (1989), "The Swiss pharmaceutical regulatory authority (IKS) provisions on product information," 1400 (April 5) 6. Scrip (1989), "UK parliamentary questions (PQ)'s on Pharmaceutical Price Regulation Scheme (PPRS) and pricing," 1391 (March 3) 6. fiarip_(1989), "UK PPRS (Pharmaceutical Price Regulation Scheme) data," 1433 (July 28) 2. 247 Sarip (1989), "US direct-to-consumer (DTC) regulations urged," 1473 (December 15) 21. Sarip (1989), ”US-Japan assembly discusses biotech regulation," 1435 (August 4) 21. Scrip (1989), "US Pharmaceutical Manufacturers Association (PMA) concerns about state biotechnology regulation," 1418 (June 7) 22. §arip (1989), "West German Federal Health Office (BGA) and regulatory policy," 1386 (February 15) 3. §arip (1989), "Will UK Pharmaceutical Price Regulation Scheme (PPRS) be obsolete in 1992?" 1464 (November 15) 3. Scrip (1990), "Drug regulation & new technology," 1487 (February 9) 2. Scrip (1990), "European Society of Regulatory Affairs (ESRA) - the Commission meets the grassroots - or divide and rule by Patricia Brown," 1505 (April 13) 6. Sarip (1990), "Food and Drug Administration (FDA) defends regulatory record," 1512 (May 9), 19. garip (1990), "New regulatory agency in Sweden soon?" 1506 (April 18), 1. garip (1990), "Second extension of comment period on US abbreviated new drug application (ANDA) regulations," 1485 (February 2) 18. Scrip (1990), "SIAR's (The Italian regulatory affairs society) regulatory affairs activities in Italy,” 1500 (March 28) 3. aarip (1990), "Swedish regulatory agency (SLA's) 1988/89 annual report," 1509 (April 27), 7. garip (1990), "UK parliamentary questions (PQs) on Pharmaceutical Price Regulation Scheme (PPRS)," 1490 (February 21) 5. Simmonds, Kenneth (1985), "The Marketing Practice of Innovation Theory." WM (UK). 11 (Autumn) 1- -6. Taylor, David (1985), "The Role of the Consumer Movement, and Its Challenge to the British Pharmaceutical Indus- try," in a u ic s ' Nicholas Wells, ed. New York: St. Martin's Press. 248 Taylor. H- and G- Voivodas (1987). The_nrieeeiznyere_3epcrtl_ , New York: Louis Harris and Associates Study 861018. Teeling-Smith, George, ed. (1972), Tha_£narmaaanriaal s , Symposium held at the Imperial College of Science and Technology, London: Office of Health Economics. Teiteiman, Robert (1988), "Wilting Lilly," Einanaial_flgr1d, 157, 10 (May 3) 36-9. Telser, Lester G., et al. (1975), "The Theory of Supply with Applications to the Ethical Pharmaceutical Industry," 1curnel_cf_Lew_enc_Eccnenice. 18 (October) 449-78- Tucker. D- (1984). The_Kcr1c_Hee1ch_nerket1__1he_£nccre_cf the.£hermecectice1_1nccecry. New York: Facts on File Inc. Varadarajan, P. Rajan (1990), "Product Portfolio Analysis and Market Share Objectives: An Exposition of Certain Underlying Relationships," Eerke_ing_§cience. 18 (Winter) 17- 29- Venkatraman, N. and John E. Prescott (1990), "Environment- Strategy Coalignment: An Empirical Test of Its Performance Implications." Stretegic_nenecement_lccrne1 (UK), 11 (January) 1-23. Vernon, John M. (1971), "Concentration, Promotion, and Market Share Stability in the Pharmaceutical Industry," Qanrnal of Ingnsrrial Eacnanics, 19 (July) 246-66. Waxman, Henry A. (1988), "The Politics of the New Drug Economics. " Bucineee_enc_neelth. 5. 5 (March). 21-2- Wells, Louis T., Jr. (1983), "Social Cost/Benefit Analysis for NNCs." in H_necinc_Effectiye1y_in_the_flcr1c Herketplece. Douglas N. Dickson, ed. The Harvard Business Review Executive Book series, New York: Chichester, England; Toronto and Brisbane, 398-411. Wells, N. (1983), The S co Ph aco R , London: Office of Health Economics.