PATENTS IN THE ANTIBIOTIC INDUSTRY MICHIGAN STATE UNIVERSITY PETER M. COSTELLO 1966 IHESIS #.-..~.—-w~l . This is to certify that the thesis entitled "PATENTS IN THE ANTIBIOTIC INDUSTRY" presented by PETER M, COSTELLO has been accepted towards fulfillment of the requirements for PH.D degree in ECONOMICS flaw /. Major professor Date August 17, 1966 0-169 . 1 if K ‘I R Y Michigan State 3 University NI .‘rn :3:p-:ition : tiustry unis is engined f to delineate Pricing, and ABSTRACT PATENTS IN THE ANTIBIOTIC INDUSTRY by Peter M. Costello The thesis explores questions of price and product competition in the antibiotic segment of the ethical drug industry under the influence of product patents. The industry is examined from the early l9h0's through 1960 in an attempt to delineate the affects of patents on technical progress, pricing, and on the competitive behavior of firms. Major sources of data were: (1) The Hearings and Reports of the Senate Subcommittee on Antitrust and Monopoly, (2) briefs and related papers from the Federal Trade Com- mission v. American Cyanamid Company, g£_al., case, (3) interviews with the research directors of several anti- biotic producers. The major conclusions of the study were: (1) the nar- row spectrum antibiotic market was highly competitive in terms of price and product innovation, while in the broad spectrum market competitive forces were suppressed. The shift in the character of the market was accomplished through the utiliz- ation of product patents to exclude entry, refusal to sell in the bulk market, forward integration into direct selling, and collusion among five of the major sellers of broad Spectrum antibiotics. (2) The discovery of the broad spectrum drugs Etectr'm as: ;-;e ‘ I 5 l d . :39 re Ve o- S. ' Tetracycii. covered, Ia p: I. u'au‘ 5‘ 1 l Uufivb We Jaay‘i hi". A;;, woe to see" 5:31“? and tries Peter M. Costello was the direct result of the invention of a method of invent- ing, and competitive pressure on profit margins in the narrow spectrum market. With the establishment of patent monopolies the rate of significant antibiotic innovations declined. (3) Tetracycline, the last major broad spectrum drug dis- covered, was probably not patentable over its substitute chlortetracycline. As a result its corporate inventor found it necessary to practice misrepresentation in the Patent Office to secure the patent thereby preventing unrestricted entry and price competition. (4) In the broad spectrum market entry in the form of substitute products caused the price leader to adjust prices downward. The resulting in- creased size of the market allowed rivals sufficient sales volume to reach an efficient level of production. The be- havior prevents the appearance of price competition and is an alternative to collusive agreements. With the innovation of tetracycline the collusive agreement was substituted for price adjustments. (5) Recent changes in public policy ought to have some effect in improving certain dimensions of the in- dustry's performance. But substantive improvement can only occur with basic changes in the patent statutes. fATBhTS IN Tfib ANTIBIOTIC lhDUSTdY PK by Peter M. Costello A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1966 ACKNOWLEDGENENTS The research and writing of this dissertation has generated a list of debts to be acknowledged. hy thanks to Professor Joel B. Dirlam, for his critical commentary throughout this work; to Professor Walter Adams, for his guidance in the area of public policy; to Professors Victor E. Smith and Paul Strassman, for their suggestions on methodology; to the research directors of several pharmaceutical manufacturers for their time in explaining the research problems confront- ing the industry; and to my colleagues for their patience in numerous discussions. Finally, I extend my appreciation to my wife and family for their endur- ance throughout this work. 111 TABLE OF CONTEhTS ACKIVOWLEDGEI'lEITTS o o o o o c o c c o c o o 0 LIST OF [PADLEJIS O O O I O O O O O O O O O O 0 LIST OF TABLES IN APPENDIX . . . . . . . . . INTdODUCTIOId O O O O O O O O O O O O 0 O O 0 Chapter I. II. III- IV. V. VI. VII. APPENDIX OKIGINS OF THE ANTIBIOTIC INDUSTJY THE TECHhOLOGY OF ANTIBIOTIC RESEAnCH THE TETdACXCLIkE PATENT PROCEEDINGS Th5 TETdACYCLINE LICERSES . . . . . Illustration--Cournot Duopoly Model iESEARCH AND DEVELOPMEhT . . . . . CONCLUSIONS AND PUBLIC POLICY . . . BIBLIOGlhiPHY o c o c c o c o c o o o c o o 0 iv Page iii vii Table l. 10. ‘11. 12. 13. in. 15. LIST OF TABLES Antibiotic Patents Held by Commercial Firms 1942—1956 0 o o o o o c o o o o o o o o c o 0 Cost of Construction of Penicillin Facilities Authorized by WPA, and Amounts Authorized for dapid Depreciation 1943-1945 . . . . . . . . Investment Costs and Capacity of Penicillin Facilities I944 . . . . . . . . . . . . . . . Production and Sale of Penicillin l9h4-l950 . . hajor Antibiotic Drugs in Current Use . . . . . Ranking of Major Antibiotic Drugs . . . . . . . Karket Shares of Antibiotic Producers from Own PrOGUCtion . 0 O O O O O O O O O O O O O 0 0 Summary of Restrictive Clauses in Foreign Licenses 0 O O O O O I O O O O O 0 O 0 O O 0 Net Profits after Taxes as a Percentage of Sales for_Four Broad Spectrum Antibiotic Producers 1946-1959 c o o o o o o o o c o o o o o o c 0 Net Operating Profits as a Per Cent of Sales before Taxes and by Products 1950-1956 . . . Computed Prices of Chlortetracycline and Procaine Penicillin on a Per Day Basis . . . Comparison of EXpected and Actual Market Shares and Prices 1949-1953 . . . . . . . . . . . . Number of Firms Producing and Selling Selected Antibiotic Drugs, 1960 O O O O I O O O O O 0 Numbers of Antibiotic Producers and Packagers . Percentage Changes in the Average Price deceived for Four Antibiotic Drugs 1946-1961. 39 51 57 :7 151 164 172 176 177 19. 20. 21. 22. 23. 24. 25. Percentage Decreases of Average Yearly Price for Antibiotic Drugs by Markets Estimate of monopoly devenue in Tetracycline 1953-1960 c o o o o o c o o c 0 Distribution of 84 New Drugs by Categories of Significance and by Years 1945-1959 desearch and Development Costs per New Drug Discovery 1951-1959 . . . . . . Ethical Drug Sales, Research and Selling Expenditures for 1958; Inventions and Innovations for Twenty Major Drug Producers 1945-1959 c o o c o c o o o c 0 Expected and Actual Values for Inventions and Innovations . . . . . . . . . . Percentage Distribution of 39 Inventions of 20 Firms by Categories of Significance and by Size Groups for 1945-1959 . . . Percentage Distribution of 70 Innovations of 20 Drug Firms by Categories of Significance and by Size Groups for the Period 1945-1959 Expected and Actual Values of Selling Expenses for Small, Medium and Large Firms for 1958 Percentage Distribution of Innovations for the five Tetracycline Sellers and the Industry for 84 New Drugs 1945-1959 . . . vi 0 189 194 210 214 213 216 217 Table l. TABLES LISTED IN APPENDIX Percentage Distribution of Manufacturer's Production of Antibiotics, 1950 . . . . . . . Percentage Distribution of Manufacturer's Production of Antibiotics, 1956 . . . . . . . Estimated Distribution of Revenue from the Sale of Antibiotics from manufacturer's Own Production, 1950 . . . . . . . . . . . . . . Estimated Distribution of devenue from the Sale of Antibiotics from Manufacturer's Own PrOduCtion, 1956 c o I o o o o o o o o o o 0 Production and Unit Value of Medical Grade Antibiotics, 1945-1960 . . . . . . . . . . . Percentage MarKet Shares for Tetracycline, 1954*].ng o o o o o o o o o o o o o o o o o 0 Major Pharmaceutical Markets and Percentage Distribution by Products, l95b and 1958 . vii 243 zuu 2&5 247 INTRCDUCTION The 1961 report by the Senate Subcommittee on Anti- trust and Monopolyl on the practices and performance of the ethical drug industry, is the fourth in a series of in- vestigations of major American industries. While the in— vestigation of the steel industry drew its importance from the basic importance of steel in an industrial economy, ethical drugs draw their importance from their position as a determinant of the health and productivity of human resources. The industry in its invention and innovation of new drugs has been an important factor in relegating certain diseases, such as typhus and typhoid, to the category of medical curiosity. In the case of many other diseases, the period of institu- tional care has been sharply reduced. In this industry where the basis of its social contribution lies in the pro- ductivity of its investment in research and development; where with the innovation of a new drug, market demand grows rapidly and is generally not discouraged by price; where the initial price may be set relative to what the market will bear before competitive substitutes enter; and where the ". . . industry uses the same promotional devices that other businesses use to insure that its new product is used as —“ 1U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly, Report, Administered Prices, Drugs, Report NO. #48, 87th Cong., lst Sess., 1961. 1 2 "2 a situation has developed widely and as soon as possible, in which the majority of the Subcommittee found much to criticize. The major conclusions of the Subcommittee can be sum- marized in the following six points: (1) The patent has been used as a device to restrict domestic and international competition in drugs. By preventing or delaying entry, the firm holding a product patent is in a position to exploit the domestic market and to divide world markets under the guise of the patent. Licensing, which results in duopoly, has often been the result of inadequate manufacturing or sales facilities, while oligopoly has often resulted from simul- taneous discoveries. Licensing does not result in price competition nor does it preclude attempts at product differentiation through advertising. (2) Specialization in research and development in combination with patents has led to a high degree of seller concentration in manufacturing, but lower concentration in selling. This, the Subcommittee felt, reflected the major firm's attempt to market a "full line” of drug products indicating that economies of scale may be present in the selling activity. This has led to cross licensing agreements to sell but not to manufacture drug products and accounts for the difference in the levels of concentration. (3) Since the mid-1950's, drug manufacturers have failed to develop significant new drugs although the number of so-called new drugs has increased considerably. 2Dr. Walter Modell, "Hazards of New Drugs," The_§all Street Journal (April 29, 1963). 3 The majority of the Subcommittee argued that this development was due to excessive allocation of research personnel to patent circumventing research and away from the higher-risk basic research. By marketing a close substitute, the firm's advertising potential could be brought to bear, thereby in- suring a share of the market. (h) Both the quality and quantity of the industry's advertising efforts came under sharp criticism. The control over the advertising to physicians through the Food and Drug Administration was termed inadequate. (5) In the matter of price, the Sub- committee compared prices in four product markets with man- ufacturing costs, prices in foreign countries, and prices of smaller firms where such existed. From such comparisons it was concluded that prices were excessive. (6) In the exam- ination of profit data, the comparison between the rates of profits of the largest drug firms with all manufacturing firms, provided the basis for the conclusion that profit rates in the industry were excessive. The majority conclusions were debated by representa- tives of the industry in testimony before the Subcommittee.3 The essence of the industry's argument was that there has been significant progress in drug therapy as a result of the 3The minority of the Subcommittee did, of course, present criticism of the conclusions of the majority. Un— fortunately, their arguments are not well-developed and so dispersed with appeal to non-economic factors (the present Italian law is a product of the Mussolini regime and is hardly one for us to emulate, Subcommittee Report, 22. g;£., p. 320), as to make them unconvincing. Hence, the industry arguments are developed largely from the Hearings. u patent system. The investment in research in a competitive economy is inherently fraught with risk. These risks fall into three categories. First, there is the risk that the research prOgram will not result in a marketable product. Second, that several firms may discover essentially the same product at the same time. Such a contingency will involve cross-licensing, market sharing, and price competi- tion. Third, that the probability is high that a new drug product will be rendered obsolete within five years of discovery. Thus,product patent protection is necessary to allow firms to recover their past research expenditures through the sale of their innovations. On the matter of selling costs, the industry argues that relatively large expenditures are necessary to provide a continuous source of information to the physician on new drug products as well as up-to-date information on older products. The size and scope of this activity were held to be essential in maintaining the rapid improvements in the health of the population. Relative to the charge that the industry's rate of new drug introductions has declined since the mid-1950's, the industry offers two counter arguments. First, that the Subcommittee lacks the technical competence to decide what is, and what is not a significant new drug. Second, in the last twenty years the numbers of new drugs marketed by United States drug manufacturers is significantly greater than by any other country. 5 Profits are admittedly high --but two factors are offered for consideration on this point. First, the risk of failure in the drug industry is significantly higher than in other industries, hence the rate of return must be greater to attract investors' funds to maintain the level of re- search spending. Second, that to compare the rate of re- turn to the firm's net worth is not a meaningful measure of the firm's profitability. The significant asset is not capital but rather the scientific staffs. Thus,the better measure is to compare earnings to sales. The logic of the argument here is that if scientific personnel could be treated as an asset this would increase the net worth of the firm and thereby reduce the rate of return. Failing this, earnings to sales are offered as a substitute which accomplishes the same end—-reduces the rate of return. The essences of the arguments focuses on the role of patents. The industry maintains that patents provide the incentive to invest in research which leads to increasingly higher levels of productivity, and that product competition provides a check on monopoly positions established through patents. The Subcommittee argues on the other hand that the patent has the effect of decreasing the level of industry productivity and that product competition is not an efficient check on patent monopolies. The role of the patent then is the heart of the issue. This study examines the role of patents in its effect on competition. It also will inquire into the traditional 6 rational of the patent as a device to stimulate investment in research and development. The issue is whether a structure of monopoly or competition tends to yield high levels of technological progress. In examining these and related questions, two ap- proaches are possible. First, the data developed by the Subcommittee could be reorganized and reworked to test the accuracy of the conclusions. The second approach is a de- tailed analysis of a particular segment of the ethical drug industry to test the accuracy and applicability of the Sub— committee's conclusions. Within the limits of the available data, the latter approach is followed. The segment of the industry chosen for intensive analysis is the antibiotic sector. The reasons for selecting this particular segment of the industry are: (1) it accounts for the largest dollar volume of sales in the industry, (2) it exhibits a clear contrast between structure and performance under competitive and monopolistic conditions, (3) the conduct with respect to patents can be clearly indicated, and finally (h) the general body of data available is considerably greater than in other segments. The plan of this study is as follows. Chapter I traces the history of the antibiotic industry from the dis- covery of penicillin in England, through commercial develop- ment in the United States under federal government sponsor— ship, and up to the discovery of streptomycin. In this period, the patent appears as a minor determinant of industry 7 conduct. While certain modifications of the basic penicillin drug were patented, and streptomycin with its modification, dihydrostreptomycin, were fully protected by patents, the market for these drugs was essentially competitive. The competitive character of the market was largely a result of the structure of the industry during this period. That is, there was generally separation in drug production from marketing. With the development of forward vertical inte— gration by producers, through merger as well as internal growth, the full potential of the patent as a device to monopolize was realized and utilized. With the discovery of the broad spectrum antibiotics, the patent as a device to exclude entry reached its height. Chapter II is concerned with the technology of anti- biotic research. Here the questions are, what are the characteristics of antibiotic research, where is the re- search activity located, who discovered these drugs and what was the significance of the various discoveries? The thesis of this Chapter is essentially that the development of a soil screening technique at Rutgers University placed at the dis- posal of the industry a research technology that was capable of being employed on a mass production basis by relatively unskilled individuals. The resulting discoveries of the broad spectrum antibiotics were a different sort of discovery than that envisioned as rewardable by a grant of monopoly. The third Chapter analyzes the patent proceedings which led to the securing of the patent on tetracycline. By 8 the mid-1950's, the value of the exclusive market had been clearly demonstrated. The appearance of tetracycline as a substitute for two established broad spectrum drugs threatened to open the market to competitive entry. The Federal Trade Commission case against five of the six major antibiotic sellers provides the information on the importance of the patent and the extent to which these firms were willing to go to insure the issuance of this patent. Questions concerning Patent Office proceedings are raised as is the question, should the tetracycline patent have been issued, and did it meet even the Patent Office's rather low standards of invention? In addition the writer raises the basic question of whether any of the broad spectrum antibiotics met the standards of invention in light of the importance of the soil screening technology. Chapter IV takes up the foreign and domestic licenses which followed the issuance of the tetracycline patent. Were these licenses restrictive to the extent that prices could be controlled? Did they restrict entry in a manner that allowed the development of a cartel in world markets? Does the existence of these agreements explain the apparent differences in drug price among different countries? Chapter V is an analysis of antibiotic pricing. A utility theory is used to explain the setting of initial Prices on new drugs as opposed to some form of target pricing. The Cournot duopoly model is used to explain the type of pricing which develops with entry. The overriding consid- 9 eration in pricing is found to be joint industry profit maximization, at least with respect to the prescription market. Here also an attempt is made to estimate the competitive price of tetracycline, which in turn is used to estimate the monopoly revenue available to these producers. The following chapter examines the technological advance of the industry as well as the antibiotic producers. A sample of the largest firms in the industry is selected to test the relationship between research and development ex- ‘penditures, technolOgical advancement, and size. Here the (guestion is, does monopoly insure a higher rate of new drug ciiscoveries? The final chapter is largely concerned with specific zpublic policy proposals that have been advanced to improve 'the performance of the industry. These proposals are ex- Ennined in increasing order of their effectiveness in improv- ‘ing industry performance. CHAPTER I caigINs_9§ THE gwgislcgig_;NDU3TBY This chapter traces the early history of the anti- laiotic industry from the discovery of penicillin in 1928 to 1:he marketing of the broad spectrums in 1956. The chapter i.s divided into four parts. The first is concerned with the Ciiscovery and development of penicillin in England up to l9hl. flihe second covers the efforts of the British researchers and t:he federal government in securing large scale production. UThe following section describes the efforts of the Committee c>n Medical Research and the War Production Board in develop- 1.ng manufacturing capacity in penicillin. The final section Cliscusses the penicillin market structure in its effects upon E>rices, technological progress, and forward integration. _pIS§CVEBXfi§NDV2§V§LOPME§E_OF_§§NIC£§§IN rww—vv The discovery and early development of penicillin can be dated from Fleming's discovery of the antibacterial properties of one of his cultures in 1928, up to 19hl when Florey and Heatley came to the United States in a search for commercial suppliers. This period is primarily concerned with Fleming's work which led to discovery, and with the work of a research group at Oxford College who advanced Fleming's findings to the stage where the clinical effective- 10 11 ness of the drug was proven. Fleming'szesearch.—-This history of discovery be- gins during the First World War in the search for an anti— septic effective in the treatment of wounds. The search by Fleming, as well as others, was for a substance that would destroy bacteria, leaving tissue unharmed. The research was unproductive in providing a solution, but Fleming continued this work after the war.1 In 1928, while employed as a bacteriologist at St. Mary's hospital in London, Fleming succeeded in isolating penicillin. His technique in this series of experiments was to grow cultures of bacteria on plates. When the growth was well advanced, the antiseptic substance was introduced and its effect observed. This set the stage for the ac- cidental contamination of a plate containing a highly developed culture. Normally contamination would have meant the discarding of the culture, but Fleming noted that the bacteria were being destroyed by the contaminating substance. Further work enabled him to isolate the bactericidal sub- stance and identify it as Penicillin rumbrum. In addition, Fleming was able to establish the range of antibacterial activity of his broth, to indicate that it was not toxic to tissue in laboratory animals, and to develop a method of producing the penicillin broth.2 —__ vi '— ._— ls. w.F1oreKWet al., Antibiotics (London: Oxford University Press, 19 9m , 31. 21bid., 631—33; also H. T. Herrick, Antibiotics (Brooklyn: Chemical Publishing Co., 19n9), pp. 9-15; and Alexander Fleming, Penicillin, Its Practical Application (London: Butterworth & Co., Ltd., 1950): 9?. 1-25. 12 There were, however, certain problems involved with penicillin research that Fleming was unable to solve. First, and of major importance, was the fact that Fleming's method of producing the broth gave quite low yields which limited the amount of the drug available for experimental purposes. Second, was the inability to extract the active substance from the broth. This, combined with the low yields per liter of broth, required large amounts of broth to be pro— duced for each experiment. Finally, the potency of the broth was highly unstable which made comparative tests of its effectiveness difficult.3 Of these problems, it was the first that regulated penicillin to the status of a laboratory curiosity. Writing in l9h0, Fleming said we have been using it in the laboratory for over ten years as a method of differential culture. It was used in a few cases as a local antiseptic, but although it gave reasonably good results, the trouble of making it seemed not worthwhile. 4 Thus, a problem in production became the critical element that remained unsolved in Fleming's research. The discovery that was to provide the foundation for a $360 million industry by 1953 lay relatively unknown in a series of articles in the British Journal Q§_Experimentalggatholqu. Florey's Research.-—In 1938 interest in antibacterial substances developed at Oxford College's William Dunn School of Pathology. By 1939, Florey and Chain had advanced the —- _‘ cum"- --v--e «- 3Florey, op. cit., pp. 633—34; Fleming, op. cit., pp. 11—12, 26-27. Florey, op. cit., p. 633. __ _____ ”— —‘~ -—--¥-~.-on4 13 research to the point where an application for funds was filed with the Rockefeller Foundation in New York.5 During this period, Fleming's work came to the attention of the Oxford group. The review of the literature turned up confirmations and extensions of Fleming's experi- ments, but in general the problems that limited Fleming's research on penicillin also limited others. The Oxford group was finally successful in extracting penicillin from the broth.6 The extraction of the active substance meant that the potency or stability of the compound was measurably increased, and with the development of assay technique, comparisons and correlations of penicillin samples against bacteria could be made.7 In May 1940, the first tests on laboratory animals were carried out which demonstrated the chemotherapeutic properties of penicillin. These results led to further in— vestigations of increased scope with the first patient treated in January 1941. But in these early clinical trials, the problems of sufficient quantities of penicillin were critical. Here the supply restraint was of sufficient importance to limit the patients receiving the drug to children, in the hope that less of the drug would be required to demonstrate and test its effectiveness. In spite of this problem, clini- cal tests did demonstrate that penicillin was effective in 5Florey, op. cit., p. 636. 6Florey, op. cit., p. 638. 7Florey, op. cit., p. 637. 14 txreating seriously ill patients in cases where surgery and otflier drugs were not effective. In addition, these trials skuawed that penicillin was non-toxic to human tissue.8 It is apropos at this point to indicate the con— ixritmtions of the Oxford group as it was their successes bfliich, to a large degree, stimulated government and com— muercial interests in the production of penicillin. In this Jregerd it could be said that this group was responsible for tune development of penicillin to the state of a "workable Clngn" Or in slightly different terms they were able to "Ieduce to practice" the discovery of Fleming. This group Pvas able to solve certain of the technical problems associated twith Fleming's discovery and as a result were responsible for t:aking penicillin out of the realm of a laboratory curiosity. Eiut with success there was also failure. The Oxford group teas still faced with the problem of low yields from their Iaroths and it was to this problem that they turned their sittention in mid-1940. In turning to the question of large-scale commercial production of penicillin, the Oxford group had developed a technique for the production of the broth. This technique required a large number of shallow pans to contain and grow the culture medium. These culture mediums were injected with the penicillin mold and growth or fermentation was allowed to take place on the surface of the medium. After a period of some ten days the culture was harvested and the penicillin 8F10rey, 0P. Cite, pp. 638-614’90 15 swilts drawn off. The process was highly labor intensive, as enach pan had to be prepared, injected with the penicillin unald, and finally harvested. In addition, the low yields gpex'ldter of broth resulted in high unit value, although we Puxve no estimate of the actual cost. The British chemical firm first approached on the question of large scale pro— dalction was evidently unable to undertake the necessary re- :search leading to mass production due to supply and personnel Jrestriction resulting from the war.9 It was in this period 1ihat the "battle of Britain" was at its height. The war in England and the resulting strains on IBIdtish manufacturing capacity and transportation system ].ed to the decision to attempt large-scale production in the United States. Thus, in July 1941, Florey and an associate, PW. G. Heatley, brought to the United States samples of the crultures used in the production of penicillin along with the tzechnical problems of mass production. It is of interest at this point to attempt an estim- aation of the costs of the early research on penicillin. Four institutions provided funds to the Oxford group and Florey Provides the amounts received from three of these institutions. First, there was the Rockefeller Foundation which between 1940 and 1945 provided $24,560. Second, the Medical Research —_ __ w —— wfi —— i 9H. W. Florey and E. P. Abraham, "The Work on Pen- icillin at Oxford," Journal of the Histor of Medicine and Allied Sciences, VI, No. 3 (1951 , 3O , cited in Federal Trade Commission, Re ort on Antibiotic Manufacture (Washington: 1958), Appendix I, p. 310. Hereafter cited as FTC Report. l6 Chauncil between 1939 and 1945 provided $33,148; and third, tflie Nuffield Provincial Hospital Trust which provided $22,584 txetween 1943 and 1945. In addition, Oxford College paid some snalaries of the research group as well as the costs of lab- f the costs of the research work of Fleming, but one would eexpect that the sum involved was quite small. Of the estimated $100,000 cost of penicillin research :111Ehgland, not all can be allocated to penicillin directly. IEn the latter years of this period the scope of the Oxford agroup's research expanded in the search for other substances twith properties similar to penicillin.11 As an estimate a :figure of $75,000 is suggested as the cost of reducing Fleming's discovery to practice. Two conclusions are suggested at this point. First, the data on the costs of penicillin research support the ob— servation that the costs of basic or pure research are gen- erally low relative to other stages in the process of in- novating.12 Here we have classified the research of the ¥ _. ___ r—w— fiv—r—w '— v—v—r—v— r—f '———— " —— W W loFlorey et s;., Antibiotics, pp. 669—70. 11Ibid. 12See for example the statement of D. Novick, in 0.8. Congress, Senate Subcommittee on Antitrust and Monopoly, Hearin s on Administered Prices in the Drug Industry, 86th Cong., 2d Sess., 1960, Part 18, p. 10513; and J. Jewkes, D. Sewers, and R. Stillerman, The Source of Inventions (London: Macmillan & Co., Ltd::”1§33), PD. 75. g2-90- 7" 17 OXTOrd group as basis. The second conclusion is more in the Inature of an observation. The process of discovery is the :xroduct of the contribution to the stock of knowledge of nuany individuals. Directly involved in the discovery process (If penicillin was not only Fleming and the Oxford group, com- :nased of some twelve individuals and ten assistants,1:3 taut others who repeated and confirmed Fleming's experiments following their publication. Indirectly involved were those iandividuals who contributed to that stock of scientific linowledge necessary to those more directly involved. DEVELOPMENT 9; PENICILLIN IN mpg; UNITED smgmss' The development of penicillin in the United States czoincides with the visit of Florey and Heatley in July 1941 11p to 1943 when significant quantities of the drug were first I>roduced. This phase of the history of the antibiotic in- Clustry involves federal government sponsorship through the (foice of Scientific Research, the War Production Adminis- ‘tration, and the Northern Regional Research Laboratory of the ZDepartment of Agriculture. Private support was provided by several drug firms as well as hospitals and universities which became involved in the clinical phases of the program. Govegnment Interest in_Penicillip.-—In June 1941, the Committee on Medical Research was formed under the Office of Scientific Research and Development. (These are hereafter referred to as the CMR and the OSRD.) The function of this 13Florey et al., Antgpiotics, p. 639. 18 (committee was to ". . . initiate and support scientific re- ssearch on medical problems affecting the national defense."1u 13y September 1941 this committee was convinced of the value c>f Florey's research and agreed to expedite the production of {Denicillin. In October, the first of a series of meetings tvas held between the committee and representatives of four EJharmaceutical firms. The purpose of the meeting was to (discuss the possibilities of a co-operative research program lbetween Merck, Squibb, Pfizer, and Lederle Laboratories, a :subsidiary of American Cyanamid, with the committee acting gas a co-ordinating agent.15 In December a second meeting was held in Which the (Zommittee was able to secure an agreement from these four :firms ". . . to put their research teams or a fraction of ‘them on the problem. As a result we began to get a trickle (bf a supply of penicillin during the early months of 1942."16 IFurther, the agreement provided that any research results or information would be turned over to the Committee for dis- semination among the firms involved in the program. This was in accord with the Committee's goal of a maximum pro— duction of information on the potential of penicillin. —'-r rrv 1”FTC, Reporp, pp. cit., p. 34. 15U.S. Congress, Senate, Subcommittee of the Com— mittee on Military Affairs, Hearings Before the Committee on S. 1237, 79th Cong., lst Sess., 1945, Part 3, pp. 461362; C, eEOI‘t, QB. 9—1.3" p. 36 and Appendix, pp. 321-26. 16FTC, Report, pp. 313., p. 36. '— fr” 19 Commergial‘Interest_i§;ngicillinp--It appears that commercial interest in penicillin did not develop until the first meeting held between the CMR and the four drug firms in October. Up to 1940 we have no indication of commercial interest. Charles Thom, in 1945, noted that survey of laboratory correspondence during the period from 1933 to 1940 shows that requests for Fleming's organism came from a number of great laboratories en- gaged in bacteriological research, but not from the pharmaceutical manufacturers. 17 But Merck in its 1943 Annual Report stated that it had begun studies of penicillin in the fall of 1939.18 And Pfizer in 1940 began the support of penicillin research at Columbia University as a result of the published reports of the Oxford group,19 The effect of this early research on these firms' ability to later produce penicillin is of course tenuous. But it can be noted that in the period 1943-45, Merck and Pfizer built two of the largest three plants for penicillin production,20 and that for the first four months of 1944 they were the largest producers in the market.21 Thus there is the possibility that this early research fr—r'. f ‘17— __ f?— w 17Statement of Charles Thom, quoted in FTC, Report pp.‘pgp., Appendix II, p. 317. It must be noted, however, that Merck in its 1943 Annual Report stated that it had begun studies of penicillin in the fall of 1939, but there is no evidence that anything came of this in the form of being in a position to begin production in the 1940-43 period. FTC, Report, pp. 3gp., Appendix II, p. 317 and n. 64. 18FTC, Report, pp.‘p$p., Appendix II, p. 317 and n. 64. 19Florey et al., Antibiotics, op. cit., p. 650. 2OSes Table 2, p. 32. ? “lFTC, Report, pp. cit., Appendix II, p. 331. 20 indicated the profit possibilities in penicillin production for Merck and Pfizer, and that this realization led in part to the decision to invest in penicillin capacity. But there is little doubt that government interest in penicillin pro— duction played a key role. While government interest, as well as the early re- search by Merck and Pfizer, provided stimulants to penicillin production, the visit to the United States by Florey and Heatley had a disseminating influence. In search of a sup- plier of penicillin to their laboratory in England, these men visited both government and private laboratories in the United States and Canada. Florey is reported to have spoken with the representatives of the following firms: Merck, Squibb, Lederle, Lilly, Johnson and Johnson, Eastman Kodak, and Connaught Laboratories of Canada.22 or these firms, the latter three did not become involved in the penicillin pro— gram that later was to develop and three firms were reported to have expressed a serious interest in the problem. They were: Merck, Squibb, and Pfizer.23 There is no evidence of how Pfizer came into contact with Florey as this firm was not included in his list. The most likely explanation is that contact was made through the Pfizer-supported research pro- ject at Columbia. The entry or interest of Squibb was based on a co-operative arrangement between Squibb and Merck to 22Ibid., Appendix II, p. 315, n. 53. 23Florey et al., Antibioticg, pp. cit., p. 650. 21 share research findings.2u Contributiongfipj phg;Nbr§pe;n Regional Research Labpratory.--In addition to visiting commercial firms, Florey also visited the Department of Agriculture's Northern Regional Research Laboratory (hereafter referred to as NRRL). It was this laboratory that he was first directed to in July 1941, prior to his conferences with private laboratories.25 And it was in this initial meeting that the solution to the problem of large scale production was first suggested. The production problem that had plagued Fleming and which Florey brought to the United States for solution was one of technology. The problem can be divided into two parts; first, the inherent productivity of the particular strain of mold initially used (productivity here defined as the ability of the mold in question to grow or multiply within a given period of time), and second, the physical conditions in which the mold is contained for fermentation. The magnitude of the production problem can be seen in a crude comparison of the yields at different periods of time. In the 1939 to 1943 period, to produce a million units of penicillin required about 210 liters of culture broth.26 By 1945 the same quantity of penicillin could be produced with one-half a liter of broth, and by 1958, by one—quarter _-_ 2“A. N. Richards to J. C. Woodruff, Merck & Co., Inc., March 12, 1942, National Archives, RC 227, cited in FTC, Report, pp. cit., Appendix II, p. 320. 25Florey et al., Anpibiotics, p. 649. 26 Fleming, fl. 9.1.1.?" p. 370 22 of a liter.27 In addition, there was a corresponding de- crease in the time required for the fermentation process. The primary causes of the increase in the productivity of the manufacturing process were the development of improved mold strains and the so-called "deep fermentation" process. The NRRL was largely responsible for both of these improve- ments. It is of interest to note that in the first meeting between Florey and Dr. Coghill of the NRRL, the deep fer— mentation process was suggested by the latter.28 This ap- proach to the problem of large scale production was undoubt- edly based on the work of the NRRL. Prior to Florey's visit in 1941, this laboratory had been engaged in research to develop a use for corn-steep liquor, a by-product of corn- starch manufacture. Along these lines of research the deep fermentation technique was used as an experimental method. With the surface culture manufacturing process, penicillin grew on the surface of the medium, feeding on the nutrient material below. The physical conditions under which the penicillin mold came into contact with the nutrient material limited the size of the vessel that could be used, thereby breaking the production process up into as many steps as vessels used. This in turn required that each vessel be separately inoculated with the initial mold, checked period- ically for contamination and growth levels, and harvested. On the other hand, the deep fermentation process allowed the 27FTC, Report, p. 119. 28Florey et al., Antibiqpics, p. 649. 23 {I‘Owth of the penicillin within the culture medium with feeding taking place on the surrounding nutrient material. Th1 s meant that the size of fermentation vessels could be increased from one liter to tanks with capacity in the two to ten thousand gallon range. The effect was to replace the " small army of unskilled workers,"29 with a few workers skilled in the operation of an automatic process. This then was one major contribution of the NRRL. A second contribution was (the examination of different types of molds in search of a higher producing strain. In co-operation with a research group at the University of Wisconsin, among others, a series of more productive strains were found and developed for commercial production.30 A third contribution of the NRRL was the discovery that corn-steep liquor was an efficient culture medium. The economic contributions of the work of the NRRL can be summarized rather broadly at this point. The ap- plication of the deep fermentation process to penicillin Production, the discovery of the improved strains, and the utilization of corn-steep liquor as a nutrient medium, had the effect of shifting outward the production possibilities Curve for penicillin productinn. Although there is no means by’ which the shift can be quantified, the magnitude of the Shift would seem to be considerable in view of the increase in penicillin output that resulted from these developments. _____ '— f —— '— ———— 29F1eming, pp. cit., p. 43. BOFTC, Re ort, pp. 347—51. 2b. Patent Influencei-—Following Florey's visit to the NRRL and commercial laboratories, two lines of reactions appeared. First, as noted above, the NRRL concentrated on the problem of large-scale penicillin production. As a re— sult of this research, the Department of Agriculture became the holder of certain key patents on the production of pen— icillin by the deep fermentation method.31 In a December l9hl meeting between the CNR, Pfizer, Squibb, Merck, and Lederle, copies of the patent application on the deep fer— mentation process were made available:32 Two of these firms had extensive knowledge of the fermentation art. Pfizer, which produced citric acid by deep fermentation,33 and Merck which had in 1938 established a fellowship for the study of citric acid production hy this method.3u Thus,at this point at least four drug producers had the benefit of the research by the Oxford group as well as the technical information of how to produce penicillin on a mass production basis, and two possessed previous experience in the technique. But it was not until June l9hu, two and a half years later, that the supply of penicillin was sufficient to release the 31Ibid., pp. 228-29. The Department of Agriculture f°110Wed a policy of licensing all qualified applicants on a royalty free basis which in the late l9h0's was a factor in the large-scale entry which took place. 321bid., p. 327, n. 98 and p. 238. See also Govern- ment Patents Board, Chemical Products and Process Patents, gpvernment Owned Inventions Available for License, Department of Commerce, Patent Abstract Series No. 2 lwashfhgton: 195M), p, 29, where the patent on this process issued to Dr. Meyer, and assigned to the Department of Agriculture, is described. H 33T. Mahoney, The Merchants of Life (New York: arper, 1959), p. 239. h 3 FTC, BQDOI‘t, Do 3320 25 drug for limited sales to civilian hospitals.35 The hypothesis that tends to explain this lag be— tween research findings and production is that significant patents either on penicillin itself or on its process of production were unavailable. Rather than devote research resources to the problem at hand, drug firms allocated their resources to attempts at synthetic production of penicillin where both product and process patents were available. The rationale of this position was provided by Merck. They reasoned that as in the case of vitamin B, where synthesis made the natural extraction process obsolete, that synthesis of penicillin would eradicate the investment in research and equipment on the fermentation process.36 Evidently the drug producers were able to convince the CMR that the SYnthesis program offered a higher probability of success than the fermentation program. Of the 56 research contracts we have data on, nine were with drug firms and involved a SYnthesis approach. In addition, these nine contracts did not call for government payments to cover the costs of this research, although such payments were offered.37 This self- f1nancing would of course give the firm a better claim to any patents on the results of their research. The patent issue forced a divergence of goals. The CMR's primary interest was in adequate production with min- imum delays. To this end they attempted to set up lines of k 35Ibid., p. 331. 3 Mahoney, 22. 92-20, pp. 197-980 37FTC’ BERG—r29 pp. “7”4’90 26 communication between the various research projects to share information on progress. The goal of the commercial lab- oratories was the production of patentable information which required secrecy. As late as the middle of 19u3 the CMR was experiencing difficulties in co-ordinating the synthesis re— search prOgrams and in having information pooled among the various firms.38 As a result of the delays in the penicillin research program the CMR made a study to determine which firms should continue the synthetic research program. Ten firms were selected and invited to enter into contractual arrangements with the OSRD.39 The patent aspect of the contract provided: (1) there is provided a complete interchange of information through OSRD of information discovered by all OSRD contractors, (ii) the work being done by the commercial firms is financed by their own funds, and (iii) the OSRD is given the right to determine the disposition, among the organizations that make contribu- tions through OSRD of valuable information or inventions, of all patents covering discoveries or inventions made under the contracts that are attributable to the inter- change of information through OSRD. In addition the Government receives royalty free licenses for military . . . and national defense purposes under all patents resulting from work done by the OSRD contractors in the synthetic penicillin field . . . Also, OSRD has the right to require the OSRD contractors that ultimately become titleholders of the patents to license either designated organizations, whether or not they are 38"Two of the seven companies are sharing information, viz: Merck and Squibb, I believe, by formal agreement. Pfizer, to some extent, shares with Merck and is contemplating sharing also with Squibb. Otherwise, there is no pooling." A. N. Richards to V. Bush, August 1h, 19h3, records of OSRD, NA, RG 227, quoted in FTC, Re ort, p. 341. 39Nine of the firms which signed contracts with the OSRD were Abbott Laboratories, Eli Lilly, Merck, Parke Davis, Pfizer, Shell Development Company, Squibb & Sons, Upjohn, and Winthrop Chemical. FTC, Report, p. 3H6. 27 contributors of inventions or relevant information upon the payment of reasonable royalties. 40 There is some evidence that the drug firms had little choice in Whether to accept the OSRD's contract or not. The War Production Board was in a position on the advice of the OSRD to deny priorities on supplies and equipment. In ad- dition to this threat the firms were informed that the OSRD had important information from University sources as well as from the English research programs that it was willing to share with the contracting firms.nl This contract provision relevant to the patent issue clearly indicates an attempt to eliminate the delays in the penicillin program and to prevent any possible abuse of the patent position if and when established. Thus9neither a legal or'illegal monopoly in penicillin was to be allowed. Recovery of’research costs then had to be through royalties or through the price of penicillin in sales to the government. What is not explained here is why, with the patent issue settled, did these firms continue to finance this re- search program from their own funds rather than from public funds. The answer here may well be that if one firm followed this approach others would have to follow to maintain the same relative position in the determination of the allocation of patent rights. From the profit point of view, it would _-—__ A noPenicillin Contract Patent Provisions, read at a meeting of the Penicillin Producers Industry Advisory Com- mittee, February 8, l9hh. Records of the WPB, N.A., R.G. 227, quoted in FTC, Report, pp. 335-336. “ISee the Richards letter addressed to nine firms as well as the Bush letter on the proposed contract in FTC, Report, pp. Buds-42 and BUB-M4. 28 make no difference in the long run, as the cost of this re- search could be recovered in sales of penicillin to the government. In determining the price, undoubtedly, this element of cost would be considered. The experience of the OSRD with the commercial firms suggests that the patent institution had the effect of allo— cating resources away from the more promising line of develop— ment. That is, from the deep fermentation process of pro- duction to an attempt at synthetic production. With the settlement of the patent issue by contract in early 19th. and the rapid progress made by the NRRL in the deep fermen- tation process, attention shifted to implementing this PIw>cess which in a few months satisfied military demands.“2 Patents and the Production of Knowledgg.--The Question raised at this point is, did participation in the SYnthetic penicillin research program result in greater productivity of knowledge for those firms under contract to the OSRD than for firms which concentrated on production? A crude test of this question is provided by the statistics 0n antibiotic patents held by commercial firms. Table 1 indicates the patent holdings of firms engaged in synthetic Penicillin research and those not in the program for the Years 1942 to 1956. Two heroic assumptions are involved here. First, that patent statistics accurately reflect re- search productivity, and second, that the knowledge gained in the synthetic research program was transferable to the ——v—7 #2 Ibldo, pa 3A7. m ?9 disczovery and production of improved antibiotics. In the ligllt of these assumptions the data in Table 1 indicates theat: on the average the synthetic research program was not a sisrrlificant producer of new knowledge. Here firms in the prwbgrram produced on the average of h? patents, while those rwrt in the program produced 5?. The tentative conclusion here is tfldat the patent did not have any positive long-run effifeects to offset its short-run misallocation of research resources. 'IABLE l.——Antibiotic patents held by commercial firms l9u2—l956 :re—vrfi "'1? '“r-sm" ma:m:r=:=.u*z=r::===:::: Numbers of Patents held by Numbers of Patents held by Firms in the OSRD Program Firms not in OSRD PrOgram Abbott lO Lederle 60 Lilly 67 Bristol 50 Merck 111 Commercial Squibb ho Solvents M8 Parke Davis b2 Pfizer NO Upjohn 20 Total 330 158 Source: Federal Trade Commission, Economic Report on Antibiotic Manufacture (Washington, 1958), p. 235. Further support for this conclusion is evident in the statistics of inventions of these firms between 19u5 and 1959. (Invention is defined as the introduction of a new drug product either from the firm's own laboratory, or as a result of its financial support of outside research.) In these terms, on the average, the firms not in the program 30 Iqaci an average of four inventions in this fourteen year pnezsiod. (Lederle leads the list with seven, and Bristol is credited with one. Commercial Solvents is not considered heexre for shortly after the war they left the industry.) The f1 rms which participated in the program averaged 2.3 inven- tions};3 This would tend to suggest that the firms which C<>rlcentrated on penicillin production and ignored the question ‘burers for resale under their own labels. data on the number of packers begins with 1950. Between 1950 and 1956 the number of firms at this stage in the dis- trltDution process grew from twenty-seven to forty-five. Table 4 indicates that the number of manufacturers declined \ 591bia., pp. 44-45. 61 .EELQ'! p. 45. Ibid., p. 162. 62Ibid., p. 66. 40 from seventeen in 1944 to fourteen in 1950.63 On a net basis it appears that from 1945 on, the number of sellers increased. The data in Table 4 relative to the net unit value of penicillin is in accord with the Federal Trade Commission's 64 conclusion that prices fell sharply in the 1945-50 period. This then is a reflection of the improvements in technolOgy and the pressure of competition. New DrugProducts.--653etween 1945 and 1950 three significantly new drugs were discovered and marketed. These Imre streptomycin, chlortetracycline, and chloramphenicol. Each of these drugs was marketed by a firm that had found the Penicillin market unprofitable. The first of these products was streptomycin, dis- covered at Rutgers University in 1944 and marketed by Merck in 1946. Merck was the firm that had bet on the synthetic Penicillin program and when forced into production used the surface fermentation process. Chlortetracycline was dis- °°Vered and marketed by Lederle in 1948. Lederle also had eritered the penicillin market equipped with the surface feI'mentation technique and had clearly been forced out of the \ 1 63Five firms left the penicillin market completely by Sg5()--Lederle, Ben Venue, Hoffmann La Roche, Sterling, and 343p and Dhone, while two entered—-Wyeth (a division of %:elfican Home Products) and the William H. Merrell Company. meg. of the five that left the market used the surface fer- alntation process, while it is possible that the other three a!) used this process. However, on this point information is absent. 6“'FTC, Re ort, pp. 162-171. th. 65The subject is taken up in the broader context of $3 entire ethical drug industry in Chapter VI. AIII-__ n a 2. 41 market by 1947. The third drug was chloramphenicol, dis- covered at Kale University and first marketed by Parke Davis in 1949. Parke Davis also appears to have selected the sur- face fermentation process but never became a significant producer and evidently left the market by December of 1944. Less significant drugs were discovered and marketed during this period. Both Lilly and Squibb improved the basic penicillin drug. Merck discovered an improved version of streptomycin by 1949, and Pfizer found a competitive sub- stitute to Lederle's chlortetracycline--oxytetracycline. The significance of this record in answer to the question raised earlier is that the firms under the more in- tense competitive pressure were the ones that engaged in innovation. Of the firms who began production with the more efficient deep fermentation process, hence under less compet- itive pressure, only minor advances were forthcoming. Forward_;ntegration and Decline of the Bulk Market.-— With the introduction of new drug products certain basic Changes took place in the structure of the antibiotic market. 131 the case of penicillin, the basic drug was unpatentable z31nd as a result unrestricted entry both in manufacture and 1a: packaging took place. Prices in turn were driven to the competitive level. Certain firms such as Pfizer, Merck, fSchenley, Commercial Solvents, Bristol, and Hayden specialized 1n.manufacturing as they lacked the sales force necessary to Contact the physician and build brand name allegiance. For these firms the bulk market was their primary source of 42 buyers. With this type of market structure, it was impossible for the manufacturer to set price. with the discovery of streptomycin, Merck received the patent on this drug under the terms of its contract with Rutgers University. But here Merck was faced with a dilemma. To exploit its monopoly position required an extensive sales force which Merck leaked. To sell through the bulk market would repeat the penicillin experience. Merck solved this problem eventually by merging with Sharp and Dhome in 1952, but in 1946 Rutgers University forced the solution. Rutgers argued that the discovery of streptomycin was too important to await Merck‘s building of a sales force. They proposed that the patent be turned over to a non-profit organization at Rutgers, and that Merck take in return a non-exclusive license and $500,000 in royalty abatement.66 Eight producers, including Merck, were licensed to 67 produce and sell streptomycin. The market results were as anticipated. Entry at the packaging level took place and Price was rapidly driven to the competitive level.68 In 1949 Merck discovered in its laboratory dihyrdros- tPeptomycin, a substitute for streptomycin. It had a clear Clainion the patent which resulted,but again lacking a de- tall force to market the drug under its own label and in ¥ 66FTC, Re ort, p. 229. 67In 1950 the producers of streptomycin were Lilly, Squibb, Pfizer, Merck, Schenley, and Heyden, the latter four Selling primarily in the bulk market. See Table 1, Appendix, 4: light of the substitutive relationship between the two drugs, Merck granted non-exclusive licenses to other manufacturers. The market experience was a repeat of the penicillin-strep- tomycin e:ther drugs with similar properties. Potassium Penicillin G, developed in tablet form by E3quibb, claims significance as the "first orally effective IDenicillin preparation."3 Prior to this Squibb development, IQenicillin was administered intramuscularly or intravenously. 'Phe advantage claimed for the oral administration was that ‘l 31bid., Part 19, p. iosuo. 52 absorption into the blood stream was slower, therefore sus- tained concentration rates in the blood were higher over a longer period. The disadvantage of this type of penicillin, as well as other oral forms, is that they are inactivated by gastric juices and intestinal bacteria. As a result the oral adnunistration of Potassium Penicillin G would require a :five-fold increase in dosage to yield the same blood levels 518 penicillin administered intramuscularly.” This would have ‘been an increase in the recommended dosage from 1,600,000 linits to 8,000,000 units per day.5 Using the 1947 price of penicillin of 30 cents per L100,000 units,6 the intramuscular route would cost $4.80 per tic properties.25 Porter, in a 1924 paper, described the inhibiting action of several bacteria. "One of the bacteria was investigated and when grown on a plate shown to produce \ 2L‘tlbidu , p. 31. 25John Crofton, "Streptomycin,“ in Penicillin, Its Practical A lications, ed. Alexander Fleming—(St. Louis: The C. V. Mosby Co., 1950), p. 400. 60 ;ac111Tfusible substance which could inhibit the growth of furugj. as far away as 2 cm."26 Laboratory observations of use irflfllflting nature of certain substances were extended 111 aai attempt to control disease. Henery, writing in 1931, "- . . reported that the addition of bacteria, actinomycetes, axml fungi isolated from the soil reduced the severity of fCKDt-rot infection caused by H. sativum, and that fungi Nexus more effective than actinomycetes and bacteria.“27 III 1932, Time succeeded in isolating numerous actinomycetes fINDm a sugar-cane field ". . . which were antagonistic to £3 Pythium parasitic on sugar-cane."28 Nakhimovakaya, in 31937, found that ". . . in certain soils the actinomycetes were capable of exerting antagonistic action, but in others they were ineffective."29 At the Rutgers University Agricultural Experiment Station, a series of studies were undertaken beginning in 1915 to determine the occurance of actinomycetes in the SOiJ.and their effect on plant life.30 These studies in- dicnated the wide distribution of actinomycetes in the soil axui their varying degree of effectiveness against bacteria. Truare was in the period up to 1939, little appreciation of true possible significance of these studies for chemotherapy. rIThis was due in part to the unstable nature of these compounds, \_ 26Florey, et al., Antibiotics, p. 39. 27Ibid., p. 44. 28Ibid., p. 53. 29Loc. cit. 30Selman Waksman, Neomycin (New Brunswick N.J.:Rutgers University Press, 1953), p. vii. 61 and 21:1 part to their highly toxic nature.31 In 1939, Waksman and others began a comprehensive resenarch prOgram at Rutgers University on the action of anti- biotics.32 Whether this research prOgram was the result of Flinning's work in the early 30's or of private communication witli members of the Oxford group, Waksman does not say. The innxression he leaves is that the inception of the program was: an independent decision based on prior work at the ex- perriment station. In 1941 Florey's work was published and ‘ther result was an intensification of Waksman's research: The rapid progress that was being made, during the years 1941 to 1942, on the production, isolation, and utilization of penicillin soon convinced us that an outstanding chemotherapeutic substance is already avail- able for combating disease caused by Gram-positive bacteria. Therefore, it was decided that the research activities be directed mainly towards a search for anti- biotics active against Gram-negative bacteria, and later, also against acid-fast bacteria. As a result of the very first survey . . . it was established that such agents are being produced by actinomycetes and it appeared that little difficulty would be experienced in isolating them from the medium, provided the proper techniques for such isolation were developed. 33 SkKDIPtly after, Waksman isolated streptothricin, which had a typlcal antibiotic character. Testing in animals confirmed thE> laboratory observations. However, ". . . the animals th£lt recovered from the infection through use of strepto- tWXricin later died because of certain latent toxicity of \ 31Ibid., p. V111. 32Selman Waksman, "Production of Streptomycin and Other Antibiotics by Actinomycetes," eds. G. W. Irving and H. T. Horace, Antibiotics (Brooklyn: Chemical Publishing Co., 1949): P- 75° 33Ibid., p. 78. 62 this antibiotic."34 This success, although limited, combined sviish the proved success of penicillin to further intensify tdie .research effort. "It was now definitely only a question <>f csontinuous search before similar compounds with less toxic properties could be found. This led to the isolation, a year later, of streptomycin."35 The impression left by Waksman's statements is ccnmtrary to the stereotyped idea of the discovery process. 'Tkne discovery of streptomycin was not a "flash of genius" winere the solution occurred to the inventor where it had not (Dcxzurred to others. Hather it was an idea, based on the wcxrk;of many other investigators, that certain microorganisms possessed antibiotic properties. And, an iteratative approach 01‘ procedure characterizes this research. This then is one Strand in the discovery of the antibiotic drugs. The second strand necessary to both the discovery arui development of antibiotic drugs was the acceptance by 'thfie medical community that a substance normally toxic to the bCKiy could be used to control disease. The great therapeutic Stuzcess of penicillin leading to medical acceptance was a necessary condition to the discovery of other antibiotic drugs. The research process developed by Waksman and others at dutgers University was basically a procedure to screen 8011 samples. This procedure can be described or divided __1 3u'ziiaksman, Neomycin, p. ix. 35Loc. cit. 63 into five steps.36 The first step was the collection of soil samples from many different locations. From the literature on microbiology, it appears that the effective antibiotic 'between the effective therapeutic dose and the lethal dose as well as determining the general pharmacological properties. Success in this fourth stage would place the drug in the hands of the physician for review and clinical testing. If the clinical testing program confirms the re- sults obtained in animal testing the drug is ready for marketing in a technical sense. It has been estimated that cnl the average only two products a year survive the clinical stage to reach the market.39 From the simplified description of the research Process two conclusions may be gained. First, it would not be ‘walid to conclude that this techn010gy is withhut its IHbelems that must be solved on a technical level. There are {of course many such problems, from the choice of culture mediums to techniques of separation, isolation, and iden- tifzioation. These sorts of problems must first be solved befRJre the program can be turned over to others for routine oPeration. Here, then, was the nature of Waksman's contri- bution. In effect he invented a method of invention.“0 ‘ ‘fl‘ OOOOOOO 38FTC, Report, p. 113. 39Loc. ci . 40This also seems to have been the case in hybrid °°rna See Zvi Griliches, “Research Costs and Social Returns: HYbrid Corn and Related Innovations," Journal of Political Econom , va11 (October 1958), pp. 419-31. 65 The second conclusion which follows is that the re- search technology invented by Waksman was susceptive to mass production Operation. The screening process was standard- ized at several stages and capable of Operation by individ- uals of limited skill and training. The tasks to be per- formed are largely repetitive and mechanical, even to step three-~observing the antibiotic action of the isolated sub- stance. It is then this sort of mass research which seems well-suited to commercial firms and which produced the majority of the advances, both major and minor, in the antibiotic drugs . The question can be raised at this point whether Waksman's research was of a basic or applied nature. Kuznets would have us distinguish between these concepts on the basis of the workability of the discovery. That is, Once it can be demonstrated that the discovery or invention Works, what follows is applied research.41 But in the case Of Waksman's discovery, what is indicated by the term "work"? The answer to this question is complicated by the fact that here we‘ have two products--the product streptomycin, and the research technOIOgy. An application of the Kuznets definition implies that both were in the area of applied research. For here it was Florey and the Oxford group which demonstrated that a microorganism found in the soil worked in the sense that it possessed the ability to control and K E ulSimon Kuznets, "Inventive Activity: Problems of Definition and Measurement," in National Bureau of Economic Research, The Rate and Direction of Inventive Activity: EEQQQmic and Social Factors (Princeton: Princeton University Press, 1962), p. 34. 66 cure disease in man. It was this group that opened the zhedical door through which could pass streptomycin and its fkillowers. Thus,by applying the Kuznets definition, Vhaksman's research effort was not inventive activity. But 111 view of the productivity of Waksman's technolOgy this conclusion appears too restrictive. Another approach to this problem is to distinguish lmetween basic and applied research on the basis of the Innibability of success of the research project in question. If“the success (here success is defined as a marketable resullt), of the project is highly uncertain then it is a tYFHB of basic research activity. On the other hand if surusess seems fairly certain, it is a type of applied re- 42 search. Applying the probability rule to Waksman's research, the isolation of streptothricin, with all the prOperties of an éaoceptable antibiotic except for the problem of latent fatwil toxicity, indicated that an acceptable drug would be foxvtkmoming with a high degree of certainty.“3 Thus,as the research project progressed the probability of success in- creased and the project was converted from one in basic re- search to one in applied research. The implication here is that basic research in some Naymakes a greater contribution to the growth of knowledge \ 1 “ZStatement of David Novick, in Hearings on Admin- W098 - . u Part 19. pp. 10510-10523. t 43 See P- 61 where Waksman is quoted on this subject Q:‘the effect that it was only now a question of continuous " aI‘oh. :7 than does applied research, although it is reOOgnized that there is a feedback relationship from applied to basic re- search.44 Here we are clearly faced with two questions: first, how are the two activities to be differentiated, and second, why does basic research have a greater economic effect? The answer to the latter question is largely determined by the definition of basic research. If the National Science Foundation approach is used then basic re- search is an activity that seeks knowledge for its own sake without any view to practical application.45 Basic research then is an activity which adds to the stock of knowledge and applied research performs the primary function of trans- lating from this stock of knowledge to practical application in terms of new products and new processes. Or in somewhat different terms it is an activity which increases the range 0f investment opportunities available in the economy. Applying this approach to the general research that PI‘Oduced penicillin and streptomycin illustrates the dif- feI‘ences in the concepts. In the development of penicillin the research undertaken by Florey was critical. His major contributions in an economic sense were to show that pro- duotion and extraction of the active substance from the \ 1 uuSee for example Burton H. Klein, "The Decision Making Process in Development," in National Bureau of Mic Research, pp. 477-97. See also Richard R. Nelson, The Simple Economics of Basic Scientific Research," Journal wplitical Economy, va11 (1959), pp. 297-306. (W “5Nationa1 Science Foundation, Third Annual Report 1 fishington: 1953), pp. 38-39; and The President‘s Scientif- ° Research Board, "The Nation's Medical Research," Science 3nd Public Polio , V (Washington: U.S. Government Printing frice: 1957’: P0 8° 68 Fleming broth was possible. Of equal importance was the clinical demonstration of the medical effectiveness of the drug. For the latter of these two contributions ensured that demand would be generated, while the former indicated that supply was at least within the realm of the possible. Again, both of these contributions increased the stock of knowledge, but each in a somewhat different sense. There was here an increase in the stock of technical knowledge available to the scientists, but there was also the knowledge created that here was an investment opportunity. It was this increased stock of Knowledge from which the OSdD could draw to convince the commercial firms to undertake research looking to large-scale production. The research prOgrams that resulted from 08:11) pressure, was basically applied research as its goal was the solution of technical production Problems necessary to exploit the investment opportunity Presented. In the same vein the development of the soil screen- ing technique did not in itself produce investment opportu- nities. But it did produce the knowledge that the operation or this technology could isolate drugs of sufficient impor- tance that investment in exploiting these drugs might be "(nth the effort. Thus the development of this technolOgy was basic research while its operation was a question of applied research. 69 aseskaH Ahu THE TdeACICLlhES It has been argued that the research technology developed by Waksman was highly susceptible to mass pro- duo tion utilization. That is, it was a type of repetitious research that required minimum scientific skill for its operation. By definition it is a form of applied research, as the primary function was the isolation of substances with antibiotic properties. The first of the tetracycline antibiotics to be isolated was chlortetracycline. The research that eventually led to this drug began in 1944 when Lederle Laboratories (a subsidiary of American Cyanamide Company) sought an improved product over penicillin. Two factors appear to account for the selection of this research project. First, it was in this period that reports on streptomycin confirmed the soil screening approach as a producer of antibiotic drUgs.46 An indication of this success was provided by Waksman's isolation of streptothricin in 1911-3. Second, Lederle's production of penicillin was, in this period, by Surface fermentation which placed them at a competitive disadvantage relative to the deep fermentation producers. To remain competitive in the antibiotic market Lederle had two alternative courses of action. They could have invested in the necessary plant and equipment to become an efficient Penicillin producer, or they could invest in the research \ uéFlorey, et al., Antibiotics, p. 1297. 7O nenaessary to isolate an improved antibiotic drug. The first ccnxrse of action, given the alternative, would not have appeared to be particularly profitable. There were no effective patent restrictions or other artificial barriers to entry, and with marked increases in productivity of pen- :iciJLlin facilities it would have appeared that the market wcnild.be well supplied. In addition, whatever research Lederle had undertaken on penicillin, it had not shared it with other firms through the 085m“? Hence, they could ex- pect no help from OSRD in receiving patent rights.48 These considerations would. seem to have served to press for the irrvestment in antibiotic research.)+9 The research program initiated by Lederle was headed by :3 Dr. Duggar, a 72 year old retired professor from the Lhriversity of Wisconsin. The research technolOgy was one of SOiJL screening where soil samples were sent to Lederle from all. parts of the world by "explorers, Boy and Girl Scouts [Bani] oil men in foreign countries."50 By 1947 Duggar had Succeeded in isolating chlortetracycline which was first marflketed in December 1948. The patent on this drug was ‘ 47Ibid., pp. 783-84 where a list of restricted pub- licertions which were provided to British scientists through 033D, is indicated. “BFTC, Re ort, pp. 48-49, where a list of these Contractors is provided. 49Tn1s is not to argue that investment in antibiotic reSearch was to the exclusion of other areas of applied re- search developed during World War II. But the availability of Waksman's method of inventing was undoubtedly an important Variable in ranking antibiotic research high on the list of Possible projects. 50Mahoney, The Merchants of Life, pp. 175-76. 71 issued to Daggar in September 1949 and assigned to the :hnerican Cyanamid Company.)1 The evidence on the discovery of chlortetracycline senems consistent with the hypothesis that the research tecn- ru310gy developed by Waksman could be operated in a rather xuechanical manner by less skilled individuals. Why was Luiggar selected to head this research project in Spite of bias age and his failure to publish work of importance in connection with antibiotic research? Florey, in his dis- <3iission of the discovery of chlortetracycline makes no reference to Duggar, nor does his extensive bibliOgraphy cziste any of Duggar's publications.52 The 1949 edition of £tnuerican Men of Science fails to cite Duggar as the dis- c oxytetracycline progressed at a leisurely pace.57 But egmxfily in 1949 the research prOgram was intensified by the eaxx31usive and successful marketing of chlortetracycline by Iiemierle in December 1948,58 and by Parke Davis' chloram- gariennicol in March 1949.59 By November of 1949 Pfizer's eifihorts yielded a substance of sufficient interest to war- rwarit a patent application on the process and product cuxgrtetracycline. Such a patent was issued to a Dr. Sobin, Ffiirxlay, and Kane in harch 1950 and assigned to Pfizerf)O The research techn010gy utilized in the isolation <3f‘ (bxytetracycline was, as in the case of streptomycin and chlxartetracycline, soil screening of samplesreceived from “Maris? parts of the world. The isolation of this drug re- qtlizred the examination of some 100,000 soil samples by a te%1fll of eleven researchers.61 With the development of oXYtetracycline Pfizer abandoned its traditional policy of Stail—ling only to other pharmaceutical firms and entered the .4 retail market with a 117.5 million advertising expenditure.“2 ¥ 57Mahoney, 22. cit., pp. 241-42. 5BFTC, he ort, p. 231. 59"Shock Treatment for Parke Davis," Fortune, Septemoer 1953, XLVIII, 108. "—— 60FTC, Proposed Findings of Fact, Conclusions of FaCtand Law, and Order to Cease and Desist, Docket No. 7221 efore the Federal Trade Commission in the hatter of American Cyanamid, et al., June 1960, p. 11. H élmahoneY’ 223 g;£., p. 242; also John Gunther, Inside Pfizer," New York Times, March 23, 1955, sec. lO,p.€. 6dmahoney, 22. cit., p. 243. 75 The research efforts that produced oxytetracycline vmere similar to those which produced chlortetracycline. The tdiree Pfizer researchers to whom the oxytetracycline patent was issued may be assumed to have been the major contribu- txars to the research effort. let their prior contribution tn: antibiotic research was nil as measured by Florey's bibliography.63 The American Men of Science lists only two c>f' the three patent holders,64 and neither is credited with d i scovering oxy te tracyc line . Of the eleven members of the research team which worked on oxytetracycline only one member, Dr. G. 0. Hobby is extensively noted by Florey with fifteen articles. It would appear that a good deal of Hobby's work was done while Connected with Columbia University in a co-operative re- search project with Pfizer on penicillinf’5 One other mem- ber of this team is noted by Florey with two articles on techniques of assay of penicillin. Thus, of the eleven mem- bers of this research team, only one could be considered accomplished in antibiotic research, and his contribution tC’ the isolation of oxytetracycline was apparently secondary to those in whose name the patent was issued. ‘ 63Florey, et al., Antibiotics, pp. 1579-1662; according to Mahoney, The Merchants of Life, p. 242, the patent was issued to Sobin, Finlay, and kane. 64Sobin is unlisted. Finley is listed as holding the M.A. degree and Kane with the B.S. American Menu-9.1: §2l2££2; 9th ed.. pp. 593 and 1002. 651"iahoney, 92. cit., p. 239. 76 With the clinical and market success of oxytetra- cycline established, the next step was to establish its chemical structure with the objective of cheaper synthetic production. This was accomplished by a. B. woodward, who headed a Pfizer research team in the summer of 1952. Woodward was able to demonstrate the chemical structure not only of oxytetracycline but of chlortetracycline.66 With the knowledge of the chemical structure of these drugs and With the knowledge that "slight alteration with such Structures may strengthen or weaken a typical action,"67 Conver, a member of Woodward's research team, was successful in removing the chlorine atom from chlortetracycline, thereby Producing tetracycline.68 After rather extensive litigation, a subject covered in Chapter III, a patent was issued to Conver in January 1955 and assigned to Pfizer.69 The choice of Woodward to head the Pfizer research team indicates the selection of an experienced antibiotic researcher. As measured by Florey's author index, Woodward, as a member of Harvard University's department of chemistry did extensive work on the synthesis and structure of peni— \ 66Federal Trade Commission, Findings as to the Facts and Conclusions of Law, Docket No. 7211, before the Federal Trade Commission in t e Matter of American Cyanamid Co., w. (Washington: August 8, 1963), p. 7. 67Harry Beckman, The Year Book of Dru Thera (Cilicago: Year Book Publishers, 1961), p. 6. 68Federal Trade Commission, Initial Decision, Pocket No. 7211. Before the Federal-Trade Commission in the Matter of American Cyanamid Co., et a1. (Washington: 1961), Do 13. 69FTC, Report, p. 253. 77 cillin. Applying the same standards to Conver, and to Minieri, Heyden's discovery of tetracycline, indicates that neither were publishers of significant antibiotic research.70 The type of research which led to tetracycline was essentially different from that which had produced strep- tomycin, chlor- and oxytetracycline in that it involved changing the chemical structure of chlortetracycline.7l Three factors seem to account for the investment in learning the chemical structure of oxytetracycline. First, as pre- viously noted, it was recognized that a slight modification in the chemical structure of a drug substance could lead to a different (improved) therapeutic effect. Second, knowledge of a substance's chemical structure was a prerequisite to any attempt at synthetic production. The earlier history of the industry indicates that synthetic production involves lower per unit costs.72 A third factor which exerted an 7OConver, in whose name the tetracycline patent was eventually issued is unlisted in The American Men of Science. 71However, it does appear that Heyden Chemical Company did produce tetracycline by fermentation in 1953, without knowledge of the chemical structure of oxytetracycline or Chlortetracycline. Commercial production of tetracycline is by fermentation, but whether it is produced from the same Strain used by Heyden is not clear. Regardless of the Hayden Contribution, tetracycline was a product of a different type of research techn010gy than that of soil screening. See for example: American Cyanamid Company, Egoposed Ewings of F3013 Conclusions of Law and Proposed Order, With Reasons ill—Supporihereo , Be ore e Fe era Tra e Commission in the Matter of American Cyanamid, et al., October 14, 1960, la 16-17 and 54. A. Leon Higginb'TF—o am, gpinion of the Commission, in the Matter of American Cyanamid, 9331., Docket No. 7211, August 1963, pp. 5, 9-10. 72hahoney, op. cit., pp. 196-98. 76 influence was the fact that rfizer's substitute for Lederle's chlortetracycline was unable to overcome Lederle's lead in the market.73 In part this was undoubtedly due to the fact that Pfizer lacked in this period the extensive marketing facilities of Lederle. A possible fourth factor might have been the advertising value of the "discovery" of a new anti- biotic. Again, there is no way to assign relative weights to these variables, but it would appear that this combination of factors would give a high rank to the investment in de- termining the chemical structure of oxytetracycline. Was tetracycline sufficiently different from chlor- tetracycline and oxytetracycline to warrant patent pro- te<.:tion?7’+ From a technical point of view it can only be concluded that it was on the basis that it was a product with a somewhat different chemical structure. From an economic point of view, it was essentially a perfect sub- sti tute for chlortetracycline. The same range of pathOgenic microorganisms were attacked and the side effects were sim- ilar. But the patent statutes recognize technical rather than economic differences, and tetracycline met the standard of the Patent Office. The last of the tetracyclines, demethylclortetra- cycline, was isolated by Lederle in 1957. It is the product —¥ 73Rough estimates of 1950 market shares for broad SPec‘otirum antibiotics indicates that Lederle accounted for :bout 70 per cent and Pfizer for about 10 per cent. PPendix, Table 1, p. 2A1. a 7“The patentability of tetracycline is discussed in ”hapter III . 79 of a mutant strain of Streptomyces aureofaciens and is pro- duced commercially by fermentatinn.75 Beyond the fact that this drug is claimed to have increased therapeutic efficiency over the other tetracyclines,there appears to be no other published information on the research that led to the iso— lation of this drug. C HLOBAMPHENIC 01. The isolation of chloramphenicol was the result of a research program similar to that which produced the majority of the other antibiotics. Paul Burkholder of Yale Univer- sity succeeded in isolating this antibiotic from a soil sample obtained in Venezuela in 1947.76 The program was initiated in 1943 when Parke Davis made a grant of £5,000 to Burkholder-with instructions to find an antibiotic effective against six types of bacteria.77 Considerable impetus was given to Parke Davis' testing and development of this drug in the late 40's by their competitive position in the anti- biotic market. As one writer stated the case, "they missed the penicillin boom that made other firms rich and later built a streptomycin plant, but the market broke before it y 0 75Paul Folsom, Ph sicians' Desk ggference (16th ed.; radell, N.J.: Medical Economics, Inc., 1961K p, 636. 76Florey, et al., Antibiotics, p. #12. 77 Mahoney, Merchants of Life, p. 66. 80 was finished."78 With the marketing of chloramphenicol in March 1949, sales; asuccess followed to the extent that Parke Davis had to turn. two Pfizer and Monsanto Chemical for supplementary pro- ductzicari. Productinn at this time was by fermentation. How- ever'jEwarke Davis‘ research was able to determine the chemical structure of this drug and chloramphenicol was the first antitaixbtic produced synthetically.79 The patent on chloramphenicol was issued in the name of fWDLLr persons in October 1949 and assigned to Parke Davis.80 Assuuaiedg that these four individuals were responsible for the IDEljor research on this drug, their previous antibiotic reseaachh as measured by Florey's bibliography reveals little Significant published research. There is one Joint article a“ltl’lrbred by the four, and one article by Bartz. Burkholder, the discoverer,81 is listed with four. The evidence on the isolation and development of ChlOI‘amphenicol fits into the pattern of the other broad spethrWam drugs. The research technology was one of soil screening initially operated in the mid—1940's by a Soierltxist with previous research in the antibiotic area. \ 78Fortune, gp. cit., p. 109. 791b1d. 80 U.S. Patent Office, Official Gazette, DCXXVII, :ztober 1+9 19499 Do 229- The Gazette lists the following Jgfizf‘ as the patentees: Harry m. Crooks, Mildred c. Bebstock, c ontroulis, and Quentin R. Bartz. 81American Men of Science, p. 342, lists Burkholder zas the discoverer. 81 The effect of the successful development of this drug was to further confirm the soil screening process and along with Lederle's success with chlortetracycline, to intensify Pfizer's research program which led two years later to oxy te tracycline . CONCLUSIONS Basic research has been defined as an activity that increases the stock of scientific knowledge, or alternatively as an activity where the probability of success is quite low. Applied research, in turn, has been defined as an activity which uses the stock of existing knowledge to create a marketable product, or alternatively, where the probability 01‘ Success is to some degree greater. In addition, there exists a feedback relationship between applied research and the stock of knowledge in which the Operation of the former actlvzi-ty can enlarge the size of the latter. This is not to argue that basic research cannot result directly in a market- able product, for with streptomycin this was clearly the case. Rather, it is to argue that the primary goals of the W0 activities differ. Following the development of basic penicillin, ten antlbiotic drugs followed through 1959. Two, on a "product 1mPOPtance criterion,"_were ranked as major breakthroughs. One, Streptomycin, was the by-product of basic research, and the second, chlortetracycline, was the PI‘OCWG'c 0f ap- plied research. Similar results are observed in the case of :32 the major advances. Chloramphenicol resulting from basic research and benzathine penicillin a product of applied research. Of the six minor advances, all appear to have been the result of applied research. From this sort of evidence it must be concluded that applied research is capable of generating significant drug products although the probability of any one invention being significant appears relatively low. In the case of the broad spectrum antibiotics, the key invention was found to be Waksman's soil screening pro- 0638- It was the operation of this technique which gen- erated two highly significant drug products—-chloramphenicol and chlortetracycline--and two close substitutes--oxytetracy- Cline and tetracycline. Tracing the discovery of these antibiotics from Burkholder to Conver, it was found that the Qualifications of the inventors, as measured by their Prior publication on antibiotics, declined. This lends further support to the importance of Waksman's technique and indirectly to the basic research which produced it. Some attempt was made to isolate the variables reSPOl'lsible for the commercial investment in research beyond a mere statement that such research was profitable. Three W111‘j-ables were apparently important: (1) the existence of a prOven research technology suitable for commercial utili- Zation; (2) the market position of the innovating firms; (3) the patentability of research results. Each of the lnnoVetting firms, prior to their major effort in antibiotic research, experienced a high degree of competition in the antibiotic i-iarket. Merck gambled on synthetic penicillin and along; with Lederle entered the penicillin market with the high cost surface fermentation process. Parke Davis missed the antibiotic market entirely. Pfizer, a leading penicillin producer, was early faced with unrestricted entry and price competition. Later, with Lederle and Parke Davis marketing broad. spectrum antibiotics, their future profitability and market position was threatened. Waksman's method of invent- ing new antibiotics provided a means of maintaining market POSitions through the isolation of more or equally effective drugs . The possibility of patents would avoid the unre- Stric ted entry situation of the penicillin market, thereby giving a measure of protection to market positions. What, in the final analysis did the patent system reward in granting patents on the majority of the antibiotic drugs? The reward was for the iterate operation of a re- sear-eh process. From a social point of view the important contribution was the research process. The commercial op- eratlon of this process produced private property rights Which could be exploited for profit. Whether the society paiq more than was necessary to elicit utilization of the pro‘3ess cannot be stated with any degree of certainty, but in a market characterized by oligolopy where one firm en- gages in the development process it would seem that others would be forced to follow. CHAPTER III THE." TETRACYCLINE PATENT PROCEEDINGS The analysis of market control in the tetracycline market through patents is divided into three parts. The first part, discussed in this chapter, considers the forces that determined the interrelationships of the major antibiotic pro- ducers and their efforts to secure a patent on the drug tetra- cycline. In the following chapter the domestic and foreign license agreements resulting from the tetracycline patent are discussed, and in Chapter V the pricing of antibiotics is ana:l-Efzed. By way of introduction the general relationship of the antibiotic to the overall drug market is briefly indicated, as is the market positions of the major producers. The question here is what were the characteristics of the antibiotic market that allowed the development of a degree of monopoly power through patents? With the isolation of tetracycline what were the alternative courses of action open to the firms in secur- ing the patent? In the litigation and manoeuvring that en- sued what was the overriding consideration? What was the re- action of two smaller firms in competition with the two largest firms for the patent on tetracycline? 81+ be In this chapter I will argue that under the assumption of initial noncollusive behavior that Lederle held a key po- sition in Pfizer's attempt to patent and market tetracycline and could have prevented the product patent from issuing. But to do so would have exposed the market to unrestricted entry . Faced with a choice of conducthederle suppressed the information which would have prevented the issue of the tetra- cycline patent and entered a collusive agreement with Pfizer to ensure the latter's securing this patent. In addition Lederle used its market power in conjunction with Pfizer in an attempt to prevent entry and monopolize the tetracycline mar Re 1: . MARKET POSITIONS The antibiotics are the largest segment of the drug market in terms of the value of shipments. In 1947, anti- biotics accounted for 9.4% of the total drug market and by 1954 had increased to 16.9%.:L By 1950, prior to the 130‘ lation of tetracycline, the three broad spectrum antibiotics accourlted for some 38.5% of the antibiotic market. By 1956, this percentage increased to 54.7%, with tetracycline account- ing for the largest share of the antibiotic market at 24.2%-2 In e‘bsolute terms the broad spectrum market in 1950 involved \ P 11958 Census 0f Manufacturers, II, Industry Statistics, Uagt 1, General Survey and Major Groups 20 to 28 (Washington: Government Printing Office, 1961). ZSee Tables 1 and 2, Appendix, p. 241 and p. 242. 56 sales at the manufacturer's level of 582 million. With the introduction of tetracycline this market by 1956 had increased to $148 million.3 The absolute size of the broad spectrum market and its relative profitability“ made entry attractive. Thus,with the isolation of tetracycline in 1952, the largest segment of the antibiotic market was threatened with entry and with the strong possibility that price cutting would develop as in the case of penicillin and streptomycin. The relative share of the antibiotic market held by the various firms is indicated in Table 7. These figures are not directly comparable as in the 1944 period only one anti- biotic was produced--penicillin. In both 1950 and 1956 some eight to nine products are included with varying degrees of competitive relationships.5 delative movements in market positions can be ex- plained to a degree in the 1944-1950 period by the existence of product patents. In this period five patented drugs were introduced to the market and in all but one case the firm in question was able to increase its relative position. Lederle moved from seventh to first with chlortetracycline, Merck with the patent on dihydrostreptomycin moved from fourth to third. Lilly with the patent on procaine penicillin, and 3See Tables 3 and 4, Appendix, p. 243 and p- 244- QFTC, Re ort, pp. 211-12. 5Tables 1 and 2, Appendix, pp. 241-242 indicate for 1950 and 1956 market shares by products. 87 TABLE 7.--Market shares of antibiotic producers from own production 1944 1950 1956 Firm Percentage Percentage_ Rank Percentage Rank Pfizer 44.4 11.7 2 23.4 2 Commercial Solvents 18.2 2.5 11 0.4 11 Squibb 14.4 10.8 5 6.7 6 Merck 4.3 11.0 3 7.2 5 Heyden 2.9 2.8 9 (b) Reichel 2.3 (b) Lederle 2.2 26.6 1 28.1 1 Lilly 1.9 11.0 4 11.7 3 Abbott 1.8 2.7 10 ‘ 1.1 10 Upjohn 1.4 0.7 15 4.4 8 Winthrop 1.4 (c) Cheplin 0.9 (c) Ben Venue 0.8 (c) Cutter 0.8 0.7 16 (c) Schenley 0.8 3.2 8 (c) Parke Davis 0.1 5.9 7 7-3 Bristol (a) 6.0 6 4.3 9 Wyeth (a) 1.8 12 5.0 7 Other 1.4 2.6 0.4 Total 100.0 100.0 100.0 (a) Entered the market after 1944 (b) Bought out by another firm (c) Left the market Sources: 1944 market shares are based on reported production for that year; 1950 and 1956 market shares are com- puted from manufacturer's own production. FTC, Re ort, p. 331 and pp. 94-95. Parke Davis with patented chloramphenicol were able to increase their relative standings from eighth to fourth, and from six- teenth to seventh respectively. Pfizer holding the patent on 88 oxytetracycline was the exception slipping from first to second. In only one case was a firm able to improve its market position without a significant product patent. This was the case of Schenley which moved from fifteenth to eighth on the basis of its production and bulk sales of penicillin. These relative movements reflect more than the patent question. In the base period certain plants were not operat- ing at their rated capacities due to the newness of the fermentation technique and the resulting technical problems. In addition these changes reflect the different efficiencies of the firms involied, their advertising and research ex- penditures, as well as the fact that some firms found the market unprofitable and transferred their capital to other markets. However, allowing for the effect of these factors, four out of the top five producers between 1944 and 1950 had marketed patented drugs. Further, in two out of these four cases production and sale of the drug was carried on ex- clusively by the firm in question. Only Merck and Lilly licensed other producers to produce and sell dihydrostrep- tomycin and procaine penicillin. In the latter case only after extensive litigation.6 Thus, while the effect of patents cannot be separated, the effect would appear to be significant. Between 1950 and 1956 relative stability in market positions was achieved in spite of the introduction of tetracycline, a substitute drug for Lederle's chlortetracycline 6FTC, Re ort, pp. 242-45. 89 7 and Pfizer's oxytetracycline. One would be inclined to ex- pect that with four firms isolating tetracycline within a year's time, where two of the four firms were bulk suppliers, that there would be some decline in concentration. However, the data on production indicates that the leading two pro- ducers in 1950 held 38.392 of the antibiotic market and 51.599 in 1956. In terms of sales,the leading two producers held 41.69g of the antibiotic market in both 1951 and 1960. However, in this nine-year period there was no change in the rankings of the leading four producers.8 Of the eighteen product markets analyzed in the Arthur D. Little Report, only antibiotics shows this type of stability. Thus it would seem apropos to conclude that competitive forces were restrained in the antibiotic segment of the drug market in the face of the successful introduction of tetracycline. To explain this observed stability in market shares, within the context of new product introduction and demand shifts,is a task of the discussion of the tetracycline patent.9 Before turning to this topic it is useful to 7Of the five firms selling tetracycline, Upjohn was able to increase its market position from fifteenth in 1950 to eighth in 1956, however this cannot be attributed directly to tetracycline. The data on production refers to sales from the firm's own facilities, but Upjohn was licensed only to sell tetracycline and not to produce it. 8Arthur D. Little, Inc., A Report on the Aspects of Concentration and Product Obsolescence in the Pharmaceutical Industry In the United States (Cambridge:”Arthur D. Littlé, Inc., 196i), p. 12. 9See, for example, Michael Gort, "Analysis of Stabil- ity and Change in Market Shares," Journal of Political Economy, LXXI (February 1963), pp. 51-61, where it was found that in- dustry growth through demand changes contributes to instability in market shares. This was not the case with tetracycline. rfo indicate certain other conditions that made cartelization of the tetracycline market possible. Up to this point two con- ditions have been indicated. First, the antibiotic market, and in particular the broad spectrum segment of this market? was characterized by fewness of producers9thereby reducing the number of agreements necessary between firms. Second, the absolute size of the broad spectrum market and the rewards available from a monopoly position in this market, were suf- ficient to make an effort at cartelization worthwhile. In addition to these conditions, there are three other major factors and two minor ones that influenced the behavior of the firms in question in their attempt to monopolize the tetracycline market. First, it can be indicated from Chapter I that the narrow Spectrum market, with many producer and sellers,was relatively competitive and characterized by declining profit margins. 0n the other hand in the broad spectrum market where production and sale were generally monopolized through patents and nonlicensing, profit margins were,in generalydeemed adequate by the firms involved. The lesson of failure to agree was certainly not lost on the firms involved in the tetracycline agreement. A second major factor favorable to a cartel scheme was the belief that the demand for drugs in general was in- elastic so that price competition would have the sole effect of reducing total revenues. Here the demand for antibiotic drugs is largely a function of the incidence of disease sus- ceptible to treatment by a particular antibiotic drug and the 91 knowledge of the physician of the range and activity of available substitutes.10 This argument is acceptable for a static world, but expenditures to inform and influence the medical community did have the effect of shifting the demand for antibiotic drugs.ll Thus the industry believed that demand was extremely elastic to selling expenditures. The point here is not a question of the shape and slope of supply and demand curves, but only that if firms believed that price competition would reduce total revenues in the short run this was sufficient to favor any scheme that would tend to eliminate this form of competition. A third major factor which favored a private settle- ment of the tetracycline diSpute as opposed to a court settlement was the legal climate relative to patent questions before the courts. As a general proposition, companies are rather hesitant to bring their patent disputes before the courts in light of judicial rulings on these questions.12 In such litigation the court consistently reviews the findings of the Patent Office as well as looking beyond the patent issue to the question of misuse. loFTC, Report, pp. 152-54. llSee Table 5, Appendix, p. 245, for output data of medical grade antibiotics of a broad spectrum character. 12For example see, Robert Levine, "The Shrunken Patent Domain in the Expanded Anti-trust Universe," Journal of the Patent Office Society, XXXIV, No. 6, 1952, pp. 436-47; Benton Baker, "Patent Rights and the Anti-trust Laws," JPOS, XKXIV, No. 9, 1952, pp. 687-89; Karl B. Lutz, "Are the Courts Carrying out Constitutional Public Policy on Patents," JPOS, “XIV, N00 10’ 1952, pp. 766‘910 92 As an example of a somewhat extreme view there is the opinion of Judge Thurman Arnold where it was said if we apply the fundamental principles of the patent law to the actual facts of the complex modern tech- n010gy of corporate research laboratories. These principles are (1) that a discovery which is the result of step-by-step experimentation does not rise to the level of invention; (2) that invention must rise above the level of accomplishment of the ordinary skilled technicians engaged in the art; (3) that the patent law must be so administered as to promote science and the useful arts. 13 Or on the question of influence of experimentation in Minnesota Mining and Mfg. Co. v. oe, where the Court held: But a showing of great industry in experimental re- search is not in itself sufficient to constitute in- vention, when the product therefore differs from a prior art only in degree, the result--no matter how useful it may be--is merely one step forward in a gradual process of experimentation. 14 In Naamlooze Vernootshafs: w. A. Sholtens Chemische Fabrieken v. Coe: Experimentation may well have been necessary to determine this. But, . . . that progress in the chemical art is reached largely through experimentation, . . . it does not follow that every new and useful result accomplished by experiment is patentable. 15 These views are not quoted to express settled law on the question of industrial research, but to indicate the risk of judicial review. It was concluded that the discovery of chlortetracycline and oxytetracycline was the result of a step-by-step research techn010gy essentially 13Potts v. 9229 145 F! End 28 (1944)° lD'Minnesota Mining and Mfg. Co. v. Coe, 99 F 2nd 989 (1938). l5Naamlooze Vernootshafs: w. A. Sholtens Chemische Fabrieken v. C22, 132 F, 2nd 575 (19h3). 93 developed by Waksman. To subject this discovery process to judicial review would have run an unnecessary risk that could be avoided through interfirm agreement. There were two other conditions present in the pre- tetracycline market that favored agreement. There was first, the general interrelationships of the chemical firms where agreements to license competitors was the usual case.16 Second, there is the indication that the costs of production of antibiotic drugs by fermentation were similar among firms thereby eliminating the problem of profit pooling.l7 Under these conditions,the simultaneous discovery of tetracycline foreshadow interfirm conflict over the patent. How this conflict was resolved for the private benefit of the group is the subject of the following section. THE TETRACXCLINE PATENT In 1950, prior to the introduction of tetracycline, the three broad spectrum antibiotics represented approxi- mately 269g of total output of all antibiotics and 38.57; of the total value at the manufacturer's level. Lederle, with chlortetracycline held 26.69% of the total antibiotic market and 69.192 of the broad spectrum market. Pfizer, with oxytetracycline held 11.79% of the total antibiotic and 12.29% of the broad spectrum market. Parke Davis, with 16 See, for example, George w. Stocking and Myron w. Watkins, Cartels in Action (New York: The Twentieth Century Fund, 1947), Chap. IX. 17FTC, Re ort, pp. 11-12. 94 chloramphenicol, held 5.99g of the total antibiotic market and 15.392 of the broad spectrum market. In the case of Lederle and Parke Davis, one product was the total contribu- tion to the antibiotic market. Pfizer on the other hand produced all but the two specialties of Lederle and Parke Davis. Lederle's chlortetracycline was the first of the broad spectrum antibiotics to be marketed in 1948, followed by Parke Davis' chloramphenicol in 1949, and Pfizer's oxytetracycline in 1950. By 1951, following what has been 18 by Pfizer in its termed a "blitz advertising campaign" marketing of oxytetracycline, this entry accounted for some 259g of the broad spectrum market, while chlortetracycline fell to some 489g and chloramphenicol increased to 279g}9 Up to 1952, these three drugs evidently were near perfect substitutes in that they were effective against the same range of microorganisms. But in 1952 the problem of the side effects of chloramphenicol restricted 20 its use. .The following discussion then is largely re- stricted to the tetracyclines. 18FTC, Report, p. 140. 19Mahoney, The Merchants of Life, p. 244. 20"Shock Treatment for Parke Davis," Fortune, XLVIII (1953), 108. The dangerous nature of the side effects of chloramphenicol made this drug an imperfect substitute for chlortetracycline and oxytetracycline in the general treat- ment of infection. Production fell from 136,676 pounds in 1952 to 9,892 in 1953, however there was no change in Parke Davis' list price on this drug. 95 desunihg the narrative of the tetracycline patent from Chapter II, where Woodward of Pfizer had succeeded in deter- mining the chemical structure of both oxytetracycline and chlortetracycline and Conover in turn had succeeded in pro- ducing tetracycline by a modification of the chemical struc- ture of chlortetracycline. This work resulted in two articles published in the Journal of American Chemistry. The first appearing in the October 1952 issue described the common chemical structure of the tetracyclines. (This article is hereafter referred to as the structure article.) In the September 1953 issue the second article appeared which described the production process of producing tetra- cycline. (This article is hereafter referred to as the production article.) In the same issue a similar article appeared written by Lederle's research staff.21 The publication of the structure article appeared as a threat to Lederle's market dominance with chlortetracycline. For with the knowledge of the chemical structure the implica- tion was that Pfizer was in the process of attempting to modify the structure in an effort to develop a new broad spectrum antibiotic. As a result Lederle resumed research in this area that had been abandoned in 1948.22 By the end 21Federal Trade Commission, Qpinion of the Commission, before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211, August 1963, pp.9-10, hereafter cited as Opinion of the Commission. 22John T. Kelton, Respondent American Cyanamid Company's Proposed Findings of Fact, Conclusions ofiLaw and Proposed Order, with deasons in Support thereof, Beforelthe Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211, October 1960, I, 15-16, here- after cited as gyanamid's Proposed Findings. 96 of 1952 Booth of Lederle succeeded in producing tetracycline by the same technique as Conover at Pfizer had used. In the course of this work Booth learned that tetracycline was co- produced in the fermentation process that produced chlor- tetracycline.23 The point that emerges here is that with the co- production of tetracycline with chlortetracycline, Pfizer could not, without infringing the Lederle patent on chlor- tetracycline, produce the end product tetracycline.24 Thus, Lederle was in an excellent position from which to protect its market against Pfizer's unrestricted encroachment. Before this issue could be settled,potentia1 competition appeared in the form of two other firms that claimed to have dis- covered tetracycline. The most potentially dangerous was the Heyden Chem- ical Company. In May of 1953, Minieri of Heyden produced a then unidentified antibiotic by direct fermentation. (That is, without first producing chlortetracycline.)25 Thus, Heyden71n producing tetracycline,would not per se in- fringe the Lederle patent. What made Heyden an even greater threat was that they were primarily a bulk supplier of anti- biotics to other firms. 231bid., p. 16. 2”Pfizer's initial patent application was in fact rejected on these grounds. See Opinion of the Commission,p.9. 250pinion of the Commission, p. 10; Cyanamid's Propgsed Findin 3, pp. 16-17. 97 The fourth firm to claim tetracycline, and the one that proved the most difficult to deal with from the Pfizer- Lederle point of view, was Bristol. Prior to 1950, Bristol had employed the soil screening process in search of an anti- biotic to compete with the broad spectrums and to offset de- clining profit margins in the narrow spectrum market. By September 1953, Bristol scientists had succeeded in isolating a substance with antibiotic properties but had not engaged in clinical testing. With the publication of the production article, Bristol produced tetracycline by the method de- scribed and compared the resulting product with their earlier product. Comparative tests indicated that these were the same products,thus Bristol filed its claim to dis- covery. But, as their "discovery" took place after Heyden's, and the product had not been clinically tested, hence not reduced to practice, it placed them at a disadvantage vis a vis Heyden.26 The relative bargaining positions of these four firms Prior to the proceedings before the Patent Office can be Summarized at this point. Lederle's position relative to Pfizer appears to have been the stronger of the two. Pfizer, in its production of tetracycline,had to first produce chlortetracycline which was patented by Lederle. Thus Pfizer would have to secure a license from Lederle with the *- 26Allen T. Klots, et al., pp}stol's_§;oposed Findings of Fact, Conclusions of Law and ppgwers to the Govgrnment s Contentions, Before the Federal Trade Commission, In the matter of American Cyanamid Company et al., Docket No. 7211, October 1960, I, 12-13; hereafter cited as Bristol's Proposed Findings. 98 latter able to control the terms of the license. Open competition between these two firms was not a likely alterna- tive, but in such a contest, Lederle's hold of some 509% of the broad spectrum market as opposed to Pfizer's 259g, would provide sufficient resources to engage in such a conflict. In terms of absolute size, American Cyanamid was roughly three times larger than Pfizer as measured by sales or assets. Pfizer's alternatives at this point were to reach an understanding with Lederle in securing a license, or to pro- duce chlortetracycline,thereby infringing the Lederle patent and press the matter into the courts. Pfizer's only con- ceivable argument here would have been that chlortetracycline was unpatentable in that its discovery did not rise to the level of invention. Selection of this latter alternative was of course highly unlikely but its existence undoubtedly pro- vided some force to Pfizer's bargaining for a license. As matters developed7a third alternative appeared which strengthened the Pfizer position. If Pfizer could secure the product patent on tetracycline,by convincing the Patent Office that co-production did not in fact occur,it would be free of the Lederle patent control on chlortetracycline. This would of course require the co-operation of Lederle in not contest- ing the co-production issue. Speculation on Lederle's po- sition under this alternative would indicate that it was to their advantage to license Pfizer to produce chlor- tetracycline and to co-operate with Pfizer in securing the product patent on tetracycline. In this way potential r A“ 77 comoet tors 110.116. te faced with two patent barriers-~on the end precinct W'ti'f/Ca’Clj-ile, €110 on the intermediate product 1 .L. 1,“ .. 1.J'1U Lit/318 L.) El chlortetracycline. lot to co- operate with Pfizer meant that a product patent would not be iss sued and that marks; t. control could only be achieved through the chlortetra— cycli‘ae: patent. This was a less preferred position for two other firms claimed to have produced tetracycline by direct fermela tation--a process that appareit y did not infringe the chlortetracycline patent. Thus, co-operation became the preferleed position instead. of allowing the entry of Bristol or Hayden. The Lederle-Pfizer bargaining positions, relative to 31481301 and irieyden, presented certain problems. Both the Lederle and Pfizer patent applications claimed the product tetracycline and its method of production by deschlorination Of chlortetracycline. Heyden and Bristol made similar claims but based on direct fermentation. The Heyden application was filed. on September 28, 1953; while the Bristol application was filed. on October 19, 1953. Fifteen days after the Heyden appllcgg—ion 1:; g filed, Lederle filed the Martin-Bohonos ap- pllcatlion, and on November 12, 1953, Pfizer was in the Patent Office with an application similar to Lederle's in that both 01 aimed tetracycline by fermentation.27 \ an nd. 27Federal Trade Commission, Findings As to the Fccts 1n\tC0nclusions of Law, Before the Federal Trade Commission, 163 flatter of American Cyanamid Company, et al., Docket Flhq7211, August 1963, pp. 8-11. Hereafter cited as FTC, in s. 100 The Lederle fermentation application resulted from their purchase of the antibiotic facilities of Heyden Chem- ical Co. Hayden announced its discovery of tetracycline on October 1, 1953 and on hovember Uth the purchase agreement was jointly announced. Thus it would appear that Lederle had immediate access to Hayden's research on tetracycline,(assum- ing that negotiations were initiated immediately following Hayden's public statement on the lst of October), to have been able to file its fermentation application by the 12th of that xuonth. The Pfizer fermentation patent application is some- what more difficult to explain, as it seems unlikely that their independent research could have discovered the strain Of microorganism responsible for the direct production of tetr‘a-Czycline in such a short period of time.28 Two alterna- tive, explanations can be offered. First, in a later develop- ment in the contest for the tetracycline patent, Pfizer was found guilty in the New York Courts of employing a John G. \ the 28A search of the briefs and proposed findings of the respondents Pfizer and Lederle turn up no statements to tet effect that they conducted research seeking to produce botl‘aoycline by fermentation. This is not unexpected as 111‘:- Pfizer and Lederle defended their cross-licensing ab angement in tetracycline with the argument that in the resence of such an agreement a blocking situation would have ofsulted and the public would have been denied the benefits by 13his discovery. If Pfizer could have produced tetracycline fr Clilrect fermentation it would not have required a license a?“ Lederle. Arthur G. Connolly, P_f_‘_Lzer's Answerin Brief, 0. Ore the Federal Trade Commission, In the Matter 0% Ameri- P 1'1 Cyanamid Company, et al., Docket No. 7221, April 1962, art 2, p. 9. 101 Broady to tap the telephone lines of Bristol and Squibb}? It is possible that through this procedure the strain of microorganiSm used by Bristol was learned. There is, how- ever, no evidence that such tactics were employed at this stage in the contest. A second alternative is that Pfizer and Lederle were at this stage engaged in a co-operative effort to exclude competition from the tetracycline market and to insure that Pfizer received the patent on this drug. Under this hypothesis Lederle may have made the deyden re- search available to Pfizer so that they could enter a claim of direct fermentation, thereby strengthening their case. Again there is no direct evidence that in fact such a co- OpeI'Eak‘lzzive prOgram was engaged in at this stage. With the elimination of Heyden only Bristol was left to be dealt with. Although other firms had expressed an interest in marketing tetracycline, none were in a po- Sitlon to undertake production in the light of the patent qU‘est.fl.on.30 Bristol's bargaining position relative to that of the Lederle-Pfizer, appeared to have been a poor one. Its fermentation patent application and date of isolation w ere preceded by the Heyden application now under the control \ Be 29Allen T. klots, et al., Bristol's Answering Brief, Qafore the Federal Trade Commission, In the Matter of Ameri- A 1:1 Cyanamid Company, et al., Docket No. 7221, April 1962, pDendix pp. 1-12. F 30Federal Trade Commission, PrOposed Finding of Dact Conclusions of Fact and Law and Order to Cease and \esist, Before the Federal Trade Commission, In the Matter f American Cyanamid Company, et 9.1., June 1960, pp. 66-68; hfireafter cited as FTC, Proposed Findings. 102 of Lederle. In absolute terms its size relative to Lederle or Pfizer was considerably smaller. Primarily a bulk sup- plier without a large sales force, an antibiotic house specializing in the relatively unprofitable narrow spectrum market, its ability to compete with its larger rivals must have appeared limited indeed. Under the assumption that Pfizer and Lederle were engaged in a co-operative program to insure that Pfizer received whatever patent might issue on tetracycline, Bristol was not included. Rather the de- cision was made to eliminate this source of competition first through the patent office and if this failed through the courts . But here full consideration was evidently not given to the determination of Bristol to enter the broad spec>trum market, nor to the aid they would receive from Other firms also determined to break into this profitable mar pie 1: . PATENT OFFICE ACTIONS AND REACTIONS This section is concerned with: (1) how the tetra- cycline patent was secured by overcoming the patent exam- iners, rejection based on the issue of co-production; (2) how Bristol was able to force its way into the market by threatening to prove that tetracycline was unpatentable over 0 hlOrte tracycline . The first patent application acted upon by the Patant Office was the Conover-Pfizer application filed on QCumber 23, 1952. This application was rejected on July 239 103 1953, on the grounds that tetracycline was not patentable in the light of the Duggar—Lederle patent on chlortetracycline. In addition, the patent examiner argued that because the chem- ical structure of oxytetracycline and chlortetracycline were similar, the discovery of tetracycline was obvious, hence its isolation did not rise to the level of invention within the meaning of the patent statutes. On October 20 Pfizer amended its application to argue that at the time of Conover's discovery the chemical structure of tetracycline was not in fact kziown.31 This was sufficient to overcome the patent examiner's initial objections and the Pfizer application was I'““1U111.'.ted for examination. The second patent application to reach the Patent Office? was the Morton-Lederle application filed on March 16, 1953' With the readmittance of the Pfizer application an interference was declared on December 28, 1953. The inter- ference was anticipated, for in November two meetings were held between the presidents of Lederle and Pfizer in which a n agreement was worked out to protect their respective po- 3 ltd-One in the broad spectrum market regardless of which firm \ 31The FTC contended that this statement to the patent Work her was a misrepresentation of the facts as Woodward's 1,113 on the structure question was complete when Conover made was discovery. What Pfizer says it meant by this statement "Orkthat the chemical structure was at the time of Conover's Part not public knowledge, but internal information that was Olin of its research input. Arthur G. Pat Olly, Pfizer's Brief in Su ort of Its Tetrac cline N, Before the FederaT Trade Commission, In the Matter octAmerican Cyanamid Company, et al., Docket No. 7211, 0ber 1960, p. 12, hereafter cited as Pfizer's Brief. ion secured the patent on tetracycline.32 The basic positions of the two firms in working out this agreement appear to be that Lederle was willing to con- cede priority of discovery to Pfizer, in light of Pfizer's earlier discovery date and the fact that Lederle had aban- doned its work in this area. On the other hand Pfizer could not produce its basic input material, chlortetracycline, without a license from Lederle. At this time Lederle was shifting its fermentation capacity into tetracycline pro- duction, thereby securing a lead in the market over Pfizer. It was then essential that Pfizer, who had no experience in chlortetracycline production, secure an agreement with Lederle to prevent them from building a commanding lead in the market. In a sense a blocking situation resulted on the assumption that tetracycline was in fact patentable. And on the additional assumption that direct fermentation was not a possible alternative that would not infringe the Duggar- Lederle patent. There are two possible views on the question of the reasons for the meetings between Lederle and Pfizer. The companies involved claim that agreement was in fact necessary to prevent a blocking situation from develop- 34 On the other hand the government claims that the ing a primary purpose of the meetings and resulting agreement was to monopolize the market by excluding competition. This was 32FTC, Findin s, p. 12. 33Cyanamid's Pgoposed Findings, pp. 20-22. 3“'Ibid” p. 24. 105 based on the fact that at the time of the meetings in November 1953, there had been two rejections of patent applications on the product tetracycline and its method of production. There was the rejection of the Pfizer application on the grounds of lack of novelty over the Duggar- Lederle patent, previously noted; and in addition there was a rejection of the Minieri-Lederle application on October 29, 1953, on the grounds of inherent co-production.35 Thus, Lederle was in a position to support the Patent Office argu- ments for rejection7thereby denying the patent to Pfizer. Here, then, the blocking situation could only develop on the assumption of the patentability of tetracycline, and this assumption was not supported by previous Patent Office actions. And, of course, it was the blocking situation, or its pos- sibility, that justified the meeting and subsequent agreement. In addition, Lederle acted to eliminate by purchase, the only known competitor, besides Pfizer, for the tetracycline patent. Further support for the government contention is provided by the fact that Lederle was instru- mental in Pfizer's entry into the tetracycline market. Part of the agreement provided that Lederle licensed Pfizer to make chlortetracycline, furnished them with the necessary technical information, and supplied bulk chlortetracycline for conversion to tetracycline.36 Thus it would appear that 35FTC, P o osed Findi 3, pp. 41, 45. 36Cyanamid's Preposed Findin s, I, 23-24; FTC, Proposed Findin 8, pp. 70-73. 106 Pfizer greatly strengthened its market position in spite of the apparent weakness of its bargaining position. The weight of evidence indicates that the purpose of these meetings was the monopolization of the tetracycline market.37 As a result of the agreement, Lederle, after an ex- change of relevant information with Pfizer, conceded priority of invention to Pfizer5thereby terminating the interference proceedings. However, on March 2, l95h, a second interference was declaired between the Pfizer-Conover and the Bristol- Heinemann applications. Bristol's initial patent application had been filed in October 1953, claiming tetracycline and its method of production by fermentation. This application was rejected first on the grounds of inherent co-production, and second, on the grounds that the publication of the structure article constituted a statutory bar to the Bristol application.38 However, even before the Patent Office rejection, Bristol evidently was unsure of its ability to secure the patent, due undoubtedly to the publication of the structure article a year before the filing of the application, and due to the 37in addition, the agreement provided that the two parties would exchange information to determine priority of invention and that the company lacking priority would aid the other in securing the patent. There was the usual cross- 1icensing arrangement and the royalty rate on chlortetracycline was set at 2-1/29g of Pfizer's net sales of tetracycline, and at 2-1/29g on net sales of tetracycline by Lederle if Pfizer secured the patent. Thus in effect, royalty payments netted out in the case of equal market sharing. Cyanamid's Progosed Findin 8. I, 23.24. 38Bristol's Proposed Findings, I, 12-14; and FTC, Opinion of the Commission, p. . l..— 107 fact that the fermentation work that led to tetracycline was part of an earlier abandoned experiment.39 In November 1953, the president of Bristol contacted Lederle to work out a cross-licensing agreement. In this meeting Lederle was in- formed that Bristol's fermentation process did not infringe the Lederle-Duggar patent on chlortetracycline. Bristol was offered a license for its own label, but this was rejected as Bristol required at least two bulk customers to reach an efficient volume of production. Bristol's own sales force was insufficient in size to achieve adequate volume.“0 The appearance of Bristol on the scene, with a pro- duction process that claimed not to infringe the Lederle- Duggar patent,must have given some degree of acceleration to the Lederle-Pfizer negotiations. But of even greater impor- tance, it increased the need to secure some form of patent protection to exclude the entry of Bristol and whatever bulk customer they would bring into the market. To admit only Bristol with its small sales force would not have disturbed 39FTC, Opinion of the Commission, p. 11. However, on this latter point, the research process appears to be one in which the collection and isolation of the microorganisms in soil samples proceeds at a faster rate than does the animal testing thereby building up a backlog of microorganisms for testing. This sort of phenomenon may be a result of the level of training required in different phases of the research tech- nology where in the latter or clinical phases greater skills are required. Thus, with the publication of the structure article, Bristol searched its backlog of microorganisms to ex- pedite the testing of those that had the characteristics of tetracycline. Thus, it is possible that the question of aban— donment was not controling, leaving only the questions of statutory bar and inherent co-production. 40FTC, Findin s, p. 11; Bristol's Proposed Findin s, 11, 186; Squibb's saIes force was estimated at 755 to 755, while Upjohn's was estimated at some 1,000. Bristol's sales force was but some 50 salesmen. Viv" 108 the market positions of Lederle or Pfizer. But to admit Bristol and its bulk customers Squibb and Upjohn would have threatened that position. In this regard, Lederle,in pur- chasing Heyden's antibiotic facilities,also acquired the Minieri patent application for tetracycline by fermentation. Thus, Lederle held a patent application competitive to Bristol's and in light of the date of this application7rela- tive to the publication of the structure article, Bristol's bargaining position must have appeared weak. But here Bristol, as the events of the case developed, did have leverage to force admittance for itself on its own terms. It did know from its own patent application rejections that the Patent Office considered co-production of tetracycline with chlortetracycline a bar to the issuance of the patent on tetracycline. Thus, Bristol was in a position to open the market to unrestricted entry. With the settlement of the first interference between Lederle and Pfizer, Bristol reacted by amending its applica- tion to,claim the salts of tetracycline. This action, as has been noted, led to the second interference being declaired. In this interference the patent examiner made what amounted to a tentative determination that the salts of tetracycline were patentable. Lederle was admitted to this interference in spite of the fact that it already conceded priority to Pfizer. In effect Lederle's arguments in these proceedings attempted to dissolve the proceedings thereby eliminating the Bristol claim. Success here would have meant an earlier 109 issuance of the patent, hence an earlier date on which Lederle would have to begin royalty payments. This action, it was contended, indicated that Lederle was in opposition to its own interests and thus engaged in a conspiracy with Pfizer to exclude Bristol, the only competitor.ul Bristol entered the market in May 1954, under its own label; con- cluding an agreement with Squibb and Upjohn in September to supply them with bulk tetracycline.42 The second interfer- ence proceedings were terminated in November 1954, with the ruling that tetracycline was not patentable on the grounds of co-production. Before discussing how this ruling was reversed by Pfizer it is well to note the behavior of Bristol in demon- strating that it would consider the interests of Lederle and Pfizer if granted a license on its terms. In later Patent Office proceedings Bristol was in a position to prove from its own laboratory experiments that7in fact7tetracycline was co-produced with chlortetracycline thereby supporting the patent examiner's argument for rejection of the tetracycline patent application. Bristol did not make this information available to the patent examiner until it appeared that Lederle and Pfizer were determined to exclude Bristol from the market.43 ulFTC, Findin s, p. 17. “ZFTC, Opinion of the Commission, p. 13. “Blt must be noted here that such action was also in the best interest of Bristol. To have informed the Patent Office that in effect tetracycline was unpatentable would have opened the market to unrestricted entry and also destroyed Bristol's leverage in forcing a license from Pfizer. 110 In addition, Bristol was faced with the problem of controlling the selling prices of its bulk customers. With its entry under the Bristol label the wholesale price was set above the prices established by Lederle and Pfizer.44 In its price agreement with its two bulk customers, Squibb and Upjohn, it set a price which "would have permitted the purchaser no profit on the initial order if the final pro- ducts were sold at prevailing market prices.”45 The agreed price was $1 million for the first 1,000 kilograms, with subsequent orders priced at 3500 per kilogram, with reductions below this price to follow as Bristol's costs of production declined. The market price at this time was $5.10 to the retailer for a bottle of 16, 250 milligram capsules.’+6 On a per gram basis this amounted to 81.275. The bulk cost of the initial order was $1.00 per gram, leaving some $.275 cents to cover the costs of packaging, waste, and the dis- tributor's margin. (Subsequent prices left a margin of $.78 per gram.) The initial price of bulk tetracycline was set atm$I,OOO per kiIOgram to allow Bristol to recover its research and development costs. But in addition this price to the bulk customers made price cutting on their part an unattractive proposition as it would involve certain losses. Thus, upon entry, Bristol by its own pricing in the wholesale 47 market, the price charged its bulk customers, and its uuBristol's Pgoposed Findin 8, II, 169. “51b1d., p. 167. “6 47Bristol in its sublicense foreign agreements on tetracycline exerted control over the prices charged by the licensee. See Chapter IV below. FTC, Re ort, p. 192. lll actions in the Patent Office proceedings, indicated that it would act in the best interests of the group if licensed. The position of Bristol's bulk customerscan be noted at this point. Given the potential value of the tetracycline market Squibb and Upjohn considered entry an attractive pos- sibility. To this end they were willing to support the Bristol bargaining position under the assumption that a patent would issue and Bristol would secure a license. (The alternative assumption being that a patent would not issue and they could enter on their own terms.) It appears that they did not question Bristol's ability to force entry. To this end Upjohn and Squibb agreed to "indemnify Bristol against all damages and losses which Bristol might incur as a result of infringement. . . ."48 The patent that could be infringed at this point was the Duggar-Lederle patent on chlortetracycline. But to produce tetracycline Bristol did not first have to produce chlortetracycline, hence any in- fringement suit on this patent would have to involve Bristol's use of a strain of microorganism identified by the Duggar patent. There was at this point no patent on tetracycline hence no infringement action. If such was the case the ‘ question arises of why should Upjohn and Squibb agree to this provision beyond the possibility that it was something Bristol demanded in the agreement? One possible answer is that it 48Gerhard A. Gesell, g£_gl., Brief and Proposed Find- ings and Conclusions of the Upjghn Company, Before the Fe eral 'Rrade Commission, In the matter of American Cyanamid Company (at al., Docket No. 7211, October 1960, p. 71; hereafter cited as Upjohn' 8 Brief. 112 placed Lederle and Pfizer on notice that Bristol had the financial support of two firms with adequate resources to support a prolonged legal contest. This would have eliminated one weapon available to Lederle and Pfizer in driving Bristol out of the market. PFIZER SECURES THE TETRACYCLINE PATENT Returning to Pfizer's actions to reverse the patent examiner's action in the second interference, the examiner ruled that the salts of tetracycline were not patentable separate from the basic product (thereby eliminating the Bristol application). That the microorganism used to pro- duce tetracycline by fermentation was of the same species as disclosed by the Lederle-Duggar patent (thereby eliminat- ing the Minier-Heyden-Lederle application). And that tetra- cycline was co-produced with chlortetracycline hence "a purer form of an old product is not inventive. . . ." (thereby eliminating the Pfizer-Conover application). These rejections left the Lederle-Duggar patent controlling which gave some measure of protection to the Lederle-Pfizer market 'positions. But it also meant that the patent barrier on the final product was potentially eliminated. The resulting scale of entry could overwhelm the legal defense based on ‘the intermediate product patent.49 Thus Pfizer decided to continue its attempts to secure the product patent. ugFollowing the FTC decision that Pfizer had secured ‘the tetracycline patent through "unfair" means several smaller :firms entered the market only to be driven out by Pfizer's cistermination to utilize the courts to protect its monopoly. 113 Pfizer, after receiving notice of the rejection of its patent application, began a series of experiments to determine the exact amount of tetracycline co-produced with cflflortetracycline. A week later this work was ordered dis- continued on the advice of Pfizer's patent attorney.5 Nevertheless this work was continued but the results were recorded separately from the regular records kept of scientific experiments. The results showed that in the production of chlortetracycline according to the Lederle-Duggar and Niedercorn patents (the latter an improvement patent in the production of chlortetracycline), from 5-1oeg of the result- ing product was tetracycline.5l Thus Pfizer,in its own lab- oratory, was aware of the correctness of the patent examiner's position in rejecting the Pfizer application. The experiment did, however, yield two bits of in- formation that Pfizer later found useful. First, it was found that the broths prepared according to the Niedercorn patent were low in antibiotic content. Second, it was found that the recovery process described in the Duggar and Pidacks patents (the latter an improvement patent in the production of chlortetracycline held by Lederle) destroyed most of the tetracycline present.52 A conference was held between the patent examiner and the PTizer patent attorney in December 1954. At this 5°FTC, Findin s, p. 23. 511b1d., p. 2a. 52FTC, Proposed Findings, p. 168 citing a Pfizer report from a Dr. Bogert who conducted this experiment. 1l4 meeting it was argued that the examiner had no reasonable basis upon which to speculate that tetracycline was co- produced with chlortetracycline. The patent examiner took the position that he would withdraw his objection if Pfizer could demonstrate that tetracycline could not be recovered in clearly identifiable form from fermentation broths pro- duced strictly in accordance with the Duggar and Niedercorn disclosures, using the strain 8. aureofaciens NHHL-2209 which had been deposited by Cyanamid with the Northern Regional Research Laboratory as part of its 53 disclosure requirements in receiving the Duggar patent. To support its position before the Patent Office Pfizer set out to prove that co-production did not take place. After obtaining the required micro-organism sample, a culture medium was selected from the Niedercorn patent, the same that had proved earlier to be a poor producer of tetracycline, and one from the Duggar patent which also proved to be a producer of low potency broth. They were represented to the patent examiner as being average producing mediums from these patents.5u In addition, Pfizer stated that the mediums were prepared strictly in accord with the disclosures in question. In fact the record contains evidence that such was not the case .55 The two broths were subject to recovery processes in 53FTC, Qpinion of the Commission, p. 18. 54Ibid.,p.l9. 55Ibid., pp. 19-20. The Niedercorn disclosure stated that tnus'best yields were obtained in a medium having a pH of 6.4 1x: 7 during the 48-hour fermentation period. Tanner, who conducted the experiment, inoculated the medium when the pH had a value of 8.1. Six and a half hours later the pH was still.£3.1 and no growth of the microorganism had taken place. The medium was then adjusted downward. 115 an attempt to determine whether tetracycline was present and in what proportions. The three recovery procedures used were the type designed for commercial recovery of tetracycline from broths with higher potency than that of the test broths.5 From this Pfizer was able to conclude that "this product was tested in a manner that he [Bogart] knows is capable of de- tecting even a small proportion of tetracycline in the presence of chlortetracycline and showed only chlortetra- cycline."57 Thus the one individual Pfizer had to convince of its argument was so convinced and the tetracycline patent was issued to Pfizer January 11, 1955. The issue of fraud in securing the patent revolves on the issue of what the patent examiner was in fact interested in relative to the co-production question. Pfizer argued that the grounds for the initial rejection of its application was that tetracycline had been in public sale more than a year prior to its filing of the patent application, and that tetracycline was but a purer form of an old product--chlor- tetracycline. What the patent examiner was attempting to establish, according to Pfizer, was whether tetracycline was co-produced to the extent that the therapeutic activity of chlortetracycline was due to the presence of tetracycline. That is, if chlortetracycline contained 503% tetracycline, to separate out tetracycline for a separate patent would fail as there would have been no advancement over the prior art. From this Pfizer argues that the examiner was not interested in 561b1d., p. 20. 57Ibid., p. 21. 116 "useless trace amounts which cannot be separated from the broth by methods now recommended for recovery of new anti- biotics."58 Thus, according to Pfizer the 5-1093 tetracycline found in its own experiments and later confirmed by other, was a useless trace amount.59 Pfizer in one of its interviews with the patent ex- aminer stated that he could assume that up to 109g tetracycline was co-produced. It was further argued that even if tetra- cycline was co-produced, this was accidental and unrecOgnized production of an inefficient amount which did not negate the novelty of Pfizer's invention. Thus, as a matter of law, there was no basis for rejection on this issue of co-production.60 But the patent examiner could not rule on the basis of an unsupported assumption. This was the very information he sought from Pfizer and which was withheld. Further it can only be concluded that as a matter of law the issue of co- 61 production was basic to the examiner's decision. The issue here is what information the examiner 58Pfizer's Brief, p. 63. 59While Pfizer, for its purposes, claimed that this percentage of tetracycline was insignificant and accidental, Lederle based its suit against Bristol for infringement of its Duggar patent on just this point. FTC,'§£oposed Findings, pp. 304-08. Bristol argued that as a matter of law there are two separate and unrelated facts and that the government was not justified in drawing any inference from one to the other of these facts. Bristol'sgggoposed Findi s, I, 135-36. 60FTC, Opinion of the Commission, pp. 44-45, citing an FTC exhibit relative to Pfizer's argument of the applica- tion of the patent code. 61FTC, Findin 8, pp. 14-15. 117 wanted from the Pfizer experiments. The Commission, based on the statements of the examiner, found that he was interested in ppy presence of tetracycline, and that Pfizer interpreted this to mean any recoverable quantities.62 The implication here is that Pfizer interpreted "any" to mean ”recoverable quantities" because its earlier experiments had shown that it could set the experiment up in a manner in which tetra- cycline could not be recovered from the resulting broth. Thus Pfizer, by withholding from the patent examiner the facts that the broth was not representative of the prior dis- closures, and that the recovery methods used were designed for broths of higher potency, misled the patent examiner in issu- ing the tetracycline patent.63 Pfizer's answers to these points were largely premised on what it assumed the examiner was concerned with on the issue of co-production. The assump- tion being that his interest was in recoverable quantities sufficient to affect the therapeutic quality of chlortetra- cycline. In Pfizer's view only such significant quantities could in any way affect the outcome of the examiner's de- cision. From this assumption it then followed that what was withheld was not material to the decision as 10 per cent tetracycline present in chlortetracycline could not affect the therapeutic qualities of the latter, hence the discovery 62113351” pp- 18-19. 20-21; FTC, Opinion of the Commission, pp. 46-47. 63FTC, Opinion of the Commission, pp. 54-57. 118 of tetracycline did rise to the level of invention.62+ But the question here is whether Pfizer's assumption was war- ranted in view of the several statements of the examiner. Given the importance of the tetracycline patent to Pfizer, as well as to Lederle, in preventing a repetition of the penicillin-streptomycin experience,the assumption was con- venient. The question that arises at this point is what would have been the patent examiner's ruling if he had known the proportion of tetracycline co-produced in the Duggar patent? In view of the examiner's rejections the question becomes one of how much co-production of tetracycline was necessary for rejection. Or on Pfizer's grounds, what part of the thera- peutic activity was in fact due to the presence of tetra- cycline? We can, of course, only speculate on this issue as the Patent Office, following its usual policy, did not allow the examiner who handled the tetracycline application to testify. Nor did the Commission find it necessary to settle' the issue directly.66 In the Commission decision there is the distinct implicaidon that the patent would not have been issued iJi'the absence of misrepresentation. If the patent ex- aminer’kwd.the same information from the experiments as Pfizer, he would have followed his original decision in re- jectixn; the tetracycline application. 64Pfizer's Brief, pp. 21-23, 62-66, 37-39. 65F‘TC, Opinion of the Commission, p. Z3, n. 7. 661b1d.’ p. 2. 119 BRISTOL SECURES A TETRACYCLINE LICENSE In the proceedings before the Patent Office Bristol was in a position to prevent any patent from issuing by making known the results of certain of its experiments which con- firmed the fact that co-production took place.67 But here it was in Bristol's best interests not to do so. By withholding this information it increased the probability that Pfizer would secure the patent. With the patent issued,Bristol was in a position to force a license from Pfizer on the basis that it could show in court that tetracycline was co-produced with chlortetracycline thereby having the patent declaired in- valid. As noted earlier, during the second interference pro- ceedings Bristol entered the market under its own label and supplied tetracycline in bulk form to Squibb and Upjohn. Both Lederle and Pfizer recOgnized that to force Bristol out of the market through Pfizer's securing the product patent on tetracycline was too slow. The alternative course of action was then adopted in which Lederle harrassed Bristol with an infringement suit based on its Duggar patent, claim- ing that in the production of tetracycline by fermentation, chlortetracycline was produced in amounts of 2 to 4 per 67FTC Proposed Findings, p. 285. This withholding of information according to the Commission made Bristol a party to the nflsrepresentation practiced before the Patent Office by Pfizer. However, from a legal point of view there appears to be no requirement that such information be made available on a voluntary basis. 120 cent.6C3 On December 9, 1954, Pfizer received notice that its tetracycline application had been allowed. In mid-December Lederle settled its suit against Bristol by granting a license to produce chlortetracycline in the production of tetracycline in exchange for royalties of 5 per cent of the 6 net sale value of Bristol's tetracycline.9 But on January 11, 1955, Pfizer sued Bristol for infringing its tetracycline patent. On the basis of these facts the government built a conspiracy case against Lederle and Pfizer to exclude com- petition from the market. The theory was advanced that the settlement between Lederle and Bristol was to avoid suspicion from having two of the largest firms in the industry attack a smaller competitor at the same time. The suit was settled when Lederle knew that Pfizer was to receive the patent. While there was no direct evidence that this information was given to Lederle, such an inference was drawn from the earlier cross-licensing agreement, and from the fact that in the a; 68Lederle's suit also had the effect of serving notice on Bristol's customers that they were open to suit as contrib- utory infringers, which with alternative source of supply avail- able at the same price would have had the desired effect of "slowing Bristol down." For a review of the tactics used in these types of patent cases see, Joseph Borkin, "The Patent Infringement Suit--Ordeal by Trial," The Chicagp Law Review; XVII ’ 1950 9 63u-652 c 69Lederle and Bristol claim that the reason for set- tlement was that Bristol allowed Lederle to examine its pro- duction culture and found that the strain was "so closely re— lated as to be . . . Lthe] . . . legal equivalent . . ." of one claimed in the Duggar patent. Cyanamid's Prgposed Find- ings, I, 48; also Bristol's Ppgposed Findin s, I, n.,136. 121 second interference Lederle acted to exclude Bristol but not its other competitor Pfizer.7O In this sequence of events Bristol may have reCOgnized that the reason for Lederle's offer of a license was due to Pfizer's securing the tetracycline patent.71 On January 3, 1955, Bristol filed with the patent examiner an affidavit to the effect that tetracycline was isolated from old samples of commercial chlortetracycline.72 According to Bristol this was an attempt to delay the issuance of the patent thereby providing them with more time in which to build their finan- cial position to contest an infringement suit. But in spite of this information the patent was issued to Pfizer on Jan- uary 11th. Here either the examiner did not see Bristol's affidavit or, if he did, found it opposed by Pfizer's state- ment of its experimental results and accepted the latter.73 Thus Bristol acted too late to prevent Pfizer from securing the patent on tetracycline. The Pfizer infringement suit was settled informally (anDecember 14, 1955, with Bristol taking a license from 7OFTC, Proposed Findings, pp. lOO-l6, 304-10. 71An alternative hypothesis is that with a license to produce chlortetracycline with tetracycline, Bristol attempted to insure its entry in the market by preventing Pfizer from securing the patent on tetracycline thereby freeing itself from patent restraints from this quarter. 72Bristol's Proposed Findin 5, pp. 20-23- 73The Commission's finding on this point is that Bristol's affidavit "did not put the patent examiner on notice of inherent production in the prior art, even if it can be as- sumed that the examiner saw the affidavit before the Conover patent issued." FTC, Findin s, p. 34. 122 Pfizer and agreeing to pay a royalty of 3-1/29g. The reason for the settlement is a matter of major dispute between the parties to this case. Bristol and Pfizer argue that the settlement was forced by the Broady conviction for illegal wire tapping. The resulting adverse publicity that would re- sult in Bristol's~injecting this matter was sufficient to 74 The government contends that the pre- cause a settlement. trial examination of Pfizer's records and personnel by Bristol convinced Pfizer that the matter of fraud in securing the patent would be proven. To allow the case to come to trial would have meant in all probability the loss of the tetracycline patent which would place the burden of control on the Duggar-Lederle patent. As Bristol already had a license under this patent, along with Squibb and Upjohn, there was nothing to be gained by judicial review of Pfizer's tetra- cycline patent or the methods used to secure this patent.75 SUhNARY AND CONCLUSIONS The claim to the discovery of tetracycline by two of the industry's dominant producers posed a threat to their respective market positions. Lederle under its Duggar patent was in a position to block Pfizer on the grounds that: (l) Pfizer had to first produce chlortetracycline to produce 74Bristol's Pr_posed Findin ngs, I, 27-29; Arthur G. Connolly, Pfizer's Answering Briei, Part 2: In Support of the Hearing Examiner's Conclusions that Pfizer was not a Party to any Conspiracy, Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211, April 1962, pp. 23-24. 75FTC, ngposed Findin 5, pp. 318-2”. 123 the end product tetracycline; and (2) that in the production of chlortetracycline) tetracycline was co-produced hence the But with the latter product was unpatentable over the former. appearance of Hayden and Bristol claiming the tetracycline patent by direct fermentation, both of Lederle's arguments involved risk. For, to deny Pfizer the patent on these grounds could have meant that tetracycline was unpatentable, thereby leading to free entry, or that if patentable but with a direct fermentation method of productionythat its Duggar patent could not be used to force a license from the patentee. Rather than assume these risks the evidence of the case lends credence to the theory that Lederle entered a conspiracy with Pfizer to monopolize an important segment of the broad spec- trum market. Here can be cited Lederle's purchase of Hayden's narrow spectrum antibiotic facilities in a distressed market, its infringement suit against Bristol and the conditions under which it was settled, the sale of bulk tetracycline to Pfizer to prevent Lederle from building a sales lead, its silence on the issue of co-production, and in the second interference proceedings the attempt to eliminate Bristol but not Pfizer ‘rom the proceedings. Thus Lederle could have blocked Pfizer's entry, but 11y by assuming the risk of competition. The alternative iurse of action was to insure that Pfizer secured. the patent ereby establishing a duopoly in tetracycline. This required 3 J elimination of competitors on Lederle's part, and for zer to design its experiment on co-production in such a ‘fl‘r -“. .—. *‘ fl. -‘-‘: ‘1' ‘ m—HW. - ._ . 124 any that the results were favorable to its case. This latter action gave Bristol the necessary leverage to force a license from Pfizer for itself, Squibb and Upjohn. This avoided Jud- icial review of the tetracycline patent as well as possible review of the whole antibiotic patent structure. In the formation of this conspiracy the patent became the legal instrument by which competition could be suppressed. But, in addition, the underlying market forces as well as the recent history of the industry, played significant roles. Lederle and Pfizer were the leading producers in the industry. Pfizer's position built on the narrow spectrum drugs and oxytetracycline, Lederle's exclusively on chlortetracycline. To maintain these positions the patent on tetracycline was essential. The entry of bulk suppliers such as Heyden and Bristol had in the narrow spectrum market resulted in price competition. Their elimination or suppression was necessary. The unfavorable legal climate on patent issues required that the tetracycline dispute be settled privately with court action a last resort. An additional lesson taught by the postwar antibiotic market was the importance of patent control to maintain market position. Only Schenley was able to make significant changes in its position without a major product patent between l9h4 and 1950. These lessons,in combination with conditions of fewness in the broad spectrum market as well as the relative value of this market7were contributing factors that led to the tetracycline conspiracy. The abuse of the patent system that was necessary in 125 securing the patent on tetracycline leads to the question of the effectiveness of this office in protecting the public interest. There is the fact that Pfizer had to convince but one individual who blocked the formation of a monopoly in an important antibiotic drug. The examiner was forced to rely on the self-serving experiment conducted by Pfizer to test his assumption of co-production. Lacking an independent laboratory to conduct such tests his resources to discover the truth of conflicting patent applications seem meager at best. Evidently the theory here is that the inter- ference procedure is an adversary proceeding in which the parties are assumed to be in competition for the patent. But there is in these proceedings nothing to prevent the formation of a conspiracy against the examiner. Nor7 as in the case of Bristol,is there anything to be gained by a competitor in making evidence available to the examiner which would prevent the issuance of the patent. On the contrary, there is more to gain by remaining silent and using such in- formation to force admittance to the market through a licens- ing arrangement. CHAPTzLI-i IV THE. TETdACYCLINE LICEKSES As a device to monopolize a market the patent and license agreement provide an unsurpassed legal instrument. The patent has the effect of carving out an area of exception from the general antitrust laws giving the holder the right to exclude competition in the product or process, or to ad- mit competitors under the terms of a license. The restrain- ing factor is that the license may not have the effect of ex— tending the monopoly beyond that which is available to the licensor under the terms of his patent. But to draw the line between actions that lie within the scope of the patent monopoly and those that lie beyond present formidable diffi- culties. In the analysis of the domestic and foreign license agreements relating to tetracycline the problem is to decide whether a particular restrictive clause had as its main ob- Jective a lawful business purpose or an attempt to suppress competition beyond the scope of the patent grant. Jki'the analysis of the domestic license agreements the main question is, do the terms of these agreements sup- port the hypothesis that a five firm conspiracy existed in 1Attorney General's Report of the National Committee to Stud the Antitrust Laws TWashington: Government Printing Office, 1955 , Chap. V. 126 127 In the case of the the manufacture and sale of tetracycline? foreign agreements was there an attempt to form or maintain a world cartel in antibiotic drugs? In the domestic market one would expect the licenses to contain covenants that con- trolled such market variables as price, output, markets, and future technical advances. Finding such restrictions would then tend to support the hypothesis of a conspiracy based on the patent license agreement. DOMESTIC LICENSE AGREEMENTS With the issuance of the patent on tetracycline to Pfizer, the cross-licensing agreement with Lederle became These two firms legally controlled the production effective. and sale of this drug--Pfizer through its Conover product patent and Lederle through its Duggar product and Niedercorn process patents. Pfizer-Lederleggicense Agreements.--The two documents executed by Pfizer and Lederle settling their patent inter- ference appear innocuous in isolation. Lederle, under its Duggar and Niedercorn patents, held a legal monopoly on the Pfizer method of production and sale of chlortetracycline. received a non-exclusive license to produce chlortetracycline as an input material to manufacture tetracycline by desch- Lederle in turn received a non-exclusive license lorenation. to manufacture and sell tetracycline under the Conover patent. As noted earlier, it was agreed that technical information would be provided to Pfizer in producing chlortetracycline and in addition that such information was to be made available 128 to each in the future thereby sharing any technology advances. It was provided that each firm could sublicense one firm in each foreign country where the licensor or its sub- sidiary did itself not manufacture the licensed products. aoyalties were to be computed as a percentage on the basis of the net sales value of the licensed products. The net sales value was defined as the gross billing price less certain de- ductions for discounts to wholesalers, transportation charges, etc. This computation was based on the prices in each country where sales took place. In essence, these license agreements did not extend the monopoly position of either firm. Pfizer could not, of course, sell chlortetracycline under its license from Lederle, but both were free to compete with tetracycline in the various therapeutic markets as well as in the agricultural markets. Thus, in the absence of an expressed or implied agreement not to compete in terms of these market variables, these agree- ments appear to have as their major objective lawful business purposes. Bristol-Lederle License Agreement.--The Lederle in- fringement suit against Bristol was settled with Bristol tak- ing a license under the Duggar patent. Bristol received a non-exclusive license to produce up to 6 per cent chlortetra- cycline in its fermentation process for the production of 2The markets here are defined as (l) therapeutic, which is further subdivided into human and veternary; (2) non- therapeutic, which is subdivided into feed supplement, food preservation, and crop spraying. 129 tetracycline. In return Bristol agreed to pay royalties of figmr cent based on the net sales value of its tetracycline. In addition Bristol agreed that its fermentation process in- fringed the.Duggar patent by using a strain of microorganism claimed in this patent. The strain involved was unidentified, being a "specied Streptomyces aureofaciens [as naturally oc- curring and/or as spontaneous or induced mutants], or suf- ficiently closely related thereto as to be its legal equiva- lent, to produce in the fermenter an antibiotic substance con- taining tetracycline . ."3 Thus, Bristol's fermentation process came under control of the Duggar patent. The terms of the license did not allow Bristol to manufacture chlortetracycline in isolated form or for conver- sion to tetracycline nor into any other substance. This in effect prevented any further modification of the basic tetra- cycline molecule which required chlortetracycline as a start- ing material. This in combination with the statement in the license which allowed Bristol only to practice its "Present Operation" (defined as its fermentation process),’4 insured that a competitive drug would not be developed by Bristol unincumbent by patent restrictions. To escape the Cyanamid patent monopoly by developing a new and competitive drug, Bristol would have to be able to prove that chlortetracycline was not involved either as a starting material or as a co- 3U.S. Congress, Senate, Subcommittee on Antitrust and honopoly, Hearings on.Administered Prices in the Drug Industry, rart 26, Appendix B, th Cong. 2d Sess., l9 0, p. 15357. a Ibid., p. 15360. 130 produced intermediate or final product. In addition the microorganism used would have to be sufficiently removed from the Species defined in the Duggar patent so as not to infringe this patent. Having taken a license under this patent Bristol could not, of course, attack its validity. An additional effect of this clause was the restric- tion of Bristol to a particular technique in producing tetra- cycline. Thus, in the event the deschlorenation process proved more efficient than the fermentation process, Bristol was denied the use of this process. There is no evidence available on this point, but if such was the case it meant that Bristol would be a high cost procucer as opposed to either Lederle or Pfizer who practiced the deschlorination process. The definition of the licensed patents under which Bristol could operate did not extend to non-therapeutic pro- ducts. Thus, Bristol could not, for example, enter the animal nutrition markets in competition with Lederle or Pfizer.5 Bristol had the right to grant sublicenses in those countries where it or its subsidiary did not practice the licensed invention. In addition, one German firm could grant a sublicense to its subsidiaries under a Bristol sublicense. Pursuing this point here, rather than in the section on for- eign licenses, Bristol had sublicensed Farbwerke Hoechst of Germany to manufacture and sell tetracycline to the drug trade ——_ 51bid., p. 15358. 131 >nly. .noexflist in turn had the right to sublicense Farben— fénxriker1}&ayer who in turn had the right to sublicense any firuxcir fddmus of its choice to sell tetracycline to the drug , L. " 0 - .- n .- 3. .. '3 - ‘ '7 grade Lnrt not in bulk in Germany or Austria. while hoechst appaanrtly could sell tetracycline to its sublicenses at prices determined independently of Bristol, there is a clause which states that Hoechst must report invoice prices on sales, in addition.to the net sales value upon which royalties were to be computed, and that Bristol had the right to itself check on the information supplied.7 The implication of this price reporting system is that there existed at least an implied agreement on the prices Hoechst would charge. Further, that Hoechst could, through the sublicensing system, control the prices charged by its licensees. But neither Hoechst nor Bayer sold tetracycline in Germany, although it would appear from the Bristol license that their manufacturing plants were located in Germany. Only Lederle and Pfizer are listed as sellers in the German Market as of 1959. In the Austrian market six sellers are listed--Lederle, Pfizer, Hoechst, and 13ayer--with identical prices of 396.02, Lepetit,8 with a price of 36.36 and Biochemie, a sublicenses of Lederle, with a price of.$5.90.9 6mm, p. 15483. 71bid., p. 15488. 8 Lepetit, an Italian-based firm, was a sublicenses of Pfizer. 91pm. Part 24, pp. 13742-137”. Biochemie was a sublicensee of Lederle which purchased. its tetracycline in bulk from Lederle of England, but evidently retained the right to set its own price. _ __._-_._—. ”a -'-———v 132 'flum it appears that hoechst received special consid- erafion tithe Lederle-Bristol license agreement with the mflersamning that they would leave their domestic market to ledenhaand.Pfizer. This point is developed further in the section on foreign license agreements. Pfizer-Bristol License Agreement.--Under the terms of'mn£:agreement Bristol was licensed to sell tetracycline tO'Mnadrug trade under its own label and in bulk only to Squibb and Upjohn. This effect was accomplished in the For sales within the royalty provision of the license. United States royalties were set at 3-1/2 per cent of the net sales value of therapeutic tetracycline sold by Bristol, "its subsidiaries and affiliates, and two existing bulk cus- tomers, and their respective subsidiaries and affiliates, to the Drug Trade . . ."lo The term "drug trade" was defined to include retail drug stores, wholesalers selling to these Therapeutic tetra- outlets, hospitals, and veterinarians. cycline was defined to mean a product sold to the drug trade thereby eliminating the agricultural markets as well as bulk sales txn packers other than Upjohn and Squibb. The royalty clause did contain a provision covering sales of non-thera- peutic ixatracycline in world markets, but this was meaning- less eus the pmbcess license from Lederle excluded Bristol from practicing the licensed invention in any but the thera- peutic market.11 102.111” Part 26. p. 151+oo. 111mm, p. 15358. 133 An additional provision of the royalty section was Hmtif Eflstol or its two bulk customers sold tetracycline to their respective subsidiaries or affiliates in foreign comfimiestme royalty liability was to be based on the net sales value prevailing on sales within the United States. There are two ways in which this clause can be interpreted. First, it could recognize variations in drug prices due to local governments setting relatively low prices as in certain of the South A::1erican countries, or7as in the case of Japan, where a degree of price control is exerted by the government through the trading companies. In addition, this provision would make the check of the royalty liability easier to compute as only one price need be used. Thus this clause can be viewed as a means of maximizing the royalty income on the assumption that prices in the United States were equal But to or higher than prices prevailing in foreign markets. for the seventeen countries for which price data is available, 12 eleven show prices higher than in the United States. A second view of this royalty provision was that it discouraged price cutting in foreign markets and encouraged setting prices higher in these markets using the United States To set or maintain higher prices, neglect- price as a base. would have no effect on the royalty ing demand. elasticities, This view requires liability9 only increasing total revenues. in turn the assumption that some form of price agreement >revailed with respect to the price in the United States F— lZIbid., Part 2L», pp. 137u2-13743. 134 market” (ltherwise this clause would operate to encourage price cutting in the domestic market thereby lowering the total world-wide royalty liability. The following chapter discusses the question of domestic prices in detail, but it can be noted here that price cutting on tetracycline did not occur. The provisions for reporting royalty liability do not appear to reveal price information on individual sales. There is here, however, the indication that to other if Bristol sold than its two bulk customers, a different report would be due. That is, if Bristol sold to a different bulk customer or to a firm outside the drug trade as defined, it had to secure a report from the buyer of his inventory po- sition and his use and resale of tetracycline, passing this information on to Pfizer. The only purpose this report would serve appears to be that it insures that Pfizer would know of such sales. Thus if there existed an implied agree- ment not to sell to such third parties, this would serve as a check on such activities. Pfizer-Squibb and Upjohn License Agreements.--The license agreements between Pfizer, Squibb, and Upjohn were essentially the same hence these agreements will be considered together. Squibb and Upjohn recieved a non-exclusive license from.Pfizer to sell only therapeutic tetracycline to the drug'mmme. Both terms were defined as in the Pfizer license tOIhflstol to exclude non-therapeutic markets, and to prevent 135 . . 13 bulk sales to other firms. In the case of sales of tetracycline within the United States,royalties were set at 3-1/29g of the net sales price. For sales in world markets royalties were based on the selling price of therapeutic tetracycline in the United States.lu These license agreements also contained a provision that neither Upjohn nor Squibb could terminate the license for a period of five years, or until 1961. This same pro- vision appeared in the Pfizer license to Bristol, while in the case of the Cyanamid-Bristol license the period reached to 1966.15 The usual provisions for termination calls for a period of 90 days following written notice to the licensor. The effect of this provision would prevent the licensed parties from terminating the license and then attacking the validity of the patent. Thus, even if the three licensed parties knew that the patent was obtained through fraud, as charged by the FTC, they would be powerless to use such in— formation for a five year period. Both Upjohn and Squibb argue that the license from Pfizer served a necessary business purpose. That is, by having a license to sell tetracycline they were not bound to purchase this drug from Bristol, or any of the existing pro- ducers. Upjohn argues that by having a license to sellJit -——.._ 13Ibid., Part 26, pp. 15427 and l54l4. ”Linen Part 26. pp. 15429 and 15416. 152219-. Part 26, pp. 15372, 15411, 15422 and 15435. 136 soulxl buyr ouldi'tetracycline from unlicensed manufacturers and.zaintain domestic prices. In addition,the fact thatrmdtmm:8quibb or Upjohn legally required a license from Pfimn'to<%ll tetracycline purchased from Bristol can be interpreted as a written agreement not to enter the agricul- 'mmadmmkem;nor to allow further entry through bulk sales to pacmers. 1 [11(1) FOAmIGn LICmnSi AdemnahTS One approach to the study of the foreign license agreements is to attempt to quantify the restrictions in these contracts with the purpose of explaining why certain features appear in one group of licenses and not in others. This would involve a detailed- study of foreign patent laws as well as the market conditions that existed in eacn country that the license operated. In addition, where the domestic producers were established abroad through foreign subsidiaries the degree of corporate control existing between licensee and licensor would have to be known. Lacking this sort of in- formation, an alternative approach is to examine these agree- ments for consistency with the hypothesis that there was an attempt to maintain or form a world cartel in antibiotics. If a world cartel existed or was to be formed with the tetracycline patent what sort of variables would. be ex- pected to be brought under control? The first requirement is control over the domestic market which was effective through the domestic license agreements. In the foreign agreements we would expect some means of controlling either price or market; shares. This could be accomplished by alloting certain :ountries to specific firms or their sublicensees. One would atpect to see some provision calling for the exchange of any uture technical information. that might be developed by the Lcensee or licensor . lbl xrovisions.--Tne nefauver hearings on drug 7 -' n ' ~ (- ,-. Hlvu'llp‘ya pricer: contain ta-z‘ent;-'—nine license agreements between the f— I\ n ’4 -‘ J“ three domestic producers and foreign firms.“ certain of the licenses are what might be called master agreements relating to chlortetracycline or general intent to license contracts, to which have been added tetracycline. There are certain common provisions of these licenses. All contain restrictions relative to the geographic area in which the licensed product may be produced and/or sold. All call for royalty payments in one form or another, although the methods of computing royalty liabilities differ. Jero- visions relating to the reporting and payment of royalties are common with minor differences. The first question here then is whether foreign prices can be set by the domestic licensor. Direct evidence on this point may be contained in the royalty rate charged the licen- see, or a clause giving the licensor the right to determine such prices. On the latter point no such clauses appear and on the former point the royalty rates have been deleted. by the Edibcorrunittee.23 Indirect evidence of price setting may be assumed where the license provides that the licensor has '— lqzzéziiég—één‘gs’ 22- 2_;_i_1_:_., Part 26, pp. 15958, part 25, hp. __ . 2311a a few cases there appears to be a difference in he royalty rate depending upon the country in which the drug a sold. For example, in the case of Pfizer's license to the :alian firm of Ledoga provides for a separate royalty rate r Argentina. (Ibid., Part 25, p. 15229.) But Ledoga does 1: appear as a seller in this market. (Ibid. Part 24, p. 742 .) A high royalty rate by Pfizer may have preserved the .__~;entine market for Lederle's exclusive supply. 111/2 um Nghtto approve the price schedule of the licensee. Suspicion of a price agreement occurs where the licensor has the right to check invoice prices or to cancel the license without cause. Table 8 indicates that in the case of Bristol with ten foreign license agreements, the licensee was TABLE 8.--Summary of destrictive clauses in foreign licenses Number Price Bulk Sale Reporting of destric- destric- Restric- firm Licenses tions tions tions Bristol 10 10 10 IO Lederle 8 6 2 Pfizer l O l (I) Where the license al 0 required to report the invoice price. contained an exclusive supply clause, the resale price had to be approved by Bristol. In the case of Lederle with eight usable foreign license agreements, four called for the report of the net invoice price. Where a supply contract was included could set the resale price. In the case the license e evidently of Pfizer, with one foreign license agreement, no requirement appears to report the invoice price. A second question is whether the licensees may sell in bulk to firms outside the patent licensing system. Evidence of the existence of this particular restriction is where the .icensed product is defined to include only sales to the drug rade, or; in finished form, or where expressly prohibited. 1813101011 of such restrictions arise in cases where the 143 require reports in sufficient detail so as to be licensor may able to determine to z-zhom the product was sold. hieach of the nineteen licenses analyzed the product " way to prevent bulk sales to is defined in a or the market This restriction takes the form of out- unauthorized firms . lightstaMHmnts to the effect that such sales cannot be mafia orsfiatements that they may be made on the approval of the licensor. The third question is whether the reporting system incamxneted in the license is the sort expected in a cartel arrangement. That is, does the licensor have the right to check on invoice prices, on quantities sold, on the form sold, ? In addition, or be informed of the identity of the buyer does the licensor have the right to approve the form in which p the package, its size, and contents. JO. the product is sold-- In the case of Bristol the reporting clause could me but interpreted as consistent with a cartel reporting scheme, only inn: of Lederle's licenses could support this interpreta- c world cartel Ob tion. The evidence then on the existence of In general the restrictive nature of the Bristol is mixed. licenses supports such a conclusion Wflile those of Lederle and Pfizer dormt. Ehcices in Foreign harkets.--A further test of the :yIMDtkM38113 of a world cartel is provided by an examination of If such an organization existed one would ex- arket prices. zeithin particular markets selling prices would be set that 144 the same. In the seventeen foreign (markets for which the admomflttce provided price data, In only five of these markets are prices in sixteen cases more than one seller appears. theseme. In.three of these markets with similar prices-- Ausualnu Germany, and Mexico--competition is between a combnmunon of the five domestic firms. But in three markets-- Canwkh England, and Venezuela--where similar conditions pre- vail,}ndces differ. In six markets where the domestic firms face competition from one or more of their licensees, prices differ. Thus, if similar prices are taken as evidence of a cartel, the evidence of differences in these prices is not consistent with this hypothesis. From the more restrictive nature of the Bristol licenses, as well as its behavior in the patent phase of the one would not expect Bristol or its licensees to be case, The data on world market prices aggressive price cutters. Lederle appears to be the price leader in bears thi s out . In nine markets where Bristol or one of the se marke ts . its:ILicensees appears as a competitor to Lederle, in only two of these markets-~India and Japan--is there a failure to fol- low Lederle' 8 price . .As noted earlier, and Lederle were listed as sellers in this market although in the case of Germany only Pfizer mo of Bristol's licensees--Hoechst and Bayer--were licensed cxr 'tkriszlnarket. In addition to Germany five other markets ZLL ., Ibid., Part 24, p. 13753. Neither Pfizer or Bristol pjicated. any price policy in foreign markets. lH-g contain licensed fir-gas who are not listed as sellers. These are Austria, Brazil, France, Italy, and lung-land, In two cases-- Germany and dugland—-Efizer and Lederle appear as exclusive sellers although at least one other firm was licensed for each of these markets. The tentative conclusion that is drawn from these rather mixed. results is that a world cartel arrangement is not in ex- istence. This is not surprising in the light of the problems that such an arrangement would face. For example, in a country such as Italy tetracycline can be produced. and sold without fear of an infringement suit as patent protection does not ex- ist. Strains of the microorganism used in the production of antibiotics can be obtained from a United States culture col- lection. As part of our disclosure requirements the Patent Office requires that such strains be deposited in this collec- tion which is available to interested parties.25 From the don- estic Italian market it is but a step to export markets.26 In any country which does not permit product patents, or where the product is not patented, the product may be sold without afringement. Even where product patents are allowed as in rtain of the South American countries, imports of the 25Leonard J. Robbins, "Patents for Microbiological nsformation--An International Problem," Journal of the Patent ice Society, XLII, 1960, 830-148. Zéln the Italian market the largest number of sellers ars at sixteen. The closest second is France with six. Of 3 sixteen, only four can be identified as licensees with implication that the balance are technical infringers. I'll _l - - l‘f’H—hfl- - -.—. u’ an—" . t- . we 27 ,mtufledpmoduct are not considered grounds for infringement. hiEumnm the patentee in enforcing a drug patent faces a aunt'wmt is inclined to view with hostility his claim for mfibrmnmnt.2 As a result it is often cheaper to license hitheiwpe of maintaining some degree of control. An additional problem faced in any attempt to organize worhiuarkets is that certain countries impose price controls directuu or indirectly through national health plans. Argentina provides an example of the former type of control while England of the latter. While it is concluded that a world cartel did not ex- ist at least up to 1960, the behavior of Bristol in world markets is consistent with its behavior in the domestic here we find rather close control of its licensees marxet. and a failure to challenge the price leader in these world markets. 27Leonard J. Robbins, Pharmaceutical Patents in Foreign <30uurtries," Journal of the Patent Office Society, XXXVII, 1955, .2 _...- 28 rehere producers in Italy "have invaded export markets p. 842, in other countries where they can only be challenged by the costly and presently uncertain procedure of individual in- See also B. Spencer and L. Chereau, ”France fringement suits." Allows Drug Patents, But on Products, Before Only Processes," gournal of the Patent Office Society, XLII, 1960, pp. semi giolxart E. Woodhams, “A Comparative Study of Patent.Laws in the Americas and the International Effects Thereof," {carnal of the Patent OJffice Society, m1, l9u9, pp. 726-42. CHAPTEd V PRICIHG OF ANTIBIOTIC DnUGS The price history of the narrow and broad spectrum drugs1nesents a situation in which different market struc- ‘hHes led to differences in the behavior of prices. With the narrow spectrum antibiotics entry was generally unrestricted leading to price flexibility on the down side as the tech- nological processes were improved. With the introduction of the first of the broad spectrums, chlortetracycline, price initially decreased with entry of competitive substitutes such as chloramphenicol and oxytetracycline, but reached a level that was uncnanged with the marketing of tetracycline. One purpose of this chapter is to explain the behavior What were the major of‘gnmice in the broad spectrum market. Was differences in the narrow and broad spectrum markets? pricezlleadership the result of explicit or implicit agree- remrt? unmet are the price policies of the various antibiotic producers and how are new drugs priced relative to established. drugs? {The first section of this chapter considers the zaestion of the price policy of several of the antibiotic yellers. The firms generally state specific profit goals rid these are examined as to actual results and to their price determination. An attempt is made to explain Tfect on 147 14% bothlnflce and market shares with the Cournot model as new antibioticcirugs are introduced. The next section is concerned with the price history of Hm:narrow and broad Spectrum antibiotics. Here the struc- ture of these markets is analyzed for the effect on price. The third section returns to the Federal Trade Com- mission charge of collusive price fixing. The different de- cisionschthe Hearing Examiner and the full Commission are discussed. Finally, an estimate is presented of the element of monopoly revenue earned in the production and sale of tetra- cycline. The following chapter examines in part the question of whether this element of monopoly had the effect of in- creasing the rate of technological progress in the industry. Price Policy.--Lederle as the dominant producer in the broad spectrum market may be characterized as the price leader.l This role was largely undertaken by default, for with the introduction of chlortetracycline in late 1948, Lederle had to set some price relative to the narrow spectrum drugs. Later, with the development of close substitutes for chlortetracycline, they were, with one exception, introduced at the prevailing price of chlortetracycline. Of the four price cflumnges observed in the prescription market, Lederle lSee Tables 1 and 2, Appendix, pp. 241-42 for Lederle market shares. 2 . . l The one exception was Pfizer's oxytetracycline intro- «iuced.srt a price above tetracycline. However, Pfizer exper- ienced.21 achieve its stated minimum goal in all but txo yeeurs, zand.Emas been ab.e to exceed its nax’mum goal in four yeetrs. 'car313 Davis has achieved its goel in only four years, :1 tku3ec2 of‘ the four cases in the last three years--1957—59. es that of (f lf‘ it: is aissnimed that Pfizer's target approxima iqts (soamoeixitxors at 12 per cent, based on their consolidated 151 data this level has been achieved in four of the twelve years. drhmolon'dm other hand exhibits the greatest amount of vanabiUtyvmich perhaps accounts for their statement that such targets cannot be set. flmzdata in Table 9 is, as noted, for overall company chmgcnmrambns. As each of the companies is a multiproduct prmhwerznm.seller, they face varying degrees of competition by omnn3drug products. Table 10 presents data showing the IABLEIML-mfiet perating profits as percentage of sales be— fore taxes and by products (1950-195s) Year Penicillin Streptomycin* Broad Spectrum** 1950 21.499 18.79g 52.1yg 1951 22.9 21.3 50.1 1952 (13.0) 0.1 43.7 1953 < 1.0) < 4.7) 43.4 1954 ( 7.4) (20-9) 37-4 1955 (17.?) 137-4) 35-1 1956 ( 5.6) (41.4) 35.5 * Inclguhas streptomycin, dihydrostreptomycin, and strep- toduocin. ** lkuxludewl chlortetracycline, oxytetracycline, tetracycline, suui chlrnnamphenicol. Source: Federal Trade Commission, chormnnio rthort on Antibiotic Manufacture, 1958, pp. 211-12. eaverenge rust (nyerating profit before taxes of three antibiotic izumgs. .HGITS arw'policy of target return pricing would have c> tain; iirto <3onsideration that the target could not be okxiexnad. iri gull products due to the competitive nature of .rtain markets. Where the target cannot be achieved on a rtxiCiLlcur golwoduct this means that in less competitive aarhets 152 price must be higher to achieve the overall rate of return Here then may be the familiar case of market power desired. used to subsidize less profitable operations in competitive 8 markets where a transfer of resources is indicated. By classifying Lederle as the price leader in the broad spectrum market implies, of course, that its competitors In general this accept- assume a role of following its lead. was of a followership role results from Lederle's competitors' This argument, in brief, view of the nature of the market. is that the market for broad spectrum antibiotics is in- elastic with respect to price, hence any price cutting in general would not expand the market at the expense of competi- tive products, and that price cuts by any one firm alone would be self-defeating as such information would be immediately reportedywith price matching the result.9 Thus, the policy of the price followers is one of accepting the price estab- lished by Lederle. The only implication at this point is that 8The reasons advanced for the continued production of unprofitable products are (l) desire to sell a full line of (2) reluctance to withdraw a drug needed by the pub- products, lic, and (3) an expectation of increased sales through market Federal Trade Commission, Economic Re art on penetration. t biotic Manufacturs (Washington: Government Printing Office, 1953), p. 213. 9Allen '1‘. Klots, gistol's Proposed Findin s of Fact, 3 Contentions, "5" swers to the Governmen Conclusions of Law and In the Matter of American Cyanamid Company, et 51., Before t 6 Federal Trade Commission, Docket No. 7211, October 1960, II, pp. 153-54, 157-59 and 161; Gerhard A. Gesell, Answering Brief of the Upjohn Com an (same case citation), pp. 23-30; Arthur Lzer s A swerin Brief (same case citation), 29-30; Allen F. Maulsby, ggopgsed Findings and G. Connolly, 2;: Part 2, pp. :L R lated Pa ers of Res ondent Olin Mathieson Chemical Corpora- tion (same case citation), pp. 115-23; See Hearin s, 92. cit., Part 24, pp. 13980-13982 for Parks Davis' similar views of the nature of the market. 153 the pricing pattern established by Lederle was acceptable to the other participants, operating in a typical oligolopistic market. The question of conscious parallel action is dis-— cussed in a later section. Pricing of Bpoad Spectrum Dpugs.--The pricing of cer- tain of the broad spectrum drugs provides an opportunity to test the hypothesis of target pricing as opposed to a policy of charging what the market will bear. In particular, in the pricing of chlortetracycline) where Lederle held a patent monopoly, if the monopoly price was set this implies that it was the market structure that was important Jrather than a consideration of the firm's overall target profit. On the other hand to set a price significantly lower than the mon- opoly price could indicate an application of target pricing. As previously noted, in the selection of the price of chlortetracycline in late 1948, Lederle had a degree of discretion in its choice of a price. According to company statements the major factors considered in the pricing de- cision were the prevailing prices of the narrow spectrum antibiotics, their overall income objectives, the element of obsolescense, and the level of research expenses.lo To this list of factors considered should be added Lederle's selling expenses which in this period were, in all probability, at I— loBalstone R. Irvine, Respondent American Cyanamid Zompanx's onposed Finding of Fact, Conclusions of Law and >roposed Order, With Reasons in Support Thereof, In the a ter 0 erican Cyanamid Company, et al., Before the Feder- 1 Trade Commission, Docket No. 7211, October 1960, II, p.96, ereafter cited as Cyanamid's Proposed Findings. 154 leastcxpml to its research expenses. In a sense the sit- uation facing Lederle in pricing chlortetracycline was sim- ilar matflmt faced by other companies in the pricing of a basically new product where the emphasis is upon the demand side of the market.14 There are two basic approaches or frameworks in which the pricing process can be explained. This decision-making process can be set in terms of internal budgetary procedures, or in terms of some sort of oligolopy model. But here the selection of an oligolopy model depends upon what assumptions one is willing to accept. Thus, the budgetary framework is used to discuss the determinants of price, as stated by Lederle) to see what assumptions seem Justified for model selection purposes. The budgetary view of the pricing process assumes that what is important is future cash flows. The product to be priced taken in isolation, gives rise to costs that are translated into cash outflows, while the combination of price and volume is translated into cash inflows. Within this framework profit maximization takes the form of maximizing the net difference between these two flows. llFTC, Economic Re ort, p. 209; and U.S. Congress, Subcommittee on Antitrust and Monopoly, Report on ong., 8 Senate, Admini tered Prices in the Dru Industr , 87th C Sess., 1951, pp. BI, 153-90. Hereafter cited as Report. leor the case of DuPont's pricing of cellophane and nylon see, A.D.H. Kaplan, Joel B. Dirlam, and Robert Lanzil- lotti, P is n in B Business (Washington: The Brookings Institute, 1953), pp. 555-53, and 96-108. 155 The element of obsolescence in the drug industry means essentially that there is a limit in time in which a product will be a substantial producer of revenue relative to its costs. Here one of two events generally takes place. First, a-competitor may develop a close substitute or the drug in question may not prove as efficient as initially thought in its development. The effect of obsolescence is to place a limit on the planning period of the firm relative to its pricing and research efforts. One firm estimates that five years is the limit in which it can expect a new drug product to be a sig- nificant revenue producer before competition drives price to "unprofitable" levels or a better product is developed.13 What this seems to mean is that on the average every five years the research effort must be able to turn out a sig- nificant new product. The influence here on pricing is that the combination of price and volume must be sufficient to support the research effort over the five dry years as well as meeting the other expenses of the firm. But here the sell- ing efforts of the firm may be able to offset the inroads of competition. The advertising expenditures associated with the 11+ marketing of a new drug are generally substantial. The purpose of this activity appears to be twofold. First, there 13H arin 3, Part 24, p. 13650. 1436 ort, Chap. IX. 156 is an attempt to gain rapid acceptance of the drug by the medical community thereby establishing the brand name as the therapeutic name. Success in this area has the effect of informing the medical community of the benefits of the drug and insures sales volume in the prescription market. An insurance feature is based upon instilling the habit of brand name prescription writing by the physician in combination with the fact that in many states the druggist cannot by law substitute a different brand although the two may be identi- cal compounds.15 The second purpose of advertising is to protect the firm from competitors' introduction of close substitutes. Until recently, on the average, some 300 so-called "new drugs" were marketed each year. In fact only a few could be classified as a new drug in that they were a marked im- provement over existing drugs, or effective against a pre- viously untreatable disease. The overall effect has been to increase the proportion of the firm's resources spent on advertising. In 1950 this activity absorbed 12.9 per cent of the firm's sales dollar, 17 and by 1959 this had increased to 25.0 per cent.18 15a ort, pp. 235-44. légséilflfiéo Part 21: P- 11628; Harry Beckman, The fear Book of Dr Thera (Chicago: Year Book Publishers, l9 0 , p. . l7FTC, Eppnom;c Report, p. 209. 18Re ort, p. 31. It 157 In terms of the firm's pricing of a new drug the price set must generate sufficient future revenues to cover the expected costs of advertising as well as other expected future costs. If the industry is spending more on this activity the innovator will be forced to increase his rate of spending to insure the adequate dissemination of inform- ation. The effect then is to increase the price that is set. The third factor to influence the pricing decision is the current and future level of research exPenditures. For the major antibiotic producers these eXpenditures ex- pressed as a percentage of sales increased from 4.49% in 1950 to 6.79% in 1956.19 A twenty-two firm average in 1959 indicated that for drug operations 7.39% of the sales dollar was spent for research.20 The size of current research budgets seems to be determined by past decisions, relative to what diseases will be investigated and the level at which these projects will be operated. Decisions relative to the areas to be investigated are based,to a substantial degree,on the firm's area of specialization. For example, one firm interviewed stated that it considered itself an antibiotic producer equipped with the specialized plant and knowledge relative to fermenta- tion technology. As a result, project proposals were formu- lated by the research staff based on an antibiotic approach. These proposals were then reviewed by management with a view 19FTC, Economic Report, p. 209. 20Hearin s, Re ort, p. 31. 158 'U)tmm potential size of the market based on the incidence of‘Uuadisease in question and available methods of treat— This was followed by an estimate of the expected The accept- ment. costs of the project and the chance of success. ance of projects was said to be largely a matter of art in which the research director makes a best guess of what will pay off. If a particular area appears promising parallel projects are accepted until one stands out as likely to succeed. At this point efforts are concentrated on the one approach with a corresponding reallocation of resources. New projects are accepted with reference to the level at which current projects are expected to absorb funds and the overall research budget. The latter figure seems to be determined by a historic consideration of past research spending relative to sales and whether this ratio has pro- duced successful results. If, as in the case of a second firm interviewed, success did not follow, then a substantial increase in the research spending was authorized, but with the idea of later bringing this figure back into line with something approaching the industry average. The basic goal in allocating funds to research is,of course,commercia1 80.00088 e21 21See, for example, the statement of Dr. A. Dole Console, former Medical Director of Squibb--"In drug re- search as practiced by the industry, commercial success is the primary goal," in U.S. Congress, Senate, Hearings on a Bill to Amend and Su lement the Antitrust Laws w;th Respect to the Manufacture ang Distribution of Dru s and for Other Purposes, Part 3, 37th Cong., 2d Sess., 1962, p. 1525. Here- after cited as Hearings on S. l5i2. 159 From the budgetary point of view the future level cfi‘research expenditures becomes a cash outflow to be covered in part by the pricing decision of a new drug. But again such decisions cannot be made in isolation. Here it would appear necessary to consider the actions of competitors in terms of their discovering significant new products as well as their ability to market minor product modifications for their value in advertising. An example of the influence of the former action is provided by the activities of at least one antibiotic producer interviewed. With the success of streptomycin in 19a? the soil screening technique was a proven isolating method. It was generally known in the in- dustry that several firms were heavily committed to the use of this technique in the search for new antibiotics. The reaction by one firm was to operate this technique so as to produce a backlog of microorganism for investigation of their antibiotic characteristics. Through this procesa,if a compet- itor was successful in isolating a significant antibiotig there was a good chance that an identical antibiotic produc- ing microorganism would be found in the backlog. In the patent interference that would follow the chances of settle- ment with a license were considered excellent. The effect here may be to produce competitive research, but how general this is within the industry cannot be stated. Some support for the reaction of research expendi- tures to competitor's success in marketing a new drug is pro- vided in Bristol's reaction to the Lederle-Pfizer success 160 In 1951 Bristol spend 7.1 per cent with the te tracyclines . In 1952 with the isolation of its sales revenue on research. of tetracycline the level of research Spending increased to 13J4per cent. Thus a competitor‘s success may have the effect of increasing the level of research spending with the view of forcing a sharing of the market under a license arrangement. An additional influence on the level of research spending involves the element of newness and its value in Here there exists a force to spend for re- advertising. If a successful search in the form of product modifications. new drug can be changed sufficiently to avoid a charge of direct patent infringement but maintain the same therapeutic efficiency, a share of the market may be gained. But these competitive reactions are in all probability random and enter the budget as contingency elements. If the event occurs the research budget can be increased and/or other projects tem- porarily curtailed in an attempt to maintain the firm's competitive position. These various factors enter the budgetary analysis along with the other costs that can be expected to be incurred over time. The exact way in which they enter cannot be stated as in the final analysis pricing. is largely a matter of art rather than a strict science. But the budgetary approach does provide the firms an estimate of the lower limit of the price charged for the new drug. A Duopoly Model.--At this point it is possible to select a duolopy model to describe the pricing of broad 161 spectrum antibiotics. Here features of the Cournot and Fellner models are used to explain market price and shares. In the Fellner model the lower limit of the price range is determined by traditional cost function, while the upper limit is determined by the intersection of marginal 22 , But wnere cost and revenue where the firm is a monopolist. rivals exist and cannot be driven from the market the demand curve facing the firm depends upon rivals' reaction to the initial price selected. Setting a price near the monopoly end of the price range does not allow rivals to capture what they think is a sufficient share of the market, thereby lead- ing to price cutting. Setting a price near the zero profit end of the price range means that some profit is lost and the leader's position as interpreter of market conditions is en- In the final analysis the price selected must be dangered. acceptable to all parties given their view of rival's strength. In terms of the pricing of chlortetracycline per se, Lederle did not initially face the problem of a rival's re- While the price structure of the narrow spectrum action. antibiotics had its influence upon the selection of a price for chlortetracycline, the gap in the chain of substituta- bility was sufficiently great so that the effect was undoubt- edly minor. But Lederle must have been aware that Parke Davis was in the process of developing a broad spectrum anti- biotic, so that its monopoly position in the broad spectrum *— ZZWilliam Fellner, Com etition Among the Few (New Augustus M. Kelley, 1968 ed.), p. 33. York: 162 market was limited. From Lederle's point of view the con- servative assumption would have been that Parke Davis' chloramphenicol was of equal therapeutic effectiveness there- by implying an equal sharing of the market,1ess the advantage of Lederle's earlier lead in marketing. Thus while Lederle could initially select the monopoly price, with the entry of Parke Davis a lower price was necessary to prevent the pos- sibility of price competition. The assumption here is that while the market demand curve is inelastic, it is not per- fectly so. In addition, in certain marketstuch as in sales to government agencies,demand is probably less inelastic as purchases in foreign countries are an available alternative to domestic suppliers. The effect of a lower price with entry would be to increase total market sales allowing the second firm to reach an acceptable volume and maintaining the sales volume of the first firm. Under dynamic conditions, where the effectiveness of advertising is admittedythe demand curve shifts outward allowing both firms to gain an increased share of the expanding market without engaging in price competition. The problem now becomes one of estimating the mon- opoly price and the cost price. If it is assumed that pen- icillin was the only substitute for chlortetracycline, al- though sufficiently limited to have but a minor influence,23 23Both penicillin and chlortetracycline are effective against gram positive bacteria, but only chlortetracycline is effective against both gram positive and negative bacteria. In.addition where the exact bacteria causing the infection cannot be identified chlortetracycline would be the drug of 163 then the utility theory of drug therapy provides some clue to drug pricing. The utility theory of drug pricing states simply that if a new drug is five times more effective in curing a partic- ular infection than the drug currently in use, then the ratio 2“ With refer- of their prices ought to be five to one also. ence to the therapeutic effectiveness of chlortetracycline over penicillin, no medical authority is available to sub- stantiate any such determination. In light of the problems involved in attempting to quantify this sort of question, the lack of such statements from the medical community is not un- But in the area of the selling of drugs such expected. From a memo from Pfizer's restraints evidently do not operate. medical department to its sales force we have an estimate that chloramphenicol is from 7 to 34 times more effective than penicillin against certain types of bacteria.25 Assum- ing the same relations between chlortetracycline and peni- cillin a comparison of the price ratios of these drugs be- tween 1949 and 1952 results in data consistent with the utility theory. In Table 11 significance is attached to the price ratio in 1948 when the price on chlortetracycline was selected, as well as the five year average of the price ratios of 31. This is especially important in the prescription choice. market where the physician often does not have access to the Laboratory equipment to identify the bacteria, hence pre- tcribes the more inclusive chlortetracycline. 2'l"'Hea1'ings, Part 24, pp. 13639-13642. 2E'Hearings on S. 1552, Part 7, p. 3366. 161+ TABLE ll.--Computed prices of chlortetracycline and procaine penicillin on a per day dosage basis Procaine Year Chlorte tracycline Penicillin Ratio 19% 6.2 5 .26 21+ 1949 4.17 .11 38 1950 2 .50 .09 28 1951 2.12 .09 21+ 1952 2 . 12 .05 42 Source: Federal Trade Commission, Economic Re ort on Anti- p_1_:o_t1c Manufacture, 1958, W In October 1948, two months prior to the marketing of chlortetracycline, procaine penicillin was listed at 513.3%, or 3.26 converted to a cost per day basis to the consumer. Applying the factor of 31 to the price of penicillin would- indicate an initial price of chlortetracycline of $8.06 on a per day basis. This price is taken as the upper limit or monopoly price. An estimate of the lower limit or cost price is $1.00.?”6 26This estimate is derived from data presented by the Kefauver Committee relative to Bristol's production costs for tetracycline. Here they estimated that the production costs, defined as the cost of bulk tetracycline, capsules, finishing, packaging and royalties, at 83.91 for 100, 250 mgs capsules, >r on a per day dosage basis of 4 capsules at $0.16. Hearings, art 24, p. 13876. Taking this figure as the cost of goods aid, the expenses of selling, research, administration, and ‘ofit, can be estimated from financial ratios for broad >ectrum antibiotics. In 1956 the following ratios applied; :search 7 per cent; administration, 7 per cent; and profit :fore federal income taxes, 37 per cent. FTC, Economic ~ ort, p. 21].. Assuming that these ratios apply in the case ' tetracycline adds some .ho to the Committee's cost figure r a total cost of $.63. This difference between this figure 165 The actual price set by Lederle was 36.25. This price can be interpreted as an offer price to Parke Davis on the assumption of equal product efficiency resulting in a degree of market sharing. This in turn implies the use of an undifferentiated dquoly model which can be extended to include more than two participants. The Cournot model is sufficient to explain the observed price and market shares which resulted. The Cournot model in its simplest form assumes zero costs of production and a failure of the participants to recognize the element of mutual interdependence. With ref- erence to Figure l, a monopolist would produce an output equal to 0D charging the monopoly price of DP, with total profits equal to ODPC. In this case 1/2 the competitive output OR is produced. If a second firm enters the market on the assumption that the other firm's output is constant, the best it can do is sell an output equal to DE, charging a price of EP', with total profits equal to DEP'F. The first firm by assuming that the second's output will be constant can increase its profits now equal to ODFB by reducing its cautput to 1/2 0K less DE. This sort of reaction continues LUTbll an equilibrium is reached,at which point each producer contributes 1/2 to the total output and each maximizes profits.27 But at the point where output is reduced to arui the estimated lower limit on price is accounted for by decreased production costs between 1956 and 1949. Z7Edward H. Chamberlin, The Theory of Monopolistic Ccun etition (Cambridge: Harvard University Press, 1962), 3th ed. Chap. III. 166 F P' i \.\H D E H K Q Figure 1 increase separate profits the Fellner concept of qualified joint industry profit maximization can be introduced.28 That is, the price remains at P' and the first firm's output is not reduced in an effort to maximize its individual profits. This is because "the potentiality of . . . warfare always exists,"29 and the market is dynamic in terms of shifting cost;and demand curves. In addition, the market share which 28Fellner, pp. cit., Chap. VI. 291bid., p. 177. 167 results from the failure to move in the direction of the Cournot solution is favorable to the first firm in the face of incomplete information relative to its marginal cost and revenue curve. Lacking such information the share of the market resulting from a price decision in the context of rivals' reactions may be a first approximation to profit maximization. But the ability of the first firm to impose a market share on its rival depends upon Fellner's factor (a) through (d). The data relevant to prices and price changes and the participants view of rival's market position can be in- jected into this model in an attempt to explain these ob- servations. The monopoly price has been estimated at $8.06 while the actual price was $6.25. It has been argued that this was sun offer price prior to Parke Davis' entry on the assumption . >354 I H z. m H a 'B:.éee.a.§.§ : 8 83+M3wo us 0 ma ma 0 a 'dQLwiHHGM4 m mu m» «'4 Q) >ad) OHOS-‘JH $4 $40 HO - .1 ~ “2:22 we: :3 :3 Markets 5), 4c}; 5.4; 00 o a. o 0 E4 «no 4H!) Prescription 29 25 25 7 5 l 0 26 4 Hospital NPA ”9 25 25 7 5 1 O 26 4 Hospital CCS 2 0 l 1 Government VA 2 4 5 6 4 Government MMSA 5 5 O 1 l *8 5 *2 Bulk 19 24 21 *Indicates an increase Source: Hearin s on Administered Prices, Part 24, pp.l3663, 13778—13782; Part 25, pp. 14209-14224, FTC Economic Report, pp. 168, 184-85, 187; Bristol's Pro osed Findings, II, 199-200; Cyanamid's Proposed Findings, pp. 220, 243. 186 prescription and NPA markets list prices are usedehile data for the CCS market are winning bid prices, as are prices for sales to the government. Prices in the bulk market are offer prices. The time period covered is roughly from 1950 to 1960. An effort was made to secure the various price series for comparable periods, however this was not possible in the case of the government market where the series begins in 1956 and runs through 1959. From this data certain tentative conclusions may be drawn. First, relative to price structures, in the narrow spectrum market price changes are reflected in all markets for which data were available. The apparent exception is in the MMSA market but herepas notedythe price series is some- what shorter. In the broad spectrum market,on the other hand,this clear relationship is not observed. The implication here is that a greater effort was made to maintain clear sep- aration of markets for pricing purposes. Or in somewhat different terms that the greater elasticities of demand were recognized in the government markets. A further implication here is that given the more elastic demand in the government ;market and the fact of secret bidding that a greater degree of price competition existed in these markets.’+9 “9Here evidence of market sharing in the MMSA market ;1s ignored. Data were developed in the Hearings and in the IUD: case against the five tetracycline producers that showed “the following market shares for awards from November 1956 through October 1959; Pfizer 46.692, Lederle, Bristol, and Eyluibb all about 179%, and Upjohn 9.59%. However, there was (iisagreement as to what awards were excluded in developing 'these percentages, especially in the latter period. See .Hearings, Part 24, pp. 13700,13715. 187 A second obvious conclusion is that the rate of price decrease was greater in the narrow spectrum market than in the broad spectrum market regardless of whether the compari- son is on a product or a market basis. In the narrow spectrum market there is a continual decline in price with certain ex- ceptions. These are also reflected in Table 15. But in the broad spectrum market the last price cut to the prescription or NPA hospitals was in 1951. Prices for all four broad spectrum antibiotics did not change until late 1960 or early 1961. For a ten year period price stability was achieved. A third conclusion from Table 16 is the stability of the agreement on the price variable. Here the broad spectrum drugs are listed in their order of market introduction. The data on price flexibility shows a general decline from the introduction of chlortetracycline to tetracycline in the prescription and EPA markets. A final conclusion from this data concerns the be- havior of price under different market structures. Under the competitive conditions found in the narrow spectrum market price shows a marked tendency to decline as the costs of production decreased. The less competitive character of the broad Spectrum market shows relatively inflexible prices. The basic cause of this difference in price performance was the use of patents to exclude competition in the broad spec- trum market and7for those operating in this market7a conspir- acy not to compete in terms of price in the prescription and NPA markets. 188 The Price Fixing Conspiracy.--In the price fixing phase of the Federal Trade Commission case against Lederle, Pfizer, Bristol and its two bulk customers, Squibb and Upjohn, the issue is basically a question of whether these firms en- gaged in a conspiracy to maintain prices or whether their actions could be interpreted as mere conscious parallelism in an oligolopist market. If, as the respondents argued, the patent on tetracycline was obtained without the benefit of misrepresentation of the facts to the patent examiner, then the fact of identical price lists indicates only the recog- nition of mutual interdependence. That for one firm to cut price in an effort to increase its sales volume would be self-defeating as its out would be promptly met by rivals. The Commission on the other hand argued these firms engaged in a conspiracy to secure the tetracycline patent thereby excluding additional entry and agreed not to compete in price. In the initial decision the hearing examiner found that the issue of coproduction revolved around "the amount or substantiality of inherent production. . ."50 involved in the production of chlortetracycline according to the Duggar- Niedercorn patents. In accepting the Pfizer argument that what it found were useless trace amounts of tetracycline and that the patent examiner was uninterested in such non-commer- DOFTC, Initial Decision, In the Matter of American Cyanami Company, et al., Before the Federal Trade Commission, Docket No. 7211, 1951, p. 20. 139 cial 1uantities, the hearing examiner disposed of the fraud issue.51 With this issue settled the examiner argued that it was highly unlikely for the respondents to engage in a protracted battle for the patent on tetracycline and at the same time enter into an agreement to fix price.52 The market behavior of the respondents in publishing similar price lists,53 in developing a system of similar customer classifications,54 marketing identical dosage forms and sizes,55incidences of identical bids to the CCS hospitals and various government agencies,56 coercion of independent wholesalers where they bid under the list price,57 did not begin with the settlement of the Pfizer infringement suit against Bristol. Thus the examiner reasoned that to observe this behavior pattern be- fore the settlement of the infringement suit where a conspir- acy could not have existed, and to observe the same pattern later proved that a conspiracy to fix price did not exist. The examiner buttressed his reasoning by citing the fact that product competition existed in the form of combin- ations of tetracycline with other antibiotics. In certain cases such competition was not met by rivals or met with a one or two year delay in time.58 The decline in the share of the antibiotic market held by Lederle with the introduction —___ 5lipiio. p- 49- 52gpig., pp. 1u3-4u, 171. 53FTC, gppposed Findings, pp. 348-50, 371-77, 568-83. 54%” pp. uoz-os, u32-uo. 55.12113, pp. 348-49. 56l2$§°s PP- “71-75- 57ipgg., pp. 526-34. 5DFTC, Initial Decision. po 132- 190 of tetracycline indicated competition not conspiracy?9 The existence of intracorporate memoranda on the subject of rivals prices was interpreted as a "strong effort to ascertain the competitive practices of other marketers."60 The prices to large NPA hospitals were in effect cut at various times by the giving of "free goods" with the placement of a large order. This behavior was not evidence that the conspiracy periodically broke down, but rather evidence of strong competition.61 Thus the interpretation of the behavior of the five respondents de- pends upon the patent phase of the case. If a conspiracy ex- isted it had to be in effect prior to the issuance of the patent on tetracycline. The full Commission in overturning the examiner's initial decision argued first that the patent examiner was concerned with whether any coproduction occurred in the pro- duction of chlortetracycline. In effect it ruled that Pfizer designed its test for coproduction in a manner that would support its statement that only "useless trace amounts were found." That isathe tests were not conducted strictly according to the Duggar-Niedercorn patents and that commercial recovery techniques were employed which were incapable of re- covering the small amounts of tetracycline coproduced. Not to have informed the patent examiner of these facts consti— 62 tuted material misrepresentation. 60 61 59Ibid., p. 135. 62FTC, Findingp As To The Facts and Conclusions of Law, In the Matter of American Cyanamid Company, et al., Before the Federal Trade Commission, Docket No. 7211, 1953, pp.21-27. Ibido, p0 luéo Ib1d0,pp0152-jéo 191 In regard to the conspiracy phase of the case the Commission stated that the law did not require that they fix the exact time when the agreement became effective. Reasoning from the 1960 Electrical Equipment case it was entirely pos- sible that while higher echelon management was engaged in a bitter centest for the patent, the sales management personnel of the respondents had entered a conspiracy for the orderly marketing of tetracycline regardless of the outcome of the patent issue.63 Thus a conspiracy to monopolize the tetra- cycline market did exist before the settlement of the infringe- ment suit and the observed behavior in both the patent and commercial phases of the case were found to be consistent with this finding.64 In the matter of price the Commission agreed that competition did not exist in the prescription market, nor ? contrary to the examiner's decision did it exist in the hosp- 7 ital or government markets. The use of free goods to secure a particularly large order in the NPA market was an example of "chiseling" which was tolerated as it did not effect prices in prescription markets where "75 per cent of the respondents' total sales were made . . ."65 In the government market while 63FTC, Opinion of the Commission, In the Matter of Of American Cyanamid Company, et al., BEfore the Federal Trade Commission, Docket No. 7211, 1953, pp. 78-81. The Commission also noted that the Broady tapping of the telephone lines of Bristol and Squibb was unauthorized by the top management of Pfizer. 64 The Commission on the timing of the conspiracy noted that it existed as of early 1955, ibid., p. 79. 65Ibid., p. 84. 192 some variation in price existed they were in fact insignifi- cant and designed to preclude an antitrust investigation.66 With regard to the shifts in the market shares which the examiner found inconsistent with a conspiracy theory, the Commission argued that such shifts were to be expected as the respondents did compete in terms of promotion efforts while not in price.67 Defining the market in terms of both broad and narrow Spectrum antibiotics, in 1950 Lederle held about one quarter of this market base on its sales of chlortetra- cycline. In 1956 it held a bit more than one quarter based chlortetracycline and tetracycline. Thus with the marketing of tetracycline Lederle's relative position in the antibiotic market was slightly improved. If on the other hand a narrow definition of the market is applied to include only broad spectrums Lederle's share moved from close to 100 per cent in l9h9 to progressively lower shares as competitive products were introduced. But this smaller relative share was of an absolutely increasing market. Between 1953 and 1956 there was a 50 per cent increase in the volume of sales of tetracycline alone. From the Cournot model employed above it is expected that the innovator's share of the market will decline in a static world as this is an alternative to price warfare. In a dynamic world of shifting demand curves this result is unlikely. For here the leader maintains his abso- lute sales volume giving up the new customer demand to the followers. J‘----.-a.u-.-.*- - 661b1do, pp. 86-870 67Ib1do’ p0 1030 193 The totality of evidence in this case would support a finding of conscious parallelism, but this the Commission rejected. To the Commission in the words of Kaysen and Turner, "The essence of the problem is not, Is there parallel action: but rather, Is there anticompetitive action?“68 The very purpose of Section 5 is to prevent "unfair methods of competition." The act does not require a showing of lessen- ing of competition, only that the techniques employed,in the opinion of the Commission, were in some sense "unfair." The formation of a conspiracy to fix price falls clearly within the context of an "unfair" market technique. Estimate of Monopgly Revenue§.--The older defense of monopoly was that it was necessary for technical progress. The current defense of oligolopy is that the simple inventions have been made and that for continued progress large size in terms of laboratories and supporting research budgets are necessary. From this it follows that the larger the profit margins the larger can be the expenditures on research and the greater the rate of technical progress. This question is examined in detail in the following chapter, but here it seems appropriate to attempt some estimate of the monopoly revenue available from the sale of tetracycline and to allocate this among the various producers. The simple hypothesis to be tested in the following chapter will be whether the firms with a larger excess profit had a higher rate of drug invention. 68Carl Kaysen and Doland Turner, Antitrust Policy (Cambridge: Harvard University Press, 1959), p. 105. 194 In the absence of cost data any estimate of monopoly revenues must be crude. What is available is the price of tetracycline, output and an estimate of sales in different 7 markets. The critical assumption in this analysis is that the gains in technology9reflected in the price of penicillin) were also available to the tetracycline producers. Between 1951 and 1960 the average decrease in the price of penicillin was about 5 per cent. This rate is applied to the 1953 price of tetracycline to produce column II in Table 17.69 By 1960 TABLE l7.--Estimate of monopoly revenue in tetracycline (1953-1960) fl e-c o Wmmgo-m‘“-—— —-.-. -— Sales, No. Estimated Total Estimated Difference of Bottles Monopoly Market Actual Competitive Col. II - 16,250 mgs devenue Value Year Price Price Col. I (000,000) (000,000) (000,000) 1953 $5.10 $5.10 4 0 3.4 I] 0 47 1954 5.10 4.85 .25 22.9 5.7 1955 5.10 4.61 .49 21.9 10.7 1956 5.10 4.38 .72 15.8 13.5 72.9 1957 5.10 4.16 .94 25.4 23.9 108.5 1958 5.10 3.95 1.15 20.7 23.8 88.9 1959 5.10 3.75 1.35 21.8 29.4 89.4 1960 5.10 3.56 1.54 21.8 33.6 69.4 Sources: Sales in terms of numbers of bottles computed from FTC, Economic figport on Antibiotic Manufacturer, 1958, pp. 73-75, for 1950 through 1956; for sales for l957 through 1960, and total market value, U.S. Tariff Commission, Synthetic Organic Chemicals, U.S. Pro- duction and Sale. 69The figure of 5 per cent produces a rather conserva- tive estimate of what the decrease in price would have been under competitive conditions. In the case of penicillin the 1946-60 rate of decrease averages 18 per cent, for streptomycin it is 17 per cent and for dyhydrostreptomycin it is 16 per cent. 195 it is estin lated that under a competitive market structure price would have fallen to 33.56.70 The reasonableness of the estimate of price in Column II can be shown by a comparison of average prices in foreign markets in 1959. In that year three foreign markets had average prices for tetracycline lower than the estimated price. These wcr Blazil with an average price of $3.10, France with .82.69, and Argentina with 81.10. In testimony before the Kefauver Coumittee Lederle stated that the situation in argentina and Brazil was explained by price controls, but that 71 was made. It would seem that in a price set by a subsidiary some allowance would be included in cost for re- search and advertising. This would appear to be especially true if such information could be used to convince the price control e.utiorities that higher prices were justified. Lack- ing direct evidence on this point it can only be concluded that the estimated conpetitive price seems reasonable as judged by iorei ign prices in three markets. The 1953 price of tetracycline can be interpreted as a price that bore a reasonable relationship to the competitive situation in tetracycline. But that over time as technOIOgi- cal gains bec'a :ue available at the rate of 5 per cent pe1 yeai, C‘J llf ti" i‘l‘ they were not ref acted in price. hus as average cos ts s A ‘ downward the margin between costs and priCe ichrea at. v— W ‘— v—w— ———————- O a/l, ‘ g -\ o 7 In 1964 the price of tetracycline actually was cut to $8.63. 71w . a ‘ gearings '. 1774 12744 1274: and 137'0. 3 pl) 2 i J a J. 9 J 0 1;; 301141 IJ is eh estimate of the year y volume of sales in the prescription market in terms of standard bottles cf / r I" o _ . . ._ ‘ o _ r -v, R: 2 h . ..l.. I.‘.. v.1“; C ‘L' M x; e O .L“ . b.) - L“ J- L.) as.) l. J 11. " l , 50 l;“' capsilss “iis 6°tln1t£ 1c L'sed or b‘C ) Commission's statement that 75 per cent of the respondents' ‘1 1 . . . . '2 sales 01 tetracycline were in the prescr1ption market. Column V is an estimate of the monopoly revenue available iron the sale of tetracycline to the five sellers. In com— parison with the total value of the market (Colun; VI), the element of homepoly revenue increased from 18 per cent in 1956 to 48 per cent in 1960. With the data available there is no method to al- locate the monOpoly revenue among the five sellers. The best that can be done here is a rough ranking based on market shares. From the data presented by Bristol73 it is clear that Lederle enjoyed the largest, although declining, share 0f the market for the period 1954-58 with Pfizer second. Upjohn ranks somewhat ahead of Squibb in third position with Bristol last. This ranking ignores the fact that Bristol was the bulk supplier of Squibb and Upjohn hence profited from the latter firm's sales. 0 7“An upward bias exists in this calculation as the total sales volume reported by the Tariff Commission is div- ided by the wholesale price in the prescription market of £5.10, resulting in the number of bottles sold. 75 per cent of this figure produces the estimate for the prescription market. Thus the assumption that all sales took place at a constant price of 35.10 overstates the case. But given the conservative estimate of techn010gical gains the bias is not likely to be seriously disrupting. 738cc Table 6, Appendix, p. 246. 'I. . Hula valid ill: 1 111A 197 while the conclusion here must be tentative,given the nature of the data and its manipulation, it does appear that profits in the tetracycline market have been substantial and sufficient to provide an inducement for the investment in research. The question in the following chapter is whether any relationship exists between the existence of monopoly power and performance in terms of significant new drugs. CHAPTER VI RESEAHCH AND DEVELOPMENT The discussion up to this point has been concerned primarily with the antibiotic segment of the ethical drug in- dustry. In an attempt to examine the record of technological progress the sample has been eXpanded to include the twenty major firms who have innovated drugs of some consequence since 1945. The purpose of this analysis is twofold. First, it at- tempts to develop a criterion of technological progress by which the record of the antibiotic producers can be Judged; and second, to raise specific questions concerning the ef- fectiveness of the research activity of the industry as a whole. In particular the following questions are asked. How many drug innovations have appeared between 1945 and 1959, and how significant from a medical point of view is this record? Is the spending on research proportionate with firm size and are the larger firms more productive of innovation? Are scale economies present? Finally, in light of the mon- opoly position of the antibiotic producersyhave they been more or less productive than the industry? 198 199 THE DATA AND THEIR SOURCES The basic data required in attempting to answer the questions posed are the record of new drugs introduced by firms in the industry, together with data on sales and var- ious categories of expenditures. A record of new drug intro- ductions was prepared by the Pharmaceutical Manufacturers Association for presentation to the Kefauver Committee.1 Data on sales of ethical drugs are provided for one year by the Kefauver Committee for fifteen firms,and it has been possible to separate out estimates of ethical sales from total sales for an additional six firms.2 Data on various categories of expenditures are provided in the Committee re- port.3 Additional data are provided on a time series basis by firms' published statements to stockholders as well as the usual financial reporting publications. £$§t of Significant New Drug§.--The PMA list of sig- nificant new drugs includes 93 entries. The apparent criter- ion for inclusion of a particular drug was its importance in current drug therapy. Current, indicates those drugs that lU.S. Congress, Subcommittee on Antitrust and Monopoly, Administered Prices in the Drug Industry, Part 19, 86th Cong., 2d Sess., 1960, pp. 108h0-1085h. 2Ibid., p. 10993. It was necessary to drop CIBA from the Committee list as the research expenditures reported also included those made in Switzerland. 3U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly, Report, Administered Prices Drugs, 87th Cong., lst Sess., 1961, p. 31. 200 had not been superseded by newer, more efficient drugs as of 1959. As such it is possible that significant drugs that were later improved upon and replaced as the drug of choice in a particular course of therapy are not included in the PMA list. One such case was found in the drug chlortetracycline which was the first of the broad spectrum family of drugs. Another case was dihydrostreptomycin, an improved version of strepto- mycin. Both of these drugs have been added to the PMA list, but it is possible that others have been missed. The second part of the PMA criterion is the importance of a particular drug. The implication is that the drug is of significant market value, but no volume figure was supplied. Again it is possible that a new drug of significance in the treatment of a disease of low incidence was not included. Thus, it is not possible to claim complete coverage for the PMA list even with the noted additions. It is assumed, how- ever, that the coverage is sufficiently complete for the re- sults that follow. In addition to listing significant new drugs the PMA also supplied data on the inventing firm or individual, the innovating firm or firms with the date of introduction, and a critique of the advantages of the drug. The FHA list was used only for the period 1945 through 1959. Entries prior to 19H5 were deleted on the grounds that it was not until after World War II that the industry was forced to look to its own resources for new drugs. Of the eleven entries prior to 1945, nine are clearly products of 201 foreign origin, and one could well fall in this category. Of the 84 remaining entries included in the list after 1945 some 33 were the product of foreign research. Classification of Significant New Drugs.--A simple listing of new drugs would fail to consider that some are of greater importance than others. A solution to this problem is provided by defining three categories of importance and classifying each drug on the basis of the PMA critique and by a standard text on pharmacology.4 The first category is defined to include drugs that were the first in a series and/or the first effective treat- ment for a particular disease or range of diseases. The second category includes drugs that are similar in their range of effectiveness as established drugs, but were either more potent and/or carried a reduced incidence of side effects. The third category includes drugs that appear to be of little or no proven advantage over established drugs. These drugs appear to be perfect substitutes. Table 18 presents the results of classifying 84 drugs into the three categories of significance as well as data for the 1951 through 1959 period indicating the total new drugs introduced to the market. The data from Table 18 provides an indicator of the industry's productivity in inventing and innovating new drugs. __ ”The text chosen was E. Krug, Pharmacolo in Nursin (St. Louis: The C. V. Mosby Co., 1963 , on he sis of clear statements of the effectiveness and side effects of particular drug in practice, rather than in terms of manufacturers' claims. 202 TABLE 18.--Distribution of 84 new drugs by categories of significance and by years (1945-1959) Categories W New* Per cent Chemical Included in Year I II III Total Entries PMA List - lllll i.---iix.....-...-..---_---..---s-i 1945 l l 1946 2 2 1947 l l 1948 l l 2 1949 4 2 2 8 1950 2 2 4 l95l 2 2 4 35 .11 1952 2 3 5 35 .14 1953 l 5 e 48 .05 l954 l 3 3 7 38 .2i i955 l 5 0 ii ~19 1956 l a 5 42 .l4 1957 l i 9 12 51 .24 1959 i i 7 9 44 .20 1959 2 lo 12 63 .19 Totals 12 23 49 84 * New Chemical Entries are the official listing of the United States Pharmacopoeia on new drugs introduced in the domestic market. Source: U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly Administered Pgices in the Drug Industry, Part 19, 86th Cong., 25 $8830, l960,ppe 10840-16854; and Pharmaceutical Manufacturers Association, Prescription Drug Industry, Fact Book, 1963, pp. 2-5. In this matter the industry argues that its performance has been more than adequate in comparison with European producers.) 5The real test here, of course, is a comparison of the rate of invention and innovation with the opportunities to in- vent and innovate. However, as the latter is indeterminant, analysis of the performance variable is forced to consider secondary indicators. 203 Since World War II research expenditures have been increased five fold. Lacking a knowledge of the industries' opportuni- ties to invent and innovate we have no way to evaluate this argument with reference to total figures on new drug intro- ductions. On the other hand it is possible to examine the composition of the aggregate figures and arrive at a Judgment of the industries' technological performance. While the data on new drug entries in the United States Pharmacopoeia (New chemical entries) and the total of new drug introductions based on the PMA list, are not strictly compar- able, as the former refer to dates of discovery while the latter are based on the dates of market introduction, the series does indicate that between 1951 and 1959, a greater proportion of new drug discoveries are significant in that they are included in the PMA list. But the doubling in the PMA inclusion rate may well be a statistical illusion. That is, the PMA list tends to include in the latter years, drugs that were still in the stage of clinical testing and were marketed as part of this testing program. Drugs that were so indicated by Krug in 1963 were classified in category III. Thus the absolute growth in the numbers of drugs in this category is perhaps slightly overstated. What does stand out here, however, is that there has been no apparent upward trend in drugs classified in category I or II. So that while there has been an increase in the num- ber of new drugs, an increasing proportion of them have been of little or no significant value in terms of technological advancement of drug therapy. 20h Industry_aesearch Expenditures.--Data on industry spend- ing for research and development is published by the National Science Foundation for the period 1956 through 1959, and in index form by the PMA from 1948 through 1959. The NSF data is considered more inclusive as all drug firms are included in its survey while the PMA includes only its members, and, in addition, the NSF definition of research and development is more restrictive in what expenditures may be so classified. Thus the NSF data is used as the relevant figures. The problem with the NSF data is that it does not re- flect price level changes. These have been in the order of 6.7 per cent per year,on the average, for research and de- velopment for the chemical industry 1J1 the period l9h5 to 1959.6 It is assumed that the ethical drug industry faced that same magnitude of price changes for research and develop- ment as the chemical industry. Table 19 presents data on the level of research and development spending by the industry adjusted for price level changes, new chemical drugs, and the average cost per drug. With such data we are in a position to examine the question of whether,over time,it has become more expensive in real terms to invent and innovate new drug products. 6 The price index for research and development expendi- tures for the chemical industry were computed from Tariff Com- mission data which show the labor costs of research and develop- ment activities and employment of scientific personnel. From this it is possible to compute average salaries and the increase in the salaries of scientific personnel. The Tariff Commission data also indicates the proportion of labor costs to total costs. Such was used to weight labor costs in determining the total increase in price. For material costs the wholesale price in- dex was used. N O \n TABLE l9.--desearch and development costs per new drug dis- covery (1951-1959) Real 3 & D Cost For New Expenditures New Chemical Drug Discovery Year in Millions Entries in Millions 1951 40.5 35 1.23 1952 38.4 35 1.09 1953 41.5 48 0.86 1954 40.0 38 1.05 1955 46.8 31 1.48 1956 52.2 42 1.24 1957 56.2 51 1.10 1958 69.0 44 1.53 1959 76.4 63 1.21 Source: B and D EXpenditures are from the National Science Foundation, Review of Data on a & D, No. 24, Dec. 1960, Washington. Price Index was computed from U.S. Tariff Commission, S nthetic Organic Chemicals U.S. Production and Salg,Second Series, various issues, Washington. If we attribute all new drugs discovered and innovated to the ethical drug industry we find that the costs in real terms have increased slightly in the nine year period for which data is available. Dividing the period into three year sub- periods, the average cost rose some 18 per cent between the 1951-53 period and 1954-56, but less than 2 per cent for the 1957-59 period. While it is difficult to associate research and deve10pment spending with new drug introductions, the 18 per cent increase in real expenditures for the first two sub- periods is associated with a decline in new drug introductions in category I and II from 10 to 6 (Table 18), and with an in- crease in category III introduction from 15 to 18. A comparison 206 of the second and third subperiod associates a 2 per cent in- crease in real research and development spending with an in- crease in category I and 11 drugs from 6 to 7, and an increase in category III from 18 to 33. Thus it would appear that while for the period as a whole real research and development ef- forts have increased, this increase has resulted in a shift in emphasis to the development of substitute drugs to compete with patented established products.7 ECONOMIES AND DISECONOMIES OF SCALE In this section certain cross sectional data are de- veloped on twenty ethical drug producers later to be used in comparing the performance of the broad spectrum producers. From this data it is possible to answer certain questions pertaining to the industry and compare the results with find- ings in other industries. In particular have the largest firms been the principle sources of innovations in the in- dustry? Do the largest firms spend a larger proportion of their assets on research and development? Do proportionate increases in research and development expenditures result in proportionate increases in new drugs? Or are their economies of scale in research and development, in selling efforts or in production costs? In addition to providing tentative answers to these questions there are certain byproducts of the analysis that 7Computation of the correlation coefficient between real 3 and D spending and new chemical entries yields a value of .74 indicating a fairly strong relationship between these two variables. 207 extend the work of others. In particular it is possible to extend Mansfield's investigation of the relationship between the size of firms and innovations to the ethical drug industry as well as his investigation of the relationship between re- search expenditures and inventive output.8 It is also pos- sible to test Hamberg's thesis that the large industrial lab- oratory will not be the major source of inventions.9 Finally, we ask whether any relationship can be discovered between patent statistics and the record of inventive output. Research spending.--Relative to the question of what determines a firm's research and development budget Mansfield's model seems to yield reasonable results.10 This model explains the current level of a firm's R & D spending as a function of the desired level of this activity and the fraction of the distance between actual levels and desired levels that the firm is able to move in a budget period. This assumes that the desired exceeds the actual level and that instantaneous adjustments cannot be made because of inefficiencies in rapid adjustments. The desired level of R & D spending depends upon the number of projects generated (a function of the size 8Edwin Mansfield, "Size of Firms, Market Structure, and Innovation,“ Journal of Political Economy, LXXI (December 1963), and "Industrial Research and Development Expenditures, Determinants, Prospects, and Relation to Size of Firm and In- vegzive Output," Journal of Political Economy, LXII (August, 19 . 9D. Hamberg, "Invention in the Industrial Research Laboratory," Journal of_§olitical Econom , LXXI (April 1963). 10Mansfield, "Industrial Research and Development Ex- penditures, Determinants, Prospects, and Relation to Size of Firm and Inventive Output," 22, cit. 208 of the firm, i.e., larger firms have larger research staffs hence generate more proposals), the probability distribution. of their expected rates of return, and the required minimum rate of return for the firm. The fraction of the distance a firm will move towards the desired level depends upon the per- centage of the firm's profits that would be absorbed in such a move and the ratio of industry research spending to profits. Thus a firm spending below the industry average and with below average profits, but with a normal distribution of research proposals would be restricted in its movement to the desired level2as such a move would absorb too large a share of the following year's profits. 0n the other hand a firm in the same situation but with a particular project with a high 7 probability of a large gain,would be less hesitant to increase 11 In addition the effects of patents on its current budget. the model would be to increase the probability of return and to increase the level of spending. The question of competitive research and the possibility of rivals being first in the Patent Office is a bit more difficult to handle within Mans- field's framework. Such situations would seem to require a side calculation (and an additional equation). Here the firm would have to have knowledge of the projects of rivals, es- timate its chances of being first or a sufficiently close second to force a cross-licensing situation, its chances of being first with rivals sufficiently close to force a cross- 11Something of this nature seemed to have occurred with Bristol. See Chapter V. 209 licensing and the effects of these possibilities upon the rate of return from the project. Thus while the Mansfield model is perhaps not complete in all details, it does provide a sufficient approximation to determine the level of a firm's 3 & D spending. Size and Inventions and Innovations.--In this section we examine the relationship between size, various categories of expenditurespand new drug introductions. In particular we test the hypothesis that the larger the firm the greater its absolute rate of spending on research and development and the greater its productivity. Hence we would expect that if there are no diseconomies of scale in the research activity that the larger the firm the greater the number of new drugs invented and innovated. To test this hypothesis twenty major drug firms were selected on the basis of their size as measured by sales of ethical drugs and the availability of data on research and development expenditures. These twenty firms accounted for about 87°0 of the industry's research and development Spend- ing in 1958 and 759g of all new drugs introduced between 1945 and 1959. The scope of the test is limited by the fact that complete data on ethical sales by firms is available only for 1958. Hence, it is assumed that research and development ex- penditures and sales are representative of the differences be- tween firms over time. Table 20 provides data on twenty major drug producers using sales as a measure of size. The sample is divided into 210 TABLE 20.--1thical drug sales, research and selling expendi- tures for 1958, inventions and innovations for twenty major drug producers 1945-1959 (in millions) —~ weighted Index of Gross Signifi- Ethical Selling Research New Inven- canoe for Firms Sales Expenses Expenses Drugs tions Innovations Lederle 187 43 12.0 8 7 12 American Home 180 43 5.8 6 l 10 Products Merck 176 32 14.2 6 4 l3 Parke Davis 172 43 8.3 3 l 7 Lilly 166 30 14.6 6 5 6 Upjohn 144 30 12.7 3 2 5 Pfizer 120 32 5.9 5 3 7 Abbott 112 31 6.3 0 0 0 Squibb 99 19 5.3 7 l 13 Sterling 71 26 2.3 l 1 l Schering 69 24 5.7 5 3 8 Smith Kline & 60 12 5.3 4 l 7 French Warner Lambert 52 14 2.7 2 2 3 Searle 33 7 4.0 4 4 9 Mead Johnson 31 9 1.8 l O 1 Carter 23 6 1.7 0 O O Vick 22 6 .8 4 2 5 Bristol 19 6 2.6 1 l 2 U.S. Vitiam 14 5 .7 2 1 2 Norwich* 8 3 .3 1 l 2 *Norwich was purchased by American Cyanamid in the middle of 1957- Source: U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly, Administered Prices in the Drug_Industry, Part 19, 1960, pp. 10993,10840-10854; Report,_Admin- istered Prices Drugs, 1961, p. 31. 211 three size groups on the basis of reported sales of ethical drugs.12 Expenditures for research and development and for selling are from the Kefauver Committee Report and apply only to drug operations for the year 1958. Gross new drugs are the total number of new drugs introduced by the firm inclusive of competitive introductions for the years 1945-1959. Inven— tions are the number of new drugs discovered by the firm in question, while innovations include both the firm's inventions and purchases of marketing rights from domestic and foreign inventors. The weighted index of new drug significant is the result of distributing each firm's innovations among the three categories of importance. A weight of 3 is assigned to drugs falling in category I, 2 for those in category II and l in caregory III. This index is on a gross basis as in several cases more than one firm is listed by the PMA as the inventor or innovator. The focus of the analysis is upon innovations rather than upon inventions. This rec0gnizes the argument of the industry that.without the extensive clinical and pre-manufactur- ing testing of new drugs performed or supported by the drug manufacturers,the rate of such innovations would be consider- ably lower in a private enterprise system. In addition, the series on innovations is undoubtedly more accurate than that on inventions as it is not clear in all cases how the PMA handled credit where the discoverer was supported by corporate funds in his research. 12The rational for the division of these firms into three categories were the gaps in the sales figures which con- veniently resulted in an almost equal number of firms in each category. 212 We are now in a position to examine the question of whether large firms spend more on research and development than do smaller sized firms. In absolute terms it is obvious that the larger the firm the larger the spending on this activ- ity. But in relative terms the smaller firms in the sample spent on the average a larger percentage of their sales dollar on research and development than did the larger firms. The average expenditure for small firms was 6.8 per cent of sales, for medium sized firms 6.1 per cent and for large firms 6.4 per cent. Major significance is not attached to this result due to the limitation or lack of time series data on research spending by firms. The data here do seem consistent with the conclusion that spending on research and development is pro- portional with size on the average. If spending on research and development is proportional with size as measured by sales, are the results of this activity also proportional? In absolute terms the answer is clear that the larger firms do on the average have a higher rate of in— novation either in terms of numbers of drugs introduced or in their average relative importance. For firms classified as large,average new drug introductions were on a weighted basis 8.6. For medium sized firms the weighted average was 5.3. For small firms the weighted average was 3.0. In terms of the weighted index large firm introductions were greater than small firm introductions by a factor of 2.87, and medium sized firms had a rate of introductions of 1.77 over small sized firms. Thus in absolute terms large firms on the average were 213 more productive in innovating new drugs over the 15-year period. A similar examination of inventions yields the follow- ing results. Large firms in the l5-year period between 1945 and 1959, averaged 5.1 new drugs on a weighted basis. Medium sized firms averaged 1.5. Small firms averaged 2.3 on a weighted basis, or almost twice that for the medium sized firms. Again in terms of the weighted index larger firms were more productive on the average over small firms by a factor of 2.22, but small firms were almost twice as productive as medium sized firms. Thus in the case of inventions the ad- vantage seems to lie with the large firms over their two com- petitive size groups. While the larger firm is more productive in absolute terms is it also more productive per research dollar? Or in somewhat different terms are their economies or diseconomies in the research and development activity? Each of the three subgroups in a sample differs in their research and develop- ment expenditures by a scale factor. Medium sized firms spend on the average 2.7 more on research and development than do small firms, while large firms spend 6.2 times as much on the average as do small firms.13 Using these scale factors Table 21 compares the expected weighted value of inventions and innova- tions with the actual average value. If large firms are as efficient as small firms then we would expect increases in 13Comparable figures based on sales are 3.6 and 7.6, which, given the nature of the data, seem reasonably close to the scale factors that apply to research expenditures. 211; TABLE 21.--mxpected and actual values for inventions and innovations Weighted Values Inventions Innovations Group ixpected Actual EXpected Actual Large 14.2 5.1 18.6 8.6 Vledium 602 1.05 801 503 Small 2.3 2.3 3.0 3.0 absolute research and development expenditures to yield pro- portionate increases in invention and innovation. In fact, the actual rate of invention and innovation falls below the expected in all cases. Thus it would appear that there are diseconomies of scale in large scale research and development. Or in somewhat different terms small firms appear to be more productive than large firms per dollar spent on research and development. Within the context of Hamberg's hypothesis are ”the proportion of all minor inventions originating in the large industrial laboratories . . . likely to exceed the proportion of all major inventions originating in these laboratories?"1’+ Table 22 sheds some light on this question for the ethical drug industry. Here we distribute 39 inventions credited to commercial firms by the PMA by categories of significance and by size groups. In terms of the total number of inventions included in the sample of twenty firms, commercial firms dis- lQhambcrg, {ournal of Political Economy, LXXI, p. 95. 215 TABLE 22.--Pcrcentage distribution of 39 inventions of twenty firms by categories of significance an by size groups for 1945 to 1959 Categories of Significance Size Group I II III Totals Large .05 .23 .31 .59 I‘ledium 000 005 01.3 018 Small .05 .08 .10 .23 Totals .10 .36 .54 1.00 covered about 55 per cent of the total number of all reported innovations. In Table 22 for both the large and medium sized firms it is apparent that inventions of little significance accounted for the greatest proportion of the inventive activity, while for the small firms inventions of some significance (categories I and II) accounted for the greater proportion of the inventive output. In terms of the total number of drugs from the EMA list, eleven can clearly be identified as discovered by non- commercial laboratories. Of these eleven, seven or 63 per cent fall in category I and II, with four or 36 per cent in category III. Thus the data seem consistent with Hamberg's hypothesis that commercial large laboratories will not account for a disproportionate share of important inventions. Table 23 is similar to Table 22 except that the dis- tribution is of 70 innovations by twenty firms. Here for the small and medium sized firm the proportion of innovations was about equal for category I and II as compared with category r- ,’ r. TABLE 23.--Percentage distribution of 70 innovations of twenty drug firms by categories of significance and by size groups 'for the period 1945-59 Categories of Significance Size Groups I II III Totals Large .10 .14 .30 .54 Medium .04 .10 .13 .27 SI.-.a.ll 003 006 010 019 Totals .17 .30 .53 1.00 III, while for the large firms the proportion in category III was greater by six percentage points. Slightly over half of the drugs marketed between 1945 and 1959 which the aha con- sidered significant were in fact substitutes for established drugs. It is difficult to Judge the technical performance of the industry from these figures. The marketing of the perfect substitute would seem to increase the degree of competition faced by the innovator and result in a net social benefit if this competition was reflected in price. But if the case of the broad spectrums-is any indicator of price competition in the rest of the industry the effect is to increase the flood of advertising in an attempt to increase market shares. Lacx- ing knowledge of the industry's Opportunities to invent and innovate perhaps all that can be concluded here is that the industry's claims to technical progress must be heavily dis- counted with the recognition that product competition and higher advertising expenditures are the result of patent monopolies. 217 Turning here to selling expenditures the question Can be raised whether there are apparent economies or diseconomies of scale in this activity. The method of analysis is similar as in the case of research and development expenditures. Table 24 compares expected and actual selling expenses for TABLE 24.--Expected and actual values of selling expenses for small, medium sized, and large firms for 1958 (in millions) Selling Expenditures Group Average Sales Expected Actual Large 164 46.8 36.8 Medium 77 21.6 21.0 Small 21 6.0 6.0 three groups of firms which differ by a scale factor based on sales of ethical drugs. Here the average level of sales for large firms differ from small firms by a factor of 7.8; medium size firms differ from small firms by a factor of 3.6. If we assume that similar differences appear for assets devoted to ethical drug production, that economies of scale do not exist in production,15 and that similar rates of return are earned on assets employed, then the average profits on sales would be similar. =ata on profits as a percentage of sales is available and shows that for the twenty firms included in the sample? profits on sales are quite similar for the three size groups. After tax profits for large firms averaged 13.8 per cent, for ljSee Federal Trade commission, Economic Report on gntibiotic Manufacture (wasnington: Government Printing Office, 1958), Chap. IV; and Report, op. cit., pp. 73-74. 21o medium sized firms 12.4 per cent, and 13.9 per cent for small firms in 1958.16 Thus, with similar rates of profits earned’ the data in Table 24 indicates that larger firms are able to generate a similar level of profits with less than propor- tionate increases in selling expenses. Then for the large firms it would appear that economies of scale in selling off- set diseconomies of scale in research. That for the medium sized firms there are no economies of scale in selling and with diseconomies of scale in research reflected in lower profit rates. And that the firms classified as small make the most efficient use of resource on the average. Findings in Other Industries.--Mansfield investigated the relationship between size and innovations in three indus- tries--stee1, petroleum, and coal. The hypothesis was that if larger firms devoted the same proportion of their resources to innovating as did smaller firms, and if they were as ef- ficient in applying results, then "one would eXpect their share of the innovations to equal their share of the market."17 He found that "the largest four coal and petroleum firms car- ried out a larger number of innovations than this, but the largest four steel producers carried out fewer."18 The share of the ethical drug market held by the four largest firms was l6Profits as a percentage of sales for 1958 obtained in the Subcommittee Report, p. 31. 17Mansfield, Size of Firmg, Market Structure, and Innovations, p. 560. 18Ibid., p. 561. 219 in 1958, 31.6 per cent.19 These firms introduced a total of 23 new drugs out of a total of 84 included in the PhA list, or 27 per cent. In terms of the weighted index of signifi- cance, these four firms had an index of 42 out of a total of 183, or 23 per cent. Thus the ethical drug producer would be classified with steel producers. Mansfield eXplains the differences in innovations in steel and in petroleum (the only two industries for which he could gather data), on the basis of the size of the market, the size of the firms in the industry, and the minimum in- vestment necessary to introduce the innovation. That is, firms must be of some minimum size relative to the market be- fore it is profitable for them to introduce the invention. Given this information it is possible to estimate the number of firms capable of innovating. If the innovation is expen- sive in terms of investment costs, this reduces the number of firms capable of making the innovation. In terms of product innovations, Mansfield found that the number of firms of minimum size were about the same for petroleum and steel, but that the cost of innovating was on the average less in steel, hence more firms tended to engage in the innovating activity thereby reducing the share of the top four. While data is lacking with which to apply Mansfield's model to ethical drugs (assets devoted to ethical drug pro- _* 19Hearings, Part 19, p. 10993; and 1958 Censusgf Manufactures, II; Industry_Statistics, Part 1; General Survey and Major Groups 20 to 28 (Washington, 1961). 220 duction over time), we can specify that similar conditions as to numbers capable of innovating are present. That is, in spite of the sharp increased costs of marketing new drugs in terms of direct mailing, Journal advertising, free samples, and detailing,20 the technical character of the product would seem to insure wide market acceptance of a significant new product. Thus in ethical drugs as in steel it would appear that size alone does not insure a high rate of product in- novation. In another test of the relationship between the level of a firm's research and development expenditures and signif- icant inventions, Mansfield found in part that in petroleum and steel,greater rates of spending by larger firms were highly correlated with weighted indexes of new inventions.21 Here Mansfield assumes that n1 = d1 [a + bdi + 081] where n1 is defined as the weighted index of the innovation of firm 1. n1 is the firm's research and development expenditure, and S1 is a measure of size. "The partial correlation of of n i on 81, holding S1 constant is .96, .98, and .70 in chemicals, petroleum and steel."22 In the sample of twenty drug firms the coefficient was found to be .58 indicating a weaker re- lationship than that found in steel. Thus while Mansfield found a rather strong long run relationship between a firm's spending on research and the significance of its innovation, 20Subcommittee Report, pp. 155-164. 21Mansfield, Industrial Research and Development Expenditures, p. 336. 22Ibid. 221 this does not appear to be the case in drugs. (Note however, that this conclusion rests on the assumption that the dif- ferences between firms in 1958 is representative of their differences over time.) In part the relationship between research and significant innovations may be due to the nature of research in these industries. Petroleum and drugs Spend about the same proportion of their research budgets on basic research, but three times more than steel.23 In petroleum basic research appears to be along the lines of inventing around an established patent to produce a competitive techni- cal process.24 In drugs it appears that various approaches to a problem are followed until one stands out as most likely to succeed. Thus the nature of the attack indicates a high rate of failure for accepted projects. In steel many of the recent innovations appear to be products of the European steel industry and are adopted in the domestic industry with spend- ing on development rather than basic research. In regard to the relationship between the weighted in- dex of innovations and size of firm7with research spending held constant, Mansfield found the relationship negative in all three industries. In drugs it was positive with a value of .62. Thus while Mansfield was able to conclude that in- 23National Science Foundation, Review of Data on Science Resources, I, No. 1, 1964. The I960 figures for drugs were 17 per cent, steel 5 per cent, and petroleum 15 per cent. 2“John L. Enos, "Invention and Innovation in the Pet- roleum defining Industry," in National Bureau of Economic figsearch, The Rate and Direction of Inventive Activit (Princeton: Princeton University Press, 1962), pp. 299-321. 222 ventive output was lower in the larger sized firms, no such clear cut conclusion is possible in drugs. The best that can be said is that the relationship between size and the proauctivity of the research dollar is weak in drugs.25 The final question in this section is that of the re- lationship.between statistics of patents issued and the record of inventive output. Here we would expect that firms engag- ing in resca~ch and development would hold a greater number of patents as their research effort increased. Data on the numbers of patents in force as of 1955 has been compiled by the Subcommittee on Patents, Trademarks and Copyrights for firms holding 75 or more pat— ents.26 From this list it is possible to identify twelve drug producers together with the number of patents held as of 1955. Is there any relationship between the number of pat— ents held and the index of significant innovations? The simple linear correlation coefficient of .12 between these two series would indicate little or no relationship. Is there any relationship between the 1958 level of research spending and patent holdings? Again the correlation coefficient of .52 would indicate a weak relationship. Thus it would appear that 25‘I'he gross measure of correlation in drugs was .76, somewhat lower than steel at .87 or chemicals and petroleum at .99. In addition the relationship between size and re- search spending was .39, again indicating a rather weak re- lationship. LUU.S. Congress, Senate, Subcommittee on Patents, Trademarks and Copyrights, Distribution of Patents Issued to ggrporations (1939155), 84th Cong., 2d Sess., I937} 223 the connection between research spending, the productivity of that spending, and patents is in general weax. Summary and Conclusions.-—The analysis of technologi- cal change in the ethical drug industry indicates four con— clusions. First, while the total number of new drugs intro- duced between 1945 and 1959 has increased, an increasing pro- portion of these introductions have been of little signifi- cant value from the technological point of view. In real terms the industry has spent increasing amounts on research and development, but not to the extent that the average costs of new drug introductions have shown a significant increase. Thus it would appear that there exists an increasing emphasis on marheting the research product and combining this with sell- ing expenditures to insure an adequate level of profitability. Second, the twenty firms included in the industry sample accounted for 87 per cent of the 1958 industry spend- ing on research and development and for 75 per cent of all new drug introduction between 1995 and 1959. Spending on research and development was nearly proportional with size as measured by sales, but the index of research productivity indicated that diseconomies of scale were present. Third, that in terms of inventions small firms seem as productive of category I inventions as large firms, but that large firms are three times more productive of category III inventions. Fourth, the evidence seems consistent with the thesis that the opportunity for inventive activity has not been greatly 224 effected by the size distribution of firms, at least for the twenty firms included in the sample. That is, basically the product is technical and that a significant invention could still be marketed in spite of the generally large expenditures in selling. THE ANTIBIOTIC P300UCEJS Turning to the broad spectrum antibiotic producers and sellers, questions concerning their productivity can be raised. These firms held a monopoly position in the broad spectrum market either as the only producer or by virtue of the patent cross-license device. The basic question is the necessity for the patent monopoly to insure a flow of new drugs. That is, if the patent monopoly is necessary for techn010g- ical advance then these firms should have been more productive than their competitors. ngductivity in Research.--In Chapter V an estimate was made of the monopoly profits earned in the tetracycline market and the five producers were ranked in their order of sharing in these profits. Between these producers and the rest of the industry it does appear that they were innovators of more significant drugs. Table 25 compares the distribution of the 64 drugs innovated between 1945 and 1959 with those produced by these five producers. Here half of the new drugs innovated by these five firms fell in category I and II, while for the industry as a whole 42 per cent were so ranked. Thus it would appear that the existence of a monopoly was a neutral factor in research productivity. 225 TABLE 25.--Pcrcentage distribution of innovations for the five tetracycline sellers and the industry for 84 new drugs (1945—1959) Categories I II III Industry .14 .28 .58 Five Sellers .08 .uz .50 Summary and Conclusions.--The analysis of the position of the broad spectrum producers relative to the rest of the industry indicates that monopoly is not necessary for advances in technology. This, of course, does not argue that the broad spectrum producers would have had the same record of new drug introduction without patent monopolies. But it does argue that some firms were able to engage heavily in research and development and in a relative sense be more profitable with- out substantial monopoly positions. The conclusion casts doubt on the hypothesis that sustained monopolies are neces- sary for technological advancement. CHAPTEd VII CONCLUSIONS AND PUBLIC POLICY Public policy recommendations in the ethical drug in- dustry rest upon conclusions of past performance and the ex- pected effect of proposed changes in public policy upon future performance. It is necessary at this point to indicate the major conclusions of the analysis of the antibiotic segment of the industry and where possible to generalize these con- clusions to the rest of the industry. The plan of this chap- ter is then to discuss each of four performance variables together with indications of the type of public policy changes which could reasonably be expected to improve the observed performance. figsearch and Technological Pregress.--The focus of analysis has been upon technical progress as measured by the innovation of new drugs. The choice of focus was due basi- cally to two factors. First, in the innovation of new drugs the industry makes its greatest net contribution to social welfare. Second, to pick the innovation as opposed to the invention variable simply recognizes the fact that it is in this area that the industry's comparative advantage lies. That is, where the major breakthroughs in drug therapy are generally in the area of basic research and where in a profit orientated firm the probability of success of basic research 226 227 is quite low, one should not expect to find a high proportion of significant inventions. It would seem that the primary function of the industry in the matter of technical progress is to provide that necessary link between the inventor and the consumer. Thus it is in the area of development research-- direction and support of clinical testing, evaluation of test results, solution of manufacturing problems-~where the in- dustry's major contribution to technical progress lies. The Subcommittee in its analysis of technical prOgress found that there had been a decline in the number of signifi- cant new drugs introduced since the mid-1950's.1 The ex- planation was that the firms had allocated their scientific personnel to development and patent circumventing research rather than basic research. In this paper the analysis has been for the 1945-1959 period. The indications are that for the industry as a whole and for the antibiotic producers in particular, the rate of technical progress has not changed significantly in this period. The rate of introduction of category I and II drugs seems rather constant in this 15 year period. However, the increase in spending on research is associated with an increase in the introduction of category III drugs. Thus in real terms the industry seems to be Spend- ing about the same proportion of its resources on basic re- search in grants to medical schools and other research insti- tutions, but the increase in spending has been for projects lU.S. Congress, Senate, Subcommittee on Antitrust and Monopoly, Report, Administered Prices in Drugs, Senate Report No. 448, 87th Cong., lst Sess., 1961, p. 126. 228 designed to produce the near perfect substitute for the established drug. It was argued that this reaction was to be expected where the rationale for research spending are profits. The affect has been to increase product competition. But the experience in the broad spectrum market indicates that there is a limit to product competition. Since the introduction of tetracycline in 1953 no significant product competition has developed. The original four producers have not been challenged by other members of the industry. This may well follow from the interpretation of the Cournot model where with each instance of entry, price must be cut to insure that the new rival receives an acceptable share of the market. The alternative to the price cut with entry,would be for the rival to institute such a cut leading to the danger of un- restricted warfare. Thus it appears that spending on re- search to develop a substitute drug is profitable only if conditions of fewness exist in the market. With increasing numbers of sellers, price approached the competitive level thereby reducing the profit potential of the market relative to other markets. In the case of the broad spectrum market it was con- cluded that the key invention was the soil screening process. This method of inventing directly resulted in three of the major broad spectrums and indirectly for the fourth. The soil screening process was essentially a product of basic research, while the operation of this technique was defined 229 as development research. while this relationship between basic and applied research, and the location of the key activity cannot be generalized to the entire industry, it does give an indication of the direction in which public pol- icy ought to move. That is, lacking evidence on the opportu- nities to innovateJthe conclusion on the observed rate of technical progress is that it has been within an acceptable range. But here the question for public policy is can this range be improved upon? Adding the experience of penicillin to that of the broad spectrums, where it was found that the key technical contributions were made by the NRHL, indicates that the non- profit research organization is more productive of basic technical change than the laboratory of the commercial firms. Thus, if the non—profit research organization is generally a greater producer of the basis for technical progress then it ought to receive a larger share of the society's available resources. A second area in which public policy can affect the future rate of technical progress is to halt a tendency to- wards increased concentration in the industry. A rough count of merger activity among the largest twenty firms indicated at least twenty-seven horizontal mergers between 1951 and 1961.2 Lederle's parent, American Cyanamid, which according to 1958 sales data was the largest producer of pharmaceutical 2U.S. Congress, House, Select Committee on Small Business, Mergers and Superconcentration 87th Cong. (Washing- ton: Government Printing Office, 1962). 230 products, was also involved in the largest number of mergers at five. One of Cyanamid's acquisitions was Norwich Phar- maceutical which was among the largest twenty producers. In the analysis of the relationship between size and technical progress a rather weak relationship was found. In the analy- sis of seventy innovations by the twenty largest firms, the largest of these also produced a greater proportion of drugs of no significance. Thus to at least maintain the current rate of technical progress in the industry calls for the application of an anti-merger policy.3 Prices and Market Structures.--In the analysis of the narrow spectrum market3the structure of this market was found to have been towards the competitive end of the spectrum with prices reflecting production costs. In the broad spectrum market on the other hand prices for over a ten year period were unchanged, reflecting conditions of monopoly. The basic cause of this difference in market structures was the use of the patent. In the narrow spectrum market exclusive positions in penicillin were not possible as the basic drug was unpat- entable, while in the case of streptomycin the patent was held by a non-profit organization which granted unrestrictive licenses. The various penicillins that were developed later as well as dyhydrostreptomycin were patented, but had to 3For a similar conclusion see William S. Comanor, "Research and Technical Change in the Pharmaceutical Industry," The Review ofgcgnomics and Statistics, XLVII, May 1965, pp.7182-9O; and D. Hamberg, '‘Size of Firm, Oligolopy, and Re- search: The Evidence," Canadian Journal of Economics and Political Science, XXX, February 196E, pp. 62-65. 231 compete in the market with the unpatented substitutes. As a result, a bulk market developed with conditions of relatively free entry. In the case of the broad spectrumsjexclusive po- sitions were maintained with patents,and to protect these monopolies the bulk market was not utilized as a source of buyers. Application of a modified version of the Cournot duopoly model to the problem of entry and product competition gave a reasonable ex post fit to the available data. Here the function of the price leader was to determine the market price which would yield new rivals an adequate share of the expanded market as an alternative to entry and price cutting. There is in these conclusions essential agreement with the Subcommittee findings that the patent has been used to restrict competition and maintain non-competitive prices.4 But with regard to the international market insufficient evi- dence existed to support the thesis that a cartel was in op- eration. The performance of price in the broad spectrum market must be judged inadequate. The root of the problem is the patent. The potential effectiveness of public policy on pricing depends on how it can change the patent system. This discussion is delayed to the section on patents. Non-Price Competition.--The Subcommittee in its anal- ysis of product competition viewed the use of resources to develop substitute products as a diversion of resources from 232 basic research thereby lowering the discovery rate of signif- icant drugs.5 Based on the analysis of the antibiotic seg- ment of the industry it is agreed that basic research is the source of significant new drugs, but it is a mistake to assume that the industry ought to conduct this research. It has not in the past, the key contributions have come from the non- profit organizations. Now it may be possible to argue that the industry has not operated the available research techniques at their optimum rate,6but the industry rate of product in- novation for the past fifteen years does not in itself support this argument. It is, of course, true that the increased spending on research has tended to be directed towards the substitute drug. But it is also necessary to recognize that in the antibiotic market in three out of four cases this sort of product com- petition did result in a fall in price. In the absence of price conspiracy in tetracycline,the record undoubtedly would have been four out of four. In the Subcommittee's view the quality and quantity of the industry's selling efforts were excessive.7 While a 5&6 ort, pp. 126-37. 6In the case of the broad spectrum market it is pos- sible to argue that with the established monopoly among the four producers the incentive to operate the soil screening process further has declined. And that for the non-producers to attempt to enter now would lower price according to the version of the Cournot model employed earlier. But to sus- tain this argument one would have to know the opportunities available to innovate new antibiotic drugs from this process. 7Ibid., Chaps. IX and X. 233 detailed analysis of this point was not made it was noted that,over time’selling expenditures as a percentage of sales had increased. This would seem to be an inherent result of the decline in price competition and profitable operations. With the conspiracy in tetracycline,competition in selling continued unabated. Perhaps because of the technical nature of the product such efforts were mutually offsetting. The concern of public policy in the area of selling expenses can only be with the quality of information associa- ted with drug innovations. The increased control over this variable available to the Food and Drug Administration is at least a step in the right direction. Patents.--Conclusions on the effects of patents in the antibiotic drugs can be divided into three areas; re- search, innovations, and markets. With regard to research it was concluded that basic research isJin the final analysithhe source of technical pregress and that in the antibiotic segment of the industry the location of this activity is in the non-profit research organization. There is no evidence available that patents have any effect in allocating resources to these organizations, nor to effect the selection of projects for study. In the case of penicillin and streptomycin the research organizations involved both received patents on their contribution but granted unrestricted licenses to applicants. Thus entry and competition were unaffected by the existence of patent pro- tection. 23a In the case of development research the available evidence indicates that patents do have an effect in resource allocation but that the results are mixed. With the early penicillin program it was concluded in Chapter I that the effect was negative. Firms participating in the program allocated their resources to projects designed to produce a synthetic penicillin rather than to utilization of the pro- duction technique made available by the NRdL and the OSdD, partly in the hope of receiving patents. Here then the ef- fect on technical progress was retardative. On the other hand Lederle and Parke Davis' establishment of patent monOp- olies undoubtedly had an effect on Pfizer's and Bristol's research programs. The success of these programs in pro- ducing substitutes was perhaps neutral in effect on technical progress, but did initially increase competition in the broad spectrum market. From this evidence one could conclude that patents were a net positive or negative forcejdepending on the weight one gives to technical progress as opposed to product competition which results in price competition. With regard to the effect of patents on the industry's rate of innovation the evidence is again mixed. There is an increasing reliance upon patent monopolies but also an in- creasing degree of product competition in the form of cat- egory III drugs, and no apparent change in the innovation of more significant drugs. In the broad spectrum market it was found that being the innovator and establishing the brand name of the drug was of greater importance in market shares 235 than the fact that a patent monopoly existed on a particular drug. On the other hand, since tetracycline no significant broad spectrums have been marketed. Whether this is due to complacency arising from monopoly, to the relative unprofit- able level of price with entry, or to the fact that the soil screening process produced the ultimate with tetracycline, cannot be stated on the basis of the available evidence. Thus to the question of the effects of the patent on invention and innovation one is inclined to agree with the Subcommittee that the "mere existence of patent protection is not a guarantee of invention, nor its absence much of a barrier."8 Perhaps the primary effect of the patent lies in the commercial strategy employed by the various broad spectrum producers. The patent prevents immediate duplication forcing rivals to "invent around" the innovation and allows the in- novator time in which to build a lead in the market. With entry through product competition,prices are adjusted down- ward to make room for the rival in recognition of the mutual interdependency of the participants. With the entry of a sufficient number of sellers, either through competitive re- search or through a bulk market, the competitive price would be approximated. If this were the only effect of the patent it could be passed off as merely a lag in the adjustment pro- cess of a competitive system. But the question is more in- volved. The evidence of the tetracycline patent proceedings 81bid., p. 120. 236 indicate that the patent can become the vehicle for conspir- acy not only to secure a patent but the basis for agreement on price. The effect is then to delay further the adjustment process giving the participants time in which to become en- trenched in the market. The present antitrust laws are adequate to attack conspiracy based upon the patent grant.9 The problem here is that the attack is upon the symptom rather than the mechanism which facilitates the conspiracy. It is in this area where recent changes in public policy as well as certain proposed changes can be effective. The 1962 Drug Industry Act contained several features that will effect the future conduct of the industry. Section 2 provides that the Department of Health, Education and Wel- fare is vested with the authority to conduct any necessary research on drug patent applications at the request of the Commissioner of Patents. This provision gives the patent examiner access to independent opinion thereby removing the major absurdity in Patent Office proceedings where the exam- iner was forced to rely' upon the self-serving statements of the applicant. If this section had been in effect during the tetracycline proceedings.it is highly probable that this drug would not have been patented and with bulk suppliers such as Heyden and Bristol in the market the broad spectrum market would have repeated the experience of the penicillin market. 9In 1963 the Department of Justice brought suit in a Southern District Court of New York to revoke Pfizer's tetracycline patent. 237 Section 10 of this act provides the Secretary of HEW with the power to designate the official name of a new drug if such action would be "in the interest of useful- ness and simplicity."10 The effectiveness of this provision upon future competition depends largely on the physician's writing of prescriptions in terms of official names rather than brand names. To reduce the importance of the brand name could reduce the degree of entrenchment of the innovator and reduce the lag in the adjustment process with the appearance of product competition. Section 7 provides an amendment to the Federal Food, Drug, and Cosmetic Act which requires that the firm not only demonstrate that its drug is non-toxic but that it is effective as well. In combination with the requirement that the firm supply the FDA with more information in support of new drug applications, the effect will be to reduce the numbers of perfect substitutes that reach the market each year. This reduction in product competition may well be offset by keep- ing inferior drugs off the market. The Hefauver hearings produced three proposed changes in public policy. The first of these proposals provided for compulsory licensing of all drug patents after a three year period of exclusive manufacture and sale. The objective here was to allow smaller firms to obtain licenses at royalties not to exceed 8 per cent of the gross sales value, and thereby 10U.S. Congress, Senate, Committee of the Judiciary, Heport on the Drug Industry Act of 1962, 87th Cong., 2d Sess., 1962, p. 6. . JH‘.‘ 238 "inject price competition into the industry and thus break down the established structure of unreasonable and usually identical prices charged by the large companies."11 In fact, however, the exclusive period would be somewhat longer as the three year period would have begun with the issuance of the patent. Given the normal delays between filing of patent applications and issuance, the actual period of protection would run closer to five years. Under this proposed legis- lation the actual period of protection coincides with the industry's experience in the market. That is, on the average under the current patent law, a new drug is considered a good revenue producer for about five years before product competi- tion forces prices down. Thus the major effect of this type of legislation would have been to prevent long run monopoli- zation of the market as in tetracycline. In addition the firm contemplating directing its research to product competi- tion could achieve the same result by simply waiting the re- quired period and entering on the basis of a license. The advantages of this sort of proposal indicate that it would be a useful direction in which to move. A second feature of the Kefauver approach was the provision that molecular modifications and combinations of drugs would not be patentable unless the Secretary of HEW de- termined that the drug in question offered a significant therapeutic advantage.12 The purpose here was first an llIbid., p. 49. leearings, Drug Industry Antitrust Act., p. 28. 239 attempt to eliminate the type of substitute drug which led to confusion among members of the medical community, and second, to remove the rational for the large advertising ex- penditures that accompany such innovations. While the ob- jectives here are commendable the effect would be a tendency to eliminate a source of competition, and in connection with the first proposal, tend to reinforce the short run monopoly position. The case of the broad spectrums provided a case in point. Product competition in the form of oxytetracycline, chloramphenicol had the effect of reducing the broad spectrum price structure. Under this proposal it is doubtful if oxytetracycline or tetracycline would have been patentable. And if they had not been patentable it is doubtful that they would have been marketed for in doing so the entire broad spectrum market would have been thrown cpen to unrestricted entry. If the analysis of the tetracycline case is correct, then the development of such substitutes would be for purposes of forcing a license from the patent holder under threat of unrestricted entry. Thus,the net ef- fect of this proposal would be to reenforce the short run monopoly position of the patentee,along with that of the licensees,where their research was successful in providing a substitute. The third major feature of the Kefauver policy con- cerned the prosecution of the patent application. Here it was proposed to amend the Sherman Act making it an unlawful restraint of trade to; (l) the withdrawal of any pending 240 patent application, (2) to concede priority of invention to another party under an agreement to Split royalties, grant licenses at discriminatory royalties, or to grant licenses only to other applicants for the patent, (3) to refrain from granting any license under any drug patent.13 The effect of this prOposal would have meant the virtual elimination of patent protection on drug products. The first and second amendments eliminate private settlement of patent disputes, while the third would insure that entry via the license ap- proach would remain open. This, of course, is what might be termed the ideal solution to the problem of patents. It insures either immediate entry and direct product and price competition or forces the innovator to initially select a price close to production costs thereby making entry un- attractive. The fact that this proposal was not incorporated into the 1962 Drug Industry Act is a tribute to the industry's political power. lBIbid., pp. 20 and 28. Arrsbex TAan l.-—Percentage distribution of manufacturer's production of antibiotics (1950) ,_.| o 94 o o :3 S H .94 H I C (D a a a o o m m H4? 5a H H .‘h >: I 54 :1 06(1) SH ma H a as m p 3 am rh4 Qr4 r4 0 o<3 no) a)m a m>a (Ia-H «4H EH 43 5.4+) 43;: 49>: up up pirm me «so so a. on. ma HH 54 Ht: ma OH «4H cu >>Q> PH 0H 0 c6 4.): OS «is: $4 $39 >30 r—{o H 490) o m Lcm o m .p rhp x:» .c:» .c 05: mm cmu mm C) am 00 00 0 an Lederle 26.6 26.6 Pfizer 1.1 0.9 0.6 0.7 4.1 4.3 11.2 Merck 0.4 1.4 2.4 l.9 4.9 11.0 Lilly 5.0 5.2 0.5 0.2 0.1 11.0 Squibb 3.3 5.9 0.2 1.4 lQ.o Bristol 1.9 3.6 0.3 6.0 Parke Davis 5.9 5.9 Schenley 1.2 1.5 0.2 0.3 3.2 Heyden 0.9 0.6 0.2 0.2 0.9 2.0 Abbott 1.2 1.5 2.7 Commercial Solvents 1.1 0.7 0.1 2.5 “yeth led lob honsanto 1.7 1.? Vick 0.1 0.6 0.1 0.3 Upjohn 0.2 0.4 0.1 0.7 Cutter 0.2 0.5 0.7 99-9* Total 16.6 24.8 4.2 3.4 11.8 4.3 26.6 7.6 99.3 * The difference in the column and row totals is due to the ex- clusion of Commercial Solvents' Bacitracin which accounted for 0.6 per cent of the total market. Source: Computed from Federal Trade Commission, Egonomic deport on Antibiotic Manufacturer, Washington, 1958, pp. 94-95. 241 n?» J: '! 242 TABLE 2.--Percentage distribution of manufacturer's production of antibiotics (1956) ‘ .D..- ’--=-— .4 o 94 o c Q :2 m H H H a, a a a» 533 g g 2, E; I a sq m g. a UJH wqa th and 43 $4» +>Q ¥>£ o m m aio «ca 5:) Qa'dil v.4 aqa m a ram: m-r-I O-r-l q-l-r-l (D >50) «Pr-l OH H 0 C6 +>s <3: tin H .Cs4 mt) r40 .p r4 +30 c>m ac» o o .p .a+s N:a .szs m .2 cs: mm mm mm m am 00 oo 9 0 9p Lederle 0.2 11.2 16.5 26.1 Pfizer 0.4 2.1 0.2 0.6 2.3 12.0 4.1 23.4 Merck 0.8 1.5 0.2 0.5 1.8 7.2 Lilly 0.3 1.1 0.1 0.7 11.7 Squibb 1.3 1.9 0.2 0.8 1.1 6.? Bristol 0.1 0.6 3.6 4.3 Parke Davis 7.3 7,3 Schenley Heyden Abbott 0.3 1.1 Commercial o,u Solvents Wyeth 0.3 0.6 5.0 Monsanto Vick Upjohn 1.0 0.4 4.4 99.6% TOtalS Li's? 8.05 006 200 601 1200 1102 703 24.2 7606 *The difference in the column and row totals of 23.0 percentage points is accounted for by other antibiotic products. Source: Same as Table l. mmamnm panama ommpamoamm .H magma sown .HmmH .GOpwaacmmz .moaamm eaooom .mamm use acaposeoam .m.: mamoaaeso oauonucmm .soummaaaoo mmanme .m.D soak omma you mpmxama on» no swam Hence ”meadow mam.ea nsm.om oam.m nsm.nm mmm.s. mam.m oma.mn emm.nm mem.samx masses Hso.a mm: mes.a happen use was nwm.a excess saw nnm.a saw sasnm amwm a .m a o Opamm :0 Amman mmwam £p0h3 :HN mm:.a omm.m mmm.n mpsmpaom Heaoamaaoo mam.m osm.m mes.m sponge mmm.a mm: mmm.a wmm.a smm.n sesame new mam.m osm.m smm.o anaemeom 58.3% 30.3” mfirwa 33mm mam mma.m 050.: www.ma acumaam mmm.m mm: smo.ma meo.s sma.mm nnasam ad Hm mm: aso.a mma.aa oas.oa omn.mm saaaq mm as.oa oso.s asa.n mam.m awe aon.mm moses 03.3» ~36 a 33.; new; a was; a mmmJ a 80.3.. you“: 256nm. “amen a 3334 m. m m mm m as as as we as I T1 K mnu J u.v "do un+ Ann 0 o 1 £5 a 7;» To In as I J a O D. d O O 8 O S u T. s Q 4 II 4 In 7:» 13 n m a 1 "do 0 .L TLu +Lt. e d 4 s s m .L .L Ttm u. 1 o 1. nA I. I. T. e B .A I 0 U u u u o o 8 I. I. .A T. d U o o T: . O .L u T. To a u 9 Adoppdao ooo .ommav composeona :30 .mamaspomhsama Scam weapoanapam no case scam escapes ho sodusnaapmae empmsapmmul.m mqmda 244 '»II.I!.|II I! .m wanes mmm "meadow an Jams~ 338. sso.no «Hm.ma emm.om www.mm mnm.ma ems.n mmo.a soo.m~ nns.~a Hmppso ono.a sas.m asm.aa emcee: soa> cpammsoz mmo.a saw osm.ma guess emo.a mpnmpaom Hmaoacssoo saw mwm.m sponge seems: amaamzom maw.mam Nam.ma mabma mxnmm osa.m mmo.a Haw oso.aa Hopmawm ome.m asa.m man oma.n mmm.s mwa.ma padsum oom.a Ham nmm.m saw mms.~m eased mmw.s smm.a men aso.s asa.m osm.ma echo: $4.: SQNQ. «£6 emerge as; $0.3 82$ SSS 33.8 one 31% mam . one men a 3m .2. a @233 1w 0 O O 0 4G 3 d3 dd dd I e u. K W x O I. a... a O 9 J 8 O O 4 TL 0.1 c“ mnw I "up. any un+ 4 .a 0 To 1 £5 a TI to Is a a J I a e o D. d cm 0 e o s T. 8.2m o B nude a. 491 a» .5 7:1. 1:3 0A m 9 9 I u 0 O T. Tu T...» P o d a. a S m T. ?L rum 3 T. U. J O Q r... T. I. I. A I a B .A J o u u u e u u _ O 9 I. u a I. .l d u n O I. . 3 0 u T. e Rumppaao ooo .onmav acaposeoaa :30 .mamaspommsamz scam moanednapsm we came on» Sony esse>ma no scandaaapmae empmsapmm|1.3 wands 1i .mmdmma macaamb .GOpwsamwm3 .mmaamm esoomm .mamm dam soaposeoam .m.D .mamoaSmso oasmmao oaumcpamm .QOammaaSoo muaame .m.D "meadow .measoa ho msamp ca measmam soaposdoaa heave .mpaaa euomwo no mcoaaaaa go magma Ca ma saaaaoaamm no esam> pad: can acaposuonms ao.asm smm am.asm _ sea nn.aam awn aa.maa so: am.amam, omm 245 a a w i W oama mama mama mama mama aama sama mama mama aama oama mama mama ssma mama asma meam> pas: acaposeoam mm mam mm one am aaa a: Nma ms was as man no as: as aom maa mam :aa aaa asa maa mam a mma de6> DHSD SoapoddOHm am awn mm osm am sea on maa an sea ma mm as as 30a 30a saa a: maa an aoa a: mam sa sea as smm.a mm aaa.ma a 03Hd> pHQD QOHPOSdOHm waa oma.wms saa amw.ams maa asa.ama asa nas.ama msa mam.wss oaa asN.ssm sea maa.aks aaa awa.asa mam ama.msa msa www.mam moa mom.mam was moa.wma ooaa a astm ooo.ms ooo.sm sas.s msam> pad: soapoSdoam oaaaoaomapoa saomEOpamapmoaemSaQ saoasopamapm *Qaaaaoanmm " II ‘I Aoamauasmaa moapoanapee weeps aeoaeoa ho osaes page see soaposeosmnu.a mamas 246 TABLE 6.-—Percentage Market Shares for Tetracycline* (1954-1958) 1954 1955 1956 1957 1958 Lederle 76.3 59.7 55.5 53.1 45.3 Pfizer 14.9 14.3 17.3 15.1 21.4 Bristol .8 1.5 2.8 4.9 5.2 Squibb 2.2 9.0 13.5 15.3 12.7 Upjohn 5.8 15.5 10.9 11.6 15.4 * Bristol indicated that these figures were derived from in camera documents. Source: Bristol's Answering Brief, In the Matter of American Cyanamid Company, et al., Docket No. 7211, New York I962, Appendix, p. 16. 247 TABLE 7.--Pharmaceutica1 markets and percentage distribution by products. 1954 and 1958 Major Markets 1954 Percent 1958 Percent Total Pharmaceutical Preparations for Human Use .31,498,918 100.0 82,256,914 100.0 Antibiotics, broad and medium range .. 252,892 16.9 361,941 16.0 Vitamins ............ 215,074 14.3 300,659 13.3 Analgesics, including narcotics ......... 194.262 13.0 256,319 11.4 Tranquilizers, seda- tives and hypnotics 42,684 2.8 236,802 10.5 Cough and cold prepar- ations 00000000.... 97,025 605 171,580 706 Laxatives and stomach preparations ...... 123,368 8.2 158,902 7.0 Hormones ............ 103,532 6.9 151,354 6.7 Antiseptics ......... 45,863 3.1 58,905 2.6 Hemantinecs ......... 32,548 2.2 39,094 1.7 Sulfonamides ........ 23,691 1.6 36,447 1.6 Anti-obesity prepara- tions ............. 21,289 1.4 29,852 1.3 Tonics and altratives 23,859 1.6 19,890 .9 Anesthetics ......... 15,111 1.0 15,620 .7 Diagnostic aids ..... n.a. 9,343 .4 Central nervous system stimulants ........ 2,000 .1 6,024 .3 Other pharmaceutical preparations ...... 305,556 20.4 404,182 18.0 Source: U.S. Department of Commerce, 1958 Census of Manu- facturers, Vol. II, Industry Statistics, Part 1, General Survey and Major Groups 20 to 28, BIBLIOGRAPHY PUBLIC DOCUhEkTS Federal Trade Commission. Economic Report on Antibiotic Manu- facturer. Washington: Government Printing Office, 1956. Federal Trade Commission. Initial Decision. In the Matter of American Cyanamid Company, et al., Before the Federal Trade Commission, Docket No. 7211. Washington: 1961. Federal Trade Commission. Findings as to the Facts and Conclp- sggns of Law. In the Matter of American Cyanamid Company, et’al., Before the Federal Trade Commission, Docket No. 7211. Washington: 1962. Federal Trade Commission. Opinion of the Commission. In the Matter of American Cyanamid Company, et al., Before the Federal Trade Commission, Docket No. 7211. Washington: 1963. Federal Trade Commission. Final Order. In the Matter of Amer- ican Cyanamid Company, et al., Before the Federal Trade Commission, Docket No. 7211. Washington: 1963. Federal Trade Commission. Opinion Accompanying Final Ordgg. In the Matter of American Cyanamid Company, et al., Before the Federal Trade Commission, Docket No. 7211. Washington: 1963. Machlup, Fritz. An Economic Review of the Patent System. Study No. 15 of the Subcommittee on Patents, Trademarks and Copyrights, Senate. Washington: Government Print- ing Office, 1958. helman, Seymour. The Impact of the Patent System on Research. Study No. 11 of the Subcommittee on Patents, Trademarks and Copyrights, Senate. Washington: Government Print- ing Office, 1958. Hestrictive Trade Practices Commission. Report Concerning the Manufacture, Distribution and Sale of Drugs. Depart- ment of Justice. Ottawa: 1963. United Nations. Commercial Statistics, Medicinals and Pharma- ceuticals. Series D, V01. XI. New York: United Nations, 1962. Z48 U. U.S. U U. U U. U U. U. S. .S. S. .S. S. .S. S. S. U.S. U.S. 249 Attorney General. Attorney General's Report of the National Committee to Study the Antitrust Laws. Washington: Government Printing Office, 1955. Department of Commerce, Government Patent Board, Chemical Products and Process Patents. Government Owned Inven- tions Available for License. Patent Atsttact Seriesi No. 2. Washington: 1954. National Science Foundation, Review of Data on Research and Development. Research and Development in American Industr , 1962, No. 40. Washington: 1963. National Science Foundation, Review of Data on Research and Development. Research and Development in the _ Chemical:;ndustry, 1956-61, No. 42. Washington: 1963. Patent Office. Official Gazette. Vol. DCXXVII, 1949. Senate, Subcommittee on Antitrust and Monopoly. Hearings on Administered Epices in the Dru Industr . Parts 14 through 26, 86th Cong., lst Sess., 1960. Senate. Report on Administered Prices, Drugs. 87th Cong., lst Sess.,—1961. Senate. Hearings on the Drug Industry Antitrust got. Parts 1 through 7, 87th Cong., st Sess., 1961. Senate. Report on S. 15523 Drug IndustrygAntitrust Act of 1962. 7 Cong., Sess., I962T' Senate, Committee on the Judiciary. Report together with individual Views to Accompany S. 1552, DruggIndustry Act of—1262. 287th Cong., 2d Sess., 1962. Senate, Subcommittee on Patents, Cepyrights and Trademarks. Distribution of Patents Issued to Corporations 1939455. 84th Cong., 2d Sess., 1957. Senate. Com ulsor Licensin of Patents--A Legislative History. 85th Cong., 2d Sess., 1958. Senate. The Patent System: Its Economic and Social Basis. 86th Cong., 2d Sess., 1960. Senate. Mana ement Surve of the United States Patent Office. 87th Cong., 2d Sess., 1962. Senate, Subcommittee of the Committee on Military Affairs. Hearings on Science Le islation, Part 1. 79th Cong., Sess., 1945. 2.30 U.S. Senate, Temporary National Lconomic Committee. Patents and Free Snterprise. 73th Cong., 3d Sess., 1941. U.S. Tariff Commission. Synthetic Clemicals , U. 8. Production and Sale, Second Series. Washing ton. BOOKS Alexander, Harry L. Reactions With Drug Therapy. Philadel- phia: Saunders, 1955. Beckman, Harry. The Year Book of Drug_Therapy. Chicago: Year Book PuBIishers, 196l} Cattel, Jacques (ed.). American Men of Science. Lancaster: Science Press, l9h9. Chamberlain, Edward H. The Theory_of Monopolistic Competition. 8th ed. Cambridge: Harvard University Press, 1962. Dirlam, Joel B., and Kahn, Alfred E. Fair Competition, the Law and Economics of Antitrust Policy. Ithaca: Cornell University Press, 1953; Edwards, Corwin. Maintaining Competition, R uisites of a Government Policy. New York: McGraw-Hill Book Co., 1949. Fellner, William. Competition Among the Few. New York: Augustus H. Kelley, 1960. Fleming, A. (ed.) Penicillin, Its Practical Applications. 2d ed. London: Butterworth & Co., Ltd., 1950. Florey, H. W., et a1. Antibiotics. Vol. II. London: Oxford University Press, 19D9. Folsom, Paul J. (ed.) Physicians' Desk Reference. Oradell, N.J.: Medical Economics,—Tnc., 1962. Irving, George W., and Herrick, H. T. (eds.) Antibiotics. Brooklyn: Chemical Publishing Co., 19H9. Jenkins, G. L., Hartung, W. H., Hamlin, K. E. Jr., and Data, J. B. The Chemistry of Organic Medicinal Products. New York: Wiley, 1957. Jewkes, John, Sawers, David, and Stillerman, Richard. Egg Source of Inventions. New York: St. Martin's Press, 1961. 251 Hahn, Alfred E. The Role of Patents, in J. P. Miller (ed.), Competition Cartels and Their Regulation. Amsterdam: North Holland Publishing Company, 1962. Kaplan, A. D. H., Dirlam, Joel B., and Lanzillotti, Robert. Pricing_in Big,Business. Washington: Brookings Institute, 1958. Kaysen, Carl, and Turner, Donald F. Antitrust Pgligy,MAn Economic and Legal Analysis. Cambridge: Harvard University Press, 1959. Krug, Elsie E. Pharmacology in Nursing. 9th ed. St. Louis: C. V. Mosby Co., 1963. Piahoney, Tom. The Merchants of Life, An Account of the American Pharmaceutical Industry. New York: Harper & Bros., l959~ National Bureau of Economic Research. Eye Rate and Direction of Economic Activity. Princeton: Princeton University Press, 1962. Penrose, Edith T. The Economics of the International Patent System. Baltimore: John H0pkins Press, 1962. Stewart, Irvin. Organizing Scientific Research for War,_The Administrative Histor of the Office of Scientificw Research and Development} Boston: Little &I Brown, 1948. Stocking, George W., and Watkins, Myron W. Cartels in Action. New York: Twentieth Century Fund, 1957. . Cartels or Competition. New York: Twentieth Century Fund, 19E8. Strassmann, W. Paul. Risk and Technolggical Innovation, Amer- ican Manufacturing Methtds During the Nineteenth Century. Ithaca: Cornell University Press, 1959. Waksman, Selman A. Neomycin. New Brunswick: Rutgers University Press, 1953. 252 Anrchss AND PERIODICALS Barker, Benton. "Patent Rights and the Antitrust Laws," Journal of the Patent Office Socie_y. Vol. XXXIV, No. 9. Washington: 1952. Borkin, Joseph. "The Patent Infringement Suit--Ordeal by Trial," The University of Chicago Law Review. Vol. XVII. No. R (1950). Cacciapaglia, Frank, Jr., and Rockman, Howard B. "The Proposed Drug Industry Antitrust Act--Patents, Pricing and the Public," The George WashingtongLaw Review. Vol. XXX, No. 5 (1962). Comanor, William S. "Research and Technical Change in the Pharmaceutical Industry," The Review of Economics and Statistics. Vol. XLVII, No. 2 (1965). Florey, H. W., and Abraham, E.P. "The Work on Penicillin at Oxford," Journal of theflgstory of Medicine and Allied Sciences. Vol. VI. No. 3 (1951). Fortune. "Shock Treatment for Parke Davis." Vol. XLVIII (September, 1953). Gilfillan, S. C. "Who Invented It?", The Scientific Monthly. Vol. XXV. (1927). Gomory, Andrew B. "Research, Development and the Patent Right," qurnal of the Patent Office Society. Vol. XXIX, No.7. Washington: l9h7. Goodman, Richard. "The Law and Monopoly: the Case of Tetra- cycline," New University Thought. Vol. III, No. 4 (1963)- Gort, Michael. "Analysis of Stability and Change in Market Shares," Journal of Political Economy. Vol. LXXI, No. l (1963). Griliches, Zvi. "Research Coats and Social Returns: Hybrid Corn and Related Innovations," Journal of Political Economy. Vol. LXVI, No. 5 (1958). Gunther, John. "Inside Pfizer," New York Times, Sec. 10, Hamberg, D. "Invention in the Industrial Research Laboratory." Journal of Political Economy. Vol. LXXI, No. 2 (1963). "Size of Firm, Oligolopy, and Research: The Evidence," Canadian Journal of Economics and Political Science." Vol. XXX, No.'l (196E). 253 Harris, Richard. ”Annals of Legislation, The Real Voice,“ The How Yorker, Karon 19, 21, and 25, 1964. Lalande, L. “A Chemist's View on Patents," Journal of the Patent Office Society. Vol. XLI, ho. E. Washington: 1959- Levine, Robert. "The Shrunken Patent Domain in the Expanded Antitrust Universe," Journal of the Patent Office Society. Vol. XXXIV, No. 6. Washington: 1952. Lutz, Karl. "Are the Courts Carrying Out Constitutional Public Policy on Patents?", Journal of the Patent Office Society. Vol. XXXIV, No. 10. Washington: 1952. Mansfield, Edwin. "Size of Firm, Market Structure, and In- novation," Journal of Political Economy. Vol. LXXI, . “Industrial Research and Development Expenditures Determinants, Prospects, and Relation to Size of Firm and Inventive Output," Journal of Political Economy. Vol. LXXII, No. 4 (1964). Nelson, Richard R. "The Simple Economics of Basic Scientific Research," Journal of Political Economy. Vol. LXVII, N0. 3 (19597: Patterson, Robert J. "Inventions in the Chemical Field," Journal of the Patent Qfifice Society. Vol. XXXVII, No. 1. Washington: 1955. Robbins, Leonard J. "Pharmaceutical Patents in Foreign Countries," Journal of the Patent Office Society. Vol. XXXVII, No. 37' Washington: 1955. . "Patents for Microbiological Transformation-—An International Problem," Journal of the Patent Office Society. Vol. XLII, No. 12. Washington: 1960. Silberman, Charles E. "Drugs: The Pace is Getting Furious," Fortune, May 1960. Spencer, R., and Chereau, L. "France Decides to Grant Patents for Pharmaceutical Products," Journal of the Patent OEfice Society. Vol. XLII, No. R. Washington: 1960. Steel, Henry. "Monopoly and Competition in the Ethical Drug Kagket," The Journal of Law and Economics. October, 19 2. Woodhams, Robert E. ”A Comparative Study of Patent Laws in the Americas and the International Effects Thereof," Journal of the Patent Office Society. Vol. XXXI, No. 10. Washington: '19Q9) 25A REPORTS Arthur D. Little, Inc. A Report on the Aspects of Concentration and Product Obsolescence in the Pharmaceutical Industry in the United States. Cambridge: 1961. Pharmaceutical Manufacturers Association. Prescription Drpg Industry Fact Book. Washington: 1962. BRIEFS Connolly, Arthur G. Pfizer's Answering Brief in Sppport of the Hearing Examiner's Conclusion that its Tetracycline Patent is Valid. Before the Federal Trade Commission, In the hatter of American Cyanamid Company, et al., Parts 1 and 2, Docket No. 7211. New York: I962. . Pfizer' 8 Brief in Support of its Tetracycline Patent. Federal Federal Federal Federal Federal Federal Before the Federal Trade Commission, In the hatter of American Cyanamid Company, et al., Docket No. 7211. New York: 1960. Trade Commission. PrOposed Findings of Fact, Conclusions of Fact and Law and Order to Cease and Desist. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7le. Washington: 1960. Trade Commission. Reply to Respondents' Proposed_ Find- ings1 Conclusions and Briefs in Sppport Thereof. Before the Federal Trade Commission, In the Iflatter of'American Cyanamid Company, et al., Docket No. 7211. Washington: l960. Trade Commission. Supplemental Reply Brief of Council Supporting_the Complaint. Before the Federal Trade Com- mission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. Washington: 1961. Trade Commission. Appeal Brief. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. Washington: 1962. Trade Commission. Reply Brief. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. Washington: 1962. Trade Commission. Supplemental Brief Pointinngut In- accuracies in Res ondents' Ar uments. Before the Federal Trade Commission, In %Ee Fatter of American gyzgamid Company, et al., Docket No. 7211. Washington: 255 Gesell, Gerhard A. Brief and Iroposed Finding s ardtpggclgitgg: of the Upjohn Company. Before the Federal Traic Con- mission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. Washington: 1960. . Answering_ Brief of the Upjohn Company. Before the Federal Trade Commission, In the hatter of American Cyanamid Company, et al., Docket No. 7211. Washington 1962. Irvine, Ralstone R. Respondent American Cyanamid Company's Proposed Findings of Factpronclusions of Law and Proposed Ordery_With Reasons in Suppprt Thereof. BefOLe the Federal Trade Commission, In the matter of American Cyanamid Company, et al., Vols. I, II, and Iv, Docket No. 7211. New York: 1960. . Notion of Respondent American Cyanamid Company for Acceptance of AnsweringBrief'for Consideration by the Hearing Qxa miner. Before the Federal Trade CommisSion, In the matter "of American Cyanamid Company, et al., Doclcet No. 7211. New York: 1961. . Answerinngrief of ReSpondent American Cyanamid Company. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. New York: 1962. .Supplemental Brief of Reapondent American Cyanamid Company in Answer to the Reply Brief of Counsel 3gp- pprtino the Complaint. Before the Federal Trade Com- mission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. New York: 1962. Klots, Allen T. Bristol's Proposed Findings of Fact, Conclu- sions of Law and Answers to the Government's Contentions. Before the Federal Trade Commission, In the hatter of American Cyanamid Company, et al., Vols. I, II, and Iv, Docket No. 7211.> New York: 1960. . Bristol's AngwertggBrief. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. New York: 1962. . Bristol's Reply Nemorandtm. Before the Federal Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. New York: 1962. Naulsoy, Allen F. Proposed Findings and Related Papers of Respondents Olin hathieson Chemical Corporati_on. Before t1e FederalJ Trade Commission, In the Matter of American Cyanamid Company, et al., Docket No. 7211. New York: 1960. maulsby, Allen F. Answering grief of Respondent Equipp. Before the Federal Trade Commission, In the hatter of American Cyanamid Company, et al., Docket No. 721l. New York: 1962. -n‘ J.- leg—an.- i! "11111111111'111111111“ 46 8403