FIGURE 6.1
Interval between the introduction of an innovation and competitive entry, 1887–86. Source: Agarwal and Gort (2001).
Voluntary Dissemination of Proprietary Technology: Private Profit, Social Gain
Conventional business wisdom says: Never let the competition know what you’re doing. But at Novell, we believe the secret of success is to share your secrets. So we established the Novell Labs program to openly share our networking software technology with other companies.
—Advertisement by Novell, Inc.,
in The Economist magazine, 21 September 1991
Pretty soon, if it continues, you’ll find that everyone’s going to have rights to everyone else’s technology, so there’s not going to be any competition.
—Julie Mar-Spinola, as quoted in the New York Times,
11 November 2001, section 3, p. 7
The previous chapter dealt with the market’s offsets to one major impediment to productivity and growth—the diversion of entrepreneurial effort and ability to unproductive activities, such as rent-seeking. In this chapter, I will examine the way in which the free-enterprise mechanism alleviates another major impediment to growth. Because firms gain competitive advantage from their possession of products and processes that are unavailable to rivals, we might expect them to do whatever they can to impede or prevent the spread of this proprietary technology. However, if every firm is denied access to the innovations that are currently employed by others, each will be condemned to activities that are at least partially obsolete, thus putting a brake on economic growth. Here I will argue that in fact competition and the pursuit of profits drive many firms to do the opposite: actually to disseminate their proprietary technology, providing it voluntarily, even to their arch rivals. Of course, they do so only if the reward is sufficient. This chapter and the next will explore the general principles underlying the resulting markets in proprietary technology and I shall offer a considerable amount of (unsystematic) evidence indicating that such technology supply occurs frequently in reality.
THE EFFICIENCY CONTRIBUTION OF RAPID DISSEMINATION
This rapid dissemination is no minor matter for the efficiency of the economy’s growth process. In his valuable article on international trade and world distribution of income, Paul Krugman (1979) takes the patterns of trade and economic growth to be governed primarily by two activities: innovation and technology transfer. The discussion in this chapter takes a similar position. It is clear that the growth of developing nations is highly dependent on the success and speed with which they can acquire and put to effective use new technology from the industrialized countries. But even growth in the industrialized economies is highly dependent on the effectiveness with which they can adopt new technology from sources foreign and domestic. That they do so to a substantial degree is surely indicated by the remarkable similarity of the technology employed throughout the industrialized world. Moreover, the vast majority of the industrial economies must rely preponderantly on technology created outside their borders because very few of them are substantial innovators themselves. For example, according to the National Science Board (2000, p. 2–40), “[t]he worldwide distribution of R&D performance is concentrated in relatively few industrialized countries. Of the $500 billion in estimated 1997 R&D expenditures for the 28 OECD [Organisation for Economic Cooperation and Development] countries, 85 percent is expended in just 7 countries [the United States, Canada, France, Germany, Italy, Japan, and the United Kingdom].” And, of these seven top R&D performers, only Japan, the United States, and Germany awarded less than 70 percent of their 1996 patents to residents of other countries (World Intellectual Property Organization). This should not be very surprising—if thirty countries all engage in invention activity, and if technology utilization is very similar in all of them, then the “average country” must be receiving 29/30ths of its innovative products and practices from abroad.
Unless technology transfer is a significant disincentive to investment in innovation, it should also be obvious that growth is promoted by the rapidity with which new technology is disseminated. The greater the share of firms and countries that make use of superior products and processes and the sooner they do so, rather than being confined to inferior substitutes, the more widespread and substantial the output and growth benefits should be. For suppose only one firm uses a new technique that increases productivity, while its competitors are forced to retain the old, less efficient technique. Then it is obvious that either output must be far lower, or (unnecessary) input use far greater, than if all of the firms engaged in the activity were able to adopt the productivity-enhancing technique.
We must, therefore, consider the market’s technology transfer process in this new world in which much of innovation has become routine. Just as competitive influences have led to routinization of innovation, there are competitive mechanisms that can make dissemination of technology a part of the regular portion of the firm’s voluntary activities. This is in direct contrast with a common view of firms as zealous guardians of the proprietary innovations in their possession, using patents, the courts, and/or secrecy to keep their technical knowledge from others for as long as possible. They supposedly do this in order to prolong the length of time during which their technological information gives them a competitive advantage and brings with it the stream of supercompetitive Schumpeterian profits.
I will show, on the contrary, that market forces frequently motivate enterprises to become active sellers of licenses for the use of their proprietary technology, or to make a variety of information-exchange arrangements ranging from implicit contracts to carefully spelled-out legal commitments. This phenomenon plays a part that is not insignificant in the continuing growth performance of the free-enterprise economies.
THE CONFLICT BETWEEN INNOVATION AND RAPID DISSEMINATION
As already suggested here, innovation and quick dissemination are two of the critical stimuli to economic growth. Each helps to increase productivity and product quality and, if one or the other were to disappear, growth would surely slow to a crawl. Yet it is widely recognized that dissemination can be the enemy of innovative activity. If a firm undertakes considerable expenditure of money and effort to carry out its innovation program, but finds that other firms, including its competitors, rapidly share in the fruits, why should that firm devote the time, effort, and funding to continue that program?
The problem is, indeed, a real one, and the amount of litigation over patent infringement indicates that it is no minor matter. Yet I will show here and in later chapters that the problem is considerably less serious than seems widely to be believed. It is relieved to a considerable degree by steps taken by many of the firms in the forefront of innovative activity to promote dissemination rather than seeking to prevent it. These firms do so voluntarily, as a source of both profit and improvement of competitive position. The result is that the market mechanism succeeds in providing a partial reconciliation of the conflict between innovation and dissemination, adding yet another powerful growth stimulus.
FIGURE 6.1
Interval between the introduction of an innovation and competitive entry, 1887–86. Source: Agarwal and Gort (2001).
Before getting to the reasons that lead enterprises to engage in dissemination of their own proprietary technology, it is worth pausing to examine some suggestive data. The evidence indicates not only that dissemination is surprisingly rapid, but that it has been growing more so with remarkable consistency for more than a century. A recent study by Rajshree Agarwal and Michael Gort (2001) examines a set of forty-six major product innovations, which they admit is not a random sample and perhaps not representative (though the list was chosen for another study and hence is not deliberately biased in relation to the current subject). They find that, in the course of a century, the average time between the commercial introduction of a new product and the entry of competitors supplying the same or similar products fell from 32.75 years at the inception of the twentieth century to 3.40 years in 1967–86. Moreover, as shown in figure 6.1, the decline was remarkably steady and persistent, and the authors report that other studies support their results.
The first suggestive implication of these striking data is that dissemination speed cannot plausibly be taken to be an accidental affair. There must be something systematic driving the trend, and in a market economy the natural candidate is profitability—as my analysis will suggest on other grounds. Second, at worst, this speedup of availability of new product technology to competitors has not brought the growth of investment in innovation by private industry to a halt. Figure 6.2 shows real investment in innovation by firms in the postwar period, and we see that the expansion of innovation investment has apparently been continuing unabated. In other words, although rapid dissemination may well constitute some disincentive for investment, the market mechanism has nevertheless provided incentives that at least do not make the one activity fatal for the other.
FIGURE 6.2
Real U.S. private R&D expenditures, 1953–98. Expenditures are deflated using the GDP implicit price deflator. Sources: National Science Board (2000), and Economic Report of the President (2001).
TWO INCENTIVE MECHANISMS FOR VOLUNTARY DISSEMINATION
Let us begin with a discussion of the incentive mechanism of markets in technology licenses. The logic of such markets is a straightforward matter. A firm that holds the rights to a valuable invention is, formally, in the same position as an enterprise that owns a facility with abundant capacity and superior quality. Either firm can use the property in question as an input to its own production process or it can choose to rent access to the input to other producers (who may be competitors), or it can do both. Which of these three options is most profitable depends on the price other prospective users of the input are willing to pay. If that price promises profits higher than the firm can obtain by using the input itself, then it will clearly pay that firm to transfer the input (or its use) to those customers. And a moment’s consideration confirms that the other firms will be willing to pay this more remunerative price for the technology license (or for the superior raw material) if and only if they are more efficient users of the item than its proprietor is. For the superior efficiency of the renters of the technology means that they can earn more from its use than the owner of the technology can. So they can still earn a profit if they pay a license fee somewhat greater than the owner would be able to earn by using the technology itself. That, in essence, is the logic of the market mechanism’s incentive for voluntary sale of technology licenses, as will presently be described more fully. Note that it tends to leave the use of an innovation to those firms that can employ it most effectively: its most efficient users.
There is another element to the explanation of the willingness of owners of proprietary technology to sell the rights to its use to others, and for others to be willing to pay the price. The purchasers of such licenses are willing to do so because such “friendly transfer” of technology is substantially more rapid than “hostile transfer,” which occurs through means such as industrial espionage and reverse engineering. Because rapid obsolescence makes speed in providing the latest model so critical in high-tech industries, it becomes worthwhile for the licensee simply to pay the price rather than expend the time and effort to “create” the technology itself.
But the licensor also receives a degree of protection in the process because even friendly transfer takes some time. A year or two can pass by the time one firm has acquired another’s new technology, learned to use it, installed the necessary equipment, and organized the requisite marketing. In high-tech industries that is no minor delay. In the extreme case of computers, models are often modified within six months of their introduction. Indeed, I am informed that, frequently, on the day the new model actually appears in retail stores it is no longer being manufactured! This means that a year or two of lag in the technology-transfer process still preserves a significant timing advantage for the innovating firm. In other words, by providing rivals with its latest products and processes, the innovating firm generally does not sacrifice all the advantage that priority in possession of the innovation can confer. But, clearly, all this takes the romance and mystery right out of the picture we have of wily, competitive firms!
The next chapter focuses on a second mechanism that leads to dissemination of technology. It provides a theoretical model that will enable us to analyze the benefits offered to a firm by membership in what I will call a “technology-sharing consortium.” The model demonstrates that the market may impose severe penalties upon any firm that remains outside such a consortium and does not share its technical information with others. That isolated firm will be able to offer products and use processes that are improved only by its own research efforts, whereas its rivals will each benefit from their combined innovative activities. Thus, exchange of technology (in contrast to pure licensing), rather than increasing the vulnerability of the participating firms to successful “enemy attack,” provides them with a degree of protection. Because each firm is given access to the others’ innovations, they are insured against the possibility that one of them, or still another enterprise, will come up with an innovation that will give its proprietor overwhelming competitive superiority. In addition, despite the lags inherent in the transfer process, each firm’s management can rest somewhat easier in the knowledge that it will not be denied access to the possibly superior technological developments of its rivals.
The model of the next chapter will also show that society derives welfare benefits from such sharing. This is partly because licensing or exchange of technological information can help to internalize the externalities of innovation activity, thereby reducing the spillovers problem. It will also be shown in the next chapter that technology-exchange consortia can be quite stable if the sharing game is repeated many times.
In sum, this chapter and the next argue that the dissemination of technology under free enterprise is considerably more pervasive and effective than the standard picture implies. It would, of course, be absurd to claim that no innovator firm imposes obstacles to the use of its proprietary information by others. Nor is the discussion intended to assert that the competitive mechanism dependably drives the economy close to optimality in its speed of transfer of technology. Rather, the economy’s performance in this area, like its performance in investment in innovation activity, is probably substantially better than may previously have been suspected.
REMOVING THE MYSTIQUE: PROPRIETARY TECHNOLOGY IS JUST ANOTHER “BOTTLENECK INPUT”
To understand what goes on in the technology-licensing markets, it is necessary to free oneself of the preconception that firms invariably have a powerful incentive to battle determinedly against use of their proprietary technology by others. Under capitalism, whether or not that is true in any particular case depends on the profitability of that retention by the firm, as compared with the profitability of supplying the technology to others.
Innovation is imbued with a mystic aura that has little basis in the logic of the objectives and behavior of firms. Innovation, of course, brings us extraordinary new products and new ways of doing things—it entertains us, makes us more productive, cures what ails us—and so, appropriately, we may view its creators with awe. But, in order to pick apart how innovation figures in the workings of capitalism, we must strip it to its essence: at heart, novel technology is simply another (durable) input to the production process, one that permits better products to be produced or that enables better processes to be used. The proprietary character of the technology means simply that this input can be obtained only from a monopolist supplier. Thus, from the viewpoint of the firm’s decision-making and of the market for the technology, such information is no more and no less than what, in other areas, is referred to as a “bottleneck input.” We can liken it to a mountain pass owned by a single railroad that other railroads, which want to compete on partially parallel routes, will need to use as an input for their own transportation service.
A railroad in the position of the owner of this bottleneck mountain pass then has three options. It can deny access to all other railroads; it can admit other railroads for a rental fee (or in exchange for services by the other railroads) and also continue to use the pass for its own traffic; or it can close down its own operations along the route and simply earn revenue as a collector of rental fees. One cannot know, without more information, which of these three is the most profitable option for the proprietor of the pass. That depends on the size of the fee, the volume of traffic of the other railroads, its own costs, its own traffic volume, and the prices it is able to charge for transportation along the entire route. In such circumstances, a firm will often find it profitable to sell access to its bottleneck facility, even to competitors that will cut into the proprietor’s own business.
The same is plainly true of the owner of a patent on a valuable innovation. The identical three options apply to this durable input, and if the market offers a license fee that is sufficiently lucrative it will pay the proprietor to go actively into the business of supplying that input. In the next section (and the next chapter), we will see that this often happens, with proprietary technology exchanged or licensed, and sometimes actively marketed.1 And the result can be socially beneficial, actually stimulating innovation by internalizing some of its externalities and by facilitating rapid and widespread adoption of improved technology. Once it is recognized that proprietary technology is just another bottleneck input, it becomes clear that there should be nothing surprising about this phenomenon.
MARKETS IN TECHNOLOGY LICENSES: IMPROVING THE TERMS OF THE TRADEOFF
All of this shows how the market helps to ameliorate the tradeoff between the incentive for expenditure on innovation and the rapidity of transfer to others of the products of that expenditure. The market does so by inducing the users of the innovations to make payments to the owners of the technology. Whether these payments take the form of money or of compensation in kind (by granting the licensor access to other inventions in exchange), they cannot be mere tokens; they must be sufficient to induce the proprietor to permit use of the inventions by others, and to do so voluntarily. The implication is that the use of inventions by others can be a profitable activity, even one that becomes the innovator’s main line of business. No longer are invention and technology transfer activities that impede one another. Rather, they become, to a substantial degree, mutually compatible.
But this is not all that the free market is able to accomplish in this arena. Among its other achievements are, first, efficiency in the selection of those firms that actually make use of an innovation and provision of an added inducement for firms to struggle to enhance such efficiency. Second, surrender of the exclusive right to use new products and processes, while providing a direct financial reward to the innovator, also leaves it with the benefits of being the first to make use of the invention, and this is an added incentive for continued investment in the process. Third, although the dissemination process does create some disincentive for investment in innovation, when done for profit it also has an influence that works the other way, at least helping to offset its recognized depressing effect. These consequences have already been hinted at in the preceding discussion, but it seems useful to end this part of the discussion with a brief review of their operation, with the three additional benefits here all described in one place.
The Market’s Choice of the Most Efficient as the Users of an Innovation
The firm that creates an invention may be more efficient than others in carrying out the innovation process, yet it may not be the most efficient of the enterprises in putting an innovation to use in the subsequent process of turning out final products. As already mentioned, the profit-seeking innovator will license to a rival only if that rival can offer it a license price that is more profitable to the licensor than the gain it could obtain by keeping use of the invention to itself. But the prospective licensee can afford to pay such a lucrative license fee only if it is a more efficient user of the technology. Thus the choice between the innovator and another firm as user of the innovator’s technology will depend on which of the two enterprises is its most efficient user. And, if there are several rival bidders for such licenses, the more efficient users will be able to make attractive offers and the less efficient will be unable to afford the license for use of the new product or process. Thus, the market will tend to see to it that the task of production using new technology will go to those best fitted to use it.
Moreover, since efficiency is not a once-and-for-all matter, rivalry for technology licenses serves as an additional incentive for continued investment in improvement of efficiency. The efficiency race can be encouraged only by the pressures contributed by competition among prospective licensees in the market for licenses.
The Seller of a License Retains the Early-Entry Advantage
We have noted that, even with all the help that the licensor can offer, the licensee characteristically needs many months or even several years before it can be fully up and running in the use of the technology that a license entitles it to employ. This means that, although the licensor gives rivals the opportunity to use its proprietary technology in competition with itself, it can depend on a considerable delay before effective competition materializes from this quarter. In the meantime, it can enjoy the advantages of sole employer of the technology, earning whatever monopoly profits are made possible by such exclusivity, seeking to develop customer loyalty for its product, and acquiring experience that improves its manufacture and distribution processes. Thus the technology proprietor can have it both ways—receiving money for its licenses and yet retaining the benefits from built-in unavailability of the innovation to others, at least in its early days, which can in some cases be the most profitable part of the innovation’s life cycle.
Lucrative Licensing as an Incentive for Innovation Investment
Finally, the very profitability of licensing or technology trading is an incentive for investment in innovation. That is, the firm that participates in this market has an incentive for innovative effort that is lacking for the firm that does not. The obvious case is the firm that specializes in the creation and licensing of innovations. Such an innovation factory will clearly drive itself out of business if it runs out of innovative products to offer for rental. An extension of this argument shows that this inducement applies to all firms for which licensing is a profitable activity. Similarly, companies that participate in a technology-exchange consortium must have a supply of new products and processes that they can offer to the other members in order to induce them to reciprocate with their own new inventions. This will be demonstrated rigorously in the next chapter. Here, again, one must not overstate the claim. A group of colluding firms may conceivably agree to mutual disarmament—to simultaneous reduction in the amount they invest in innovation. In the absence of such an agreement, however, the market forces can confidently be expected to run the other way, to induce the firms that license or exchange their technology profitably to spend more on innovation than they would have otherwise.
The conclusion, once more, is that the market mechanism has influences that are not obviously available to non-capitalist economies. These tend to make it profitable to engage simultaneously in the innovation “arms race” and in the licensing of any new inventions obtained in the process. Imperfect though this mechanism may be, it still seems remarkably effective (see figures 6.1 and 6.2 again), and is yet another of the features that differentiate the free-market growth process from any other that we know of.
THE MARKETING OF PROPRIETARY TECHNOLOGY IN PRACTICE
I turn next to some real illustrations of the marketing of proprietary technology, to show that it is no figment of the theorist’s imagination and to suggest that it is surprisingly common. First, consider some indirect evidence. A number of studies have found that innovation is disseminated far more quickly than one might expect from the hypothesis that the firm, aided by the legal system, does everything in its power to prevent or delay it. For example, a study by Mansfield, Schwartz and Wagner (1981) indicates, from data provided by a hundred American firms in a variety of manufacturing industries, that “[i]nformation concerning development decisions is generally in the hands of rivals within about 12 to 18 months, on the average, and information containing the detailed nature and operation of a new product or process generally leaks out within about a year” (p. 217). This and other evidence of impressive speed of dissemination imply that, if each of the world’s private firms is striving to keep its innovations entirely or largely to itself, these enterprises are uncharacteristically ineffective in their pursuit of this goal.
In reality, the opportunity entailed in technology dissemination can and often does lead firms to engage energetically in the marketing of licenses for their technology, treating this activity as a prime profit center for the firm.2 This is the goal of private industrial laboratories, of which Thomas Edison’s firm was perhaps the most publicized example. Other firms, even those that use the technology themselves, also engage in the sale of licenses as a substantial business activity. For example, the industry producing polypropylene resin, one of the most widely used plastic raw materials, has a hotly contested market in package licenses. The package includes the basic technology, the crucial catalyst that is compatible with the selected technology, a commitment by the license supplier to update the process and catalyst technology, and a commitment to provide required inputs and technical assistance. The two principal suppliers of such package licenses are Montell (wholly owned by Royal Dutch Shell) and Union Carbide Corporation. The revenue the latter derives from the sale of package licenses to resin producers throughout the world is a substantial proportion of the revenue it obtains from its own resin production (though that is not true of Montell).
It is, then, incorrect to depict the typical firm as a determined guardian of its technology against all use by others. Licensing has become a widespread business, as illustrated by the profusion of conferences, websites, and organizations devoted to technology transfer, along with wide media coverage of licensing agreements between individual companies. Some examples include a series of conferences, beginning in 1992, organized for business firm participants by the MIT Enterprise Forum on the subject of “Entrepreneurial Technology Transfer.” The initial conference offered participants the opportunity to “learn from the leaders in taking innovative technology to market.” The Licensing Executives Society, an organization that claims nearly 10,000 members from over sixty countries, holds “meetings, seminars, and training sessions for education and the exchange and dissemination of knowledge and information on licensing and intellectual property … the Society [monitors] domestic and international changes in the law and the practice of licensing and protecting intellectual property and [encourages] articles, reports, statistics, and other materials on licensing and protecting intellectual property.”3 The Technology Transfer Society is also active in disseminating information about licensing. A recent conference run by that organization on “Leveraging Technology for Competitive Advantage” featured presentations by business executives on technology transfer as a strategy for competitiveness in large corporations. Representatives from such companies as Texaco, Advanced Micro Devices, Hewlett-Packard, and Northrup Grumman gave papers centered around the themes of industrial competitiveness, organizational technology transfer, technology transfer success factors, and R&D commercialization.4 And there is a profusion of Internet websites offering a range of resources for technology transfer. Examples include Patent and License Exchange, Inc. (www.pl-x.com), Technology Access (www.uventures.com), CorporateIntelligence.com, Yet2.com, Pharmalicensing.com, PatentAuction.com, TechExchange Online (www.teonline.com), Knowledge Exchange Auction (Knexa.com), and Cool License (www.AnIdea.com). Surely all this implies eloquently that enterprising distribution of technology has become a widespread feature of business reality. Indeed, the National Science Board (2000, p. 2–56) reports that, for the period 1980–98, U.S., European, and Japanese firms collectively entered into almost 9,000 strategic technology alliances.5
ON TECHNOLOGY-EXCHANGE CONSORTIA IN PRACTICE
Several studies have examined firm behavior directly, and they confirm that businesses are often prepared to share their proprietary technological information with others, their horizontal competitors included.6 By the nature of the subject, these investigations rely heavily on samples and case studies. For one cannot expect to find systematic statistics on the extent of technological cooperation among otherwise independent business firms.
Eric Von Hippel (1988, chapter 6), for example, examined a sample of eleven (of the forty-firm U.S. total) American steel mini-mills. These enterprises, which use electric arc furnaces to recycle scrap steel, are considered world leaders in labor productivity. They have outperformed Japanese rivals and now provide a very substantial proportion of the steel output of the United States. Von Hippel found, through a series of interviews, that all but one of the firms in his sample regularly and routinely engaged in the interchange of information with the others: “reported know-how trading often appeared to go far beyond an arm’s length exchange of data at conferences. … sometimes, workers of competing firms were trained (at no charge), firm personnel were sent to competing facilities to help set up unfamiliar equipment, and so on” (p. 79, my emphasis). The know-how traded was, indeed, valuable. It often entailed exchanges with direct rivals and, though engineers and technicians normally carried out the exchanges, it was done with the knowledge and approval of management. The implicit arrangement described in this study is interpretable as a predominantly ex ante exchange. Each firm stands ready to reveal technical information to the others, with the implied understanding that the others will reciprocate by providing it with information on the new technology they acquire in the future, technology whose nature is as yet likely to be unknown.
Several other studies of the subject are reported by Von Hippel. For example, Allen, Hyman, and Pinckney (1983) examined a sample of more than a hundred Irish, Spanish, and Mexican firms and found exchanges at trade shows followed by plant visits and direct supply of technical information in response to inquiries. Such studies show that technical information exchange does actually occur. However, their limited samples and the relatively small sizes of the enterprises involved may leave open the possibility that the phenomenon is comparatively isolated. There is nonetheless reason to believe that it is much more than that.
Since I began studying the general subject of technology transfer some years ago, I have made it my business to raise this issue with every business firm that has engaged me as a consultant. This has involved perhaps twenty firms, including some of the giant enterprises of the American economy. In each case, the existence of some form and degree of technology exchange between the firm and its rivals was readily acknowledged by company management. In several cases, managements indicated that they agreed to it rather reluctantly, ascribing it to regular exchange of information among the scientists and engineers employed by them. The managerial personnel giving this report indicated that they were uncomfortable or even indignant at the giveaway of valuable information produced at company expense. However, they claimed to have little choice since, according to them, retention of their competitive position required the employment of scientists and engineers of high quality. And such able persons were unwilling to work for companies that did not permit communication with their counterparts in other enterprises (compare the quote by J. B. Say at the beginning of the next chapter).
In another instance, a retired vice president of one of America’s most innovative firms described a technology-sharing arrangement with a Japanese enterprise. Representatives of the two companies meet annually to settle their “balance of payments”; a sum decided by negotiation is paid by the firm providing the less valuable innovations to its “consortium” partner, as compensation for the latter’s more valuable contribution. A curious feature of this arrangement is that the two firms bargain not over the value of the innovations provided in the current year, or in the previous year, but over the innovations each expects to provide in the following years. A possible explanation of this ex ante orientation of the arrangement as a means to discourage cheating will be offered in the next chapter.
Another consulting assignment, in which I was engaged as this chapter was being written, entailed work for Perkin-Elmer Corporation, a firm that manufactures and sells analytical (scientific) instruments (notably those using precision optics) throughout the world. Since World War II, Perkin-Elmer has entered into many agreements with domestic and foreign firms for the ex post transfer of technology under license. The most interesting of these (for our purposes here) was an agreement in 1960 with the Hitachi Corporation for the systematic exchange of technical information.7 In the Perkin-Elmer/Hitachi contracts, the two firms undertook to supply regularly to one another a full menu of technical developments that had come into their possession. Each firm was authorized to produce, with full technical assistance by the other, any product on any such menu it had received, paying a royalty rate of 6.0 to 7.5 percent for items that had been invented by the other firm, with lower royalty payments on items to which the licensing firm had made a smaller innovating or development contribution. In 1971, the contract was modified to restrict the amount of information each firm was expected to supply along with its menus before the other firm had committed itself to its product selections. Evidently, the firms felt that they had been giving away too much information without compensation.
Excerpts from their contract of 10 December 1968 illustrate the spirit of the agreements:
Perkin-Elmer … and Hitachi desire to insure a continuous exchange of technical information in agreed-upon fields of analytical instrumentation. … 3.1 To the extent permitted by other agreements to which one of the parties hereto … is a party, and in conformity with government security or other restrictions, (a) Perkin-Elmer agrees to make available to [Hitachi] and (b) [Hitachi] agrees to convey to Perkin-Elmer all information each has available to it during the term of this Agreement within the Product Principle List. … 3.2 Upon request [the firms will supply to one another] copies of all drawings available to each containing information relating to such Product Principles. 3.3 Each party may send technical representatives to the other party’s premises at its own expense and at times convenient to the other party to obtain technical information related to such Product Principles. 3.4 Upon request [either firm] will send technical experts to [the other] at times and places mutually agreeable for periods not to exceed thirty (30) days. … In each case, the requesting company will be required to pay all travel and maintenance expense of such technical experts but not salaries. (pp. 1, 5–6)
The contract of 8 April 1971 also states:
Each party covenants and agrees that it will at all times hereafter … take all reasonable steps to keep all Technical Information communicated to it by the other party secret and confidential and will not divulge any of the said Technical Information to any person or corporation other than those sub-licensed hereunder. (p. 8)
In other industries the technology-sharing arrangements differ, but they are often there. The vice president of IBM in charge of its patent portfolio8 wrote to me that many firms in his industry licensed their patents, either in cross-licensing contracts with one another, or in limited exchanges with smaller participants in the industry. The contracts usually cover some defined field such as semiconductors or input–output devices. They are formulated in a bargaining session in which the two negotiating firms compare what each has to offer the other (larger firms being expected to provide more patents but to benefit more from each patent provided by its smaller partner), and the difference in the value of the patent offerings of the two firms is made up by a monetary “balance payment.” The contract, which normally runs over several years, entitles each firm to use the other firm’s current patents in the field covered (the ex post component of the arrangement), as well as other patents issued later during the life of the contract (the ex ante component). The contract covers patents but not know-how (i.e., information about the best ways to put a patent to use), which remains proprietary. According to the supplier of this information, the reasons such cross-licensing contracts are felt to be important by IBM are that they “level the playing field,” save wasteful costs such as outlays on reverse engineering or on inventing around patents of a rival, and enhance the firm’s freedom of action by permitting it to introduce new products without fear that it will be accused of infringing someone’s patents.
Other examples are easily found. United Technologies’ Pratt and Whitney, which manufactures aircraft jet engines and associated products, has acted as a partner with one of its principal competitors, General Electric, “in the development of an engine to power a very high speed future aircraft.”9 And, in a very different industry, a newspaper report tells us that wine making in California has been characterized by a “spirit of cooperation and communication. … While different wineries compete fiercely for sales, winemakers have commonly shared their technical knowledge and tricks of the trade, to their mutual benefit” (New York Times, 2 July 1992, p. 1). Ernesto Gallo told me that his winery helps to finance the winemaking program at the University of California, at which the latest techniques that have evolved in the company are taught to students. He also said that technical communication among the vineyards continues—facilitated by the very large proportion of master vintners at rival vineyards who were trained at Gallo—at least partly to make jobs at Gallo attractive to capable individuals. Finally, even in the fierce race to produce practical high-definition television technology, two of the rival groups (one consisting of General Instrument Company and MIT, and the other consisting of Zenith Electronics Corporation and AT&T) agreed to pool their efforts and to share the rights that would be provided by the U.S. patents that resulted from their combined efforts (New York Times, 8 May 1992, p. D1).
It is clear, then, that firms, large and small, do in many cases exchange technological information, sometimes reluctantly, perhaps more often very deliberately.10 And a good deal of this is carried out through inter-firm exchanges.
THE PRICE MECHANISM AND RECONCILIATION OF PRIVATE AND SOCIAL GOALS IN DISSEMINATION
As elsewhere in the economy, the price mechanism can help to deal with the dissemination dilemma for society—the problem that easy dissemination reduces the incentive for innovation but facilitates its general use. Licensing and rental fees can enable the economy to attain something like the optimal compromise in this tradeoff. They can help to preserve the inventor’s reward, while permitting others to use the ideas. It is this price that induces some firms to go beyond grudgingly permitting the use of some of their technology by others when forced by circumstances to do so.
If the price is right, it will pay the firm that owns any such input to permit others to use it. The question, of course, is: What is the right price? It is also natural to ask whether the price that is right for the licensor is also the one that is best for society. In particular, do there exist features of the market mechanism that drive the license fee toward the socially optimal level? We will return to these questions in chapter 13, where it will be shown that there is such a thing as an economically efficient license fee. I will also argue that this efficient fee offers benefits to both the licensor and the licensee, but that market forces support its adoption only imperfectly.
APPENDIX
SOME OTHER PERKIN-ELMER PATENT-SHARING ARRANGEMENTS
Besides its contracts with Hitachi, the Perkin-Elmer Corporation has had about a hundred other such arrangements since World War II. The following are a few examples illustrating very briefly how much their terms differed:
CSIRO. Spectroscopy uses instruments of two general types: infrared spectrophotometers, which are particularly useful for analysis of the composition of a sample whose makeup is not known in advance, and atomic absorption (AA) instruments, which are better adapted to precise measurement of components previously known to be contained in a sample. The AA technique was invented in Australia and was patented in the early 1950s by a firm whose acronym is CSIRO. Before the availability of commercially viable AA equipment, CSIRO began to license its patents, first on an exclusive basis, then, after 1959, to companies throughout the world. Perkin-Elmer was among the licensees, agreeing to pay what was CSIRO’s standard royalty (depending on the country of spectrometer sale) of about 2 percent of the value of the instrument. The relatively low fee reflected the substantial development work that the licensees had to carry out to produce a salable AA instrument.
Advanced Radiation. In the 1970s, Perkin-Elmer patented a process for the manufacture of the “Microalign instrument,” which uses a mercury capillary lamp to etch circuit patterns on photo-resistant semiconductor wafers. Having been a successful sole producer of the item for some period, in 1977 the firm encountered pressure from its customers to arrange for a second supplier of the Microalign lamps to ensure their continuous availability. Perkin-Elmer selected Advanced Radiation Corporation (ARC), which it described as “a small, flexible supplier,” and entered into an agreement to allow ARC to produce the lamp and sell it directly to customers. Because of its blocking patent, Perkin-Elmer was able to arrange for a royalty of 48 percent of the value of ARC sales, with the arrangement apparently also proving extremely profitable to the licensee.
Laser Precision. In the late 1970s, infrared spectroscopy underwent a major improvement permitted by the growing power of computers. The new approach utilized the Fourier transform to make more effective use of the data generated by the spectroscope. Perkin-Elmer undertook its own development effort in the late 1970s, seeking to design a Fourier-transform instrument of its own. Because the effort was going too slowly, and to avoid a major gap in its product line and tardy entry into the field, in 1981 Perkin-Elmer began negotiations with the Analect Division of Laser Precision Corporation for access to Analect’s Fourier-transform instrument. At the time, Analect was a tiny concern with three full-time employees, but it held the patent on a viable technique for the purpose. The license called for a payment of $1 million plus a royalty of 4 percent of sales. However, Perkin-Elmer never exercised its option to acquire the license. Instead, it chose to buy the finished instrument from Analect until Perkin-Elmer came up with a later-generation instrument of its own.
Perkin-Elmer Puerto Rico. In 1970 Perkin-Elmer founded a wholly owned subsidiary in Puerto Rico, and licensed it to manufacture a number of the products for which Perkin-Elmer held the patents. The Puerto Rican output was sold entirely to the parent company, which carried out their distribution. The transfer prices seem to have approximated the figures in the license agreement with other firms, though the U.S. Internal Revenue Service, predictably, challenged them.
___________
1. Note that voluntary dissemination via a licensing market still may require the firm to rely on patents and secrecy. Otherwise users might be able to acquire the technological information without paying the proprietor for its use, and the market for licenses might then shrivel or break down altogether.
2. The profits offered by licensing can be substantial: “The budget for IBM Research, which rose about 10 percent last year, has been growing faster than … the company’s overall revenue, which grew just 1 percent, to $88.4 billion in 2000. IBM’s profit from licensing its inventions has been growing even faster [reaching the] $1.7 billion it received last year” (B. J. Feder, “Eureka! Labs with Profits,” New York Times, 9 September 2001, section 3, pp. 1 and 12). Comparing total company profits with these licensing revenues, which by their very nature constitute virtually pure additions to profit, IBM’s licensing fees for 2000 amounted to slightly more than 20 percent of the firm’s total profit!
5. Some recent examples include patent and licensing deals between Apple Computer and Microsoft (see Chris Ward, “Rivalry that Ended in Friendship,” The Times, 13 August 1997); a technology licensing agreement in the electrical components industry between Maxwell Technologies and Siemens Matsushita Components GmbH (see Bruce V. Bigelow, “Maxwell Technologies Expects $30 Million in Device Royalties,” San Diego Union-Tribune, 7 October 1998); a pharmaceutical technology licensing partnership between Pfizer and ArQule (see Ronald Rosenberg, “ArQule, Pfizer in 4½-Year Accord; Drug Deal May Be Worth $117 Million,” Boston Globe, 22 July 1999); strategic technology transfer alliances in the medical imaging market made by Standard Imaging (see Maria Carlino, “Standard Imaging,” Journal of Commerce, 23 February 1996); technology licensing of power conversion products between Vicor Corporation and NEC Corporation (see “Vicor, NEC Agree on Technology Licensing,” Boston Globe, 5 March 1998); and technology licensing in print technology between Pitney-Bowes and Hewlett-Packard (see “Pitney-Bowes and Hewlett-Packard Settle Litigation, Announce Business and Technology Agreements,” PR Newswire Association, Inc., 4 June 2001).
6. “This is not to say that there are not situations where firms license their direct market competitors. However, these seem to be industries where licensees do independent R&D, proprietary gains come directly from a head start in any case, and there is an implicit or explicit reciprocity about licensing certain kinds of technology” (Nelson, 1996, p. 70).
7. For more on the technology-sharing arrangements involving Perkin-Elmer, see the appendix to this chapter.
8. I am deeply grateful to Vice President Howard G. Figueroa for giving me the information described here, and for editing this portion of the manuscript for accuracy.
9. Letter to me by the vice president in charge of Management Information Systems at Pratt and Whitney.
10. Note also the following observation by George and Joll (1981, pp. 231–32) about practices in the United Kingdom: “In addition a group of firms in research-intensive industries may operate a patent-pooling and licensing arrangement by which all the firms agree to license one another but no outside firms. Indeed, Taylor and Silberston (1973) found that in the pharmaceutical industry the most important advantage claimed for the patent system was that it gave the firms something to put into such a patent-pooling system so as to gain access to the other firms’ patented drugs.” This is a clear example of internalization of the externalities of the R&D process that is emphasized here.