Antitrust and Technological Innovation
The courtroom drama of U.S. v. Microsoft, now playing in Washington, D.C., has drawn hyperbolic press notices. Some observers portray the trial as the first test in the dawning Information Age of antitrust law made in the now-past industrial age. Microsoft CEO Bill Gates plays John D. Rockefeller in this construction, and Assistant Attorney General Joel Klein reprises the role of a Progressive-era trustbuster. The analogy is not entirely fanciful; Microsoft is indeed an important case. Its novelty, however, is overstated. The objective of using antitrust law to create “a democratic high-technology system” (as David Cushman Coyle put it in 1938) is deeply rooted in the U.S. political and legal tradition. Antitrust enforcement has been a hidden dimension of science and technology (S&T) policy over the past century, and it deserves to be brought squarely into the view of the S&T policy community today.
The impact of antitrust law on research and innovation is indirect. The law shapes industrial competition and the terms of cooperation among firms; these in turn influence firms’ incentives to undertake R&D, to strive for productivity growth, and to bring new products to market. “Indirect,” though, does not mean “unimportant.” In some sectors, antitrust policy has been far more consequential for research and innovation than the federal R&D spending policies that have attracted far more attention from analysts and policymakers. As the funding and performance of scientific and technological activity increasingly shift into the private sector in the coming decades, the relative importance of antitrust policy will continue to grow.
Antitrust law is not the only dimension of S&T policy that has been obscured by the conventional wisdom in our field. Tax policy, trade policy, labor law, and regulation are also important but largely overlooked influences on the scale and scope of research and innovation in the private sector. None of these areas have been wholly neglected, yet none have received the effort they deserve. More important, we lack the conceptual framework to consider these instruments of public action in conjunction with one another and in conjunction with public R&D funding, which is the way that they appear to the firms and people whose behavior they influence. Our conceptual disarray mirrors the fragmented nature of the policy process. At best, the various policies that influence research and innovation are uncoordinated; at worst, they are contradictory.
The difficulty of integrating antitrust enforcement into S&T policy is particularly acute. The effective use of antitrust law to enhance scientific and technological progress requires careful case-by-case analysis and delicate implementation. There are no simple rules to be applied. Prosecutors, judges, and juries, who are not typically seen as central to the S&T policy process nor are they experts in it, make many of the critical decisions. By law and custom, their work cannot be brought fully into the purview of the legislative and executive branches, nor, judging from common sense and experience, should it be. A degree of decentralization and nonexpert administration in this realm provide a check on the political authorities; moreover, the system has worked reasonably well in the past. But it can be improved. Cautious efforts to train analytical attention on the interaction of antitrust and innovation and to enhance the linkages among analysts, enforcement authorities, and elected officials are warranted.
The “postwar consensus”
According to the conventional wisdom, S&T policy after World War II reflected a consensus about the appropriate role of the federal government. In this rendition of history, the “postwar consensus,” first articulated by Vannevar Bush, legitimated financial support for academic research and the pursuit of R&D-intensive government missions, such as national defense and space exploration. The economy serendipitously benefited from these federal investments, but without much further help from public policy. If publicly funded academics or contractors had commercially useful ideas, the market would find its own ways to develop and diffuse them. The consensus view gave public R&D spending substantial credit for economic growth, even though the mechanisms by which this occurred remained nearly unexplained. In effect, this view essentially equates S&T policy with federal R&D spending.
This conventional wisdom is not so much wrong as it is incomplete. Perhaps the best way to see the omissions is to think about how public policy factors into private decisions related to research and innovation. An R&D manager evaluating a project proposal or a venture capitalist mulling a high-tech investment may well recognize that the scientific or technological opportunity grew out of federally funded work. More concretely, however, the dominant considerations will be the likely response of competitors and customers, the legal and regulatory hurdles that will need to be cleared, the tax liability to be incurred, and so on. Courts, regulatory agencies, trade negotiators, and internal revenue administrators may well figure more significantly in these decisions than the more conventional subjects of S&T policy analysis. This suggests the need to broaden our conception of S&T policy beyond that handed down by the consensus view.
One major government activity to which the conventional wisdom has been blind is antitrust policy. Antitrust enforcement agencies shape industrial structure in the United States by restricting the business practices of firms with great market power; reviewing mergers and acquisitions with an eye toward limiting the accumulation of market power; and, on rare occasions, precipitating the breakup of companies. The consequences of their work for research and innovation are complicated. Firms in highly concentrated industries, for instance, have an incentive to slow the pace of technological change in order to increase their profits from existing products. On the other hand, they also have an incentive to invest in long-term, large-scale R&D efforts, because they can appropriate all the benefits of these investments and do not have to worry about imitators. By promoting competition, antitrust policymakers erode both incentives. At the other end of the spectrum, in highly fragmented industries, the incentives are reversed. By countenancing cooperation among otherwise competing firms, antitrust policy can expand the scope and extend the time horizon of scientific and technological investments by firms in such industries, but it can also allow them to collude in order suppress potential advances. The ultimate technological result of the tradeoff struck by antitrust policy between competition and cooperation, even in these extreme cases, can be assessed only with empirical evidence.
The antitrust tradition
Technology was very much in the minds of the original advocates of antitrust laws. The emergence of giant corporations in the late 19th century hinged on innovations in communications, transportation, and production. In attacking big business, antitrusters also expressed an interest in regulating the pace of innovation and compensating farmers, workers, and small business owners for its deleterious effects on them. Yet, as David Mowery has shown, the major antitrust statutes, the Sherman Act of 1890, and the Clayton and Federal Trade Commission (FTC) Acts of 1914, along with contemporaneous interpretations of patent law, inadvertently enhanced the incentives for large firms to invest in new technologies. Forbidden to collude, they consolidated instead. The emergence of central corporate laboratories in these consolidated firms during the first three decades of the 20th century, a signal development in the history of the U.S. national innovation system, owes much to this unanticipated effect of public policy.
Antitrust policy went through a major transformation in the 1930s and 1940s. Again, technology played a significant role in the thinking of reformers and, again, the policy change had important long-term consequences for research and innovation. The catalyst for the change was the so-called “Roosevelt recession” of 1937-1938. In their search for the cause of this sharp economic downturn, some New Dealers focused on the concentration of economic power. They alleged, among other things, that large corporations used their control of patents to inhibit technological innovation, thereby choking off economic growth and causing unemployment. Assistant Attorney General for Antitrust Thurman Arnold, who was appointed by President Franklin D. Roosevelt in March 1938, initiated a series of actions on these grounds against some of the nation’s best-known high-technology companies, including Standard Oil of New Jersey, DuPont, General Electric, and Alcoa.
Arnold’s efforts faced substantial opposition, particularly from military authorities during World War II; indeed, he was forced from his position in 1943 and was succeeded by his deputy, Wendell Berge. Berge persevered, however, and, after the war, thanks to judicial appointments made by Roosevelt, won key cases in the Supreme Court. The Court in essence gave antitrust law precedence over patent law in cases where they conflicted, and it legitimated compulsory licensing of patents as a remedy. This remedy was employed in such important areas of technology as semiconductors, computers, aluminum, color film, pharmaceuticals, and synthetic fibers in the early post-World War II period. Research by F. M. Scherer shows that the direct effects of compulsory licensing were positive or, at worst, neutral for the industries affected.
Although instances of overt enforcement declined as the precedents aged, the policy insinuated itself into corporate technology strategies, leaving a lasting indirect imprint. Case studies suggest that new firms, whose founders might have been deterred from starting them under prewar conditions or that might have been snapped up by larger firms, were given space to grow by this policy in the 1950s and 1960s. DuPont, for instance, refrained from purchasing potential new competitors as aggressively as it had in the past, instead investing its cash in in-house R&D. Even more interestingly, the antitrust policy promoted by the Department of Justice (DOJ) combined with the massive expansion of Department of Defense (DOD) R&D and procurement spending in this period to foster a vibrant array of high-technology startup companies. DOD deliberately diffused new technologies and sought out alternative suppliers for them. Sometimes, startups even secured DOD commitments before they secured initial venture capital. Yet the positive interaction between DOJ and DOD policies was largely accidental; the residue of wartime bitterness between the two departments remained substantial.
U.S. versus IBM and AT&T
The policy change initiated by Arnold and completed by Berge and the Supreme Court faded into the background as the Cold War deepened. Analysts interested in the impact of government on research and innovation focused their attention on large-scale federal R&D spending, initially by DOD and the Atomic Energy Commission (AEC), later by the National Institutes of Health, the National Science Foundation, the National Aeronautics and Space Administration, and the AEC’s successors. The focus made good sense; the federal share of national R&D spending peaked in the early 1960s at about two-thirds of the total. In the same period, economic reformers turned away from the New Dealers’ concerns about industrial structure and monopoly power and, under the influence of Keynesian economics, toward fiscal and monetary policy issues. The two trends fit together in studies correlating aggregate R&D spending with long-term economic growth. This line of thought suggested to policymakers that such spending, divorced from purposes and mechanisms, was all that mattered. Of course, policymakers continually faced choices about the whys and wherefores of spending and other uses of government authority, and academic interest in microeconomic phenomena related to scientific discovery and technological innovation never disappeared entirely. Nevertheless, the connection became much fainter.
Although no longer the object of as much attention as in earlier decades, DOJ and the FTC remained aware that their efforts could have major effects on research and innovation as well as on the more traditional variables of price and market share. The twin cases filed against technology behemoths IBM and AT&T, which were both resolved in 1982, exemplify the point. The IBM case, the third by the government against the company since the 1930s, was filed on the last day of the Johnson administration in 1969. Among other things, DOJ charged that the company used its market power and its control of standards to deter innovations by competitors and to extend its dominance into new market segments in which it did not necessarily offer the best products. IBM, for instance, was alleged to have engaged in practices such as bundling and predatory pricing that were detrimental to manufacturers of peripherals and software, depriving consumers of advances that these competitors might have made. IBM was accused as well of promising customers “paper machines” to keep them from ordering real machines from competitors (the hardware equivalent of what Microsoft critics today call “vaporware”).
DOJ ultimately abandoned the case, but not before IBM’s market position had begun to erode in the face of new competition. Some of these competitors were Japanese firms whose questionable practices with respect to intellectual property might have been pursued more vigorously by the U.S. government had DOJ not been locked in conflict with IBM. New competitors closer to home offered mini- and microcomputers, products that IBM was reluctant to offer in part because they cannibalized its mainframe computers, but in part because it feared antitrust recrimination. Once IBM decided it had to get into the personal computer market, the company made a series of decisions that eventually ceded the bulk of the profits from this highly successful effort to Microsoft, Intel, and other firms. Whether antitrust concerns influenced these crucial choices is a matter of debate. What seems clear is that the antitrust case changed the terms of competition, contributing to the pursuit of a greater variety of technological paths than IBM would probably have pursued had it been left to its own devices.
U.S. v. AT&T, filed in November 1974, built on a number of precedents, too. The Truman-era DOJ had accused AT&T of illegally crushing its competition in telephone equipment manufacturing. With the support of DOD, AT&T settled that suit on favorable terms in 1956, maintaining its major technological assets, including Western Electric and Bell Labs. However, the consent decree did compel it to license its entire patent portfolio and to stay out of nontelephone markets. The foundational patents for the semiconductor industry were among those licensed; the strictures on AT&T competition in this area facilitated the growth of new firms that later became household names. The Federal Communications Commission (FCC) also shaped AT&T’s business and technological environment in the 1950s, 1960s, and 1970s, permitting competitors to offer innovative products and services while limiting AT&T’s responses. The most dogged of these competitors was MCI, which ultimately filed a private antitrust case against AT&T while pressing its advantage in the FCC.
In its 1974 case, DOJ reiterated its concern about AT&T’s dominance of the equipment market, suggesting that competition would unleash a burst of technological innovation. The Modified Final Judgment in the case provided an opportunity to test that contention, because it forced AT&T to divest its local telephone operating companies and lifted the 1956 consent decree. The technological consequences of the AT&T case, like those of the IBM case, remain disputed. Some observers attribute the accelerated deployment of fiber optic lines, the development of the wireless industry, and even the growth of the Internet to the breakup, whereas others lament the downsizing of Bell Labs and the chaos in the management of the national communications system. Again, what seems clear in this case as well is that antitrust policy altered the spectrum of technological opportunities in an important sector by expanding the number of players and shifting the relationships among them.
Mixed messages
The 1980s brought renewed attention to questions of competition and cooperation in research and innovation, and to the technological implications of antitrust policy in particular. Although the IBM and AT&T cases attracted public interest, the emergence of Japanese economic and technological competition was the primary trigger for this interest. Japan posed a conundrum for the conventional S&T policy wisdom-it spent relatively little on R&D and yet achieved impressive results. One influential interpretation of the Japanese model held that industry-wide cooperation fostered by the Ministry of International Trade and Industry contributed centrally to Japanese firms’ technological achievements. Yet such cooperative efforts were deterred in the United States, scholars and executives argued, by the threat of antitrust enforcement. In 1984, Congress enacted the National Cooperative Research Act (NCRA) with this concern in mind, relaxing the antitrust sanctions against cooperative R&D ventures of otherwise competing firms. Some 575 such ventures were registered in the ensuing 10 years.
Just as DOJ’s cases were accomplishing its objectives of enhancing competition in computing and telecommunications, the NCRA was facilitating broader cooperation. Other 1980s policy experiments also pushed in divergent directions. The Bayh-Dole Act of 1980, for example, gave federal grantees control of intellectual property resulting from federally funded R&D and permitted them to issue exclusive licenses to it, a policy that would have appalled Thurman Arnold. On the other hand, the Small Business Innovation and Research program forced agencies to set aside a fraction of their R&D funds for small firms, presumably enhancing their ability to compete with their larger brethren.
The 1990s have witnessed a strengthening of both the cooperative and competitive threads of federal policy. The Clinton administration has made fostering R&D partnerships a central element of its S&T policy, even as DOJ and the FTC attack such high-technology giants as Microsoft and Intel. Viewed in the context of the history of antitrust policy, the most surprising of these endeavors is the Partnership for a New Generation of Vehicles (PNGV), which involves the “big three” U.S. automakers along with a host of suppliers and related firms. PNGV’s goal of collaborative R&D in pursuit of environmentally sound designs echoes that of the auto industry’s ill-fated cooperative technology development efforts of the 1960s. These efforts prompted an antitrust investigation, which alleged collusion in the suppression of new technologies and which was settled by their disbandment in 1969.
On the competition side of the ledger, although the big antitrust cases have generated interest and attention, DOJ and the FTC have also attempted to clarify the underlying doctrine guiding their use of antitrust law to promote technological innovation. Officials at the two agencies have offered a new framework for the review of proposed mergers, for instance, which considers whether the merger will reduce R&D competition in any field. This “innovation market analysis” has been applied to such mergers as Roche-Genentech (with respect to human growth hormone and potential AIDS treatments) and between Ciba-Geigy and Sandoz (gene therapy). Another area of antitrust enforcement interest is standards. The enforcement agencies have pressed for open and transparent standard-setting processes. One example is the FTC’s case against Dell Computers in 1996, which alleged that the firm tried to game the process in an attempt to sabotage standards allowing peripherals and CPUs to interoperate. Intellectual property is a third major technology-related interest of the two agencies, which jointly issued guidelines in this area in 1995. Although intellectual property rights can stimulate competition and innovation under many circumstances, they are not supposed to be used to extend monopoly power, nor should firms with market power attempt to deny competitors their rightful intellectual property protection, as Intel has been accused by the FTC of doing.
The nexus of technological innovation, standards, intellectual property, networks, and globalization is fomenting energetic debates in the antitrust legal community. In reviewing his agency’s extensive hearings on “competition policy in the new high-tech global marketplace,” FTC Chairman Robert Pitofsky identified these issues as the central challenges for the foreseeable future. Yet the characterization of Clinton administration policy as “reverse Schumpeterian”-implying a prejudice against large and powerful firms-is an overstatement. Both DOJ and the FTC profess a strong bias against intervention, and, as we have seen, other agencies have favored cooperation over competition in research and innovation, even among big firms. The apparent paradox is personified in one of Washington’s power couples, Anne Bingaman, who headed the Antitrust Division of DOJ during the first Clinton term, and her spouse, Senator Jeff Bingaman of New Mexico, who has been instrumental in drafting partnership-oriented legislation since he was first elected in 1982.
Implications
To some extent, the tensions between fostering cooperation and enhancing competition are illusory. The systems of innovation in different industries vary considerably, and it is both possible and sensible for public policy to be tailored to solve different sorts of market failures in different industries. The government should nurture research consortia when the threat of free riding by competitors deters investments in new technologies, and it should also stimulate competition when powerful firms choke off promising alternative technological paths. However, although both tendencies are present in the current policy, whether the policy conforms to this rational interpretation may reasonably be doubted. A more plausible reading is that we are observing the imperfect results of a highly decentralized policymaking process. One agency does not necessarily know or care what another is doing; they may have quite different, and equally legitimate, instructions from masters at opposite ends of Pennsylvania Avenue. Regulators, courts, and private litigants are deliberately insulated from political influence. As Philip Areeda, an eminent scholar of antitrust law, observed, “There is no other country in the world in which such important, national economic decisions are made on such a decentralized, undebated, and largely nonexpert basis.”
Ironically, this fragmentation has not necessarily been a bad thing. The unintended interactions of independently operating arms of the government in S&T policy have sometimes yielded positive results, as in the case of the DOJ/DOD interaction in the 1950s. The antitrust case against Microsoft might combine with federal R&D funding programs to create this sort of “positive interference” pattern today; browsing software, it is worth remembering, first emerged from the federally funded supercomputer center at the University of Illinois. This supposition does not imply that Microsoft has violated the law nor that punitive action will be taken against it. The history described above suggests that changes in behavior induced by antitrust policy in both the target firms and their competitors can be economically and technologically significant even if the prosecution fails. Government R&D funding might play a catalytic role for new and small firms in these circumstances.
However, we should not be so sanguine as to assume that positive interference patterns will outweigh negative interference patterns, in which well-intentioned, independently initiated, and contradictory policies cancel one another out. It is one thing to laud experimentation in what is admittedly an unpredictable and complex interaction between government and business, and still another to praise chaos, contradiction, and confusion, even when the results are occasionally good. At a minimum, the policy analysis community must broaden its vision of S&T policy and consider more carefully the interactions between R&D funding and antitrust policies, and more broadly among all the policy instruments, obvious and hidden, that shape the environment for corporate research and innovation. In an era in which the share of the nation’s R&D funding that comes from private sources is 70 percent and rising, this set of tasks deserves a higher priority than it has had to date.
To call for analysts to take on new problems is simple; to devise a more integrated policy process is not. The risks of overcentralization are substantial. The decisions entailed in this realm of policy are difficult, and fragmentation allows for a certain degree of hedging of the public’s bets. Mistakes made by one arm of the government can be compensated for by another. Moreover, a more centralized process could impose a simplistic, one-size-fits-all set of rules that are inappropriate to the circumstances. In addition, a more integrated policy process supervised by elected officials would be more prone to corruption and unprincipled political manipulation. Improper influence is a common charge in antitrust cases in particular and is well worth guarding against.
Nonetheless, cautious steps in the direction of greater integration can and should be taken. The FTC has already acknowledged a responsibility to serve as something of an early warning system, bringing critical issues to the attention of other policymakers. The Technology Administration in the Department of Commerce might be assigned a similar monitoring and advisory role, keeping tabs on the impact of public policy on research and innovation at the industry level. The National Economic Council in the Executive Office of the President is the appropriate venue for coordinating interagency efforts in this area, and the antitrust enforcement agencies ought to participate in them, at least to some degree. The Council of Economic Advisors can bring its expertise to bear on matters relating to antitrust and technology, as it has on some occasions in the past. All such deliberations must be open to the extent possible, and the grounds for decisionmaking must be clearly articulated to allay fears of crude politicization.
Although decisions linking the hidden and less-hidden sides of S&T policy should be informed by legal, economic, and technological expertise, they are, in the final analysis, fundamentally political and that may be the most important justification for tinkering with the system. To put it another way, these decisions involve the use of coercive power and the allocation of public resources under conditions of uncertainty, and they have significant consequences for the national and even global economy. A more integrated policy, but one that is still fragmented by constitutional design and historical practice, should mean greater accountability for the outcomes of these decisions without jeopardizing their integrity.