Proceed with Caution: The Application of Antitrust to Innovation-Intensive Markets
Calvin S. Goldman , Q.C., Richard F.D. Corley
and Michael E. Piaskoski of Blake, Cassels and Graydon LLP
The explosion of the information economy, combined with the rapid introduction of the associated new technologies, hardware, software and shareware, etc., has challenged many of the traditional tools of legal protection and enforcement, causing regulatory authorities, law enforcement agencies, lawyers, judges and lawmakers to play 'catch-up' with this dynamic industry. As a result, established legal understandings of copyright, jurisdictional borders, illegal content and contracts must be revisited in this era of innovation, information and e-commerce.
So too of antitrust and competition law. Antitrust authorities are finding themselves faced with the challenges posed by markets that refuse to stand still and shun the traditional trappings of competition law that use a 'snapshot' of product and geographic markets, market shares, monopolies and barriers to entry. As the products and services of the information economy become increasingly complex and change occurs almost daily with successive innovations, antitrust authorities are (and will continue to be) faced with questions that many of their traditional understandings of markets and methods of analysis and enforcement. For example, as a variation on the 'chicken and the egg' theory, one might ask, 'does innovation spur competition or does competition spur innovation?' The answer to this conundrum is important as any attempt to limit one may have significant effects on the other. Further, is it acceptable from an antitrust perspective to hinder the creation of a widely-accepted technological standard (and a therefore dominant market position for the company creating the standard) if consumers and society as a whole will benefit from the ongoing innovation and wide use of the standard? Should the protection of consumers from the static effects of short-term price increases trump the long-term effects of dynamic gains arising from innovation? The answers are not straightforward and continue to be discussed and reviewed by antitrust authorities around the world.
Keywords: Competition law, antitrust, information economy ,intellectual property, network effects, low barriers to entry, minimal marginal costs, vertical and horizontal integration, complementarity, traditional product and service markets, IP markets.
Citation: Piaskoski, Corley and Goldman, 'Proceed with Caution: The Application of Antitrust to Innovation-Intensive Markets ', 2004 (1) The Journal of Information, Law and Technology (JILT). <http://elj.warwick.ac.uk/jilt/04-1/piaskoskicorleyandgoldman.html>. New citation as at 15/06/04: <http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2004_1/goldman/>.
In this paper, we attempt to draw attention to the unique characteristics of high innovation markets that make it difficult for antitrust authorities to know when and how to intervene. We do not suggest, however, that antitrust has no role in the information economy. Nor do we suggest that, in general, the authorities have done a poor job of assessing mergers and conduct in these industries. Indeed, many antitrust authorities have made considerable efforts to study, analyze and otherwise understand the competition issues related to high innovation markets. In fact, in 2002, antitrust enforcement agencies from ten different jurisdictions participated in a Roundtable on Merger Assessment in High Innovation Markets chaired by the Competition Committee of the Directorate for Financial, Fiscal and Enterprise Affairs for the OECD (the 'OECD Roundtable'); many of their insightful comments and observations have been included throughout this paper. 
Rather we argue that antitrust authorities must formulate a clear picture of the industry and technologies under their review, acquire an enhanced understanding of some of the underlying economic principles of the new economy, and respect intellectual property laws that act to encourage and facilitate innovation. We suggest that developments and changes in information technology markets have ameliorated the conditions which have in the past prompted antitrust enforcement action against companies such as Microsoft, and raised serious questions about the desirability of the antitrust enforcement actions taken in those cases. Most importantly, we believe that antitrust authorities must act with restraint and caution in dealing with high innovation industries and appreciate the consequences of their actions on this economy and, in particular, on the incentive to innovate.
While firmly established in economic theories such as the protection of consumer welfare and the efficient allocation of resources, the objectives of antitrust and competition law have been influenced over the years by the evolution of 'economic activity' in society in general. Early competition law was concerned with the concentration of wealth and the effects of wealth transfers on consumers and farmers. As western society became more highly industrialized and the field of industrial organization economics developed, the academic literature, legislation and judicial decisions focussed increasingly on the promotion of allocative efficiency  as the principal objective of competition law enforcement, downplaying the importance of other objectives, such as the maintenance of fragmented industries and markets.
More recently, competition law economists and policy makers have recognized that dynamic change and innovation, the pillars of the information economy, are important sources of economic activity; they can lead to long-term efficiency and productivity gains which will ultimately benefit consumers. Accordingly, they are gaining increasing acceptance as very important sources and indicia of competition.  However, despite acknowledgement and acceptance of these new indicia of economic activity and competition, the traditional underlying theories of and analytical approaches to competition are challenged by the innovation market paradigm. Teece and Coleman state that:
…we do not believe the antitrust agencies anywhere in the world are at present well equipped to deal with competition policy in high-technology industries.…The very nature of competition, the definition of industries, the basis of competitive advantage, the effects of 'restrictive' practices and the nature of economic rents are all different in the context of innovation. The costs of error are great … 
The reasoning of the United States Court of Appeals for the District of Columbia Circuit in the Microsoft case  highlights the challenges faced by both antitrust authorities and independent arbiters who are increasingly confronted with issues arising from the interface of competition law with industries undergoing rapid technological change:
Antitrust scholars have long recognized the undesirability of having courts oversee product design, and any dampening of technological innovation would be at cross-purposes with antitrust law.
We suggest here only that the limited competence of courts to evaluate high-tech product designs and the high cost of error should make them wary of second-guessing the claimed benefits of a particular design decision.
Apart from the lack of textual support, we think that a balancing test that requires courts to weigh the 'synergies' of an integrated product against the 'evidence of distinct markets,' is not feasible in any predictable or useful way. Courts are ill equipped to evaluate the benefits of high-tech product design, and even could they place such an evaluation on one side of the balance, the strength of the 'evidence of distinct markets,' proposed for the other side of the scale, seems quite incommensurable. 
Robert Pitofsky observes that the importance of high innovation industries to the economy, combined with limited antitrust experience regarding the uses of intellectual property, should cause antitrust authorities to pay careful attention to ensure that overall economic growth is not compromised by the abuse of private market power.  Some (including the defenders of Microsoft) believe that antitrust should have no role in the information economy at all as it is dangerous to challenge and undermine investment incentives in these dynamic industries.  Others suggest that, due to the rapid pace of innovation in high-tech industries, regulatory authorities should refrain from intervening 'unless certain that doing so will benefit consumers and the economy.' 
The basic question then is whether the information economy 'is so different in kind from the economies of the past as to warrant the view that practices that restrict output, create or enhance monopoly power, or raise artificial barriers to new competition are no longer the sorts of things the law should strive to prevent'.  Further, should antitrust’s conventional role as enforcer of competitive principles be abandoned because it may be unnecessary or counter-productive, or, as some would believe, harmful? Should basic antitrust principles continue to apply, after taking due notice of different characteristics of the high-tech sector of the economy, or should a whole new set of rules be created for the information economy?
Advocates of a 'hands-off' approach rely on the theories of Joseph Schumpeter that posit that monopolies tend to innovate more vigorously due to their scale economies in R&D and greater opportunities to exploit the value of their innovations  and that, in any event, high-tech monopolies will be eroded by the effects of 'creative destruction', under which innovation destroys old business models or entire industries and creates new ones. Based on these theories, those opposed to antitrust involvement in the high-tech sector argue that: (1) traditional antitrust rules may discourage innovation by diminishing the rewards incumbent firms expect to gain from their inventions and developments; (2) incumbent firms may reduce their investments in innovation unless allowed greater freedom to exploit their market power and respond to new technologies  ; (3) change is too rapid; (4) the markets are self-correcting; and (5) in any event, the markets are too complex and the technology too difficult to understand.
The Schumpeterian theories have been challenged on a theoretical level by those who believe that the high-tech monopolist is unlikely to dilute its own monopoly rents through 'creative destruction' and self-generated inventions that will render its products obsolete and cannibalize its own sales. Further, there appears to be little empirical evidence or consensus that the application of ordinary antitrust rules will retard innovation generally. Rather, studies show that unchecked market power will impair rather than enhance long-run innovation. 
Jacobson believes that competition in innovation markets is well within the scope of legitimate antitrust protection: 'Exempting all acquisitions that are consummated before a foreseeable market has developed would risk preventing new discoveries from ever coming to market and allowing significant price increases for those that do — while spurring no incremental innovation.'  He observes that, in the Microsoft case, two different opinions regarding antitrust intervention formulated by the court of appeals. While the court specifically rejected the innovation exemption when evaluating whether Windows gave Microsoft monopoly power  , it appeared to accept such an exemption when determining whether Microsoft had unlawfully tied its Internet Explorer browser to the Windows operating system. In particular, the court held that software markets were too new and the courts were too inexperienced in dealing with them to apply the same per se rule for tying arrangements that had been adopted in other industries. 
The experience of the European Commission (EC) does not support the proposition that there is a need for an innovation exemption - for example, it has not found that competition is largely 'for the market' and that market power is undermined quickly by other innovations. It has, in fact, intervened in a number of merger cases that threatened to create or strengthen a non-transient dominant position in markets that can be characterized as high innovation markets.  It believes that antitrust authorities have shown that they are well-suited to monitor the information economy by, for example, relaxing their normal preference for structural rather than behavioural solutions to perceived problems.
The U.S. Federal Trade Commission (FTC) and the Department of Justice (DOJ) believe that claims that antitrust has no relevance in high-tech markets are overstated. Market dominance can still exist due to a variety of reasons, including large sunk costs, network effects, an installed base of customers, and other barriers to entry. For example, once a network monopoly is established, it is relatively easy for the monopolist to exclude potential competitors by adopting a policy of allowing only its own products to connect with the existing network. The monopolist may be able to monopolize successive generations of product, or complementary products or services. This means that the best products or services may not necessarily succeed and potential competitors, recognizing the tremendous difficulties of challenging the incumbent monopolist, may lose incentives to compete. Antitrust, therefore, can still play an essential role in order to prevent abuses of market power. Healthy competition, after all, spurs further innovation.
Therefore, assuming that there is a role for antitrust in the information economy, the characteristics of high-tech markets still require that antitrust laws be applied cautiously, informed by the facts of the specific situation. It is important that antitrust authorities know when to intervene and when to forbear; it may be extremely difficult to judge when a market will be sufficiently 'innovation-intensive' to justify non-intervention.
The unique economic characteristics of the information economy require that antitrust authorities consider carefully the implications of their actions for future investment, research and innovation. They must keep in mind that overly static efforts to promote (or protect) competition within the framework defined by existing markets may pre-empt or inhibit innovation and future competition for the development of new products, services and markets; any efforts to promote static efficiency by protecting certain competitors from the normal economic consequences of vigorous competition in markets characterized by demand-side network externalities are likely to inhibit the achievement of the dynamic efficiency gains associated with the development of new products and markets. In the context of the information economy, innovation is the key determinant of competition and should not inadvertently be impeded by the enforcement actions of antitrust authorities.
Some of the many competition law challenges posed by high-innovation markets can be illustrated with reference to the economic characteristics of the segment of the economy commonly referred to as the 'information economy', i.e., the segment of domestic and global economies that is engaged in the creation, processing and communication of information. The movement toward IP and information as a primary source of value, network effects, the rapid rate of industry change, falling barriers to entry and the other factors discussed in this paper suggest that dynamic change and the long-term increases in efficiency that flow from innovation are more important to competition than are short-term static efficiency gains. As suppliers compete to develop products, systems and standards that deliver greater value to consumers, a number of factors will encourage innovation, including the relatively low marginal costs of producing and reproducing information, low barriers to the marketing and distribution of information-based products over the Internet, the durable nature of information products (which requires that all suppliers, including those with high sales shares, innovate to compete against the installed incumbent base).
As a result, the underlying economic assumptions, sources of market power and conduct that may be considered anti-competitive differ in the information economy from those applicable in the traditional industrial economy. The rapid rate of change in such markets and the inability of antitrust authorities (or anyone else for that matter) to assess or predict the development and market impact of new inventions suggest that antitrust authorities should err on the side of caution so as to avoid causing any unnecessary harm to innovation and competition in such markets. Consequently, information-based economies may be expected to present the following implications and challenges for antitrust enforcement.
B. High-Tech Markets are Complex and Uncertain
Former FTC chair Robert Pitofsky describes the hurdles faced by 'technologically-challenged' antitrust authorities:
Many high-tech industries involve questions that are challenging for lawyers and judges who typically lack a technical background. For example, defining relevant markets, i.e., the process of identifying those firms that compete so closely with other firms that they can substantially influence the exercise of market power, is difficult enough under any circumstances. But it can become far more difficult in high-tech industries such as biotechnology, where products that might curtail the market power of a dominant incumbent firm are not in existence yet, and will not reach the market for several years; or in the cable industry where the essential question is when satellite transmission will become a real competitive force in the cable market. Similar problems arise with respect to telecommunications, a sector of the market where many believe competition for local operating companies will eventually be offered by electric utilities through their existing grid and electricity wires into the home. Each of these issues raises questions in the realm of science and technology that often will be difficult to address. 
However, technological and specialized expertise is generally required in all merger cases, regardless of industry. Antitrust authorities must familiarize themselves with the unique characteristics of the relevant industry. We believe this should be no different in high innovation industries.
The job of antitrust authorities is made more difficult by the uncertainty associated with the success of any given R&D project, the evolution of a particular market and the introduction of new competitors, new technologies and new downstream products. The joint venture between BSkyB and the Hilton Group demonstrates the difficulties associated with the uncertainty of future innovations. The case involved the establishment of a 50:50 joint venture between BSkyB and Hilton Group to develop a fixed odds betting business linked to Sky Sports channels on BSkyB’s digital satellite platform. The joint venture proposed to offer interactive 'in-vision' betting in which viewers could place bets on the very events they were watching. One of the reasons the merger was referred to the UK Competition Commission was the lack of certainty about new developing innovations and whether the joint venture would have an adverse effect on the evolution of competition within this market.  In advising the UK Secretary of State for Trade and Industry, the Director General concluded:
This proposed joint venture presents a dilemma. In some respects it may enhance innovation and competition in the wider betting market. But its exclusivity provisions pose possible risks to the development of competition in interactive betting (which is forecast to be a large market within the next few years) and in the acquisition of sports rights. These risks require a more thorough examination which the Competition Commission is best placed to undertake.
While the parties subsequently abandoned the proposed joint venture before the Competition Commission could complete its review of the static and dynamic effects of the merger, antitrust authorities will have to make predictions and assessments and draw conclusions on the expectancies of future markets.
C. The Interplay With Intellectual Property
The principal source of value in the information economy is knowledge and information, rather than tangible resources or capital. As a general rule, information may copied freely, endlessly and virtually costlessly, and can be used simultaneously by an unlimited number of individuals. Information is exclusive only to the extent that it is kept confidential by the owner and/or protected by applicable intellectual property (IP) laws, including domestic copyright, patent, trade-marks and trade secret laws and under applicable international treaties.
While it is understood, at least at an abstract level, that IP laws and competition law are both intended to promote the development of an efficient economy, each employs fundamentally different mechanisms to achieve this common long-term goal. In general terms, IP laws foster long-term dynamic efficiency gains by providing incentives for investments in the development of valuable works. IP laws create legally enforceable private rights that protect (to varying degrees) the form and/or content of information, expression and ideas. IP laws confer on an IP owner the right to unilaterally exclude others from using that property and to maximize its value through exploitation and exchange in the marketplace, enhancing the incentive for investment and future innovation. The primary purpose of these laws is to define the scope of these rights and determine under what circumstances they have been infringed upon or violated.
Competition law, on the other hand, is principally concerned with the achievement of static allocative efficiencies by preventing the inappropriate accumulation or exercise of market power with respect to existing products or services offered by existing competitors.  The pursuit of shorter-term allocative efficiency goals could adversely affect the incentives of companies to invest and innovate, and thereby reduce the long-term benefits to society that flow from research and development (R&D) activities and the protection of IP rights. Since the right to exclude is necessary for efficient, competitive markets, the enforcement of competition law should not interfere with the exercise of this basic right and should only be applied to conduct associated with IP that impedes the efficient production and diffusion of goods and technologies and the creation of new products.
Each of the U.S., the EU and Canada have developed intellectual property guidelines that attempt to find a balance between IP and competition laws and define the role of the relevant enforcement agencies when reviewing IP licensing arrangements.
D. Dynamic Versus Static Efficiencies
The role played by innovation in the information economy highlights the differences, and sources of potential conflict, between static and dynamic efficiency goals. Static efficiency is epitomized by the economist’s model of perfect competition, under which a large number of perfectly-informed suppliers compete with one another to supply an undifferentiated product to a large number of perfectly informed consumers. Dynamic efficiency is achieved as a result of the development of new products or new processes for producing existing products, through innovation and technological change. Innovation is antithetical to perfect competition to the extent that innovations are generally created with a view to developing new markets within which the innovator (for a certain period of time) hopes to be the sole supplier, rather than to assist the innovator to compete more effectively with other competitors within existing markets.
Antitrust authorities need to recognize and appreciate the types of innovation that may occur in a high innovation industry. For example, 'drastic innovation', which involves an innovation that is so superior to existing products and processes that a smaller new market entrant will be able to 'leapfrog' over the technological leader, can negate the alleged market dominance of a firmly-entrenched technological incumbent, even pushing the incumbent out of the market.  In that sense, significant market power by one firm may be quickly diluted. Further, some markets may be continually dynamic while others experience an innovative phase before maturing. These characteristics illustrate the need to understand and assess the nature of innovation before determining what impact a merger will have on dynamic and static competition.
As discussed above, advocates of the Schumpeterian school of thought believe that a more concentrated market will provide the profit incentive firms need in order to innovate  , meaning that a trade-off between static and dynamic efficiency can exist. This would suggest that antitrust authorities should refrain from making decisions in the interest of promoting static efficiency that may constrain competition in innovation/R&D markets. However, more recent theory and empirical evidence suggests that competition in the product market can enhance dynamic efficiency based on the notion that competitive firms will have a strong incentive to innovate so that they can 'escape' competition. Therefore, in order to determine whether a trade-off between static and dynamic efficiency exists, merger cases will need careful, in-depth analysis on a case-by-case basis, including determination of whether an innovation is drastic or non-drastic. 
E. Network Effects, Competition 'For the Market' and Tipping
The phrases 'network effects' or 'demand-side network externalities' are typically used to describe the manner in which the addition of further users of a product or technology will increase the value of the product or technology to the benefit of all other existing users.  For example, the very first telephone on a network would have no value at the time it was purchased because there would be no other telephones to call. However, each telephone subsequently added to the network adds to the value of the first telephone and all other telephones on the network, thereby conferring benefits and value to the other members of the network that are external to the purchaser of the additional telephone. The same type of network externalities arise whenever users adopt a product or technology which becomes more valuable to existing users and, perhaps, less expensive to provide.
Networks are not defined only by the interconnection of physical assets such as twisted copper pairs, fiber optic cables or radio spectrum. 'Virtual networks' are comprised of users of compatible technology who are not linked by a physical network, but who generate network externalities as a result of their adoption of the common technology of the network. The users of Windows-based computers, VHS cassette players, audio CDs, Playstation and Xbox games, MP3 files and the Internet protocol comprise examples of five highly successful virtual networks where the widespread adoption of the particular technology has increased the value to consumers of the compatible products which they had previously purchased in essentially the same manner as the increase in the number of telephone users increased the value of that network. 
Network externalities and the minimal marginal cost of duplicating information challenge fundamental economic assumptions that relate value to scarcity and increased quantity to diminishing marginal value. However, it is important to note that the principal beneficiaries of network effects, in most cases, are consumers rather than suppliers.
The effects of network externalities played an important role in the U.S. Department of Justice’s review of WorldCom Inc.’s acquisition of MCI Communications. In particular, the review focused on the 'Internet backbone' market and the effect of the merger on the Internet industry. The merger contemplated the combination of two of the four U.S. or world-wide Internet backbones used for wholesale Internet transmission services to retail internet service providers dedicated high-speed internet access for large multinational firms. Given the complex and highly technical web of independent relationships, and the dynamic nature of a market characterized by rapid technological change, it was difficult to define a relevant product market.  In particular, each of WorldCom and MCI, as a backbone provider, were dependent on each other and, therefore, had an incentive to support and implement efficient interconnections to ensure no degradation in quality that might result in losing customers to the other backbone networks. Following their merger and the integration of their networks, MCI/WorldCom would have far less need to depend on the other backbones than those backbones would have to depend on it. 
High innovation markets are often characterized by competition 'for the market ' (i.e., based on performance) rather than competition 'in the market' (i.e., based on price). Competition for the market results in highly differentiated products, and buyers who are, for the most part, unresponsive to a small post-merger price increases. Paul Geroski characterizes 'competition for the market' as follows:
At bottom, it is about establishing a standard and imposing it on the market. This means, in effect defining what the product is, what it works well with and what consumers should expect when they use it. A firm that is able to do this and keep control of that standard is often referred to as a ‘first mover’ and is usually thought likely to benefit from various ‘first mover advantages’ which protect it from imitative entrants who, in due course, try to compete in the market. In the context of a standard, first mover advantages arise whenever control over the design plus accumulated production experience or scale related cost savings give rise to cost differences between incumbents and entrants, when pre-emptive investments in product specific plant or in controlling scarce inputs raise entry hurdles for followers, and when network externalities create large collective switching costs for consumers who may, in any case, have formed habits of purchase or a degree of brand loyalty to the first mover. 
The exponential growth and positive feedback characteristics of markets with strong network externalities that are characterized by competition for the market can make such markets dynamically unstable and subject to 'tipping' in favour of the perceived market leader. 'Tipping' refers to the propensity for markets with network externalities to become 'winner take most' if not 'winner take all.' By definition, network externalities imply that, all else being equal, consumers would rather have the same product that most other consumers have. In these markets, any small early advantage that a seller might acquire, whether real or perceived, that leads consumers to think that it will have a plurality of sales, could translate into a dominant market share as everyone wants to jump on the same technological bandwagon.
For example, the purchasing decisions of computer users will be affected by their expectations and perceptions of which hardware, operating systems and other software will be the most popular and widely disseminated (and compatible with their customers and suppliers), and therefore the most effectively supported by wide varieties of innovative competitively-priced, compatible products and accessories. Purchasers will tend to buy the system they perceive to be the 'likely winner', while ignoring less popular (but not necessarily inferior) computer systems. Similarly, the expectations of software suppliers will shape their decisions on what support software products to develop, i.e., they will attempt to develop compatible software for which there is likely to be a significant potential market. 
Firms attempt to exclude competitors in order to 'tip' the market towards their products and shorten the period of time for 'intense' competition. Therefore, as Brennan observes, attempting to interpret pricing tactics and other activities when competition is for the market is not straightforward. Pricing below cost might be considered predatory in ordinary markets, but will be the expected 'first mover advantage' in the development of a tipping-prone market, when the expected result is that the winner will get a monopoly. Essentially, firms seeking future monopoly rents may compete them away by offering first-time below-cost discounts to early adopters. 
In some cases, the propensity of a market to tip may warrant additional antitrust vigilance. In ordinary markets, a firm may have to exert considerable effort to appropriate market share from its competitors. If the market is subject to tipping and characterized by network externalities, a firm’s tactic that creates what would normally be considered a relatively small and harmless competitive advantage could prevent competitive entry altogether. 
A related concern is that in a tipping-prone market, the 'right' monopoly will not necessarily emerge to be the next industry standard. For example, standards may last too long because of inertia as network externalities discourage users from departing from the incumbent standard, even if the adoption of the newer standard would be better for all concerned. Standards may also be adopted too easily if every consumer decides to adopt because they expect other consumers to migrate, regardless of whether the original standard objectively was the preferred standard. 
F. Low Barriers to Entry and Minimal Marginal Costs
The growth of electronic networks, in particular the Internet, has substantially reduced barriers to entry and continues to reduce the costs of processing, distributing and acquiring information through low-cost automation of these functions. The Internet and Internet protocol extranets are being used to create highly efficient markets for a vast array of products and services, including consumer-to-consumer auction sites (such as those operated by eBay), business-to-consumer Internet sites (such as Amazon.com) and sites which facilitate direct business-to-business sales of commodities, manufactured goods and all other forms of products and services. In addition, the principal productive resource required by information economy companies is skilled human capital rather than the costly physical facilities and raw materials typically required by companies operating in the industrial economy. 
The initial cost of creating information and information-based products is generally a sunk cost and the marginal cost of producing additional copies of such products is minimal. For example, much software is distributed at minimal cost and for no charge over the Internet or licensed for relatively nominal prices using 'shareware' products. The cost/performance ratios of other products and services that have a substantial information content are also continuing to fall. The falling costs of computer processors, memory and disk drives, and telecommunications services reflect the high percentage of low marginal cost information content incorporated in these products and services, scale economies and the other factors discussed in this section.
Minimal marginal costs mean that the average cost of information-based products will fall continuously with increasing volume. Therefore, at a conceptual level, continually declining average and marginal costs are inconsistent with the economic assumptions underpinning the classical, micro-economic analysis upon which domestic competition laws are based. In other words, marginal costs will not rise with increasing quantity to intersect the demand curve at an equilibrium price and quantity.
Open access to markets through efficient electronic networks such as the Internet has allowed both suppliers and customers to bypass more costly and constrained traditional distribution channels. This phenomenon is perhaps best demonstrated by the rapid development and worldwide distribution of successive generations of software products, such as media players, document viewers and browsers using the Internet.
Information and information-based products (such as software) are durable goods, which may in most cases be used indefinitely without any additional costs or expenditures. In order for a supplier to sell replacements for such products, users must be persuaded that the new features and enhancements incorporated in the replacement not only justify the cost of the upgrade, but are superior to other new and competitive products. Consequently, information economy suppliers are frequently in the position of having to compete against the prior generations of their own products, as well as the products and upgrades offered by their competitors.
In high innovation markets, antitrust authorities must determine what barriers to entry exist (i.e., what resources are necessary to innovate successfully, how quickly they can be obtained and assembled, and on what terms). Careful consideration must be given to how easy it would be for consumers to switch from the products of the merging parties to those of new entrants. It may be, for example, that network effects, when combined with significant switching costs, act as powerful barriers to entry to new competitors. On the other hand, it may be that the resources needed to successfully innovate can be obtained quickly by hiring away qualified staff from other firms.  Further, when assessing market power in the information economy, an important distinction must be drawn between the barriers to entry created by unique physical facilities and those which have been created by a widespread consumer preference for or acceptance of a particular product or service offering. While the former may give rise to circumstances in which market power is being exercised inappropriately and intervention is required, the latter should not be viewed as a source of market power, insofar as consumers were free to purchase an alternative product at the time of the original acquisition and are generally free to switch, and in practice will switch, to a superior product if one is offered.
Competition in high innovation industries typically depends upon the implementation of ideas that have little respect for geographic borders or entrenched market power. Therefore, competitive problems that arise in high innovation industries may be self-correcting through the rapid and seemingly perpetual introduction of new products, especially through 'drastic innovations'. 
G. Exclusion of Competitors: Vertical and Horizontal Integration
One of the most common ways that mergers in innovation-intensive markets can exclude competitors is through vertical integration and 'gatekeeping'. Innovation-intensive markets often involve complementary products in vertically-related markets. Dominant companies in one particular market may seek to merge with a company operating in a vertically-related market. The merged entity may be in a position to exploit complementarities between its own portfolio of products in order to foreclose the downstream market to other players. This will have a particularly significant effect where competition is for the market and may cause the market to 'tip' towards a firm that would not otherwise have gained a position of dominance. 
Gatekeeper effects and foreclosure in high innovation markets have been a major source of concern for the EC.  For instance, in the AOL/Time Warner case, the EC was concerned that AOL, through its existing European joint venture with Bertelsmann and its proposed merger with Time Warner (which in turn had planned to merge its music recording and publishing activities with EMI), would have controlled the leading source of music publishing rights in Europe. Further, the merger created the first Internet vertically-integrated content provider, distributing Time Warner’s branded content (music, news, films, etc.) through AOL’s network.  Because of AOL’s structural and contractual links with Bertelsmann, the EC determined that the merged entity would have preferred access to Bertelsmann’s large library of music. Therefore, by having de facto control over the leading source of music publishing rights in Europe in connection with its ubiquitous network, the merged entity could become the dominant player in the market for Internet on-line music delivery, effectively becoming the ‘gatekeeper’ and dictating the conditions for the distribution of music audio files over the Internet. A further concern was that the merged entity may be able to format Time Warner’s and Bertelsmann’s music such that it would be compatible with only AOL’s music player (Winamp), and not competing music players. (Winamp, however, would still have the ability to play the music of competing record companies using non-proprietary formats.) 
Competitors may also be excluded by horizontal mergers that use the enhanced position in the primary market to foreclose vertically-related markets. For example, Vivendi’s proposed acquisition of an interest in satellite pay-TV broadcaster BSkyB contemplated the merger of their two different conditional access technologies (SECA and NDS, respectively), which together would have accounted for more than 60 per cent of all pay-TV subscribers within the European Union.  Another example of foreclosure by horizontal integration is the prohibited MCI WorldCom/Sprint merger involving the combination of two operators’ networks. The EC’s investigation concluded that, through the combination of the two networks and customer bases, the merger would have led to the creation of a 'company of such absolute and relative size compared to its competitors that both competitors and customers would have been dependent on the new company for universal Internet connectivity'. As a result, the merged entity would have been in a position to behave independently (in respect of both its competitors and customers), control technical developments, raise prices and discipline the market by selective degradation of its interconnections with competitors. 
Complementarity (or compatibility) among components is crucial in high innovation economies such as the computer industry. As discussed above, computer operating systems and applications software that are compatible with each other form a 'virtual network', exhibiting increasing value per unit sold as total sales of compatible goods increase.  One of the benefits of complementarity is the standardization and interoperability among components, known as an 'open systems architecture.' A good example is the Intel/Windows PC structure, where many brands of computers conform to the same technical standards and can utilize the same operating system. Once an operating system is functional, users can take advantage of a variety of applications that are compatible with that operating system, thereby reaping the benefits of complementarity.
From the point of view of a high innovation firm, complementarity and compatibility can cut both ways. On the one hand, a firm can benefit from the externality of the total sales of all compatible firms. On the other hand, compatibility implies more similar products, and therefore more intense competition among the firms that produce compatible products. (To avoid the more intense competition, a firm may want to be incompatible with others.) As a result, it is society that reaps the benefits from the increased competition in an environment of compatibility, and not the firms. 
I. Role of Standards and Open Specifications
Competition in information economy markets frequently revolves around the development of a popular specification or standard, whether developed as a de facto standard by industry or as de jure standard by government. As is particularly apparent for the Internet, minimal marginal costs and the desire of many companies to establish their new products or services as a virtual network standard have led to a veritable explosion in the range of products and services which are offered at no charge over the Internet.
The acceptance of standards reflects certain product characteristics that competitors are willing to agree on for their mutual benefit of growing the size of the market. Characteristics which are not the subject of standardization are used by companies to differentiate their products and provide the basis for competition. 
The success of open specifications and standards, such as those underlying the global telephone and fax systems, the Internet and the IBM-compatible personal computer, have clearly demonstrated the positive network externalities that result from open access to specifications and networks, as well as the development of fully compatible and interoperable products. Increasing recognition of the economic value of ubiquity and the minimal marginal cost of reproducing information-based products are increasingly leading customers to further demand, and suppliers to provide, products that comply with clearly documented specifications and standards. Another result is the formation of industry coalitions and joint ventures to develop and support open standards in order to facilitate the development of larger interdependent networks and markets for new types of interoperable and compatible products. The adoption of industry standards provides manufacturers with some measure of assurance that compatible new products will find a ready market. Firms have no doubt also noted that delays in the development of a consensus regarding the technical specifications of a new type of product or service may impede or block the introduction of that product or service. 
Standard-setting has the potential to create market power and enhance the market value of a technology by reducing or eliminating the selection of close substitutes. Firms may seek to make their technologies the standard of choice by participating in 'competitions' (whether in the market (de facto) or before standard-setting bodies (de jure)). 
Ernest Gellhorn notes that, on the one hand, standard-setting can be an efficiency-enhancing tool and can support innovation and market competition by providing an agreed-upon 'base' for further product development and ensuring product quality, increasing output, fostering innovation and reducing prices. On the other hand, if misused, standard-setting can block entry and increase costs by excluding new innovative products and services, especially if such standards are incorporated in governmental codes.  While express standard-setting agreements among competitors may occur, abuse of the standard-setting process can be more subtle; achieved through nonpublic cooperation/coordination among incumbent producers (for example, through their trade associations).  Therefore, Donald Deutsche of Oracle Corporation recommends that, in order to achieve broad acceptance of standards and the resulting largest possible market, it is important to include all stakeholders (large and small and from all technical/philosophical camps) in the standards development process. As history has repeatedly shown, attempts to seek competitive advantage through standardization by precluding major stake holders result in failure – the ignored constituency can merely initiate a competing standards effort that confuses the market and ensures that there is more than one 'standard' option. 
While market definition and market power must remain pillars of antitrust analysis, caution needs to be exercised to ensure that the unique characteristics of high innovation markets are reflected in the market definition analysis and, hence, the resultant determination of market power. While antitrust authorities may rely upon traditional product and service market definitions for certain sectors of the information economy, more 'elusive' technology sectors may demand a more creative approach.
It would appear that the various sectors of the information economy can be described by three possible types of market definitions: (1) traditional product (goods) and services markets; (2) technology or IP markets; and (3) innovation markets. Indeed, the joint DOJ and FTC 1995 Antitrust Guidelines for the Licensing of Intellectual Property (the 'US IP Guidelines') address each of these three types of markets as they may relate to IP licensing arrangements:
If an arrangement may adversely affect competition to develop new or improved goods or processes, the Agencies will analyze such an impact either as a separate competitive effect in relevant goods or technology markets, or as a competitive effect in a separate innovation market.  [Emphasis added]
A. Traditional Product and Service Markets
Product and service markets are the markets with which antitrust has been traditionally concerned; they include such readily identifiable markets as pharmaceuticals, hardware, software, computer chips, telecommunications services, Internet services or computer services.
An example of the application of the product and service markets is the market for 'the provision of universal Internet connectivity', which was first defined in the EC’s review of the 1998 WorldCom/MCI merger. There, the EC found that the merged entity’s network would have created a dominant position in this market. When the merging parties notified their subsequent merger with Sprint in 2000, they argued that recent changes in the European wholesale Internet market negated any new claims of dominant position. These changes included the liberalization of the EC telecommunications markets (and the resulting increase in the number of European ISPs and content providers), content delivery techniques controlling the flow of Internet traffic, and lowered leased line prices. As a result of these technological developments, the parties argued that the Internet could no longer be considered hierarchical in nature and that European ISPs were no longer dependent on the largest Internet connectivity providers to obtain global connectivity. The EC’s investigation concluded, however, that none of the above factors had had any significant impact on the structure of the market - even the largest European Internet connectivity providers were still dependent on the top-level global connectivity providers and lacked the ability to impose any competitive constraints on these providers. 
In high innovation product and service markets, pre- and post-merger concentration levels may not always serve as useful indicators of a transaction’s possible effects on competition as they will only be able to capture static effects. For example, in biotechnology, pharmaceutical or next generation aerospace cases where the product has yet to reach the market, market share information will be of little relevance and use. Faced with an absence of proper concentration information, the FTC has stated that it may look at other factors, such as the merging companies’ investment levels, R&D progress, experience in previous generation products, and success in related markets.  Situations such as these have led authorities to adopt technology or innovation markets where competition is for the market rather than in the market, and market share data and concentration levels may not provide an accurate picture of actual market power.  Firms may offer temporarily low prices and incur nominal profits or even losses in an attempt to acquire the whole market and enjoy the all-important 'first mover advantages'. Under these circumstances, the application of a standard market definition approach may result in a misleadingly narrow market.
While market power is normally attributed to market concentration and barriers to entry, market power in the information economy depends on other unique factors, including the intensity of network effects, the size of the installed base, users’ switching costs, and the degree and ease of compatibility granted to competitors and their products. 
In some high-tech industries, network effects can make competition between different systems/networks fierce in a dynamic sense. As Stenborg argues, signs of relaxed competition using static measures (e.g., high concentration ratio, high price-cost margin, etc.) may in fact conceal vigorous competition in the dynamic sense. Further, conclusions drawn from static measures such as market share may result in a distorted conclusion of strong market power, even though dynamic competition is healthy and may ultimately control static market power. Indeed, the expectation of significant market power for some period of time is a necessary condition for dynamic competition. Under conditions of healthy dynamic competition, the existence of short-run market power is not a matter for concern. Instead it is a consequence of information economy characteristics such as network effects.  Accordingly, equating high market share with dominance can be potentially dangerous to innovation.  In this regard, the minimum antitrust authorities should be concerned with is the possibility of entry and other potential competition. 
Determining market shares was a challenge in the 1998 investigations of the U.S. DOJ and the EC of MCI/WorldCom because, at that time, there was no commonly-accepted method upon which to rely. The DOJ and EC examined market shares using a variety of methods, including: (1) shares of overall Internet industry revenues generated by ISPs connected to a specific backbone; (2) percentage of ISPs connected to a specific backbone versus the total number of ISPs connected to all of the backbones combined; (3) the volume of Internet traffic originating, terminating or transmitting on a backbone’s network; (4) the number and type of Internet points of presence (POPs) on a backbone’s network; (5) the number of circuits connecting customers to a backbone; (6) the density of a provider’s network and web of customers; and, finally, (7) the number, type and significance of each network’s customers. However, under any of the foregoing measures of market share (none of which was ideal), it was determined that MCI/WorldCom would be the dominant player and substantially larger than any other player. 
A further challenge faced by antitrust authorities is that market definition can be guided by the characteristics of future products that are at an advanced stage of development.  Therefore, market power not only refers to a strong position in today’s market, but also to strength in a future market characterized by present 'first mover' advantages and network externalities.  This results in current market power being maintained, rather than being transient. The existence of these dynamic effects makes it difficult to formulate a proper method to ascertain acceptable pre- and post-merger market shares.  In other words, a currently high market share is not necessarily indicative of the presence of market power today and a currently low market share may change to market power in the near future.
In the BSkyB/Hilton Group merger, the difficulties associated with relying on pre-merger market shares were described in the Director General’s advice to the Secretary of State for Trade and Industry:
Given that iDTV betting is a new and evolving market segment, market share figures are not reliable and it is more difficult to determine the likely effects of the JV than with a merger in a more mature market. 
Faced with these difficulties, antitrust authorities have turned to more 'innovative' methods of defining elusive and ephemeral high tech markets.
B. Technology or IP Markets
Some jurisdictions, including the U.S., EU and Canada, have introduced the concept of 'technology' or 'IP' markets, i.e., markets in which companies compete in the licensing of intellectual property, have been introduced by many antitrust authorities. For example, the U.S. FTC and DOJ will analyze the competitive effects in technology markets when rights to intellectual property are marketed separately from the products in which they are used. 
In the Draft EC Guidelines on the application of Article 81 of the EC Treaty to technology transfer agreements  (the 'Draft EC Technology Guidelines'), the EC describes the use of a separate 'technology market':
Technology is an input, which is either integrated into a product or a production process. Technology licensing can therefore affect competition both in input markets and in output markets. An agreement between two parties which sell competing products and which cross licence technologies used for the production of these products may restrict competition on the product market concerned. It may also restrict competition on the market for technology and possibly also on other input markets. For the purposes of assessing the competitive effects of licence agreements it may therefore be necessary to define relevant goods and service markets (product markets) as well as technology markets. 
Under this approach, technology markets consist of the licensed technology and its substitutes, i.e., other technologies customers could use as substitutes. Technology markets can be defined using the same principles as used for product markets. In other words, starting from the relevant technology, identify those other technologies to which customers could switch in response to a small but permanent increase in relative prices, i.e. the royalties.  Alternatively, the market for products incorporating the licensed technology may be used.
The Canadian Intellectual Property Enforcement Guidelines (the 'Canadian IP Guidelines'), which use the 'IP market', expands on the EC’s approach. They provide that the relevant market may be defined based on: (a) the intangible knowledge or know-how that constitutes the IP; (b) processes that are based on the IP rights; or (c) the final or intermediate goods resulting from, or incorporating, the IP.  This means that markets are not defined based on R&D activity or innovation efforts alone, but also the effects on price or output.
In order to violate the Canadian Competition Act, a firm must engage in anti-competitive conduct that 'creates, enhances or maintains market power', i.e., the ability to price independently of market forces.  This suggests that antitrust intervention should occur only where market power has been created, enhanced or maintained as a result of an arrangement between independent entities (whether in the form of a transfer, licensing arrangement or agreement to use or enforce IP rights), and not just from the mere exercise of an owner’s IP right. Under this principle, antitrust authorities would only intervene where an IP owner licenses, transfers or sells the IP to a firm (or a group of firms) that are (or would have been) actual or potential competitors without such an arrangement. 
An example of a U.S. case involving technology markets was the proposed joint venture of Montedison and Royal Dutch Shell.  There, the FTC was concerned that the joint venture would lessen competition in the licensing of polypropylene and polypropylene catalyst technologies. On a product market basis, the combined market share of the two firms for polypropylene production was relatively modest. However, the technologies controlled by the firms represented over 80% of production capacity pursuant to technology licence. 
Another example is Dow Chemical Company’s acquisition of Union Carbide.  In this case, the FTC required Dow to divest and license intellectual property for the production of polyethylene to BP Amoco, its former partner in developing the technology. The FTC alleged that the merging firms’ control over polyethylene production technology would lead to anti-competitive effects, both in the licensing of the technology and the market for polyethylene, as the two firms would control over almost all commercialized catalyst technology.
Calculation of IP market shares and market power incorporates many of the principles used in traditional product and goods markets. In Canada, market power is established in the context of many factors relevant to an innovation-intensive market, including the level of concentration, the existence of effective substitutes for the IP in question, barriers to entry, the rate of technological change and the ability of competitors to enter into the market by 'innovating around' or 'leap-frogging over' incumbent and dominant technologies.  Often, conditions of entry will be more important than market concentration. In other words, a high degree of market concentration is not sufficient to justify a conclusion that the transaction or conduct will create, enhance or maintain market power.
The Draft EC Technology Guidelines suggest that there are two methods to calculate market shares in case of technology markets: (1) on the basis of each technology’s share of total licensing income from royalties, representing a technology’s share of the market where competing technologies are licensed; or (2) on the basis of sales of products incorporating the licensed technology on downstream product markets. 
C. Innovation Markets
1. Innovation Markets Defined
In simple terms, innovation markets are markets in which firms compete in research and development. The 'innovation market' approach to antitrust analysis, which was first introduced by Gilbert and Sunshine  , attempts to introduce dynamic efficiency considerations into merger enforcement, recognize the importance of innovation as a means of non-price competition and a source of welfare gains, and prevent mergers that would reduce competition in innovation.
Gilbert and Sunshine believe the innovation market is necessary because a merger that reduces the level of competition in innovation will harm consumers through adverse changes to prices, product characteristics and the rate of new product introductions in downstream product markets. However, since innovation is hard to quantify, the innovation market approach focuses on the level of R&D expenditures, the primary innovation-creating activity of modern firms.  Therefore, under this approach, reductions in R&D expenditure are treated as an expression of market power. 
Many antitrust authorities believe that the innovation market approach is well-grounded in economic theory and that traditional antitrust analysis cannot adequately address anticompetitive effects of mergers on the incentive to innovate, particularly where there is no existing market. 
The 'innovation market' concept has been adopted in the US IP Guidelines, which define an innovation market as follows:
An innovation market consists of the research and development directed to particular new or improved goods or processes, and the close substitutes for that research and development. The close substitutes are research and development efforts, technologies, and goods that significantly constrain the exercise of market power with respect to the relevant research and development, for example, by limiting the ability and incentive of a hypothetical monopolist to retard the pace of research and development. The Agencies will delineate an innovation market only when the capabilities to engage in the relevant research and development can be associated with specialized assets or characteristics of specific firms.  [Emphasis added]
The Guidelines further explain:
A licensing arrangement may have competitive effects on innovation that cannot be adequately addressed through the analysis of goods or technology markets. For example, the arrangement may affect the development of goods that do not yet exist. Alternatively, the arrangement may affect the development of new or improved goods or processes in geographic markets where there is no actual or likely potential competition in the relevant goods. 
The application of the U.S. innovation market approach is limited to merger cases where specialized 'innovation assets' or capabilities can be identified. Under this approach, antitrust authorities must be able to identify 'markets in which R&D [is] directed toward particular new products or processes [that] require specific assets that are possessed by identified firms.'  Such specialized assets most often include physical assets, experience, production capability, and intellectual property.
In EU case law, the innovation market idea has not been used; rather, the effects on innovation are analyzed in an ad hoc manner, often using the notion of potential competition.  In the Draft EC Technology Guidelines, the EC states that it will attempt to confine itself to examining the impact of technology transfer agreements on competition within existing product and technology markets. It notes that only in a limited number of cases would it be useful and necessary to define and use innovation markets. Such cases may include where an agreement affects innovation aiming at creating new products and where it is possible at an early stage to identify R&D poles. In these cases, it can be analyzed whether, after the agreement, there will be a sufficient number of competing R&D poles left for effective competition in innovation to be maintained. 
2. Criticisms of the Innovation Market Analysis
Beyond the expected additional cost of enforcement both to antitrust authorities and to merging firms in R&D-based industries, there is a considerable body of criticism as to why the innovation market concept is not workable.  The discussion below summarizes some of these criticisms.
(a) The state of knowledge regarding innovation is insufficient to permit the assessment of the innovation effects of mergers. - R&D competition is generally more complicated than price competition, and the incentives, path of progress and outcomes are much harder to predict. Rapp states that the basic problem is a lack of means for judging whether innovation is harmed or served by a merger when there are as yet no products to evaluate: 'The more removed the R&D projects in question are from issuing commercially viable products or processes, the less predictable and more fragile is their state. Just where the innovation market approach is most apt, it is most dangerous.' 
(b) The connection between concentration and output on innovation market is problematic - Innovation market theory is based on the beliefs that (1) an increase in R&D concentration is likely to reduce the level of R&D undertaken and (2) a reduction in the level of R&D is likely to diminish innovation. However, Rapp believes that there is little basis in fact or theory for these beliefs. While it may be accepted that smaller firms in unconcentrated markets which do not have the economies of scale to support R&D efforts may not engage in R&D, a monopolist will generally have less incentive than a new entrant to engage in R&D that may lead to a substitute for an existing product or that may lower the cost of producing an existing product. This is because such R&D efforts may cannibalize the monopolist’s supra-competitive rents and make its current sunk investments obsolete. 
Several studies have attempted to find statistical relationships between firm size or market concentration and various measures relating to R&D and innovation (e.g., R&D expenditure, R&D productivity, patent counts and counts of significant innovations). However, the conclusions appear to be that there is no correlation between the number of firms engaging in related R&D projects and the level of total R&D or innovation activity in the market.  Further, there are many historical business examples of large firms with significant R&D 'assets' losing the competitive race for innovation to much smaller firms with limited R&D resources. 
(c) R&D expenditure is not a suitable proxy for innovation - Innovation is intangible, uncertain, unmeasurable and often even unobservable, except perhaps in retrospect. There are no market transactions in innovation, only in the inputs - the labor of scientists, engineers and entrepreneurs and the capital to build research facilities and to fund R&D in advance of returns - and the outputs - technology and products. Thus, R&D expenditure is an input to the process of innovation.  To measure market shares in terms of R&D expenditures or R&D capacity requires an assumption that innovation and R&D are indeed the same thing and that there is a positive functional relationship between the rate of R&D expenditure (or the amount of R&D capacity) and the quantum of innovation produced by a firm. 
Rapp argues that the level of R&D expenditure is a poor proxy for the level of innovation at the market level; there are several market factors that influence a firm’s R&D expenditures, including diminishing returns to R&D investment, the opportunity cost of funds and expectations about the appropriability of the gain. Even if one could predict with confidence that a merger would lead to a reduction in R&D expenditure, there is no principled basis for saying whether this is good or bad for competition and innovation  Rapp concludes that, since there is no principled way to distinguish good cutbacks in R&D from bad ones, an innovation market approach, even performed on a case-by-case basis, would be a mistake. 
(d) The capacity to innovate is hard to monopolize - There is little in the history of modern technology to suggest that firms are able to monopolize innovative capacity. Although patented technology can be monopolized, the normal inputs to R&D - research scientists, engineers, software developers, laboratories, computer centres, etc. – may be acquired by competitors. If the main inputs to innovation are continually on the market for sale, there is no opportunity to corner the market for innovation. 
3. Product and technology markets are sufficient to address all significant competitive harm
The general consensus appears to be that there are, in fact, very few cases in which the innovation market approach should be used. First, cases where there is a clear competitive overlap between the merging firms at the product market level, the innovation market analysis can presumably be dealt with using a traditional product market analysis. Second, for cases where one or the other firm is a potential entrant into the product market of the other (with both entrant and incumbent engaging in R&D), traditional 'potential competition' analysis will generally suffice. Presumably, firms that are engaged in overlapping R&D projects can threaten to enter the product markets in the overlapping areas of R&D. In other words, R&D makes one or both firms either potential competitors or entrants in conventional product markets. Since the release of the US IP Guidelines, the FTC has brought complaints against several additional mergers on the grounds that innovation markets would be harmed and has obtained relief either in divestiture or compulsory licensing. However, many, including the US FTC, believe that, given the nature of the direct product market competition in these cases, the enforcement actions employing the 'innovation market' concept could have legitimately been brought using the traditional analytical tools such as potential competition theory, without resort to innovation market theory. 
Perhaps the only situation that warrants the innovation market approach is where no products have emerged and the only competition subject to analysis is R&D competition. Because there are as yet no actual downstream products, the assessment of the anti-competitive effects must be based upon a determination of whether the combination will advance or retard technical progress. 
An example of properly-used 'innovation markets' is the U.S. FTC’s review of American Home Product’s 1994 acquisition of American Cyanamid.  In this case, the FTC was required to allege that competition in the market for the R&D of a Rotavirus vaccine would be harmed since neither of the merging firms manufactured a Rotavirus vaccine at the time of the merger. Since the product market was non-existent, no argument for potential competition could be made.
The innovation market approach attempts to capture the principles of product market analysis but, in doing so, misses the mark. We note that in product markets, the optimal product price is based on marginal cost and any increase in price away from this benchmark will be bad for consumers. On the other hand, in innovation markets, the optimal level of R&D is unknown. Moreover, it is impossible to predict the effect of a change in the structure of an innovation market on the level of R&D activity. 
Many of the jurisdictions participating in the OECD Roundtable, including the EC, the UK, Australia, the Netherlands, Japan and Korea, suggested that the traditional competition law principles can be applied to high innovation industries and there is no need to adopt or develop a new set of principles or analyses. For example, the EC considers that the general nature of its merger rules when applied to the factual circumstances of a particular case allows the rules to keep pace with technological developments in a way that more specific regulatory frameworks cannot, regardless of how quickly industries develop or change.  ) The UK’s non-statutory merger guidelines do not refer to specific arrangements, approaches or checklists to deal with mergers in high innovation markets.  In the UK’s view, it would be difficult and confusing to have different and rigid approaches to merger review, i.e., one for static industries and one for innovative industries.
The Netherlands Competition Authority does not favour a modified approach to high-tech mergers because it is difficult to define the situations under which such an approach would be apply – the following significant obstacles would have to be overcome: (1) the criteria to be applied would have to be clear and transparent; (2) the situations and relevant markets to which they apply would have to be determined; and (3) given that every transaction is different, one would have to decide if a specific deal qualifies according to the specified criteria. 
The overly ambitious application of competition law can deter socially desirable investments in innovation and technology transfers; therefore it is important that antitrust agencies understand the likely impact of their enforcement efforts on innovation markets. As stated above, where innovation is a key determinant of competition, it should not be inadvertently impeded by the intervention and enforcement actions of antitrust authorities. In other words, competition policy should be used to promote, not hinder innovation. As U.S. Federal Reserve Chairman, Alan Greenspan, was quoted in a 17 June 1998 article in the Wall Street Journal, 'I would feel very uncomfortable if we inhibited various different types of mergers or acquisitions on the basis of some presumed projection as to how markets would evolve.' Mr. Greenspan went on to state that, particularly in technology industries, 'history is strewn with people making projections that have turned out to be grossly inaccurate.' Therefore, as noted by many of the countries in the OECD Roundtable, a somewhat customized approach is needed in high innovation markets, requiring a case by case approach, with due consideration to standard concentration measures and barriers to entry.
Given the inherent uncertainties in predicting how a market is going to evolve, remedies for mergers in innovation-intensive industries can be particularly difficult to design and the precise effect the remedy is largely uncertain.  Appropriate remedies in high innovation markets should therefore be customized to consider the specific nature of the markets concerned. An example of a custom-made remedy can be found in the EC’s review of the Vodafone/Mannersmann case, where the Commission was faced with the problems associated with the creation of a seamless pan-European mobile telecommunications service market and the markets for pan-European wholesale roaming. These concerns were addressed by undertakings provided by the merged entity designed to give other mobile operators the possibility of providing their services using the integrated network of Vodafone Airtouch/Mannesmann.  This remedy is particularly interesting because due to anticipated developments in the sector, it was limited to a period of three years. This case illustrates that antitrust authorities do not have to accept situations of temporary market power and they can respond with temporary remedies.
One possible remedy to a 'network monopoly' is to break up the network. However, this is a drastic measure that may not be feasible and may result in substantial injury to consumer welfare.  Another remedy is to mandate the network operator to provide access to the network for all qualified suppliers, customers and competitors on fair, reasonable and non-discriminatory terms (as has been imposed on 'bottleneck monopolies' such as local telecommunications facilities or cable televisions high-speed Internet network).  However, there is still a need for antitrust enforcers to proceed cautiously in breaking up or mandating access to an existing network, even when that network is dominant, until the origins and significance of network efficiencies are clearly understood. Rather, antitrust authorities should concentrate on their traditional role of ensuring that network monopolies are achieved through legitimate competitive conduct and that network dominance is maintained only through superior skill, foresight and industry, and not by unnecessarily exclusionary conduct. 
To minimize any adverse effects on innovation, the focus of antitrust remedies should, where appropriate, shift from structural correction to behavioural correction and monitoring. In its submission to the OECD Roundtable, Brazil indicated its preference to use time-limited 'contingent' remedies: 'In general, due to the monitoring costs incurred in the enforcement of behavioural remedies, antitrust agencies favour structural ones. However, innovative markets are a dangerous terrain to keep allegiance to traditional practices, because the changing nature of technology can easily overcome any apparent market power arising in the short run.'  In Canada in circumstances where the potential efficiencies and anti-competitive effects of a merger are very difficult to ascertain and weight, a merger may be assessed ex-post using behavioural monitoring, coupled with a specific right of the competition authority to take further action (including ordering divestiture) should anti-competitive effects appear in future. In the appropriate circumstances, this is preferable to taking immediate and drastic action that may have the effect of inhibiting innovation and market growth.
In summary, while the authors of this paper take the view that the principles of traditional antitrust and competition law may be applied to high innovation markets, they must be informed by, and customized to take into account, the realities and sensitivities of these markets. Accordingly, it is suggested that antitrust authorities apply the following principles when dealing with high innovation markets. In doing so, they should be guided by the following two questions. First, are the resources being applied to protect the process of competition or a particular competitor/complainant? Second, is the application of enforcement resources in the particular instance truly in the interests of consumers? 
(a) When deciding whether innovation competition will be harmed, the authorities must first understand the market-specific, firm-specific facts about innovation. The drivers of innovation must be identified, whether they are participants, customers, component suppliers, collaborators or independent R&D entities. Further, the nature of the current market players must be understood and whether they are 'leaders' or 'followers' in innovation.
(b) Antitrust enforcement should be careful to avoid deterring research or innovative activities by inappropriately interfering with IP rights.
(c) Antitrust authorities must determine if a merger will affect incentives to perform R&D, if overall R&D spending will decrease, and whether the merger will affect the diversity of likely innovations.
(d) The authorities must be cognizant of the rapid rate of innovation and technological change which is resulting in competitors entering areas in which they did not previously compete, and leading to the development of new systems and markets which are competing with and displacing existing market leaders, entire markets or technologies.  Since antitrust considers predictions of future effects, the more uncertain the future is, the greater the need for caution rather than intervention. Similarly, the more dynamic the market, the more certainty should be required before intervention. 
(e) Antitrust analysis has traditionally focussed on high market shares in the context of 'perfect competition' featuring a large number of relatively small firms. Static, perfectly competitive market structure cannot persist in many information industries. The promotion of innovation and the resulting long-term dynamic efficiency benefits should be favoured over static efficiency goals and short-term market effects.
(f) Consumers should not be prevented from enjoying the benefits of the network externalities resulting from the widespread consumer adoption of common systems and standards even though this may result in universal acceptance of a single product and potentially market dominance). 
(g) Where network externalities and the effects of 'tipping' appear to render an industry closer to oligopoly or monopoly, the authorities must exercise caution and not act based on imprecise or contradictory information regarding emerging technologies.
(h) The marketplace should determine which new features and products should be introduced to the marketplace. Similarly, the marketplace should determine which standards succeed and whether open standards are to achieve widespread acceptance.  As the focus of competition will shift from competition among existing products to competition to develop new systems and standards which are of value to consumers, consumers should not be prevented from enjoying the benefits of competition between competing systems and standards and competitors should not be protected from the adverse consequences of consumer decisions to adopt competing systems and standards. 
(i) The authorities must take into account, where applicable, the low barriers to entry associated with the intangible nature of the underlying products, the almost costless deliveries of products and services over the Internet and the absence of any physical restraints or constraints on the ability of competitors to quickly duplicate and deliver competing products.
(j) The authorities must be wary of the gate-keeping effects of vertical integration and foreclosure.
(k) The concept of innovation markets should be used sparingly and not where traditional concepts of product markets, IP markets and potential competition can be applied.
Finally, as a passing thought, we draw attention to the borderless features of technology and global networks, as well national protection of IP rights and technology licences, requires caution and co-operation. The globalization of markets and the movement of commerce into cyberspace may be expected to increase the occurrence of extra-territorial anti-competitive conduct that adversely affects the domestic markets of other countries. Any attempts to constrain such conduct become difficult where IP rights and related technologies are licensed from abroad and are subject to the laws of the foreign jurisdiction. Interjurisdictional issues relating to inter-agency cooperation, the exchange of confidential information and the requirement that national agencies work together in a spirit of positive comity must be addressed more frequently and expanded to high-tech industries subject to IP licensing. 
The appropriate application of competition law in high innovation markets requires that antitrust authorities understand the characteristics of these markets and the rationale for and economic implications of the particular conduct or transaction under consideration. The movement toward IP and information as a source of value, network effects and the other characteristics of the information economy, require that antitrust authorities focus on dynamic change and long-term increases in efficiency rather than short-term static efficiency gains. The rapid rate of change in such markets and the inability of the authorities to assess or predict the likely market impact of new inventions strongly suggests that they should err on the side of caution in order to avoid unnecessarily generating high cost errors which, in the end, could do more harm to competition in the marketplace by pre-empting innovation and the development of standardized products which would be of great value to consumers.  If antitrust authorities move too quickly without a full appreciation of the potential future consequences of any challenge, be it to a merger or other non-criminal conduct, there can be a chilling effect on the market, which can lead to significantly less investment, innovation and growth and a very real loss of potential benefits to consumers. It can also cause potential entrants to be wary of investing in the market, unless and until the uncertainty is corrected.
 The resultant issue paper, 'Merger Review Emerging High Innovation Markets', DAFFE (2002) 20 ('OECD Innovation Paper'), was published on January 24, 2003 and is available at:
 Allocative efficiency is achieved when suppliers are selling products at a price that is equal to the marginal cost of those products. Micro-economic theory predicts, in a competitive market in which numerous small suppliers are competing to sell a scarce but undifferentiated product to a large number of well-informed customers, that the market will converge at a price equal to marginal cost. By way of contrast, a monopolist will increase the price charged until the marginal revenue equals the marginal cost, thereby capturing the maximum amount of monopoly rents and, incidentally, creating dead weight losses due to the fact that certain consumers have purchased less appropriate alternatives or foregone consumption due to the unnecessarily high prices.
 Innovation increases economic activity and long-term efficiency by creating new products and services which are superior to and/or less expensive than existing alternatives and more efficient processes for the manufacture and/or delivery of products and services. For example, the use of packet-switching technology such as that employed by the Internet has facilitated the development of innumerable new products and services and has given users access to communications services for a small fraction of the cost associated with the use of dedicated private circuits to perform the same function.
 D. Teece and M. Coleman 'The meaning of monopoly: antitrust analysis in high-technology industries' (1998) 43:3-4 The Antitrust Bulletin ('Teece and Coleman') 801 at ·.
 United States v. Microsoft Corp., 147 F.3d 935 (D.C. Cir. 1998).
 Id. at 948-952.
 R. Pitofsky, 'Antitrust Analysis in High-Tech Industries: A 19th Century Discipline Addresses 21st Century Problems' (Prepared Remarks for the American Bar Association, Section of Antitrust Law's Antitrust Issues in High-Tech Industries Workshop, 26 February 1999) ('Pitofsky'), available at
 See, for example, C. Ahlborn, D. Evans and J. Padilla, 'Competition Policy on Internet Time: Is European Competition Law up to the Challenge?' (2001) European Competition Law Review 5; T. Bresnahan, 'Competition, Cooperation, and Predation in Innovative Industries' (Paper presented at 3rd Nordic Competition Policy Conference, Stockholm, 2000); D. Evans and R. Schmalensee, 'Some Economic Aspects of Antitrust Analysis in Dynamically Competitive Industries', NBER Working Paper W8268 (2001).
 Teece and Coleman at 845.
 J. M. Jacobson, Do We Need a New Economy Exception for Antitrust?' (Fall 2001) Antitrust ('Jacobson'), available at http://www.akingump.com/docs/publication/6.pdf.
 See J. Schumpeter, Capitalism, Socialism, and Democracy (1942) at 81-106 (1942). See also R. T. Rapp 'The Misapplication of the Innovation Market Approach to Merger Analysis', (1995) 64 Antitrust L.J. 19 at 22 ('Rapp'), available at http://www.ftc.gov/opp/global/rapp2.htm.
 M. Stenborg, 'Do We Need New Competition Policy in the 'New Economy'?' 2002(2) The Finnish Economy and Society 49 at · ('Stenborg'), available at http://brie-etla.org/_pdf/stenborg_fes_2002-2_pp49-60.pdf.
 234 F.3d at 48-50.
 Others however, have suggested that the problem was not really the 'new economy' aspects of software integrations, but the per se tying rule itself. See J. M. Jacobson & A. Qureshi, 'Did the Per Se Rule on Tying Survive Microsoft?' (14 August 2001) N.Y.L.J. at 1.
 OECD Innovation Paper, Contribution from the European Commission ('EC Contribution') 161 at 161.
 OECD Innovation Paper, Contribution from the United Kingdom ('UK Contribution') at 128.
 See G. Roberts and J. Putnam, 'Allegations of Harmony Between IP and Competition Policies: In Search of the Lost Chord' (Address to the Canadian Bar Association Competition Law Conference, Ottawa, 1 October 1999) at 3.
 'Non-drastic innovation' will simply involve a step change to permit technological laggards to catch-up but not to overtake the incumbent technological leaders. UK Contribution at 134.
 This is a different argument to the need for high profits in existing markets in order to fund innovation.
 UK Contribution at 127.
 J. Rohlfs, 'A Theory of Interdependent Demand for a Communications Service', (1974) Bell Journal of Economics and Management Science 5.
 To expand on a few of these examples, the success of Windows-based computers has greatly expanded the range and ability of, and continues to reduce the price of, compatible hardware, software and services available to the owners of these computers; the ubiquity and dominance of VHS cassette players created a massive market for compatible media and video rentals (now being replaced by another virtual network of DVD players); and the success of the Internet has greatly increased the value of e-mail services, HTML pages and other Internet facilities.
 OECD Innovation Paper, Contribution of the United States ('US Contribution') 141 at 144-145.
 Id. at 145.
 P. Geroski, 'Competition for Markets', in Finnish Competition Authority (2002) Workshop on Market Definition – Compilation of Papers from the 4th Nordic Competition Policy Conference (Helsinki, Finland, 5 October 2001) at 66.
 T. J. Brennan, 'Do Easy Cases Make Bad Law? Antitrust Innovations or Missed Opportunities in United States v. Microsoft' AEI-Brookings Joint Center for Regulatory Studies (May 2002) at 46. Also published in (2002) 69:5-6 George Washington University Law Review ('Brennan').
 However, tipping may be of limited competitive significance in most cases in the information economy as low barriers to entry, minimal marginal costs and positive feedback effects permit a superior new product or service to rapidly displace an inferior, incumbent product, notwithstanding the initial success and widespread acceptance of the inferior product.
 Brennan at 46.
 Studies comparing software producers to other manufacturing companies have found that the percentage of total revenues accounted for by labour costs are approximately five times higher for software producers than for other manufacturers. Conversely, the cost of goods sold expressed as a percentage of total revenues is more than three times as great for manufacturing firms as it is for software producers.
 Secretariat Background Note at 27-28.
 See Pitofsky. For example, the acquisition by Microsoft of a 24% interest in the UK cable operator, Telewest, involved two vertically-linked innovative sectors; the supply of software for digital TV set boxes and the supply of cable delivery platforms. It was not possible to assess the case by using references to 'potential' competition, as it was uncertain as to how the industry would develop and how exactly potential competitors would enter the market. Entry through innovation in this quickly evolving industry could change not only with respect to the number of competitors in the market, but with respect to the relevant market itself. Therefore, the static idea of a 'contestable market' is not necessarily relevant nor easy to apply in these circumstances. UK Contribution at 131.
 UK Contribution at 130-131.
 OCED Innovation Paper, Summary of the Discussion ('OECD Summary') at 170-171.
 EC Contribution at 164.
 UK Contribution 131.
 EC Contribution at 163.
 N. Economides, 'Competition and Vertical Integration in the Computing Industry' in Competition, Innovation, and the Role of Antitrust in the Digital Marketplace, J. A. Eisenach and T. M. Lenard, eds. (Kluwer Academic Publishers, 1998) (Presented at the conference Competition, Convergence, and the Microsoft Monopoly: The Future of the Digital Marketplace, Progress and Freedom Foundation, 4 February 1998) ('Economides') at 2.
 D. Deutsch, 'Intellectual Property Strategies in Standards Activities' Remarks before the FTC/DOJ Hearings: Competition and Intellectual Property Law and Policy in the Knowledge-Based Economy, (18 April 2002) ('Deutsch') at 2.
 For example, delays in the development of widespread support for one of the alternative specifications for digital radio and high definition television have substantially delayed the introduction and widespread adoption of these technologies.
 By contrast, when the invention would dominate the alternatives in a technology market on its own inherent merits, ratification of the market outcome by formal standard setting is an afterthought; it changes nothing. R. Rapp and L Stiroh, 'Standard Setting and Market Power' (FTC/DOJ Hearings: Competition and Intellectual Property Law and Policy in the Knowledge-Based Economy, 18 April 2002) ('Rapp & Stiroh') at 3-4.
 Prof. E. Gellhorn, 'Standard-Setting' (Remarks before the FTC/DOJ Hearings: Competition and Intellectual Property Law and Policy in the Knowledge-Based Economy, 18 April 2002) ('Gellhorn') at 1.
 Id. at 2.
 Deutsche at 2-3.
 US IP Guidelines at ¶ 3.2.3.
 EC Contribution at 162.
 US Contribution at 143.
 UK Contribution at 129.
 Stenborg at 55.
 Id. at 57.
 C. Ahlborn, D. Evans and J. Padilla, 'Competition Policy on Internet Time: Is European Competition Law up to the Challenge?' (2001) European Law Review 5.
 Stenborg at 57.
 US Contribution at 143.
 EC Contribution at 163.
 UK Contribution at 129.
 US IP Guidelines at ¶ 3.2.2.
 Available at http://www.les-ch.ch/Media/TTBE_Guidelines.pdf.
 Draft EU Technology Guidelines at ¶ 18.
 Id. at ¶ 20.
 For example, a merger between two firms that individually license similar patents to various independent firms that, in turn, use them to develop their own process technologies may reduce competition in the relevant market for the patented know-how if the two versions of that know-how are close substitutes for each other, if there are no (or very few) alternatives that are close substitutes for the know-how and if there are sufficient barriers that would prevent the development of conceptual approaches that could replace the know-how of the merging firms. This last condition may hold if the scope of the patents protecting the merging firms’ know-how is sufficiently broad to prevent others from 'innovating around' the patented technologies, or if the development of such know-how requires specialized knowledge or assets that only the two merging firms possess and that potential competitors could not develop or obtain in less than two years. Canadian IP Guidelines at 11.
 However, there will be no violation if market power was attained solely by possessing a superior quality product or process, introducing an innovative business practice or other reasons for exceptional performance. Canadian IP Guidelines at 12.
 For example, the Canadian Competition Tribunal has held that the mere exercise of the IP right to refuse to license was not an anti-competitive act: competitive harm must stem from something more than just the mere refusal to license. However, conduct that goes beyond the unilateral refusal to grant access to the IP could warrant enforcement action. For instance, if a firm acquires market power by systematically purchasing a controlling collection of IP rights and then refuses to license the rights to others, thereby substantially lessening or preventing competition in markets associated with the IP rights, the Competition Bureau could view the acquisition of such rights as anti-competitive. Canadian IP Guidelines at 7.
 Montedison S.p.A., 119 F.T.C. 676 (1995).
 The Dow Chemical Company and Union Carbide Corp., FTC Dkt. No. C-3999 (15 March 2001).
 Canadian IP Guidelines at 12.
 The EC considers the second approach as a good indicator of the strength of the technology. First, it captures any potential competition from undertakings that are producing with their own technology and that are likely to start licensing in the event of a small but permanent increase in the price for licenses. Secondly, even where it is unlikely that other technology owners would start licensing, the licensor does not necessarily have market power on the technology market even if he has a high share of licensing income. If the downstream product market is competitive, competition at this level may effectively constrain the licensor. An increase in royalties upstream affects the costs of the licensee, making him less competitive, causing him to lose sales. A technology’s market share on the product market also captures this element and is thus normally a good indicator of licensor market power. Draft EC Technology Guidelines at ¶ 21.
 R. J. Gilbert and S. C. Sunshine, 'Incorporating Dynamic Efficiency Concerns in Merger Analysis: The Use of Innovation Markets' (1993) 63(2) Antitrust Law Journal, 569-601 ('Gilbert & Sunshine').
 Rapp at 20.
 Gilbert & Sunshine at 594-597.
 Morse at 6.
 US IP Guidelines at section 3.2.3.
 Gilbert & Sunshine at 596.
 Stenborg at 57.
 Draft EC Technology Guidelines at & 22.
 See generally Rapp and Morse. See also G. A. Hay, 'Innovations in Antitrust Enforcement' (1995) 64 Antitrust Law Journal 71 and R. Hoerner, 'Innovation Markets: New Wine in Old Bottles', id at 49.
 Rapp at 45.
 Monopolists may indeed engage in incremental innovations to existing products and processes, quickly copy innovations introduced by smaller rivals, or engage in other defensive R&D, but are less likely to pursue 'disruptive technologies' that threaten their products and dominance. Therefore, empirical studies show that 'drastic' or 'paradigm-shifting' innovations are most often created by niche firms and new entrants. See Morse at 6.
 Rapp at 28.
 R. B. Starek, III, Commissioner, Federal Trade Commission, 'Innovation Markets in Merger Review Analysis: The FTC Perspective' (Prepared Remarks before Continuing Legal Education Seminar, The Florida Bar, Orlando, 23 February 1996) ('Starek'), available at: <http://www.ftc.gov/speeches/starek/orlando.htm>.
 Rapp at 36.
 Id. at 27.
 The evidence from financial markets is that firms often invest in projects - including R&D projects - which have negative discounted cash flows, i.e., 'losers.' One reason is imperfect foresight: managers cannot always tell beforehand which investments will pan out and which will not. Miscalculation and dashed hopes, however, are not the only reason for wasteful R&D investment. See Rapp at 34.
 Id. at 36.
 Gilbert & Sunshine at 570, 599.
 See, for example, US Contribution at 149, Morse at 8 and Rapp at 37-46. Examples include: the FTC’s 1997 challenge of Ciba-Geigy’s proposed acquisition of Sandoz, Baxter Int’l, Inc., Dkt. C-3726, 123 F.T.C. 904 (1997); Upjohn Co., Dkt. C- 3638, 121 F.T.C. 44 (8 February 1996); the 1997 transaction between Digital Equipment Co. and Intel and the Halliburton/Dresser merger; Glaxo plc., File No. 951 0054 (16 March 1995); Sensormatic Electronic Corp., File No. 941 0126 (4 January 1995); Wright Medical Technology, Inc., C-3564 (23 March 1995); Boston Scientific Corp., File No. 951 0002 (24 February 1995) Shell-Montedison, File No. 941 0043 (11 January 1995).
 Rapp at 44.
 American Home Products, 119 F.T.C 217 (14 February 1995)
 Rapp observes that an R&D project can have four different results: (1) a paradigm-shifting blockbuster innovation that transforms whole technologies and markets; (2) a curve-shifting innovation that saves substantial costs or creates important new products; (3) a minor improvement over current technology; or (4) a waste of money, none of which can be predicted with certainty. Rapp at 47.
 EC Contribution at 161.
 UK Contribution at 125.
 OECD Summary at 168.
 UK Contribution at 134.
 EC Contribution at 164.
 Denial of mandatory access may encourage applicants to challenge the dominant firm by establishing their own competitive network or joining smaller rival networks. However, these solutions may be no longer viable where an incumbent network passes a 'tipping point' leading to a single dominant network. Pitofsky. In this scenario, there will be no viable alternative networks that can be organized or joined.
 OECD Innovation Paper, Contribution from Brazil at 76.
 These points are addressed more fully in an article by William Gates, Chairman and CEO of Microsoft, which was published in The Economist on 13 June 1998. The detrimental effects on consumers that are likely to result from inappropriate antitrust intervention in technology markets are also discussed in an article written by Professor Paul Kedrosky of the University of British Columbia, which was published in the Wall Street Journal on 10 June 1998.
 See US Contribution at 146.
 This characteristic of the information economy is demonstrated by the rapidity with which apparently dominant software products can rise and fall in popularity as better products are developed and quickly take their place; for example, Wang word processing systems, Visicalc, Wordstar, dBase, Lotus 123 and earlier video game products.
 Lind argues in favour of adopting a 'substantial lessening of competition' threshold for merger review in high innovation markets: '…it provides a more useful focus than that of dominance because often markets will have a dominant firm in any case. For many mergers in the new economy the relevant question is whether a merger will strengthen or weaken competition in the process of determining the dominant firm.' R. Lind, P. Muysert and M. Walker, 'Innovation and Competition Policy' (Office of Fair Trading, Economic Discussion Paper 3: A report to the Office of Fair Trading prepared by the Charles River Associates, March 2002) at 129.
 As discussed by the D.C. Circuit Court in Microsoft, any dampening of innovation which resulted from courts overseeing product design would be at cross-purposes with antitrust law.
 These benefits may include below-cost pricing and other incentives offered by competitors, each of which is seeking to have its product or system adopted as the basis of a virtual network. The desire of suppliers to capture some of the benefit of potential network externalities means, in many cases, that competing innovators will in effect be 'bidding for the field' in an effort to confer network externality benefits upon their customers and thereby ensure the success of their products. Such conduct is not only economically rational but also inevitable and desirable in view of the economic characteristics of the information economy.
 The OECD and the International Competition Network are making considerable efforts in these areas.
 This is not markedly different than the principle that antitrust authorities have recognized over the years of applying considerable caution when deciding whether to challenge mergers or other conduct that involves a prospective analysis of likely effects.