TY - JOUR
AU - Costa-Cabral,, Francisco
AB - Abstract Innovation has so far been handled by competition law according to market structure, that is, by assuming that market power also allows undertakings to evade the competitive pressures that spur innovation. This structural approach has fitted innovation into a tried-and-tested analytical and normative framework. Its limits have nonetheless become apparent as competition law is increasingly hemmed in by a static outlook and is called on to apply no harm to innovation unrelated to market power. As such, this paper proposes complementing a structural approach with two advances from strategic management studies. The first advance is the ‘resource-based view’, which connects competitive advantage with firm heterogeneity. Since undertakings do not have the same capabilities, the exit of innovators from the market might not be compensated for by the entry of equally innovating undertakings even if barriers to entry are low. Harm to innovation is thus centred on assets granting ‘innovation capabilities’, such as intellectual property or pipeline products. Cases of abusive refusal to license and mergers of parallel research show that rival claims over these assets are to be resolved based on differences in those innovation capabilities. The second advance is the theory of disruptive innovation, which explains major changes in consumer preferences and production methods. Strategic management has established that an inefficient start is an integral part of disruption, allowing disruptors to be ignored until their productive efficiency increases enough to shift the market. This contrasts with the notion of competition on the merits allowing the exclusion of less efficient competitors. Competition law must therefore adapt to strategies which do not show an effect on market structure, notably the higher prices of market power, but which are aimed at preventing disruptive innovation from occurring. I. Introduction ‘Prediction is very difficult, especially about the future.’ Niels Bohr1 It is an article of faith of European Union (EU) policy that almost any problem, from economic growth to climate change, can be solved by innovation.2 Massive funding schemes and multiple legal regimes are conceived or adapted to facilitate these benefits of innovation, from intellectual property (IP) law3 to the Digital Single Market.4 Contrary to this legislative activity, competition law in the EU still relies on pre-digital prohibitions of collusion and abuse of dominance, as well as a scarcely more recent system of merger control.5 Competition law is nonetheless rooted in economic theory, where it is uncontroversial that innovation is the main driver of economic growth.6 Despite all the incentives at EU and public level, economics also hold that innovation is mostly spurred on by the competitive process.7 This suggests a central role for competition law as the prime regulator of that process. It might therefore come as a surprise, to those unfamiliar with the discussion,8 that competition scholars are chiefly divided between those criticizing competition law’s shortcomings in addressing innovation9 and those arguing for restraint in doing so.10 The European Commission (Commission) and EU Courts (the Court of Justice and the General Court)11 have proven less reluctant: in time-honoured fashion, broadly construed provisions have allowed the Commission to act vigorously in relation to innovation without being pinned down to a definite framework. This paper seeks to address this enforcement reality, giving due weight to doctrinal preoccupations but picking up the gauntlet of a theory justifying—indeed, demanding—the application of competition law to innovation. Standing in the way is the market structure methodology usually employed in competition law. Roughly speaking, this methodology focuses on market power—the ability to ignore competitive constraints and raise prices—and requires the definition of relevant markets. Market power correlates with concentrated markets, justifying the normative preference for competitive markets. However, the relationship between market power and innovation is much more ambiguous. Whether innovation is better served by competitive markets in the traditional sense or by overlapping ‘creative destruction’ by monopolists is one of the classic discussions in economics.12 This paper will remain agnostic on this discussion, focusing instead on two consensual points: that innovation is given a powerful incentive by the advantage that it grants in the competitive process13 and that innovation can be stifled by anticompetitive action.14 The problem is that competition law, it seems, would only be able to tackle harm to innovation when its effect is similar to a price increase from market power.15 The scholarly criticism, divided in purpose, converges on this limitation. One set of authors accuses competition law of privileging ‘static’ over ‘dynamic’ efficiency, that is to say, simply keeping existing markets competitive.16 The gist of the criticism can be understood by considering that innovation may well occur outside, or create very different markets from, the market structure policed by competition law.17 This does not deter a second set of authors from defending that, to the extent of competition law’s competence, innovation would be protected by keeping markets competitive.18 What these other authors criticize is departing from such enforcement logic, which in their view would leave no other method to work with. Depending on the perspective, therefore, a market structure methodology would either hamper the ability to address innovation or imperil the legitimacy of competitive enforcement. The issue nevertheless goes far beyond doctrinal disagreement. Competition on innovation—in other words, not only on price and quality but also in terms of obtaining the advantages that innovation grants—is real, and there has never been a shortage of calls for regulators to intervene. Competition law has one way or the other influenced the salient discussions about innovation, from digital interoperability to pharmaceutical research and passing by every major tech firm acquisition. The forced divesture of research and development (R&D) in the Dow/DuPont merger,19 a multi-billion deal at the heart of worldwide agriculture, is a stark example of how competition law can set the pace of innovation in an industry with multiple links to EU policy. The Commission did so, moreover, by appealing directly to harm to innovation. In view of the above, this paper will attempt to build a bridge between enforcement pragmatism and theoretical disagreement. Both the Commission and scholars are right, if only partially. The Commission is right in intervening, in line with the authors wanting to address dynamic competition, if innovation is thwarted under the remit of competition law. It would nevertheless be ill-advised, as the authors arguing for restraint maintain, to stretch a structural approach to encompass unrelated harm to innovation. This paper holds that the legal concepts of competition law already provide the necessary room. What matters is that incumbents exclude or acquire rivals in order to avoid competing on innovation, regardless of market power concerns. The paper suggests that, when market structure cannot explain intervention, more promising economic grounds can be found in strategic management studies.20 These studies are the source of the concept of disruptive innovation,21 which has proved to be central for an understanding of innovation and which provides much of the thrust of this paper. This concept originates from the other explanatory device used by this paper, the so-called ‘resource-based view’:22 in addition to incentives at the level of the market, to which a structural approach is limited, the resource-based view connects innovation with resources at the level of the firm. According to those resources, only certain undertakings will be able to innovate in response to incentives. The exclusion of those innovators may therefore not be compensated by new innovators even if the market remains open and competitive.23 As this paper will show, this explains the importance already given in abuse of dominance and merger control to assets with innovation capabilities. The paper makes three main claims. The first is that the dynamic assessment required by innovation can be dealt with by competition law’s existing framework. The paper will therefore start by clarifying the notion of efficiency under static and dynamic perspectives, and how the resource-based view relates to innovation (Section II). The second claim is that harm to innovation can be framed by rival innovation claims. The resource-based view will be used to adjudicate those claims based on innovation capabilities, as already reflected in current enforcement, and to discuss how appropriating investment in innovation conditions intervention (Section III). The third claim is that the current tests of competition law fail to protect disruptive innovation. The paper will examine how the inefficient start of disruptive innovation is at odds with allocative and productive efficiency standards, and guidelines will be given on how to correct this (Section IV). II. Static and dynamic perspectives It must be made clear from the start that, by proposing that competition law should directly address innovation, this paper does not suggest a radical change. Competition law already covers innovation concerns. It is no sleight of hand that its legal concepts provide this latitude, as competition law is designed to accommodate a variety of economic themes. A useful bridgehead is the notion of efficiency. Innovation has been connected with dynamic efficiency, as just described, and the failure to address it with an excessive concern for static efficiency. Efficiency has traditionally been used to link different aspects of competition law, and dynamic efficiency can anchor a notion of dynamic competition encompassing innovation.24 Dynamic competition has been associated with the interface with between IP and competition law.25 The starting premise is that the legal monopoly granted by IP rights embodies the incentives for innovation. Competition law should thus defer to IP rights, even though they may confer market power, under the risk of harming those incentives. Some object to this deference, as IP rights can also be used to block innovation,26 while others take it as confirmation that competition law should stay away from innovation.27 The case law on abuse of dominance would illustrate such deference by only considering the refusal to license IP as abusive in ‘exceptional circumstances’.28 The deference of competition law to IP rights is nonetheless overstated. No exceptionality has been raised regarding anticompetitive agreements or mergers involving IP rights. While several explanations can be summoned to link all the disparate applications of competition law to IP,29 in practice this happens as’ seamlessly as it does to contractual freedom and property rights. Competition law does not set out to undermine these rights but, if it must regulate economic activity, it must necessarily override those rights when necessary.30 Even the supposed deference in abuse of dominance is set aside, as will be discussed at length, for such scarcely exceptional reasons as ‘technical development’ or allowing a ‘new product’. This paper therefore moves away from interpreting competition law’s role in innovation as the result of its interface with IP rights. As innovation is identified with dynamic efficiency, it is contrasted with static efficiency—namely, with allocative and productive efficiency.31 In essence, allocative efficiency relates to higher output at lower prices, while productive efficiency refers to cost savings. Dynamic and static efficiency thus appears to differ on their subject matter: innovation for dynamic efficiency, productive factors such as output, prices, and costs for static efficiency. However, economically significant innovation cannot but be reflected in productive factors. A better difference is the degree of change involved: static efficiency would be the best configuration of factors as they presently exist, dynamic efficiency the best combination considering how they might be improved.32 If dynamic and static efficiency are different perspectives on change,33 they can both be applied to productive and allocative efficiency—and to innovation itself. All that is necessary is to consider the degree of change. Costs can be rationalized but also move radically with new production methods; output and price can adjust to existing market conditions as well as to new products and competitors; and even the dynamic improvements brought about by innovation have their origin in R&D and other static factors. This conceptual adjustment, discussed in this section, allows framing dynamic competition as a combination of innovation, allocative, and productive efficiency. This section further proposes considering dynamic competition under a resource-based view. The basic idea, taken from strategic management studies, is to examine competitive advantage from the point of view of individual firms. Such competitive advantage goes beyond what is allowed by market structure, as it is said to be based on possessing certain resources.34 It is no coincidence that this is similar to productive efficiency—the resource-based view is intended to orientate managers towards achieving it.35 However, breaking away from a static perspective, the resource-based view applies to building ‘dynamic capabilities’ intended for innovation.36 A. Allocative and productive efficiency Competition scholars are well aware of the law’s concern for allocative efficiency. Producing as long as cost is met (and there is consumer demand) is what is supposed to happen in competitive markets, while lower output and higher prices are the hallmarks of market power. Competition law thus covers undertakings37 pooling their market power by colluding (restrictions of competition under Article 101 of the Treaty on the Functioning of the EU (TFEU)), acting unilaterally from a position of significant market power (abuse of dominance under Article 102 TFEU), and growing their market power by acquiring other undertakings (merger control under the Merger Regulation).38 Harm in those situations varies, but they indelibly connect competition law with market structure. Productive efficiency is also associated by scholars with ‘efficiencies’. Although the emphasis is placed on cost savings (using fewer resources), efficiencies also include improving quality (more utility from the same resources). This bears no relation to market structure: productive efficiency is an advantage in competitive markets, but also increases margins in concentrated ones.39 Harm to productive efficiency is thus not generally seen as an autonomous reason for competition law to intervene. The customary function of productive efficiency is to act as a shield, trading harm to allocative efficiency against gains in cost and quality: Article 101(3) TFEU exempts restrictions of competition that contribute to ‘improving the production or distribution of goods or to promoting technical or economic progress’, the case law on abuses of dominance considers efficiencies as objective justifications,40 and such efficiencies may quell structural concerns in merger control.41 The notions of allocative and productive efficiency are coloured by the static perspective of the TFEU provisions and their enforcement.42 The symptomatic competitive offence is collusive price-fixing and market sharing, so-called ‘cartels’. Although cartels are restrictive by object under Article 101 TFEU, dispensing proof of their effects, the Court has made clear that the concern is their immediate consequences for allocative efficiency.43 Cartels are prosecuted even if, as economists suggest, they eventually fail or their members defect.44 It is therefore no defence that, under a dynamic perspective, a particular cartel may ultimately prove harmless or even beneficial for allocative efficiency.45 Because the concern is static, cartels are remedied by voiding the agreements and letting undertakings act independently again. Even though cartels are commonly used as the example of a static perspective,46 the same can be said of abuse under Article 102 TFEU. Exclusion or exploitation is assumed to happen concurrently with the abusive behaviour, and it is irrelevant whether it manages to succeed or persist. Predatory prices notably ignore whether the dominant undertaking is able to recoup below-cost sales by subsequently raising prices,47 while margin squeeze and rebates are configured as inexorably harming ‘as efficient competitors’—drawing such efficiency from static costs and price relationships.48 By refusing to consider whether abuses may endure changing conditions, the Court ends up blurring the distinction between static and dynamic competition. Furthermore, efficiencies have been defined in a particularly static perspective. Incentives to productive improvements do have a dynamic dimension, as discussed below, but legal standards require proof that efficiencies will materialize in the short-term, including their quantification.49 The purpose of such a requirement is to compare productivity gains with allocative efficiency losses, as part of the proportionality assessment, thereby assuming that they are on the same static plane. If they are not, static allocative efficiency is given preference insofar as the Commission finds that restrictions like cartels cannot be considered proportional.50 Despite this static outlook, the notion of dominance introduces an important dynamic element by considering potential competition.51 Dominance starts from the static exercise of market definition,52 assessing market power based on short-term substitution in reaction to small price increases.53 However, a dominant position further requires protecting this market power against potential competition, that is to say, against longer term expansion of capacity and market entry.54 Hence, allocative efficiency is ultimately determined by barriers to expansion and entry. Because potential competition demands thinking about the ability to overcome those barriers, and the changes that this might in turn set off, it can open the door for innovation.55 The impact of this dynamic turn is mixed. A dominant position is necessary for applying Article 102 TFEU, but it is also disconnected from the tests of abuse.56 Furthermore, a claim of potential competition must overcome the presumption of dominance from (static) market shares above 50 per cent.57 The most important consequence has therefore been for merger control,58 since creating or maintaining a dominant position is incorporated in its substantive test.59 In the Horizontal Merger Guidelines, the Commission confirms it will consider unilateral and coordinated effects based on market barriers.60 Another dynamic element is the (limited) use of harm to productive efficiency as the basis for a competitive offence. Productive efficiency is unrelated to market structure, and if it is directly addressed so too can innovation.61 This is not obviously a question of second-guessing undertakings’ decisions on cost and quality. Competition law may intervene when undertakings degrade quality,62 for example to segment the market or avoid certain costs, and when undertakings impose ‘naked’ costs on their competitors, such as paying distributors not to carry or delay their products. Limiting quality may constitute a cartel or an exploitative abuse,63 while ‘naked restrictions’ are considered abusive exclusion.64 None require consequences for allocative efficiency,65 even if they have so far adopted the usual static approach of cartels and abuse. B. Innovation An essential characteristic of innovation is its uncertainty.66 Although it is reasonable to expect novel improvements, particularly from technology, it cannot be anticipated exactly if, when, and how they will take place. Strategic management studies have provided an important insight by distinguishing between ‘sustaining’ and ‘disruptive’ innovation: sustaining innovation takes place within the value network of firms, giving customers more or better in the attributes they already value; disruptive innovation takes place outside that value network, generating new preferred attributes.67 Productive efficiency is open to the improvements of both sustaining innovation and efficiencies. As legally defined, efficiencies reflect a static perspective: productivity gains have to be estimated to such a degree that they cannot be said to be truly innovative, only unimplemented. Thus, it is preferable to limit efficiencies to combining or reconfiguring existing resources68 and use another term—‘follow-on innovation’—for dynamic productivity gains (both to cost and quality, and not only the attributes emphasized in sustaining innovation). In summary, productivity would be improved statically by efficiencies and dynamically by follow-on innovation,69 while disruptive innovation would bring about new cost and quality relationships.70 Innovation is typically used interchangeably with dynamic efficiency,71 but it may also be seen from a static perspective. The uncertainty of innovation can thus be contrasted with the certainty of R&D expenditure. R&D is usually lumped with efficiencies and subject to the same burden of proof.72 R&D can nevertheless also provide a competitive parameter that can be affected by behaviour falling within the scope of competition law. As a possible source of innovation, R&D is - in and of itself - worthy of being protected from anti-competitive action. This method fits particularly well for industries where new products demand a substantial investment.73 Market structure logic can therefore apply to R&D, and in this way indirectly to innovation, by assuming that market power is a reflection of R&D investment.74 If competitors are excluded by merger or anticompetitive behaviour, the competitive constraint of R&D is lessened;75 if the market becomes more concentrated or coordinated, overall R&D diminishes.76 These assumptions can be superimposed to existing or future product markets,77 or lead to separate ‘innovation markets’ that identify sources of innovation.78 Such an analysis is appropriate for what it sets out to do: link concentration with—all else staying the same—a decrease in R&D. However, this static perspective cannot fully account for the changes brought about by innovation.79 R&D should be seen together with other static contributions to innovation. Follow-on innovation usually targets advantages in existing markets, benefiting from the insights of undertakings already competing there. Thus, in addition to R&D, it can result from ‘learning-by-doing’—the same process which leads to efficiencies.80 Moreover, contrary to what its definition as new attributes might suggest, disruptive innovation also accords a role to efficiencies. It is central to the notion of disruptive innovation that new attributes are not immediately valued by consumers—in other words, disruptors start out inefficient.81 As will be discussed in a separate section below, disruptors require an increase in productive efficiency in order to fulfil their promise. The picture that emerges is of multiple links between static and dynamic factors. Table 1 summarizes the static and dynamic perspectives examined so far: from competition in defined markets to potential competition, but also from R&D to follow-on innovation and from efficiencies to disruption. Those links become fuzzy in a dynamic perspective: static competition spurs (all kinds of) innovation; R&D can also lead to market entry and disruption; and productive efficiency affects the costs of entry, as well as follow-on innovation by learning-by-doing, in addition to its role in disruption. Table 1. Summary of static and dynamic perspectives Allocative efficiency Innovation Productive efficiency Defined markets R&D Efficiencies Potential competition—Follow-on innovation—Disruptive innovation Allocative efficiency Innovation Productive efficiency Defined markets R&D Efficiencies Potential competition—Follow-on innovation—Disruptive innovation Table 1. Summary of static and dynamic perspectives Allocative efficiency Innovation Productive efficiency Defined markets R&D Efficiencies Potential competition—Follow-on innovation—Disruptive innovation Allocative efficiency Innovation Productive efficiency Defined markets R&D Efficiencies Potential competition—Follow-on innovation—Disruptive innovation The links might seem trivial, but it is nonetheless important to formalize them. The confidence in competition law’s ability to deal with allocative efficiency, productive efficiency, and even R&D, contrasts with the hesitation is some quarters regarding innovation. Moreover, these links underlie the Court’s statement that a dominant position is not illegal in itself.82 Dominance is supposed to result from higher efficiency, an idealized running of the market that the Court calls ‘competition on the merits’.83 Nonetheless, a dominant position is defined by safeguards against a competitors productive efficiency (market power) and against dynamic allocative efficiency (market barriers).84 The logical conclusion is that competition on the merits depends on innovation in order to contest dominant positions.85 This puts innovation squarely within competition law’s mandate. C. The resource-based view Strategic management studies offer several theories of competitive advantage, the resource-based view being only one of them. The most famous is Porter’s ‘five forces’, which seeks to best position firms according to industry characteristics. This ‘positional school’ was inspired by industrial organization economics.86 However, despite this link with some of the economics that underlie competition law,87 strategic management has not greatly influenced competition law in academic circles. This has prevented competition law from benefiting from strategic management’s wider economic and social sciences background.88 The resource-based view sits on the economic side of that spectrum,89 making it particularly easy to add to the competition law toolkit. The resource-based view starts from the concept of Ricardian rents: some firms may be more efficient but, for some reason, not have the capacity to supply the whole market.90 This capacity limitation allows the entry of marginally less efficient competitors until the market is full. Such entry affects market price, as is normal for allocative efficiency, but the more efficient firms remain in a position to extract a rent due to lower costs or higher willingness to pay—higher productive efficiency.91 Because the rent is ascertained by comparison with what structural competition allows,92 the resource-based view is complementary to positional schools.93 Its advance is in developing competitive advantage in relation to firm heterogeneity. Starting from the premise that undertakings differ from one another due to their resources, the resource-based view improves on the assumption that they respond in the same way to market structure.94 It is a dynamic perspective, as befits a theory aimed at orientating management, since what matters is the potential to achieve a rent and not if it is actually realized.95 Schumpeterian rents are explained by a resource-based view, the high efficiency disruptor being followed by marginally less efficient imitators.96 Resources are therefore used to estimate innovation capabilities,97 similarly to how it will be seen that competition law links innovation with certain assets. Two clarifications about the resource-based view are nonetheless necessary. First, resources are considered in bundles and not as individual assets.98 This means that, for fully assessing the innovation capabilities of assets such as data99 and IP, they should be considered together with the other resources that make up the undertaking. Bundles may include intangible resources such as the undertaking’s management, strategy, or company culture.100 It is nevertheless better to approach a resource-based view starting from assets, since competition law is already familiar with their ability to influence allocative and productive efficiency.101 Second, the resource-based view pays considerable attention to the resource characteristics which affect competitive advantage. The main characteristic is difference in resources, since that is the source of higher productive efficiency. However, the advantage will also depend on whether the resource is reproducible, mobile, and favourably acquired.102 First, it is clear that whether the resource can be copied will influence the advantage it grants. Second, advantage can also be shifted when the resource moves between undertakings. Third, the conditions for acquiring or producing the resource can erode the rent extracted, including temporal arbitrage (trading costs against benefit at different points in time). With these precisions in mind, the resource-based view adds to the understanding of dominance.103 A dominant position is defined by the Court as allowing an undertaking ‘to behave to an appreciable extent independently of its competitors, customers and ultimately of its consumers’.104 According to a market structure explanation, this independence is from substitution. Yet, the advantage captured by the resource-based view grants independence from demand on top of what the market structure allows. Factors often considered to be market barriers—eg IP, consumer goodwill, or cost advantages105—can be seen in this perspective: not as being necessary for entering the market, since limited production allows less efficient competitors, but as being necessary for competing with the dominant undertaking on an equal footing. In other words, resource advantages do not prevent competitors from expanding output, but from taking market share. The notion of dominance should therefore be open to a resource advantage. A dominant position based on or complemented by this advantage is more coherent with the robust rivals and medium market shares that are often found.106 Indeed, it is an overall better fit with the ‘special responsibility’ that sets dominant undertakings apart107—market barriers shelter all undertakings, not only those with high market shares. The Court has already attributed dominance to technical and commercial advantages alongside, and distinct, from market shares and market barriers.108 In a dynamic perspective, a resource advantage can become a market barrier if less efficient competitors are pushed out by increased production or if the market contracts. More importantly, as this paper discusses, such an advantage can be used to harm innovation. The resource-based view can further help to explain abusive refusal to supply. As noted above, abuses usually do not go beyond static exclusion. In an abusive refusal to supply, however, the Court demands that refused input be ‘indispensable’.109 Not only is the input an asset, but indispensability is characterized in the lines of the resource-based view. First, the input is different and immobile. Second, the Court examines whether the input can be replicated, considering the incentives for the dominant undertaking and its competitors to invest in its production.110 In other words, the abuse examines whether the refusal can be effective in a dynamic environment. III. Rival innovation claims Having established that competition law can—and should—address innovation as part of dynamic competition, and having introduced the resource-based view for that purpose, this section will explore how to adjudicate cases of harm to innovation. It is, however, useful to deal first with the argument that competition law should only address innovation indirectly via market structure. The discussion so far, despite painting a richer picture of dynamic competition, does not yet disprove this argument. The emphasis placed on potential competition might even embolden it—if market barriers take innovation into account, then a competitive market structure would reflect innovation. This could lead to the conclusion that competition law either refers to market structure or wrongly ignores it. Such a dichotomy is nevertheless false. Addressing innovation directly—that is to say, demanding harm to innovation instead of assuming it from market power—is not opposed, but iscomplementary, to an indirect approach. Structural concerns were enough to assume that innovation would suffer in Deutsche Börse,111 and the merits of this approach are evident and this approach gets a lot right. Preserving competitive markets ensures that successful innovation will not be for naught,112 channelling undertakings’ efforts towards competition on the merits. Since it is important to define what counts as competitive markets. The criticism that innovation should not be used to relax structural requirements is theoretically justified, as long as the concern is indeed allocative efficiency.113 Overall, an indirect approach is correct in that market structure provides a tried-and-tested framework protecting innovation against market power—but only that. There is no reason for complacency. The Commission has started a public consultation on enlarging merger control to innovation related assets such as data and IP, as mergers of free-service providers risk escaping turnover thresholds.114 By not advancing a parallel adaptation of its substantive guidance, the Commission seems confident that it can address innovation under the current status quo of competition law—so much is apparent from its Merger and Innovation Policy Brief.115 However, as will become apparent in this section, undertakings can harm innovation in other ways. Innovation capabilities can be snuffed out, regardless of market barriers, by control over certain assets. Moreover, static competition can thrive in the absence of innovation. Despite the merits of an indirect approach as a primary method of analysis, there is more to competition law. Competition provisions refer expressly to the protection of innovation, and they have been applied directly to innovation. The indirect approach subordinates competition law to market structure, when it is methodology that should adapt to normative ambition. This section will therefore start by examining cases where competition law has been applied to harm innovation proper. It will be seen, roughly speaking, that the innovation capabilities of certain assets give rise to rival claims: on the one side undertakings seeking access to the asset or forced divestures in mergers in order secure competition on innovation, often called a ‘level playing field’, on the other, undertakings wanting to retain control of the asset as a return for investing or acquiring it, the problem of ‘appropriability’. The question is how competition law can settle these claims. It will be seen that market structure methodology must conjure markets out of assets, and delegate appropriability to IP law; a resource-based view can focus on innovation capabilities without the intermediation of market power, and assess appropriability based on actual competitive circumstances. A. Cases of harm to innovation The cases calling for a resource-based view are those where harm to innovation is not covered by market foreclosure. The two harms are caught together insofar as competition on innovation is negatively affected by market concentration. In those situations, competition rules only have to apply as normal. For example, an exclusivity obligation prohibited under Article 102 TFEU might deprive rivals of the income to invest in R&D,116 in the same way that a horizontal merger removes pressure to innovate between the merging parties. This is the indirect approach just referred. It is clear that (market foreclosure) exclusion can harm innovation; a different question is whether harm to innovation can constitute (another sort of) exclusion. The legal concepts of exclusion, understood as hindering rivals in any manner other than competition on the merits,117 and a significant impediment to effective competition, going beyond dominance based on market power,118 provide room for innovation alongside structural and productive concerns. What matters is that innovation is part of rivalry beyond structural bounds, and that such rivalry is negatively affected by exclusion.119 So much is confirmed by a number of cases finding harm to innovation without delving into (the absence of) structural concerns, namely on low market barriers and strong residual competition. Competition authorities might obscure this by clinging on to structural appearances. Harm to innovation might be taken as an indication of market power instead of the other way around, particularly if resource advantages are confused with market barriers—as commented, both grant the independence characteristic of dominance. This obfuscation is only natural since, as will be developed below, the Commission has only provided guidance for intervention based on market structure. Ultimately, intervention can always be painted as structural since market power may also affect innovation.120 Despite such pretences, innovation has taken centre stage in three groups of cases. The first group is the case law on abusive refusal to license IP. As discussed in relation to the resource-based view, indispensability involves a dynamic analysis. Magill, and later IMS, further demand that the refusal eliminates all competition in a given market and prevents a new product for which there is consumer demand.121 The condition of eliminating all competition is static, since exclusion operates along market definition. The condition of preventing a new product brings indispensability and exclusion to bear on harm to innovation. This case law took an important turn when the General Court relaxed the conditions set by the Court of Justice in Microsoft. Exclusion was lowered to the usual standard of competitive disadvantage, and preventing a new product was effectively substituted by an impediment to technical development.122 As such, the General Court did not check whether the refusal of interoperability data would do more than prevent competitors from continuing to develop innovative features.123 Although Microsoft was not appealed to the Court of Justice, the Commission has taken it as an accurate statement of the law.124 The second group of cases is based on Commission merger decisions involving parallel R&D. The Commission’s guidance touches on this issue. The Horizontal Merger Guidelines start by associating market power and harm to innovation,125 but then depart from market structure by ignoring concentration levels if the merger involves ‘important innovators’.126 The guidelines thus acknowlede that some undertakings influence dynamic competition beyond their market shares, only to add ‘in particular when the market is already concentrated’.127 After this candid demonstration of the lack of relationship between innovation and market structure,128 the guidelines get to the innovation issue: ‘pipeline products’.129 The concern (unstated in the guidelines) is that pipeline products might be abandoned post-merger,130 notably if there is a duplication of R&D efforts or the merged entity already commercializes a successful competing product.131 In order to avoid such harm to innovation, the Commission has forced the divesture of pipeline products on several occasions.132 It did so most recently in Dow/DuPont, abdicating from market definition altogether and considering pipeline products (and even earlier research) under ‘innovation spaces’ according to discovery targets.133 The third group of cases relates to Commission practice on vertical relationships. In the Non-Horizontal Merger Guidelines, the Commission repeats both the structural association of innovation and market power134 and the disregard of concentration levels if the merger involves innovators.135 The guidelines go on to state that vertical mergers136 can lead to concerns over ‘input foreclosure’. The reasoning is the same as in abusive refusal: denying rival access to an input in order to gain a competitive advantage downstream. However, contrary to abuse, the guidelines do not require indispensability. Instead, the Commission examines the ability, incentives, and likely impact of the foreclosure.137 Although the guidelines do not mention innovation in relation to vertical foreclosure, some Commission decisions have made this connection. Notably, Intel/McAfee guaranteed that downstream competitors (in security solutions) would innovate in the same terms as the merged entity (when using Intel’s chips).138 Related innovation also seemed to be a concern in Google Search: in the initial stages of the investigation, the Commission stated that rivals’ incentives to innovation would be harmed if they could not get the same prominence in Google’s search engine as Google’s own products.139 That is nevertheless close to the argument, broadly accepted, that (any sort of) exclusion prevents undertakings from reaping the rewards of successful innovation. Perhaps this is why innovation is no longer referred in the final decision.140 These three groups of cases have in common assets with innovation capabilities: IP in abusive refusal to license; pipeline products in parallel R&D; and inputs in vertical relationships. IP and data are also the basis of the consultation on the changes to the Merger Regulation.141 It can therefore generally be concluded that competition law has associated harm to innovation to specific assets.142 Assets are only valuable in bundles that allow exploiting them, as the resource-based view holds and divestment remedies confirm.143 As such, in cases of harm to innovation it is also necessary to examine the undertakings concerned. B. Adjudicating innovation claims Intervention has not been just about assets, but about rival claims over them. These claims are split along keeping control over the asset and having it (forcibly) shared. In abusive refusal and vertical relationships, the asset is controlled by a dominant undertaking or a merged entity while competitors seek access to it. In parallel R&D, the asset is bound to be controlled by the merged entity unless competition authorities intervene to divest it. These claims are argued and settled over which alternative is better for innovation.144 Therefore, they are claims of innovation. Innovation claims require a dynamic assessment, often putting competition authorities in the position of having to act ex ante.145 This is so in merger control, where effects have to be anticipated. Conversely, abuse usually treats exclusion ex post—Magill and IMS were about innovative products already being commercialized. Thus, although the new product condition is dynamic, the Court could handle it like static exclusion. In Microsoft, however, interoperability data were claimed to be necessary for future innovation. This may be why the General Court abandoned the new product condition in favour of (a more generic) impediment to technical development—taken from the letter of Article 102 TFEU and, not coincidently, also appearing in the Merger Regulation.146 Innovation yet to occur is one of those difficult predictions about the future which, as quoted in the introduction, Bohr jested about. Market structure allows anticipating innovation through incentives to compete, but it must be remembered that these incentives are considered in a static perspective.147 As such, technical development may be interpreted in the static manner of harm to productive efficiency,148 merger control has relied on static investment in R&D,149 and vertical foreclosure is based on a static concept of exclusion which does not require indispensability.150 This facilitates intervention, and as already mentioned competition in defined markets provides an important incentive to innovation,151 but it would be odd if dynamic competition boiled down to the same incentives used for static analysis. The point is that competition law should focus on capability to innovate, not only incentives. If innovation is beneficial in both competitive and concentrated markets,152 a reasonable starting assumption is that all undertakings have some incentive to innovate regardless of their market power.153 This is, after all, what is desired by a level playing field to compete on innovation. In contrast with the similarity of incentives, however, not all undertakings have the capability to innovate.154 This difference in capability, not structure, is what ends up determining acquisitions and exclusionary strategies. Differences in innovation capabilities are precisely what the resource-based view aims to capture. Applying the assumption of firm heterogeneity, the question can quickly progress from whether innovation is at stake—if the asset is found to have innovation capabilities, by definition it is—to examining which undertaking has a better innovation claim. Since the claims concern the same asset, the legal standard should be whether undertakings can do something different from each other. This requires examining, according to the onus of the legal test at issue and backed up by economic expertise as necessary, how the asset integrates with the resources that make up the undertakings in question.155 This legal standard—difference—is the answer to the problem of anticipating innovation. It is no static assessment, therefore does not require predicting exactly how innovation will shape out, only how differently it will according to the undertakings concerned. Difference thus integrates the counterclaim that keeping control of the asset might be better for innovation. This counterclaim is the starting point in merger control, which examines the effects of the acquisition of the asset.156 However, as discussed in more detail in Subsection III.C, such counterclaim has (wrongly) been understood as a justification under abuse. Comparing the innovation capabilities of dominant or merging undertakings and their competitors is compatible with the standard of technical development in Articles 101 and 102 TFEU and in the Merger Regulation, which necessarily points towards a gain for the whole industry. The breakthrough allowed by the resource-based view is connecting the legal standard of difference with the factual assessment of integrating the asset in the undertakings concerned—which will naturally vary according to these undertakings’ resources. Such resources can be examined directly or their potential assessed through previous innovative efforts.157 Thus, if the integration would lead to mere reproduction, a refusal to license will not be considered abusive.158 In the absence of a new product to latch onto in Microsoft, the General Court stated that previous novel features ‘spoke volumes’159. Dow/DuPont also concentrates on ‘assets and capabilities’, its concerns being the loss of rivalry of integrated R&D players as shown by previous innovation in the same spaces. This is enough to anticipate that pipeline products would be discontinued post-merger without having to predict exactly which.160 Dow/DuPont further clarifies previous Commission practice in this regard. Thus, only the innovative capabilities of integrated R&D players are considered, as opposed to less-capable or non-innovative competitors.161 Others cases would suggest that preventing any competing product from emerging would harm innovation.162 However, a product may be new but not different. Parallel R&D and compulsory access may address a need already being served,163 that is to say, compete on price or another parameter besides innovation.164 Addressing structural concerns remains fully within the Commission’s competences, of course, but it also benefits the resource-based view to keep them separated.165 C. Appropriability Grounding harm to innovation on difference, tout court, may seem like a policy prescription to share assets whenever rivals are able to improve on their use. This is bound to raise a concern for appropriability, that is to say, how investment in assets is incentivized by the ability to appropriate their returns. Some have argued that the current level of intervention already chills innovation,166 since it would interfere with the calibration of incentives set by IP rights. This goes to the supposed deference of competition law to IP rights, signalled at the outset, which would only allow intervention in case of significant structural concerns. While IP rights have been widely associated with appropriability,167 this does not follow from the market power incentives which feature so prominently in competition law discourse: rather than competitive markets, the incentive to innovate would lie in legally-protected expectations.168 It is argued that, while appropriability explains a more limited intervention than the definition of harm to innovation would suggest, competition law does not outsource this function to IP law. The legal expectations of competition law are as much of an incentive as IP rights, since they protect the value of innovation from being emptied by anticompetitive action.169 Concerns over appropriability are often raised when refusals to supply are found abusive.170 However, such concerns should be dwarfed by the consequences of defining markets based on assets. An individual monopolist was found for each of the schedules in Magill,171 the same occurring with the data format used internally in IMS.172 This makes innovation appear at the mercy of ‘asset monopolists’, increasing the risk of intervention exponentially by opening the way for structural concerns.173 This would endanger appropriability enormously were it not overshadowed by the dynamic requirement of indispensability—which is what, in the end, is crucial for finding an abuse. Regardless (or in spite) of asset-based markets, the question remains of how appropriability keeps intervention in check. Appropriability does seem to be behind the statement, already referred to, that a refusal to license IP is only abusive in exceptional circumstances. However, the conditions of abuse are not strikingly exceptional: the IP of one product can be used to create a competing product, that IP thereby becomes indispensable, and exercising the IP right is enough to completely remove such competition. This is the story of every unauthorized IP adaption. Furthermore, exceptionality is not available for other interventions affecting IP rights. Appropriability came to the fore in Microsoft, where it was argued that because IP rights protected incentives they should be treated as an objective justification.174 Although it was not spelled out, this would make abuse truly exceptional: only appearing when the exercise of the IP right would be disproportionate.175 The General Court did not perceive this, stating that such justification would prevent the abuse from ever occurring.176 It nonetheless examined another justification: whether access would have a ‘negative impact on [Microsoft’s] incentives to innovate’.177 Not only did Microsoft fail to prove so, the General Court pointed to two factors against such a negative impact: there was no cloning or copying, and it was industry practice to share interoperability data.178 There is much to take from the General Court’s analysis of incentives in Microsoft. Appropriability does not stop at the door of IP179: cloning or copying, as well as industry practice, have an obvious impact on incentives. For good reason this was also examined regarding preventing technical development.180 This double role—for a finding of a restriction as well as objectively justifying it—lines up incentives to innovate with ancillary restraints under Article 101 TFEU.181 As such, rather than a justification being for the dominant undertaking to prove, appropriability should be part of establishing harm to innovation to begin with182—as it already was in Microsoft.183 Once appropriability is freed from IP rights to focus on incentives, the resource-based view can connect it with competitive advantage.184 Market structure suggests that intervention does not harm incentives, since the advantage of market barriers remains.185 The resource-based view goes further by examining the resource characteristics which, regardless of market structure, can create a competitive advantage. A resource-based view moves from competition on the merits, which merely registers productive improvements,186 to whether a rent is created from investing in those improvements. It is precisely this incentive that appropriability aims to capture. The resource-based view thus explains why intervention is limited: it would affect the rent drawn from competitive advantage. This is not a question of preserving market power since, as already commented, that makes for poor appropriability. More simply, innovation involves risk187 so that it is hard to determine the adequate rent.188 The resource-based view is well aware of the causal ambiguity surrounding this rent,189 so instead it concentrates on competitive advantage. The resource characteristics affecting competitive advantage are, as already referred to: difference, replicability, mobility, and favourable acquisition. Difference was found to be the legal standard for harm to innovation. Replicability and mobility are how legal regimes condition the expectation of appropriability.190 Favourable acquisition may thus be, of all the characteristics, the one that links intervention with appropriability. This characteristic examines whether, in a dynamic context, the rent is nullified by the cost of acquiring or producing the resource. If the rent is preserved, the incentive will hold. Those situations are quite limited insofar as ex ante and ex post competition for the asset will erode the rent.191 Abusive refusal to license, however, shows this might not always be the case. In Magill the schedules were a by-product of broadcasting activity, IMS involved a de facto standard to which the whole industry contributed,192 and Microsoft referred to the industry practice of sharing interoperability information. In principle, the rent from controlling these assets was not reflected in their cost.193 Moreover, intervention seems to privilege market returns and discount other kinds of appropriability—even if these other kinds are equally, if not more, effective at rewarding investment. The benefits of vertical integration are set aside for creating a level playing field for all undertakings to innovate in the market. Divestment that guarantees market presence is also favoured, even if innovators are amply compensated by acquisition and duplication negates the returns from the (sometimes only) innovative product. This preference for market returns may explain the lopsided number of abuse and mergers cases: a dominant undertaking will stay in the market, advising against intervention, while an acquired undertaking trades this option away. The role of appropriability remains largely unstated, and a resource-based view is but a first approach. Privileging market returns is normatively coherent with competition law’s emphasis on competition on the merits. In addition, the circumstances behind favourable acquisition are as varied as dynamic competition.194 The absence of formalized rules and scattered Commission practice suggest that incentives to innovate are evaluated on a case-by-case basis. Considering the burdens of (inadequately) framing appropriability as an objective justification, and that incentives are already incorporated in harm to innovation, any further role for appropriability is indeed currently better left to enforcement priorities. IV. Disruptive innovation Rival innovation claims, as examined so far, have not differentiated between follow-on and disruptive innovation. Neither has competition law in general. Follow-on innovation normally lowers costs or improves the quality of existing products, falling in line with competition law’s concern for consumer welfare.195 Attempting to reverse the market changes brought about by disruption will appear equally detrimental to consumers. However, there is no visible loss if disruption is frustrated before those changes occur—markets continue to operate as competitively as before. Thus, although disruption is the most emblematic form of innovation, it is much harder to capture. The root of the problem is that—as the theory of disruption emphasizes—disruptors start off as less efficient. Although disruption brings to mind superior products that make whole markets obsolete, this is only after disruption has succeeded. In reality, strategic management studies have found that disruptive products emerge at the fringes of the market, or related markets, which are neglected by incumbents.196 Because disruptive attributes are not valued initially, disruptors start by competing through low prices.197 This ‘disruption from below’ can be contrasted with ‘disruption from above’: expensive products which start by serving only a minority of consumers.198 It has additionally been argued that the firm itself can support disruption, shifting the emphasis from demand to supply.199 Like disruptive products, production methods can depart from, and eventually overtake, the preferred ways of doing things. However, also like products, disruptive methods start off inefficiently in comparison with the incremental improvement that established methods have benefited from. Demand and supply can be analysed in isolation but are often related, as the business model of disruptors also differs from incumbents.200 In any case, the theory of disruption requires an increase in productive efficiency. The unpromising start is what allows disruptors to go on unscathed, as incumbents are naturally occupied with getting the most out of preferred attributes and production methods.201 If an undertaking comes to the market and immediately beats incumbents on their terms, this is a particularly successful case of follow-on innovation.202 Yet, if disruptors are to fulfil their promise, they must grow out of their starting inefficiency.203 Disruption from below must increase utility in order to challenge preferred attributes,204 disruption from above must decrease price to be mass-marketed, and disruptive methods must be assimilated by the undertaking to achieve a cost advantage.205 This increase in productive efficiency is a necessary, but not sufficient condition. Whether new attributes become more valued by consumers, or any other kind of disruption comes through, is shrouded by the uncertainty of innovation. As such, tests of harm must be satisfied with the potential for disruption and focus on strategies that frustrate it. Market power gains are moot for those strategies. Disruptive innovation is very different from preserving a competitive market structure, or even the conditions for follow-on innovation, and should be given priority over them. This section will discuss whether exclusion can incorporate the above elements that are particular to disruptive innovation. A first difficulty comes from strategic management studies’ scepticism on whether disruption can be sufficiently anticipated so as to be stopped. The application of competition law to disruptive innovation will be examined afterwards, namely whether existing tests based on allocative and productive efficiency do not prove counterproductive. The section concludes with guidelines for adapting those tests to harm to disruptive innovation. A. The inevitability of disruption? Strategic management studies are doubtful of incumbents being able to ward off disruptive innovation. Disruption seems inevitable ex post and, more importantly, impossible to predict ex ante. Disruptive products can come from anywhere, and knowing that they will have new attributes not currently valued is of little help. This is so even if strategic management avoids market definition: disruptive products end up defining their own market,206 but they might start in the same one which they will eventually disrupt or in a separate, related market.207 The scope of potential disruption is therefore enormous, as practically any firm will have less efficient rivals and imperfect substitutes. Incumbents are said to be caught unaware by disruption, as already remarked, because they offer what the market wants and use the most productive methods. Not only that, it makes sense for incumbents to concentrate on the real and present danger posed by rivalry and substitution. Many potential disruptors—if not all—will fail to make the necessary increase in productive efficiency.208 It has been suggested that incumbents can create separate internal units, so as to force them away from the currently favoured attributes and production methods.209 Incumbents with a large R&D budget can also spread research from incremental improvements to market-changing technologies.210 However, none of these tactics has changed the generalized view that disruption will arrive unexpectedly and irresistibly.211 This fatalistic view contrasts with the narrative of competition law. Market power would both rob the incentive to innovate—the ‘lazy monopolist’212—and grant the means to prevent disruption: acquiring or foreclosing disruptive innovators.213 Harm to innovation would thus be identifiable ex ante, or at least sufficiently ex post for competition authorities to act in a timely fashion.214 The main obstacles would be focusing enforcement on static concerns such as price-fixing215 and, as raised recently, the acquisition of low business volume innovators falling below merger thresholds.216 The enforcement record is somewhat mixed. Microsoft does not seem to have been disrupted despite the decisions against it, and by the time Magill was decided Magill itself had gone broke. However, the deterrent effect of these cases is hard to gauge. Merger intervention can more easily be painted as success, as the Commission’s account of its innovation policy does,217 but no particular claim is made there regarding disruption. Strategic management could point to the survival of Microsoft’s products and the disappearance of the single-broadcaster guides of Magill as proof that disruption breaks through, or not, regardless of intervention. In any event, there is dearth of cases where intervention is credited with safeguarding disruption.218 The view of strategic management changes, but does not nullify, the need for intervention. A tempting move towards conciliation is to say that, even if disruptive innovation is inevitable, consumers should not suffer in the interim. This is certainly valid from a static perspective, where innovation is a given and the only aim is to maximize its benefits. However, competition law does have a word to say about securing dynamic competition. Disruption appears inevitable after the fact, but its inefficient start gives credence to the belief that it can indeed be frustrated. To begin with, incumbents can devise a rational policy against potential disruptors. It is true that disruption can come from an unknown number of sources, but that does not prevent incumbents from trying to anticipate them.219 Incumbents can project which new attributes and production methods have disruptive potential and, rather than trying to nip them all in the bud, stay on the lookout for those that show signs of accelerating productive efficiency. Even if these turn out not be disruptive, as most will, a prophylactic policy might be cost-effective considering the dire consequences of successful disruption.220 Such a policy is also relatively easy to set in place. Many undertakings engage in constant acquisitions which can accommodate (and hide) measures against potential disruptors.221 Moreover, the inefficient start of potential disruptors makes them vulnerable to competitive pressure. The incumbent does not have to supress the disruptive technology but, as will be discussed, only limit markets so as to prevent potential disruptors from gaining enough productive efficiency.222 The failure of a potential disruptor will be taken as another inefficient undertaking exiting the market, deterring others from engaging in similar investments. The ability of competition authorities to identify the exclusion of potential disruptors has thus been overestimated. Innovators are assumed to fit neatly into structural competition, their smothering by market power for all to see, when this is only the case for disruption which has already been successful.223 Nevertheless, if incumbents can attempt to anticipate potential disruption so can competition authorities. Acquisitions of parallel R&D are a natural place to start, followed by the targeting of potential disruptors or disruptive markets. All that is necessary is to fine-tune tests away from market structure, as will be seen next. B. Lack of specific tests Disruptive innovation sabotages many of the tests used by competition law, as the exclusion of potential disruptors can be reflected in market structure just as well as slip by unnoticed. Sometimes an increase of market power will be made at the expense of a potential disruptor and this will be enough to find an infringement. At other times the loss of the disruptor will leave plenty of residual competition and low market barriers, or incumbents will point towards their higher productive efficiency, and disruptive innovation will be successfully frustrated. The same happens regarding follow-on innovation. Sometimes it will allow potential disruptors to increase their productive efficiency, but at other times these will simply be different types of innovation. Follow-on innovation, despite not reflecting market structure, is a closer fit with competition law’s drive for technical development. In contrast, disruptive innovation requires countenance for inefficiency, as potential disruptors do not offer productive improvements and may not even fulfil their potential. As such, in case of rival claims, follow-on innovation may end up being given priority over disruptive innovation. Harm to disruptive innovation seems best caught by the test of abusive refusal to license, namely the condition of preventing a new product for which there is consumer demand. Magill was historically a case of disruptive innovation: the new product, an all-broadcaster guide, shifted preferences away from single-broadcaster guides.224 Framing the test as the Court did—and not as leveraging to a separate market created by disruption—turned out to be very appropriate. The new product condition catches novel attributes with some disruptive potential (the consumer demand) but without demanding successful disruption (a fully separate market).225 The problem with an abusive refusal to license is that follow-on innovation can also be considered a new product. This is what happened in IMS: the sales data format was sought because it had become the de facto industry standard,226 so by definition the new product was an incremental advance. The same reasoning applied in Microsoft, even as the new product condition was dropped: interoperability data were necessary for improving certain ‘parameters which consumers consider important’,227 the very definition of sustaining innovation. As a result of IMS and Microsoft, abusive refusal to license was no longer about disruptive innovation. This move has had implications for how innovation has been understood in competition law. The emphasis is currently placed on eliminating competition which might result in follow-on innovation. This is the case not only with IMS and Microsoft but also with vertical foreclosure in general. However, preventing disruptive innovation does not mean competition is eliminated. On the contrary, it is disruption which is liable to do so by overtaking existing markets, together with any healthy competition on follow-on innovation that exists there.228 Merger control reinforces the lack of a specific concern for disruptive innovation. The Non-Horizontal Merger Guidelines characterize vertical foreclosure in structural terms, namely by the ability to affect a downstream market by denying access to inputs or customers.229 Foreclosure may thus be adapted to catch whole market impediments to follow-on innovation, but not exclusion circumscribed to potential disruptors. Furthermore, the guidelines cover mergers where there is no market overlap—‘conglomerate mergers’—and where the acquisition of disruptors in related markets should rank the highest. Nonetheless, the only concern shown is for static tying arrangements.230 Disruption could in theory fall under the Horizontal Merger Guidelines’ reference to ‘important innovators in ways not reflected in market shares’.231 However, contrary to the detail dedicated to foreclosure, the guidelines do not specify what constitutes an important innovator beyond recent entry and evading coordination.232 These characteristics have been associated with so-called ‘mavericks’,233 and so too has disruptive innovation.234 The latter is nevertheless still to be confirmed in practice. The Facebook/WhatsApp merger provides a case in point. WhatsApp offered communication services with an innovative privacy and data protection policy which, were it to evolve into social networking, could disrupt Facebook’s business model based on monetizing personal data.235 WhatsApp’s acquisition was deemed not to raise any competitive issues since there was no static overlap between online communication and social networking markets.236 Post-merger, Facebook revised WhatsApp’s policy,237 while maintaining it as the leading communications service. As a result, no disruptive competition on data protection developed, nor did WhatsApp challenge Facebook in social networking.238 Judicial review of merger decisions appears equally limited. In Cisco, an appeal of the Commission’s approval of Microsoft’s acquisition of Skype,239 the General Court considered the possible integration of the parties’ communication products. They were found to be in different markets, so the General Court started by reaffirming the case law of the Court of Justice that the assessment of conglomerate mergers involves a ‘prospective analysis’ of ‘uncertain and difficult’ causality.240 Despite this case law waiving at dynamic competition, the General Court chose to emphasize that a ‘significant impediment to competition [must be] the direct and immediate effect of the concentration’.241 Such certainties are only appropriate for static competition. Even though the merger in Cisco did not involve an impediment to disruptive innovation, a standard of direct and immediate effect is bound to nullify any such claim. Curiously, Cisco did examine disruption, but from the side of the merged entity. The General Court rejected the argument that Microsoft could foreclose its competitors by integrating Lync with Skype242 as uncertain and in the future.243 The General Court also found that the prospective advantage was vague, lacked ‘real and significant demand’, and would not become a ‘must have’ since alternatives were available.244 This is precisely how disruptive innovation looks ex ante. The General Court even considered that Lync’s low market shares denied it the power to exclude.245 Of course, a merger should not be objected to because it might lead to disruption.246 However, Cisco might very well be used against the reverse claim. In conclusion, tests which could catch harm to disruptive innovation suffer from a lack of distinction with follow-on innovation and an excessive connection with market structure. These tests can still work if competition authorities use their discretion to prioritize investigations or raise merger concerns involving potential disruptors. Exclusion does not need to be grounded on preventing disruption, even if that is its purpose. For instance, leveraging market power into a potentially disruptive market is a typical defensive strategy, which can be caught by the leveraging alone.247 However, this can only go so far. Not all leveraging is anticompetitive,248 and not all exclusion will raise a structural or follow-on innovation concern to grab hold of. C. Guidelines for disruption The remainder of this section will propose guidelines for adjusting tests to prevent disruptive innovation, namely in relation to disruptive potential, exclusionary strategies, and limiting principles. The first, disruptive potential, relates to the capability to shift preferred attributes or production methods. The emphasis should not be put on a novelty since, as abusive refusal to license shows, this is as likely to catch incremental improvement. Disruption is not really ‘new’: it is already favoured by some consumers or undertakings, but not yet enough to shift the market. Since market shifts are surrounded by the uncertainty of innovation, disruptive potential has to be defined in the negative. Tests should start by excluding competition on preferred attributes and production methods. This adapts the legal standard of difference to the particularity of disruption. If undifferentiated innovation claims are made based on the same asset, as in Microsoft and Intel/McAfee, they are unlikely to be disruptive. Contrary to follow-on innovation, a disruptive product may as plausibly emerge from sharing an asset as from innovating to no longer need it.249 In addition, disruption will appear unlikely to succeed because it does not focus on current market trends. That disruptive potential hinges on its shortcomings is, as one author remarked, ‘incredibly counterintuitive’.250 It is not surprising that Cisco failed to appreciate that lack of demand, speculated from existing preferences, allows for disruptive potential rather than disproves it. There must nevertheless be some positive indication of this potential. Strategic management points to niche consumers which are not served by incumbents.251 However, potential can also be inferred from incumbents’ reaction, namely paying high prices for low turnover acquisitions252 and adopting other defensive strategies. Those strategies will necessarily involve exclusion, but will only be related to market structure by coincidence. Their purpose is not to gain or preserve market power, but to supress disruptive potential. They may therefore result in market concentration or, just as well, operate in apparently competitive conditions—particularly if they involve the control of assets with disruptive potential or other resources that disruptors need to operate. Competition authorities can rely on an undertaking’s intent253 and behaviour254 in order to identify these strategies. They fall under three basic categories: denying resources needed for disruption, preventing an increase in productive efficiency, and pre-empting markets with disruptive potential. First, incumbents can limit the disruptor’s access to certain resources. This can take the form, as seen for rival innovation claims, of refusing to supply an asset or acquiring parallel R&D with disruptive potential. Another way is to promote product differentiation and proprietary technology.255 Rival innovation claims have tended to measure technical development in comparable improvements in cost and quality. Thus, the prospects of follow-on innovation will always outweigh the starting inefficiency of potential disruption.256 Since disruption should be given priority in terms of competition policy,257 counterclaims of follow-on innovation by the dominant undertaking or the merged entity should only be allowed subject to strict proportionality. Secondly, incumbents can prevent potential disruptors from achieving the productive efficiency necessary for disruption. This can be done by playing with incumbents’ positional advantage: (regular) inputs can be refused, customers pried away, and operating margins reduced to the point of lasting inefficiency. These advantages can be rallied to support a product aimed, implicitly or explicitly, at matching the disruptor. A more straightforward strategy is simply to acquire potential disruptors and subsequently ‘mothball’ them258 or, as happened after FaceBook/WhatsApp, defang them. The defining trait of all these strategies is that they target undertakings for exclusion based on their disruptive potential. If disruptive innovation is to receive any meaningful protection from competition law, the notion of maverick must be defined to include—or, for the benefit of clarity, be limited to—potential disrupters.259 This would add to the limited protection already granted against abusive discriminatory tactics,260 such as selective pricing261 and naked restrictions.262 Furthermore, any tactic known to be anticompetitive should be considered abusive if it targets potential disruptors. This includes situations where, even though the tactic is potentially or nominally applicable to all other undertakings, only potential disruptors are significantly affected. That could be the case with predatory pricing263 and refusal to supply.264 Thirdly, a market showing signs of disruption can be subject to containing measures which do not individualize particular undertakings. The incumbent can attempt to occupy as much of the market as possible, ‘crowding out’ potential disruptors. It can also try to stunt the growth of the market in order to lower its productive efficiency. Because they operate at market level, such strategies may be caught by abuses like exclusivity, rebates, and margin squeeze. However, the trend of limiting those tests to the foreclosure of an ‘as efficient competitor’ is very problematic.265 Potential disruptors are less efficient by definition. Even though they may eventually prevail,266 in the interim they are left without protection from tactics that lean on static advantages. Enforcing competition law against these types of strategy would naturally be subject to limiting principles. Appropriability remains available, as discussed in the previous section,267 and rival innovation claims also. As already commented, a counterclaim of follow-on innovation should only be admitted if its likelihood and benefit manifestly outweighs the possibility of disruption. However, harm to disruptive innovation is not always tied to an asset—incumbents may recognize the disruptive potential of assets but move to exclude disruptors in other ways. The two main limiting principles thus come from the negative definition of potential disruption, examined above, and from questioning the needed increase in productive efficiency. A defence could be raised that, regardless of the exclusion of potential disruptors, they would not be able to increase their productive efficiency.268 Since this increase is necessary for disruption, the onus is on competition authorities to show it as part of harm to innovation. However, this must be adapted to the tests in question. Some require likely effects, most notably merger control, while others are content with harm in the abstract, such as abusive discrimination. A related issue is whether the acquisition of potential disruptors (or their assets) would actually favour an increase in productive efficiency. It is well accepted that acquisitions might provide the means to develop disruptive advances.269 Although this may appear as a justification, it is in effect a case of rival disruptive innovation claims about the acquirer and the acquired. It should therefore be adjudicated under the usual condition of proportionality, that is to say, that the acquisition is both necessary and adequate (not going beyond what is necessary) for disruption to succeed. V. Conclusion This paper has attempted to answer the question of what competition law can do for innovation. First and foremost, by keeping markets open and competitive, it ensures that competition on innovation will not be hindered by market power. Whenever innovation is listed alongside price and quality, a powerful signal is sent that undertakings can innovate their way to success. The benefits of this indirect approach have for a long time delayed the realization that, if competition law is to truly regulate dynamic competition, it must also address harm to innovation directly. This paper has proposed tapping strategic management studies for a complementary approach, the resource-based view. This approach formalizes firm heterogeneity as part of the competitive discussion, and allows two significant advances in relation to structural incentives. First, it explains why certain practices are subject to enforcement: those which limit the use of assets with innovation capabilities. Second, it provides an initial proposal for why enforcement does not harm appropriability: because, in certain circumstances, the rent from the competitive advantage is preserved. This is reflected in abusive refusal to license and merger divestment of parallel R&D. The paper has sketched a legal standard to evaluate rival claims over such assets based on how undertakings might use them to innovate differently. The paper further examined the implications of the theory of disruptive innovation. Despite its importance being generally recognized, disruptive innovation has so far received scant attention in competition law. This is perhaps because, under the theory of disruptive innovation, disruptors start off as less productively efficient than the incumbents they will displace. The exclusion of potential disruptors may be mistaken for the efficient functioning of a market which remains competitively open—but only for undertakings that play by the same terms. The paper proposed guidelines to adjust for this, focusing tests on the targeting of potential disruptors and on measures aimed at containing potentially disruptive markets. Strategic management is characterized by a multitude of sources, across the spectrum of social sciences, and by an assumed lack of consensus on many issues. The resource-based view has been presented in such a way that purposely complements industrial economics. This was enough for the present purposes, as was the consensual view on the theory of disruptive innovation. Further steps should aim for equally well-established theoretical ground. Concrete cases must be solved based on intelligible criteria that can be translated into normative guidance and reviewed in a judicial context, as all technical contributions to competition law should be. Nevertheless, these contributions promise to change how competition law operates. Economic advances have concentrated on market structure at the expense of innovation, but this trade-off is not inevitable. The normative ambition of competition law is to regulate innovation, its legal concepts are open to it, and competition authorities are willing. Economics are now in a position to catch up, as Dow/DuPont shows, and more cases of harm to innovation are to be expected. It is those cases, and not the policing of market power, which will define the EU’s policy on innovation. TILEC has received funding from Qualcomm Inc., and the research conducted there was in accordance with the Royal Dutch Academy of Sciences (KNAW) Declaration of Scientific Independence. This article refines the ideas set out in ‘Innovation in EU competition law: The resource-based view and disruption’, Jean Monnet Working Paper 2/17, NYU School of Law up to spring 2018. I thank the reviewers and Professor Ioannis Lianos for their instructive comments. Footnotes 1 <www.economist.com/blogs/theinbox/2007/07/the_perils_of_prediction_june> accessed 22 November 2018. 2 One only has to look at the breadth of Horizon 2020, the EU Framework Programme for Research and Innovation: <http://ec.europa.eu/programmes/horizon2020/en/what-work-programme> accessed 22 November 2018. 3 ‘A Single Market for Intellectual Property Rights: Boosting creativity and innovation to provide economic growth, high quality jobs and first class products and services in Europe’ COM(2011), 287. 4 ‘Mid-Term Review on the implementation of the Digital Single Market Strategy: A Connected Digital Single Market for All’ COM(2017), 228. 5 State aid control is also part of competition law, but these and other issues of public competition law—despite their undeniable impact on innovation—will not be covered by this paper. 6 Eg, Shapiro, ‘Competition and Innovation: Did Arrow Hit the Bull’s Eye?’ in Lerner and Stern (eds), The Rate and Direction of Inventive Activity Revisited (Chicago: University of Chicago Press, 2012) 362. 7 Porter, ‘Competition and Antitrust: A Productivity-Based Approach to Evaluating Mergers and Joint Ventures’ (2001) 46 Antitrust Bulletin, 919 (Revised 30 May 2002) <http://www.isc.hbs.edu/Documents/pdf/053002antitrust.pdf> accessed 22 November 2018, 4. 8 See, eg, Lianos and Dreyfuss, ‘New Challenges in the Intersection of Intellectual Property Rights with Competition Law—A View from Europe and the United States’, CLES Working Paper 4/2013 32. 9 Authors such as Sidak and Teece, Wu, and Kerber have advocated a major shift but, as Lianos notes, most competition scholars would agree that innovation is not the focus of competition law and that a move in that direction is desirable. Sidak and Teece, ‘Dynamic Competition in Antitrust Law’ (2009) 5 Journal of Competition Law & Economics, 610; Wu, ‘Taking Innovation Seriously: Antitrust Enforcement if Innovation Mattered Most’ (2012) 78 Antitrust Law Journal, 328; Kerber, ‘Competition, Innovation, and Competition Law: Dissecting the Interplay’, MAGKS Joint Discussion Paper 42-2017 12; and Jenny, Lianos, Hovenkamp, Marshall, and Thambisetty, ‘Competition Law, Intellectual Property Rights and Dynamic Analysis: Towards a New Institutional “Equilibrium?”’ (2014) 42013 Concurrences, 13, 2. 10 Wright and Petit question whether competition law has (yet) acquired the toolset for handling innovation and emphasise casuistic analysis. Wright, ‘Antitrust, Multi-Dimensional Competition, and Innovation: Do We Have an Antitrust-Relevant Theory of Competition Now?’, in Manne and Wright (eds), Regulating Innovation: Competition Policy and Patent Law Under Uncertainty (Cambridge: Cambridge University Press, 2011) <http://ssrn.com/abstract_id=1463732> accessed 22 November 2018, 32, and Petit, ‘Significant Impediment to Industry Innovation: A Novel Theory of Harm in EU Merger Control?’, ICLE Antitrust & Consumer Protection Research Program White Paper 2017-1 21. However, few have argued as expressly (and convincingly) as Ibáñez against addressing innovation directly. Ibáñez Colomo, ‘Restrictions on Innovation in EU Competition Law’ (2016) 41 European Law Review <http://ssrn.com/abstract=2699395> accessed 22 November 2018. 11 Unless differentiated as ‘General Court’, ‘Court’ will refer to the Court of Justice. 12 For all, Kerber (fn 9), 6. 13 Wright (fn 10), 20. Shapiro calls this ‘contestability’, a view which is compatible with any relationship between market structure and innovation. Shapiro (fn 6) 364. 14 Hovenkamp, ‘Antitrust and Innovation: Where We Are and Where We Should Be Going’ (2011) 77 Antitrust Law Journal, 751 and Kerber (fn 9), 4. 15 That is the gist of listing innovation alongside price, quality, and other competitive parameters, as part of the definition of market power. Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings, OJ 2009, C 45/2 (Article 102 TFEU Guidance), 11 and Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ 2004, C 31/3 (Horizontal Merger Guidelines), 8. 16 For all these authors, see Wu (fn 9), 328. 17 This would affect the analytical framework but also competition authorities’ enforcement priorities. 18 For all these authors, see Ibáñez (fn 10), 8. 19 Case M.7932 Dow/Dupont. 20 Kerber (fn 9), 20. 21 Christensen, Raynor, and McDonald, ‘What is disruptive innovation’ (2015) Harvard Business Review <https://hbr.org/2015/12/what-is-disruptive-innovation> accessed 22 November 2018, 4. 22 Eg, Peteraf, ‘The Cornerstones of Competitive Advantage: A Resource-Based View’ (1993) 14 Strategic Management Journal, 179. 23 Audretsch and others make the interesting observation that the Chicago school assumes that the supply of entrepreneurs is infinite in the long run, while the Austrian school argues that there is a ‘scarcity of entrepreneurial resources’. Audretsch, Baumol, and Burke, ‘Competition Policy in Dynamic Markets’ (2001) 19 International Journal of Industrial Organization, 619. 24 Kerber (fn 9), 5. 25 Hovenkamp (fn 14), 749 and Wu (fn 9), 314. 26 Wu (fn 9), 325. 27 Ibáñez (fn 10), 22. 28 Joined Cases C-241-2/91 P, Magill, C:1995:89, 50. 29 Petit, ‘The Antitrust and Intellectual Property Intersection in European Union Law’ (2016) <http://ssrn.com/abstract=2796670> accessed 22 November 2018, and Lianos and Dreyfuss (fn 8). 30 Lianos, ‘Competition Law and Intellectual Property Rights: Is the Property Rights’ Approach Right?’ in Bell and Kilpatrick (eds), (2006) 8 Cambridge Yearbook of European Legal Studies. 31 Kathuria, ‘A Conceptual Framework to Identify Dynamic Efficiency’ (2015) 11 European Competition Journal, 321. 32 Dynamism comes not only from change but also from how market actors foresee and adapt to it. Sidak and Teece (fn 9), 603. 33 This is a ‘broad’ view of dynamic efficiency, as categorized by Kathuria, namely one which focuses on (existing versus changed) knowledge. Kathuria (fn 31), 330. 34 As Peteraf defines it, the resulting rents ‘cannot be attributed to an artificial restriction of output or to market power’. Peteraf (fn 22), 181. 35 Peteraf (fn 22), 180. 36 Teece, ‘Intangible Assets and a Theory of Heterogeneous Firms’, Tusher Center for Intellectual Property Management Working Paper 4 19 (2015) and Kerber (fn 9), 10. 37 The subjects of competition provisions in the TFEU, understood as entities engaged in economic activity regardless of their legal form (which sets economic activity as the object of competition law). 38 Regulation 139/2004 on the control of concentrations between undertakings, OJ 2004, L 24/1 (Merger Regulation). 39 Larger firms can even have greater incentives because of their larger volume of production. Shapiro (fn 6), 366. Even if the threat of exit in competitive markets would carry a more powerful incentive, it would still not be enough to correlate productive efficiency and market structure. 40 Case C-209/10, Post Danmark I, C:2012:172, 41. 41 Horizontal Merger Guidelines (fn 15), 76. 42 Audretsch and others comment that Article 101(1) TFEU is ‘a useful device for students seeking to memorise the list of possible anti-competitive acts of a static monopolist’. Audretsch and others (fn 23), 627. 43 Case C-67/13 P, CB, C:2014:2204, 51. 44 Whish and Bailey, Competition Law, 8th edn (Oxford: Oxford University Press, 2015), 560. 45 The latter claim could be made for cartels dealing with sector overproduction, but the Court rejected the idea of examining the ulterior benefits of stabilizing the sector. Case C-209/07, BIDS, C:2008:643, 34. 46 Wu (fn 9), 316. 47 Case C-202/07 P, France Télécom, C:2009:214, 110. Without recoupment the allocative inefficiency is merely temporary and to the benefit of consumers. 48 Case C-52/09, TeliaSonera, C:2011:83, 32 and Case C-413/14 P, Intel, C:2011:83, 136. 49 Guidelines on the application of Article 81(3) of the Treaty, OJ 2004, C 101/97 (Article 101(3) TFEU Guidelines), 36 and Horizontal Merger Guidelines (fn 15), 86. 50 Article 101(3) TFEU Guidelines 46. 51 Sidak and Teece (fn 9) 614. 52 Kerber (fn 9) 4. 53 Notice on the definition of relevant market for the purposes of Community competition law, OJ 1997, C 372/03 (Notice on Market Definition), 15. 54 As well as from countervailing buying power. Article 102 TFEU Guidance 12. 55 Another instance of dynamic reasoning which incorporates potential competition is the theory of ancillary restraints under Article 101 TFEU, as discussed in Section III.C. 56 Case 85/76, Hoffmann-La Roche, C:1979:36, 91. 57 Case C-62/86, Akzo, C:1991:286, 60. 58 Although Continental Can, the judgment which opened the way for merger control by considering the acquisition of an undertaking as a possible abuse under Article 102 TFEU, is typically static: the Court was concerned with to avoid a contradiction between allowing such acquisition and cartel enforcement under Articles 101 TFEU. Case 6/72, Continental Can, C:1973:22, 25. 59 Concentrations are caught by the Merger Regulation according to turnover and jurisdictional thresholds which, although sizeable, are no guarantee of market power. The substantive test is whether the concentration significantly impedes competition, in particular by the ‘creation or strengthening of a dominant position’. Merger Regulation (fn 38), Art. 2(2). 60 Horizontal Merger Guidelines (fn 15), 70. 61 Productive inefficiency related to market structure is ‘x-inefficiency’—the costs of maintaining market barriers, for example policing a cartel or lobbying for market barriers—insofar as such costs are borne by undertakings (they may also involve negative externalities, such as corruption or excessive regulation). 62 It is assumed that, if the price does not move, quality degradation is consumer surplus which is (inefficiently) sacrificed. 63 Stucke and Ezrachi, ‘The Curious Case of Competition and Quality’, University of Tennessee Legal Studies Research Paper 256 (2014), 9. 64 Article 102 TFEU Guidance 22. 65 The ability to degrade quality or impose costs may depend on market power but, as already remarked, this is not linked with the tests of anticompetitive behaviour. 66 Schilling, ‘Towards Dynamic Efficiency: Innovation and its Implications for Antitrust’, (2015) 60 Antitrust Bulletin, 206 and Kerber (fn 9), 9. 67 Streel and Larouche, ‘Disruptive Innovation and Competition Policy Enforcement’, TILEC Discussion Paper 2015-21 2. 68 Kathuria (fn 31), 338. 69 Kathuria (fn 31), 325. 70 This is similar to the definition of innovation adopted by Waller and Sag: ‘any change in the status quo that (i) allows one to do something one could not do before or (ii) allows one to do something already possible while using fewer resources than were required before’. Waller and Sag, ‘Promoting Innovation’ (2015) 100 Iowa Law Review <https://ssrn.com/abstract=2479569> accessed 22 November 2018, 3. 71 The ‘narrow’ perspective of dynamic efficiency. See fn 33. 72 Article 101(3) TFEU Guidelines 54 and Horizontal Merger Guidelines (fn 15), 81. 73 The Commission’s Merger and Innovation Policy Brief, accounting for its practice on innovation, gives the example of the pharmaceutical and medical device sectors, which allows competitors to be identified at an early stage. <http://europa.eu/rapid/press-release_IP-16-3337_en.htm> accessed 22 November 2018. 74 Kern, ‘Innovation Markets, Future Markets, or Potential Competition: How Should Competition Authorities Account for Innovation Competition in Merger Reviews?’ (2014) 37 World Competition, 198. 75 Federico and others propose a model where total innovation decreases in mergers between close-competitors due to a reduction of R&D by the merging parties. This considers a possible positive incentive from higher coordination. Federico, Langus, and Valletti, ‘Horizontal Mergers and Product Innovation: An Economic Framework’ (2017) <https://ssrn.com/abstract=2999178> accessed 22 November 2018, 2. 76 Haucap and Stiebale present a model where a merger has a negative effect on both the merged entity and non-merging parties’ R&D in an industry with high R&D, backed up by data of pharmaceutical mergers. Haucap and Stiebale, ‘How Mergers Affect Innovation: Theory and Evidence from the Pharmaceutical Industry’, DICE Discussion Paper 218-2016 3. 77 Petit (fn 10), 8. 78 Kern (fn 74), 178 and 180 and Kerber (fn 9), 13. 79 Hence, Dow/DuPont analyses the effects of previous market concentration on R&D as a complement to other factors. Dow/DuPont V.8.5. 80 What Wu calls ‘evolutionary rather than planned’. Wu (fn 7), 316. 81 Streel and Larouche (fn 67), 3. 82 Case 322/81, Michelin, C:1983:313, 10. 83Post Danmark I 21. 84 As such, an allocative efficiency explanation of why dominant positions are allowed does not hold: the rents from market power do provide an incentive for entry, but market barriers prevent it. This nevertheless still conditions competition authorities to pursue exclusionary abuses in preference to exploitative ones. 85 This is also the conclusion of the US Supreme Court: ‘[t]he opportunity to charge monopoly prices—at least for a short period—is what attracts “business acumen” in the first place; it induces risk taking that produces innovation and economic growth’. Trinko, 540 U.S. 398, 407 (2004). 86 Porter (fn 7), 14 and Porter, ‘The Five Competitive Forces That Shape Strategy’, (2007) Harvard Business Review reprint 1. 87 Porter recognizes the link but favours dynamic competition, as strategic management studies do, and criticizes the static approach of US antitrust. Porter (fn 7), 15. 88 Greene and Yao, ‘The Influences of Strategic Management on Antitrust Discourse’ (2014) 59 The Antitrust Bulletin, 790. 89 As does the positional school just mentioned. Greene and Yao (fn 88), 793. An evolutionary and behavioural approach can also be applied to resource capabilities. Sidak and Teece (fn 9), 697. 90 Peteraf (fn 22), 180. 91 Peteraf and Barney, ‘Unravelling the Resource-Based Tangle’ (2003) 24 Managerial and Decision Economics, 315. 92 Peteraf and Barney (fn 91), 313. 93 Peteraf (fn 22), 186 and Peteraf and Barney (fn 91), 312. 94 Greene and Yao (fn 88), 796. 95 Peteraf (fn 22), 98 and Peteraf and Barney (fn 91), 313. 96 Peteraf and Barney (fn 91), 318. 97 Peteraf (fn 22), 93 and Teece (fn 36), 22. 98 Peteraf and Barney (fn 91), 317. 99 Kerber (fn 9), 19. 100 Peteraf (fn 22), 94 and Teece (fn 36), 8. 101 Kerber (fn 9), 18. 102 What Peteraf calls the cornerstones of competitive advantage: heterogeneity, imperfect mobility, ex ante limits to competition, and ex post limits to competition, are adapted here (without a strict correspondence) to the conditions mentioned. Peteraf (fn 22), 185. 103 Since the resource-based view is dedicated to productive efficiency, it can naturally add to the analysis of efficiencies. That is not covered by this paper, which focuses on harm. 104 Case 27/76, United Brands, C:1978:22, 207. 105 Whish and Bailey (fn 44), 195. 106 Dominance has been found below th 50 per cent market share, and even that presumption leaves ample room for non-short-term capacity adjustments. 107Michelin 57. 108Hoffman-La Roche 48. 109 Case C-7/97, Bronner, C:1998:569, 40. This is done separately from the condition of excluding all (static) competition. 110Bronner 46. This is connected with incentives for innovation regarding IP, as examined next. 111 T-175/12, Deutsche Börse, EU:T:2015:148, 157–179. 112 Kern (fn 74), 194. 113 Ibáñez (fn 10), 15. 114 The Commission does not mention innovation expressly, but refers to ‘high market potential’ and ‘products under development’. <http://europa.eu/rapid/press-release_IP-16-3337_en.htm> accessed 22 November 2018. 115 <http://ec.europa.eu/competition/publications/cpb/2016/2016_001_en.pdf> accessed 22 November 2018. 116 The decision leading to Intel therefore found that an exclusivity obligation had ‘a negative impact on AMD’s ability to recover its investments in research and development and thus on its incentive to engage in similar activities in the future’. Ibáñez (fn 10), 22. 117 From market foreclosure to acquiring rivals and raising their costs, exclusionary abuses under Article 102 TFEU and restrictions of competition under Article 101 TFEU cover a wide range of competitive disadvantages. This is mostly due to competition law’s static outlook, for which such disadvantages are enough to negatively impact competitive parameters. 118 See fn 59. 119 Shapiro (fn 6), 383. 120Dow/DuPont thus refers to market power event though, as mentioned below, it does not consider harm to innovation in relation to defined markets. Dow/DuPont 2004. 121 Case C-418/01, IMS, C:2004:257, 38. There is also a third condition: that the refusal is unjustified. Subsection III.B will discuss objective justification. 122 Case T-201/04, Microsoft, T:2007:289, 563 and 647. In Clearstream the General Court also lowered the exclusion required in Bronner for abusive refusal to supply an input, and confirmed its indirect effect on innovation. Case T-301/04, Clearstream, T:2009:317, 149. 123Microsoft 656. 124 Article 102 TFEU Guidance 78 and 81. The Commission follows Microsoft’s loose interpretation by analysing (all) refusals to supply under the lower level of exclusion and ‘consumer harm’. 125 Horizontal Merger Guidelines (fn 15), 8. 126 Horizontal Merger Guidelines (fn 15), 20(b). In general, the guidelines advise interpreting market shares according to the dynamics of innovation and growth. Horizontal Merger Guidelines (fn 15), 15. 127 Horizontal Merger Guidelines (fn 15), 37. 128 The guidelines state that a merger might, as an alternative to concerns over pipeline products, result in a better ability to innovate and an incentive for competitors to do the same. Horizontal Merger Guidelines (fn 15), 38.The Merger and Innovation Policy Brief tries to solve this ambiguity by leaning on market structure, but its requirements of competitive constraint and market barriers are neither explicit in the guidelines nor in the cases. Merger and Innovation Policy Brief (fn 73), 3. 129 Again emphasizing that this is unrelated to market shares. Horizontal Merger Guidelines (fn 15), 38. Pipeline products are supposed to be ‘related to a specific product market’. 130 Merger and Innovation Policy Brief (fn 73), 4. 131 A different example of harm through the ‘acquisition’ of parallel R&D would be the Commission’s ongoing investigation of Google’s copying of rivals’ web content—known as ‘scraping’—which might lead those rivals to abandon further efforts to innovate that content. However, it remains to be seen if this novel abuse will be found. <http://europa.eu/rapid/press-release_IP-16-2532_en.htm> accessed 22 November 2018. 132 The Merger and Innovation Policy Brief points to Medtronic/Covidien, Novartis/GSK oncology, and Pfizer/Hospira, where several drugs undergoing clinical trials were divested. The brief also mentions General/Electric Alston, where in addition to a pipeline product there was also the divestment of assets for improving an existing product. Merger and Innovation Policy Brief (fn 73), 4 and 5. 133Dow/DuPont 350. There are indications that a similar method was followed in Bayer/Monsanto. <http://europa.eu/rapid/press-release_IP-18-2282_en.htm> accessed 22 November 2018. 134 Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ 2008, C 265/7. Non-Horizontal Merger Guidelines (fn 134), 10. 135 Namely, if one of the parties is likely to expand in the near future due to ‘recent innovation’ (which can also be read as interpreting mergers in dynamic markets). Non-Horizontal Merger Guidelines (fn 134), 26. 136 Mergers in related markets (conglomerate mergers) will be analysed in Section IV on disruption. 137 Non-Horizontal Merger Guidelines (fn 134), 11. This is described as creating market barriers, but the guidelines define foreclosure as competitive disadvantage. Non-Horizontal Merger Guidelines (fn 134), 29. 138 Case M.5984, Intel/McAfee 342. In addition to Intel/McAfee, the Merger and Innovation Policy Brief refers to ARM et al. JV, Telefonica UK et al. JV, and Intel/Altera. Merger and Innovation Policy Brief (fn 73), 6–7. 139 <http://europa.eu/rapid/press-release_IP-15-4780_en.htm> accessed 22 November 2018. 140 <http://europa.eu/rapid/press-release_IP-17-1784_en.htm> accessed 22 November 2018. 141 See fn 114. 142 For example, divestment of pipeline drugs has included manufacturing equipment, IP and other rights, technology, scientific and regulatory material, and staff. Merger and Innovation Policy Brief (fn 73), 4 and 5. 143 Kerber (fn 9), 16. 144 Framed like this, the fate of the asset is a question of allocative efficiency; as discussed in Section II, the different types of efficiency and innovation are joined in a dynamic context. 145 Kern (rightly) links ‘specialised assets’ with the discoverability of innovation sources, but still considers those assets as market barriers and not resources granting a competitive advantage. Kern (fn 74) 197. 146 As ‘technical … progress’, regarding the appraisal of concentrations, Merger Regulation (fn 38), Art. 2(1)(b). 147 Sidak and Teece are sceptical that this can be averted. Sidak and Teece (fn 9), 585. 148 See fnn 63 and 64. 149 Petit (fn 10), 9. 150 See fn 137. 151 Gans, ‘When Is Static Analysis a Sufficient Proxy for Dynamic Considerations? Reconsidering Antitrust and Innovation’ (2011) 11 Innovation Policy and the Economy, 55. 152 As noted above, a monopolist benefits from increased productive efficiency. See fn 39. Furthermore, concentrated markets may (or not, depending on the facts) supress price competition and therefore increase incentives to innovate. Federico et al. (fn 75), 2. 153 The contestability of a market may vary, as Shapiro examines, and it should not be assumed that the incentive is the same for every undertaking—it may, as a structural approach holds, diminish with market power. In any event, Shapiro also differentiates between incentives (contestability and appropriability) and ability (synergies) to innovate. Shapiro (fn 6), 365. 154 Kerber (fn 9), 15. 155 The bundles of resources object of the resource-based view, as referred to above. This is also close to the ‘synergies’ analysed by Shapiro, which relates to the ability to innovate. Shapiro (fn 6), 365. 156 Not only for the parties but also for their competitors. 157 In this manner innovation is also assessed ‘indirectly’ but, contrary to the assumption that market power can negatively affect innovation (judicially validated in Deutsche Börse), the legal standard remains direct harm to innovation. 158IMS 49 and Microsoft 657. 159Microsoft 654. This would not be the case if these were the expected results of market structure incentives. 160Dow/Dupont 349–350. 161Dow/DuPont V.8.6. 162 Thus, the Merger and Innovation Policy Brief bundles ‘pipeline products that would likely have entered existing markets or that would have created entirely new product markets’, considering that there is a loss of innovation in both cases. Merger and Innovation Policy Brief (fn 73), 4. 163 Kern argues, against the seminal research by Sah and Stiglitz on the wastefulness of parallel research—and in line with this paper but not linked with the resource-based view—that firm heterogeneity may lead to different outcomes in pharmaceutical research even if the therapeutical goal is the same. Kern (fn 74), 200. If sufficiently proved, this could counter claims of lack of difference. 164Novartis/GSK oncology examined ‘new products that will be developed for the same product market’, raising concerns for both price and variety. Case M.7275, Novartis/GSK Oncology business 110. Ppfizer/Hospira stated that reduced incentives to innovate would ‘lessen price competition’. Case M.7559, Pfizer/Hospira 58. In contrast, Medtronic/Covidien concerned a product that would become ‘a strong contender for the market, including for indications for which [the acquirer’s] device is not currently approved’. Case M.7326, Medtronic/Covidien 249. 165 Vertical relationships are problematic in this regard: if the Commission follows the Non-Horizontal Guidelines to the letter, it will apply a (structural) test of foreclosure which, as already remarked, does not even mention innovation. Intel/McAfee nonetheless considered ‘rapid innovation’ at the level of market definition. Intel/McAfee 109. 166 As Baker notes, appropriability can be argued as a defence against an offence, as a reason to interpret competition rules narrowly, and in relation to remedies. Baker, ‘Evaluating Appropriability Defenses for the Exclusionary Conduct of Dominant Firms in Innovative Industries’ (2016) 80 Antitrust Law Journal, 431. 167 Appropriability is highlighted by Shapiro as an incentive for innovation, and connected to IP rights. Shapiro (fn 6), 364. 168 Sidak and Teece (fn 9), 592. 169 Baker argues that the harm to appropriability from intervention can be surpassed by the added incentive for rivals to innovate due to protection against exclusion. Baker (fn 166), 437. This applies to both an indirect approach and the harm to innovation argued in this paper. 170 Article 102 TFEU Guidance 75. 171Magill joined the appeals from the General Court of two of them. 172 The Court stating that ‘[i]t is sufficient that a potential market or even hypothetical market can be identified’. IMS 44. 173 In addition to abuse, the analysis of every asset of an undertaking would consume merger control, with the risk of having to divest internal assets that have a significant market value. 174Microsoft 669. 175 This would explain why the abuse was seldom pursued, but still leave significant room—namely where IP protection is extended beyond its duration or the products covered, as was indeed later found abusive. Case C-567/14, Genentech, C:2016:526 and Case C-385/07, Grüne Punkt, C:2009:456. 176Microsoft 690. An objective justification can also be based on preserving the integrity of the IP, as it is generally accepted for the refusal of physical inputs like infrastructure. 177Microsoft 696–698. This negative impact on incentives is examined in itself, not in a balance with industry incentives to innovate, as the General Court rightly rejected. Microsoft 710. 178Microsoft 700–702. 179 This was highlighted incidentally when the General Court discussed that the protection of secret information—another way to secure appropriability—was on a par with public IP. Microsoft 693. 180Microsoft 654 and 657. 181 Whish and Bailey (fn 44), 135. Some restrictions of competition, like exclusivity or non-compete obligations, have been found necessary for market entry or business acquisitions—in other words, for the appropriability of those investments. The Court has stated this quite clearly in relation to investments in IP. Case 258/78 Nungesser, C:1982:211 and Case 262/81, Coditel II, C:1982:334. 182 If the ongoing investigation of Google results in finding scraping abusive this would be perfectly illustrated, as the issue is denying Google’s rivals the appropriability of their investment in content. 183 The distinction is quite fine since ancillary restraints also require objective necessity. Whish and Bailey (fn 44), 135 and 136. The conceptual difference is that objective justification assumes there is harm, ancillary restraints conclude there is not. In any event, it should not be up to the dominant undertaking to prove that there will be copying or cloning, since they are not responsible for rivals’ behaviour. 184 Lianos calls for internalizing IP values, notably the promotion of incentives to innovate, in competition law enforcement. Jenny and others (fn 9), 3. 185 Baker (fn 166), 437. 186 See fn 83. 187 Schilling (fn 66), 193. 188 The valuation of risk has traditionally been a problem for finding exploitation in competition law. Whish and Bailey (fn 44), 760. 189 Peteraf (fn 22), 182. 190 As such, IP law is mostly about replicating and transferring the IP. This does not mean there is no appropriability problem if copying or taking the asset is free—competition law can itself step in to secure an incentive, as could be the case for scraping. 191 Ex ante competition can take the form of bidding for the asset, and ex post competition may lead to an increase in its supply. Peteraf (fn 22), 182 and 185. As such, favourable acquisition is not so much about allocative inefficiency as of preventing markets for the resource from appearing. 192IMS 29. 193 Wu argues, in relation to standard setting and platforms, that ‘[t]he platform that declares itself closed from the outset does not gain the advantages of inviting development on an open platform. The problem is with platforms that gain dominance based on a practice of serving as the entire industry’s basis for innovation and then later use that position to destroy any threats to their dominance’. Wu (fn 9), 324. 194 Former public monopolies have long been associated with the willingness to mandate access in the EU. Whish and Bailey (fn 44), 747. The Court states that legal monopolies should be taken into account in dominance and the Commission considers these monopolies, as well as the use of State resources, as indications that incentives will not be harmed by intervention in an abusive refusal to supply. Post Danmark I 23 and Article 102 TFEU Guidance 82. 195 Whish and Bailey (fn 44), 19. 196 Christensen et al. (fn 21), 5. 197 Owings, ‘Identifying a Maverick: When Antitrust Law Should Protect a Low-Cost Competitor’ (2013) 66 Vanderbilt Law Review, 344. 198 Streel and Larouche (fn 67), 3. 199 Gans, ‘The Other Disruption’ (2016) Harvard Business Review <https://hbr.org/2016/03/the-other-disruption> accessed 22 November 2018, 2. 200 Christensen and others (fn 21), 7. 201 Streel and Larouche comment that disruptive innovation ‘comes from the blind side of incumbent firms’. Streel and Larouche (fn 67), 3. 202 Christensen (who coined the theory of disruption) and others argue that Uber, often given as an example of disruption, is one such case: ‘Uber’s service has rarely been described as inferior to existing taxis; in fact, many would say it is better’. Christensen and others (fn 21), 6. 203 The so-called ‘second phase’, becoming mainstream (after the ‘first phase’, the inefficient start). Streel and Larouche (fn 67), 4. 204 Christensen and others (fn 21), 5. 205 The many instances of disruption also exemplify an increase in productive efficiency: streaming and private rentals needed to become as convenient as dvds and hotels, and cars and mobile phones as affordable as horses and land lines. This goes to show that the border with disruption through the firm is fluid, since production methods play an important role in increasing the utility or lowering the cost. 206 Streel and Larouche (fn 67), 6. 207 See fn 196. In market definition terms, disruptive products start in the same market if they substitute incumbent products by compensating lower value with lower price, and in a different market if the value or price gap does not allow substitution. After disruption the market is redefined and marginalizesthe incumbents, either within the same market or in a reduced separate market. 208 Christensen and others (fn 21), 8. 209 Christensen and others (fn 21), 11. 210 Schilling (fn 66), 194. 211 It is however also accepted that public power can be used to prevent disruption, notably regulation protecting incumbents. OECD, Key Points of the Hearing on Disruptive Innovation, DAF/COMP/M(2015)1/ANN8/FINAL, 3. This could come under the scope of Article 106 TFEU, which extends competition rules to undertakings with special or exclusive rights. This paper does not discuss the case law on Article 106 TFEU, which has indeed covered the prevention of technical progress. 212 Curiously, management studies speak of ‘organisational slack’ as enhancing the experimentation and risk taking necessary for innovation—another example of the lack of a relationship with market structure. Schilling (fn 66), 198. 213 Wu remarks that ‘innovation and exclusion are alternative responses to an external challenge’. Wu (fn 9), 319. 214 Waller and Sag (fn 70), 5 and Streel and Larouche (fn 67), 7. 215 Waller and Sag (fn 70), 19. 216 Streel and Larouche (fn 67), 9. 217 The acceptance of remedies means that the Commission’s claim of harm to innovation, as well as the adequacy of the remedies, goes unchallenged. Such a challenge was unsuccessful in Deutsche Börse Brief (fn 73), 6. 218 The main (and sometimes only) example is Microsoft’s tying of Windows with Internet Explorer, which led to a judicial decision in the United States and commitments in the EU. Waller and Sag (fn 70), 8 and Streel and Larouche (fn 67), 8. The claim is that browsers could have disrupted operating systems by shifting processing from the personal computer to the internet (which also did not happen). 219 Waller and Sag (fn 70), 2. 220 Highly priced mergers might still be loss-making but for the premium of preventing disruption. Streel and Larouche (fn 67), 9. As disruption is uncertain and the potential disruptor is inefficient, the premium may not be fully reflected in the acquisition price. 221 Streel and Larouche (fn 67), 6. 222 Audretsch and others comment how ‘excessive innovation’ may be used to prevent resource-constrained competitors from evolving. Audretsch and others (fn 23), 627. 223 If a disruptive product is as efficient as incumbents it will have already surpassed them, making acquisition or exclusion much more difficult. There is a good argument there for competition law not to intervene to safeguard innovation but to remain vigilant on market structure. 224Magill might appear as a disruption without an inefficient start: as Magill’s multi-schedule guide would be immediately preferable to the broadcasters’ single-schedule. However, broadcaster magazines offered entertainment content which Magill did not—it only combined schedules. Hence, Magill’s guide would have to improve its content in order to be successfully disruptive (as television guides have since done). 225 This condition is enough for disruption, and does not require reading the Court’s reference to ‘the development of the secondary market to the detriment of consumers’ as the market for the disruptive product. IMS 48. The primary market in the Court’s reasoning is not the incumbent’s product but, as discussed earlier, the market for the asset. 226 See fn 192. 227Microsoft 656. 228 Therefore, there is a strong incentive to collude not to disrupt a (still) competitive market. This is what happened in Wouters, where a bar association prohibited lawyers from practising together with accountants, as these ‘one-stop-shops’ would have a disruptive advantage over traditional lawyers. The Court duly found a restriction of technical development under Article 101 TFEU but, instead of becoming a leading case on disruption, Wouters also considered such restriction necessary for the ‘proper practice of the legal profession’ (not coincidentally, similar to the kind of public regulation that keeps disruption at bay). Case C-309/99, Wouters, C:2002:98, 90 and 107. 229 Non-Horizontal Merger Guidelines (fn 134), 48. 230 Non-Horizontal Merger Guidelines (fn 134), 93. 231 Horizontal Merger Guidelines (fn 15), 20. 232 Horizontal Merger Guidelines (fn 15), 37 and 42. Entry and non-coordination are however only the expectations of a competitive market, so that they add little to a market structure approach. Indeed, the concept of a maverick—an undertaking which ignores market incentives—is paradoxical to such approach. 233 Owings (fn 197), 328. 234 Waller and Sag (fn 70), 17 and Streel and Larouche (fn 67), 10. 235 Many users left on the news of Facebook’s acquisition, showing that it served niche consumers on privacy—an indication of disruptive potential. Case M.7217, Facebook/WhatsApp footnotes 79 and 174. 236Facebook/WhatsApp 107, 158, and 165. 237 As Facebook stated that it would not do so during merger control, the Commission fined it for providing misleading information. <http://europa.eu/rapid/press-release_IP-17-1369_en.htm> accessed 22 November 2018. This was the only option available, since the Commission did not consider WhatsApp’s innovative competition on data protection and Facebook’s move to scupper it. Costa-Cabral and Lynskey, ‘Family Ties: The Intersection of Data Protection and Competition Law in EU law’ (2017) 54 CML Rev, 38. 238 Communication services can evolve into social networking, as has happened with the leading social network in China (WeChat). 239 Case T-79/12, Cisco, T:2013:635. 240Cisco 117, citing Case C-12/03 P, Tetra Laval, C:2005:87, 44. 241Cisco 118. The General Court further connects this to future decisions ‘made possible and economically rational by the alteration of the characteristics and the structure of the market’. 242Microsoft 119. 243Microsoft 121. 244Microsoft 125–127. The reference to ‘traditional telephone communication’ is particularly telling. 245Microsoft 134. 246 As Hovenkamp observes, the claim that innovation is exclusionary is one ‘that antitrust wisely either rejects or else limits to situations where the innovation at issue is no innovation at all, but only an attempt to contrive incompatibility with the complementary products of rivals’. Hovenkamp (fn 14) 751. 247 As confirmed in Microsoft regarding the tying of Windows with media players. Microsoft 1088. This may, however, have influenced Microsoft’s decision not to bundle a search engine with the next version of Windows, which according to Streel and Larouche allowed the disruption of operating systems by Google making the search engine ‘the central stage in the ecosystem’. Streel and Larouche (fn 67), footnote 12. 248 A product may be designed from the ground up to be integrated. Thus, in order to find tying, Microsoft referred to several instances of separate commercialization. Microsoft 917 and 925. 249 Ibáñez (fn 10), 21. 250 Owings (fn 197), 344. 251 Owings (fn 197), 349. 252 Hence, adjusting merger thresholds to innovative assets may pass by complementing turnover with the value of the transaction. Streel and Larouche (fn 67), 9–10. 253 This may be spelled out by incumbents but does not have to be, as purpose can be inferred from contextual factors. Case C-549/10 P, Tomra, C:2012:221, 20. 254 Streel and Larouche (fn 67), 11. 255 Audretsch and others (fn 23), 627. 256 Even though the tying of Windows with Internet Explorer has been understood as a defence against disruption, only the practices which Microsoft could not show to be ‘innovation for the benefit of its customers’ were prohibited. Waller and Sag (fn 70), 10. 257 Streel and Larouche (fn 67), 6. 258 Streel and Larouche (fn 67), 5. 259 Owings argues that constraints on price and coordination are not merger specific. Owings (fn 197), 343. In any event, these constraints are sufficiently represented in tests of structural harm. 260 This includes the obligation, in merger control, not to create conditions ‘in which abusive conduct is possible and economically rational’ (found in relation to discrimination). Tetra Laval 79. 261 Although granting (above cost) favourable prices to rivals’ customers is not objectionable in itself, this can be abusive if it is part of a strategy to ‘drive that competitor from the market’. Post Danmark I 29. 262 Paying to delay, cancel, or restrict the marketing of a rival’s products. Case T-286/09, T:2014:547, Intel, 212. The presence of IP can conceal or complicate this. Hovenkamp (fn 14), 753. 263 First, prices below average variable costs are presumed to have an exclusionary intent, and second, such intent will also make prices below average total costs abusive. France Télécom 109. Therefore, targeting potential disruptors would prevent rebutting the presumption and establish abuse, respectively. Selective predatory practices are in any event easier to implement. Article 102 TFEU Guidance 72. 264 The Commission states that it will examine refusals ‘to punish customers for dealing with competitors’ according to the market foreclosure demanded for exclusivity. Article 102 TFEU Guidance 77. However, this is barely different from paying customers to raise obstacles to competitors, for which the Commission does not require foreclosure. Article 102 TFEU Guidance 22. 265Post Danmark I 25. 266 Therefore, the Court’s definition of competition on the merits as ‘the departure from the market or the marginalisation of competitors that are less efficient and so less attractive to consumers’ should not be read literally in relation to potential disruptors. Post Danmark I 22. 267 Ibáñez argues that the ‘exceptionality’ of a refusal to supply should apply to discrimination. Ibáñez Colomo, ‘Exclusionary Discrimination Under Article 102 TFEU’ (2014) 51 Common Market Law Review, 154. 268 The increase in productive efficiency can again be projected based on a resource-based view, in particular by the mobility and replicability of resources. 269 Streel and Larouche (fn 67), 4. © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals-permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Innovation in EU Competition Law: The Resource-Based View and Disruption
JO - Yearbook of European Law
DO - 10.1093/yel/yey019
DA - 2018-01-01
UR - https://www.deepdyve.com/lp/oxford-university-press/innovation-in-eu-competition-law-the-resource-based-view-and-NYuZSdLnnG
SP - 305
VL - 37
IS - 
DP - DeepDyve
ER -