Merger remedies: fostering innovation?†

This article describes to what extent EU merger control practice takes into account the specificities of different sectors when assessing innovation effects, in the context of applying remedies. The pharmaceutical, ICT and telecom sectors are exemplified via a number of influential cases. It will be stated to what extent innovation goals are similar across sectors and why it should be sufficiently evaluated whether such goals will be accomplished by a certain merger control decision.


. Innovation and competition law
In recent years, it has been researched to what extent innovation is a goal of competition law. As regards the EU, the nexus between competition and innovation has been specifically described in the field of intellectual property rights (IPRs) and Article 101 and 102 Treaty on the Functioning of the EU. 1 The Technology Transfer Block Exemption Regulation and the Technology Transfer Guidelines seem concerned with creating an innovative environment. 2 Moreover, the EU Commission and Courts have tried to stimulate innovation by providing, under exceptional circumstances, for compulsory licencing of IPRs, in order for competitors to build further upon this knowledge. 3 Thus, in recent years, these institutions have become more aware of the effects of their decisions on technological development. The above-mentioned aspects of the practice of the Courts and Commission are now well known; however, it has not yet been sufficiently researched to what extent the approaches in other fields of EU economic law are in line with these findings.

Innovation and merger control
A field for which it is not clear yet to what extent it contributes to innovation, is merger control. There is still not enough transparency on why the remedies applied would be optimal for competition and for innovation. 4 Remedies can be too restrictive on contract freedom, or, on the other hand, not protect competition enoughboth restricting innovation. The hypothesis is that it is still not sufficiently explained how innovation effects are taken into account. What specifically needs to be researched more, is how approaches can differ per sector. 5 A concentration can remove incentives to innovate for the merging partiesfor example, two pharmaceutical companies working on competing medicines. A merger can then lead to one type of medicine not being developed anymore. This is something the Commission wants to prevent; a remedy can be selling off a particular part of the business. 6 In the ICTsector, the analysis focuses more on other companies' incentives to innovate. Here, the Commission looks at the merged firm's ability to foreclose competitors. 7 Specific concerns arise in areas such as the Internet, where it is easy for consumers to switch to a competing service. The net neutrality debate is an example of this. 8 The sectors thus seem to need differentiation.

Merger remedies and innovation
1.2.1. Interoperability There are multiple ways in which mergers in innovative markets can lead to competition concerns. Therefore, the Commission often tries to resolve these concerns by only accepting a concentration when remedies are put in place. A sufficient extent of interoperability can be a solution, for example. Interoperability means: "the ability to exchange and use information", which is not an absolute ability, but a matter of degree. 9 Merging parties often have to ensure that their future products will still work well together with those of competitors, because otherwise, the merger will not be cleared. The competitors will then not be driven out of the market; demand for their products will not decline for this reason. A commitment to provide interoperability information can thus be beneficial to competition, since it can facilitate new entrants or improve sales of competing products. Although not as intrusive as having to divest a certain product business, these interoperability remedies impact heavily on a firm's contract freedom. That is why they have attracted critical remarks as well. According to Hoehn and Lewis, firms are in this way subject to regulatory oversight for a longer period of time. Such behavioural commitments are given for the long term, and it has to be monitored whether they are complied with, so measures of this kind are even said to be "quasi-structural". 10 1.2.2. Intellectual property issues regarding interoperability remedies Interoperability standards and protocols are frequently based on IPRs. Often, IPRs owned by the notifying parties have to be licenced before rival products can sufficiently work together with those of the merged entity. Other possible solutions are: divestment of intellectual property (IP) to an independent body, or a commitment to implement a certain protocol on existing and/or future products (see, e.g. the Cisco/ Tandberg case). IP can be licenced on a royalty-free basis or for a royalty payment. Often, licences have to be granted on "fair, reasonable and non-discriminatory" terms. Despite elaborate discussions in the literature, it is still not entirely clear what this means, leading to legal uncertainty. 11 Merger remedies of this type thus have to be examined with care, taking into account the specificities of each case.

Structure of the article
This article will evaluate whether the differences between sectors are sufficiently taken into account by the Commission when assessing merger remedies. The following research question will be answered: "To what extent does EU merger control practice take into account the specificities of different sectors when assessing innovation effects, in the context of applying remedies?" In paragraph 2, a general overview is given on the possible effects of mergers on innovation. In the third paragraph, it is explained which approach the EU Commission and Courts generally take when assessing innovation effects of merger remedies. In paragraph 4, the pharmaceutical, ICT and telecom sectors (as a network industry) are exemplified via a number of influential cases. Lemley has already described why competition law should differentiate between sectors when trying to foster innovation; this article will test this theory for the EU Commission's practice. 12 Paragraph 5 compares, and explains the differences between sectors. The conclusion states to what extent innovation goals in these different areas are linked and why it should be sufficiently evaluated whether such goals will be accomplished by a certain merger control decision.

The relevance of innovation for merger analysis
The innovativeness of a market can be relevant for merger analysis in various manners. A merger can provide a firm with a combination of assets that allow it to come up with a new product first. This is an effect on "competition in innovation". Innovations can comprise markets in themselvesan improvement of a certain product may result in a new market. The merged entity's technology is thus relevant when defining the market, but it also needs to be taken into account when assessing its market power. Access to an important technology increases a firm's bargaining power. However, if there is much product innovation in a given market, then even a company with a high market share in the current generation of products may not be considered dominant. 13 2.1.2. When does a merger impede competition? It is important to look whether there is competition in the market, or for the market. Competition for the market means that there is mostly a near-monopoly, but over the years, different (sometimes newly established) firms enjoy this position. The structure of the market then "changes radically". If the market will probably change radically in the short to medium term, a merger is not likely to harm consumer welfare. It can be problematic, though, when the merger itself alters the conditions of competition and makes entry more difficult. In Philips/Agilent Health Care Solutions, 14 the Commission explained that innovation was needed for a strong market position, and that market positions were not stable. In Hoffmann-La Roche/Boehringer Mannheim, 15 the Commission came to a similar conclusion. If one of the merging parties is not yet active in a certain market, but is developing a new technology, a merger may significantly impede effective competitionsince an important potential competitor may disappear. 16 The effects of mergers on innovation can thus vary significantly. Therefore, it is necessary to identify different types of cases in which they play a role.
2.2. Different types of transactions in innovative markets 2.2.1. Incumbent firms, new entrants, current and future markets To better understand the effects of mergers in innovative markets, they can be divided into categories, for example those described by Glader. One category comprises transactions between actors already active on a concentrated product/technology market where innovation is an important means of competition. A second category of cases relates to transactions between incumbent firms and potential new entrants. Successful innovation is often needed for entry. A third scenario is competition for future product markets. There are a number of alternative research and development (R&D) routes that can result in a particular product or technology. Possibly, there is currently no close substitute for the product or process that is being created. R&D can result in a completely new product, or a "new generation" product which renders the current version obsolete, for a large part. 17

New technologies which could belong to different markets
Apart from the previous, there are upstream R&D transactions which can create various new technologies and products, which could belong to different downstream markets. Such situations are rather complex to assess, because it is difficult to define the exact downstream market and its characteristicsat least at the time that the Commission is involved. In these cases, the merging firms often have a combination of R&D assets that are crucial for the development of various markets. IPRs are thus important for the Commission to keep in mind. Important IPR combinations may fall in the same hands after a merger (the same can happen in patent pools, which will not be considered in this article Innovation, R&D and IPRs can thus affect the analysis of a concentration in different ways. 3. The Commission's general approach to mergers and innovation 3.1. The Commission's test 3.1.1. Merger guidelines Having looked at the ways in which mergers can affect innovation, we will now see how the European Commission deals with such situations. It is mostly difficult to predict how technology (and thus the market) will develop. Usually, a certain "critical mass" of R&D and investment is needed to successfully launch a new technologya process which may take much time. However, once this process is successful, innovation leads to better finance and vice versa. The fact that the Commission needs to look at (still ongoing) R&D processes, entails legal difficulties. After all, before R&D is completed, it is not possible to tell with certainty what its effects will be on the marketwill it be a success or not? 19 Nevertheless, there are a number of guiding principles. The Horizontal Merger Guidelines 20 explain the way in which the Commission assesses concentrations. It looks at innovation when defining the relevant market, when describing the structure of the market, and when looking at barriers to entry. Both past and (possible) future innovations are relevant. When substantial investments in R&D are needed, this is viewed as a barrier to entry. On the other hand, offering an innovative product can be a way to enter into a market. The Commission looks at the extent to which the industry is concentrated, the level of specialization, the way the industry is developing globally, economies of scale/scope and IP portfolios. Even a firm with a high market share is less likely to be dominant when it has (many) competitors with very active R&D divisions. The Commission is primarily interested in the market leader and its closest competitor (or the most qualitative or innovative one) and according to Laskowska, this is a good starting point, but the Commission does not examine companies' R&D strength in a sufficient manner: its argumentation is too general and "insufficiently supported by elements of proof". 21 In many situations, the Commission does not clearly investigate the specific circumstances of each case and the specific effects on the competitors concerned. Instead, it only looks at what is likely in theory, while actual market circumstances might be different. 22 These will not be considered in this article, because they do not fall within the definition of 'concentration', and will thus not be scrutinized ex ante by the Commission.

Strong points in the Commission's analysis
However this criticism seems justified, the Commission's analyses contain many sound patterns of reasoning. When competition between leading firms declines because they show less innovation, this does not necessarily mean that there is less competition between leading innovators, since these might be different firms. Firms may show different types of innovation: one firm can focus on improving quality, while another prefers to enhance its advertising. This does not matter, as there is still competition causing innovation. 23 The Commission does find it problematic when merging firms carry out activities or own assets which are to a large extent complementary. If the new entity has a large amount of market power, this firm can for example refuse to licence its technology to competitors. 24 The Commission is especially concerned when important structural links between firms are created, for example when they divide IPRs strategically. When certain crucial patents are brought together in a single firm, this can lead to a dominant position. However, the parties can claim that efficiencies arising from the merger outweigh the drawbacks. The Commission is likely to approve the merger when innovation will still be important for competition, when there will still be enough actual and potential competitors, and when the acquired firm is currently not the acquiring firm's most important competitor with regard to R&D. When a merged firm has large financial resources, it is easier for this firm to be innovative, since it can more easily take the risk that a new product might fail, or use these financial means for research. It may be harmful, however, when competitors as a reaction limit their R&D activities or leave the market entirely. 25

Difficulties in the Commission's analysis
Mergers of R&D-intensive companies can thus lead to synergies, economies of scale, and improved R&D for the merged firm, but they may also give rise to competition concerns. According to Laskowska, the Commission "does not clearly show concrete harm to innovation" 26 in such cases. The Commission often merely assumes that the combination of two important innovators will have a negative impact on R&D, and that competitors will not sufficiently countervail this strong position. 27 In some cases, it discusses a very wide variety of factors, while in other cases only one or two are given attention. There are no clear explanations for this. In some cases, the Commission takes into account overlaps with 23 Ibid 5. other markets because there could be synergies, but in other cases it does not, again without a clear explanation. The dynamic character of many markets is taken into account, however, according to Laskowska, again in an unsystematic way. 28 In her opinion: […] analysis of harm to long-term competition is generally not an examination of potential harm but rather a statement. 29 […] leading innovation cases have not been identified, and […] the Commission disregarded basic economic insights in some cases, and inconsistent and perfunctory treatment of innovation concerns is not rare. 30 Thus, the Commission certainly shows awareness of the importance of innovation, but it has also received criticism for not being clear and consistent enough.
Other cases show the difficulties of assessing precise effects as well. In UTC/ Goodrich, 31 the Commission concluded that the concentration would lead to a near-monopoly on one of the relevant markets, and looked whether the merged entity's incentives to conduct R&D would be diminished. 32 In such cases, it does not have to be proven directly that innovation has suffered. When a rival firm is excluded from the market, or when a near-monopoly is created that is likely to last for a long time, this already means that competitors have fewer incentives to innovate, 33 and Ibáñez Colomo states that in this sense: […] the assessment of the effects of a merger between pharmaceutical companies […] is not fundamentally different from the acquisition of Aer Lingus by Ryanair. 34 This seems to imply that sectoral differences do not have a major role here. It is, however, debatable whether this statement is always correct, since very different interests are at stake. The threshold might be met more easily in one sector than in another.

Will competitors be excluded?
It is important to remain careful. When a merger would give a competitive advantage to the merged entity, this is in itself not a sufficient reason for the Commission to take measures. After all, when a firm has a strong position on the market, this can mean different things for competition. It can give other firms an incentive to be equally innovativeon the other hand, it can cause them to leave the market. When the merged entity's position is merely strengthened, without a near-monopoly being established, the Commission additionally has to show that this entity will not face sufficient competitive constraints after the merger. A more problematic situation can arise when the merger specifically enables the merged firm to come up with a new product. This first-mover advantage will however only significantly impede effective competition when the large amount of market power involved is likely to last for a longer time. 35 Admittedly, this is difficult to predict. That is why it is very important for the Commission to make use of the most recent economic theories on innovative markets. In this way, it can best assess whether a merger will still allow competitors to do their best to come up with new products, or whether the merged firm will have such an important combination of assets that it can exclude others from the market. Viewpoints for this are developing in case law, as we will see in the following paragraphs.

The Commission's solutions 3.2.1. Interoperability remedies
Despite the complexities of assessing the precise effects of mergers, there is a diversity of solutions to their possible negative effects on competition. Both behavioural and structural remedies are important. The Commission has much flexibility in determining how long commitments should last. Commitments that are shorter in duration allow for more adequate reactions to fast developments in the market; long-lasting commitments are a heavy burden on firms and are often not even necessary, according to Hoehn and Lewis. When there is a long-term interoperability remedy, this could limit the merged entity's return on investment. For competitors, incentives to invest and innovate may be reduced as well. They can continue selling the products at issue, so they will not be inclined to develop innovative technologies of their own. 36

Criticism on interoperability remedies
Hoehn and Lewis are also very critical regarding the practice of appointing a Monitoring Trustee to ensure that commitments are complied with. After all, this is a very intrusive measure. When a company is monitored in this way, it could make different decisions. For example: to alter a product, to stop developing or delivering a certain product, or not to introduce a new/next generation product. innovation. 37 Having to licence on royalty-free terms can remove incentives to invest for a merged firm, but on the other hand, competition can increase if the beneficiary could previously not gain access to the market. 38 There are thus, again, dilemmas to be dealt with.
The Commission thus makes use of various economic insights in order to assess the innovation effects of mergers, however, it still faces difficulties in applying a case-by-case approach.
4. Commission practice 4.1. Early decisions on mergers and efficiency 4.1.1. Problems with the Commission's reasoning When analysing the criticism on the Commission's merger practice, we can see to what extent it is appropriate, why shortcomings have come into being, and whether the Commission is succeeding in improving it. In order to do this, the Commission practice will be placed into a historical perspective.
A number of early cases show lines of reasoning that are not entirely convincing. In Ciba-Geigy/Sandoz, 39 the Commission explains that the R&D divisions of the companies concerned are large, causing synergies and other important advantages which competitors do not have. In Shell/Montecatini, 40 the merging parties argue that the market is developing quickly and new firms are likely to enter. The Commission only explains that R&D processes usually take long in this sector and that it is uncertain whether new parties will show upit does not further specify its reasoning for the concrete case. In Procter & Gamble/ VP Schickedanz, 41 a number of firms could potentially enter the market, but the Commission pointed out that it was unlikely that they would do so in time. In Tetra Laval/Sidel, 42 some competitors were the merged firm's equals in quality and innovation, but did not have such an extensive product line. The Commission did not indicate which differences in circumstances would influence its assessment. 43 So, these early cases show that economic circumstances were not dealt with in a consistent manner. 37 Hoehn and Lewis (n 9) 10. In Intel/McAfee, a commitment to provide interoperability information was made for a specific term, as usually. In GE/Instrumentarium, an unlimited term was given for licencing obligations.

Scope for improvement
So, how could it do better? First, some general remarks need to be made about innovative markets. Fatur's analysis shows some characteristics of the ICT sector. Merger control should be based on the same principles as in other sectors, but some aspects have to receive more emphasis than in traditional sectors. Concerns about innovation and its effects on the competitive process are more importantas is the case in some other sectors. While this is very difficult, predictions of how the market and technologies will develop must be as realistic as possible, in order to prevent both Type I and Type II errors. 44 There are often IPRs involved, adding questions on the scope and validity of these rights. The above-discussed examples of (compulsory) licencing can indeed be beneficial to competitiona competitor will be able to operate more effectively on the market. 45 As has been stated, it is important to determine whether there is competition in the market or for the market, and whether new products and product functionalities are means of competition. Economies of scale and network effects (caused by a concentration) can lead to first-mover advantages, possibly resulting in anti-competitive foreclosure. It is, however, important to keep in mind that a merger usually also enhances the innovative capacities of the merged entity. 46 The following paragraphs will discuss a number of examples showing that the Commission is making progress in better taking into account the specific circumstances of concentrations in technological sectors. However, it will also be pointed out that there is still room for improvement.

4.2.
Recent developmentstelecom sector 4.2.1. Hutchison/Telefónica UK When looking at some recent cases, we see that there is still much room for improvement. Nevertheless, the Commission seems to have developed more convincing lines of reasoning, for example with regard to the proposed Hutchison/ Telefónica UK concentration. Not every provider has its own network, so some (so-called "virtual providers") have to use another firm's network to offer their services. As a result of the takeover, there would have been fewer mobile network operators willing to host virtual providersweakening the negotiating positions of the latter. The merged entity would have been part of both network sharing arrangements, Mobile Broadband Network Ltd and Beacon, that are currently there in the UK. The merged entity would know everything about its competitor's network arrangements. The Commission feared that this would have negative effects on innovation in mobile infrastructure, for example new 44 A Type I error occurs when a merger is blocked which would not be anti-competitive. A Type II error occurs when a merger is allowed which is anti-competitive. 45 Andrej Fatur, EU Competition Law and the Information and Communication Technology Network Industries (Hart 2011) 217. 46 Fatur (n 45) 218. F. Kartner technology such as 5G. Hutchison offered remedies, which however did not resolve the concerns. A remedy was proposed to host one or two virtual operators on the merged entity's network. However, even then, mobile virtual operators would have been "commercially and technically dependent" on the merged entity. A promise was made to give virtual operators access to 4G and future technologies. The Commission however concluded that this would not prevent the expected adverse effects; this is why it decided not to allow the transaction. 47

Vodafone Airtouch/Mannesmann
In Vodafone Airtouch/Mannesmann, 48 the Commission referred to a future market: "the provision of advanced seamless pan-European mobile telecommunications services". Even though such services could not be provided yet for technical reasons, there was consumer demand. The merged entity, the Commission concluded, would be the only company that could provide these services in the short to medium term. This could incentivise it to refuse access to its network to rivals. 49 Nevertheless, the concentration was approved, since the merging firms' intention was to give third parties non-discriminatory access to the network. Similar cases in new markets caused concerns as well, but these have mostly been solved via commitments. 50 In other cases, the Commission referred to the evolving nature of the market as well; for example in the medical cases GE/Instrumentarium 51 and Bertelsmann/Burda/Springer-HOS-MM; 52 and the utilities case TXU Europe/EDF-London Investments. 53

Hutchison 3G UK/Telefónica Ireland, Telefónica Deutschland/E-Plus
In horizontal cases, increasingly use is made of use of behavioural remedies, such as access to infrastructure. This is exemplified by the Hutchison 3G UK/Telefónica Ireland 54 and Telefónica Deutschland/E-Plus 55 cases. According to Cook, Frisch 47 Hutchison/Telefónica UK (Case COMP/M.7612) [2016] OJ C357/08. Commission press release on this case: europa.eu/rapid/press-release_IP-16-1704_en.htm. and Novak, this is a "subtle refinement of the Commission's previous approach". The Commission's approach in earlier years was to let network-owning companies offer access via a pay-as-you-go model. In these more recent cases, new entrants had to make up-front payments for access to a dedicated "pipe" for voice and data traffic from the network. The reason behind this is to stimulate the new entrant to make optimal use of the network capacity that it acquires. 56

Liberty Global/Ziggo
A case in which innovation is directly addressed, is Liberty Global/Ziggo. 57 It deals with the only two linear premium pay TV film channels in the Netherlands (Film1 and HBO). A concern was that the merged entity could charge too high wholesale prices to other pay TVoperators, or even refuse to supply them certain channels. The main innovation problem, however, lies with the increased market power of the merged entity vis-à-vis TV broadcasters. According to the Commission, this could hamper the development of audio-visual over-the-top (OTT) services (the delivery of audio, video and other media over the Internet instead of the service provider's network). These services compete with traditional pay TV services. The merged entity could decide to supply certain channels, or give access to its network, only when the other party promises to limit its OTT activities. Broadcasters had already signed agreements which limited their possibilities to offer such OTT services. These risks were again tackled via commitments, including the divestiture of Film1. Most importantly, the merged entity will not contractually limit its rivals' OTT activities anymore, at least during 8 years. Apart from this, it has to maintain an adequate interconnection capacity through at least three uncongested routes into its Internet network in the Netherlands. 58 The Commission thus attaches great importance to a sufficient level of access to a network. In this way, it tries to prevent that companies owning a network exclude new entrants from the market. It is important, however that it also keeps in mind that the network-owning companies should have enough incentives to invest and innovate as well.

Intel/McAfee
In the ICT sector, the Commission focuses more on whether a concentration would harm the ability of the merged entity's competitors to innovate, instead of the merged entity itself. In the Intel/McAfee case, 59  being compatible with Intel's CPU's 60 and chipsets. The case was solved via the following remedy: Intel promised not to stop other providers of security software from operating on its chips or coming up with innovative solutions. In this way, competition and an innovative environment are protected, but the benefits of tighter integration of chips with security software are also attained. 61

ARM
A similar case is ARM. 62 This firm supplied IP architecture for application processors used in for example smartphones and tablets. It wanted to set up a joint venture (JV) in order to develop hardware-based security solutions for e-mail and mobile banking, inter alia. The Commission feared that, due to its strong position, ARM could diminish the interoperability of this hardware extension with software solutions competing with the JV, or refuse/delay the communication of technical information to competitors. The remedy applied made sure that it would not do so. The case Telefónica UK/Vodafone UK/Everything Everywhere seemed to cause fewer problems in the Commission's view. It concerned a JV for a mobile wallet platform to be set up by three out of four UK mobile network operators. The Commission did not think it likely that competitors would be foreclosed, and alternatives already existed. 63 A similar conclusion was reached in the Intel/Altera case. Altera made so-called field-programmable gate array (FPGA) chips that could speed up some tasks of Intel's CPU's, such as recognizing images or search. The Commission wanted to know whether Intel could foreclose Altera's competitors by not licencing its proprietary technologies that connect the CPU to the FPGA. The conclusion was that there was a useful alternative, namely a certain open standard interconnect technologyand Intel had already licenced (or made offers to licence) the technology at issue to FGPA competitors. 64

Specificities of ICT markets
Some specific aspects of ICT markets have to be kept in mind. 65 There can be switching costs for consumers. These costs make it more difficult for a new 60 Central Processing Units (used in computers). company to win customers, but also more profitable, since the new customer is then (to a smaller or larger extent) "locked in". When there is a radically new technology, this makes joint dominance less likely. Since there are typically high fixed costs, but low marginal costs, often only a few companies survive in these sectors, due to fierce competition in early phases. 66 High market shares are not decisive, since they are often transitory. This can be exemplified by the Logica/CMG and HP/Compaq cases. 67 The Commission can estimate how successful products under development will affect the market. Possible competing products under development are also taken into consideration. 68 Admittedly, the result after some time may be different in the "real world"but it is all that can be done at this point.
Network effects are taken into account as wellthey are common in the ICT sector. 69 The Commission looks whether the expansion of smaller firms is hindered or if their ability to compete is otherwise restricted by such effects. 70 When the market is already concentrated and the acquired company, although much smaller than the acquirer, is an important competitor, the risk of consumer harm is high. It is relevant whether a new firm on the market will probably be influential in the future. The Commission looks whether new entrants have "essential" access to R&D or IPRs. 71 The previous lines of reasoning specifically apply to horizontal effects. When there is a non-horizontal merger, other factors become relevant. A common concern in the ICT sector involves two-level entry theory; the presence of a vertically integrated firm can make it difficult for a new firm to enter at only one level. The Commission is primarily concerned when it is significantly more difficult to enter at one level than at multiple levels. This is often the case when a non-vertically integrated company's component does not work (optimally) with a component produced by the vertically integrated merged entity. Another danger is that a dominant firm can often leverage its position to another market via a concentration. 72

Google/DoubleClick
Conglomerate mergers usually entail lower risks for competition, because the products are not as closely connected. An example is the Google/DoubleClick 66 Ibid 220. Network effects mean that a service becomes more useful, the more people are making use of it. For example, telephones or social mediathe more people are using them, the more people you can contact. merger. 73 Google could use this merger to become the dominant intermediation platform for online advertising. DoubleClick's market power was however not problematic, since there were strong rivals. Prices were thus kept lower, and customers could easily switch to other suppliers (they frequently did so). The combined customer basis could enable the merged entity to improve its service by better targeting of advertisements. However, the specific type of targeting at issue was contractually prohibited. Moreover, multi-homing 74 was common. The Commission did not think it probable that AdSense would become dominant after the concentration. 75 Competitors' prices of ad serving would have to increase significantly in order for AdSense to gain a strong dominant position, which was very unlikely. So, vertical integration can provide efficiencies that compensate for its negative consequences. 76

Cisco/Tandberg, IMS Health/Cegedim
In Cisco/Tandberg, 77 both companies operated on the market for videoconferencing solutions. As a merger remedy, they agreed to divest their rights in the Telepresence Interoperability Protocol (TIP) to an independent bodythis protocol might become the industry standard. In this way, other manufacturers of videoconferencing solutions could also participate in updates to the TIP protocol. For the meantime, licencing commitments were made. Moreover, the companies published a source code library and promised to implement TIP on a number of products and successor products. 78 IMS Health/Cegedim 79 needed a slightly different approach. The activities of these companies were dependent upon each other. The software made by Cegedim could only operate properly in combination with brick structure data (a system of proprietary information which segments geographical areas into units for tracking pharmaceutical sales). IMS Health was dominant in the market for this. It had to sell off its primary market research business and promise to continue to conclude (on a royalty-free basis) "Third Party Access Agreements" to its brick structure. 80

Divestiture remedies
In other sectors, the analysis and explanation of competition and innovation effects has improved as well. The Commission states that a horizontal merger can hinder innovation, because some products which are being improved, or still under development, will not enter the market anymore. It is especially concerned with these problems in the pharmaceutical and medical equipment sectors, given the potentially large effects on the health and life of persons. The Medtronic/Covidien merger, for example, was conditionally approved. Medtronic was the leading firm on the market for drug-coated balloons to treat vascular diseases, and Covidien had almost completed developing an effective competing treatment (called Stellarex). According to the Commission, the transaction could have caused an important rival firm to leave the market. That is why it only cleared the merger after Medtronic agreed to sell off the Stellarex business. 81 The Novartis/GSK case was similar. The Commission identified that Novartis would have in all likelihood stopped developing two innovative drugs, because it would acquire drugs with the same working mechanism from GSKwhich would have resulted in duplicate clinical programmes. A divestiture remedy was put in place. 82 A similar case is Pfizer/Hospira, 83 a concentration which was also conditionally approved. It concerned a biosimilar drug for treating certain autoimmune diseases. Biosimilars have the same therapeutical mechanism as original patented biological pharmaceuticals, but they are not exact copies, unlike generics. They are much cheaper to produce, which can ease access to medicines for patients and/or reduce burdens on public healthcare systems. The Commission feared that Pfizer would have delayed or stopped the development of the biosimilar infliximab, in order to focus on Hospira's similar product. Or, Pfizer would have returned Hospira's product to its co-developer Celltrion. Again, a divestiture remedy was put in place. 84

Interoperability remedies
The Siemens/Drägerwerk merger 85 raised concerns on the markets for ventilators and anaesthesia equipment. This is not a pharmaceutical case, but it is interesting to take it into account, given the fact that it concerns medical equipment and health considerations could thus play a role. The merged entity would gain a very high market share in the market for anaesthesia delivery equipment; it could decide not to provide necessary information for competing monitors to interoperate with those created by Siemens. To resolve these concerns, Siemens agreed to divest Siemens Life Support Systems; the purchaser would then compete with the merged entity. It also committed to provide interoperability information on a royalty-free basis, to allow rival products to communicate with these and other relevant products manufactured by Siemens. 86 As described by Hoehn and Lewis, the companies in the cases described above might have given away interoperability information for free because this was standard industry practice. Before the merger of Siemens and Drägerwerk, and GE and Instrumentarium, healthcare technology companies already showed this practice, in order for their products to be widely used. 87 According to Lundqvist: Patent law and copyright law could be of importance in upholding competition by supplement and by diversity. Patent rights force firms to innovate and innovate around, while competition law should still prevent mergers and joint ventures that concentrate on the innovation process. That would be the formula for creating innovation while upholding the goals of competition 88 In his view, R&D-intensive firms should be able to exit long-term agreements with large firms when it can be proven that these large firms are trying to buy innovative firms out of the market, to prevent a radically innovative product from being introduced, or show similar behaviour. When there is a joint R&D venture or merger, it would then have to be evaluated whether specific research lines will not be stopped or severely hindered. 89 It is again important to be careful when applying these insights, as the large firms still need incentives to invest and innovate as well.
The Commission thus seems able to take into account differences between sectors when assessing the innovation effects of a concentration.

. Differentiation between types of markets
In recent years, the Commission seems to have improved its analysis of the innovation effects of mergers. It has recognised that innovation can lead to an increase in market power, and even result in new markets. However, the fact that a market is rapidly developing can also make it more difficult to assess whether a merger will create a dominant position, and if so, whether this is problematic. After all, a new product or a significant improvement of a product, by for example a new entrant, can mean that even a firm with a very high market share might not be dominant. Looking at these figures only, is thus likely to be misleading. It is important that the Commission becomes even more aware of the different types of situations that innovative markets can be in. Is the market already concentrated, how likely is it that there will be a new entrant, is there a product under development for which there is currently no substitute? 90 This is often difficult to assess, so the Commission would be well-advised to make even more use of recent economic insights. After all, if an incorrect economic analysis is applied, this is likely to lead to decisions that hamper innovation.

Intrusive remedies
Even quite recently, the Commission has received criticism on its merger control practice, 91 but these remarks do not seem to be entirely valid anymore. Its approach in the beginning, easily finding that competition is harmed, seems to have arisen out of necessity. After all, a decision has to be taken in each case and the risk of both Type I and Type II errors is very high. The Commission seems to mind Type II errors more. Most of the time, errors can only be discovered with the benefit of hindsight. This is largely still true, the Commission often attaches very strict conditions before clearing a concentration, but it now seems to better explain the reasons for this. Economic research is pressing further as well and will lead to even more knowledge on how innovative markets develop. So, we can be hopeful that the Commission's analysis will further improve.
We see that the Commission is taking an activist position, not afraid to require quite intrusive remedies. Criticism that it would go too far in doing this, is especially heard in the ICT sector. IPRs in this sector are often essential for other companies to use or build further upon, and one product can contain many patents. 92 So, it can lead to anti-competitive foreclosure when a (merged) firm with a large market share can freely decide not to licence this IP. Indeed, the Commission often assumes that a merged firm has to give competitors access to essential technology as well. Merger commitments in these sectors often entail that certain information has to be licenced. information can enable a new firm to enter a market, leading to more competition. The most important examples of this can be seen in network industries such as telecom. The fact that only few firms own a network, can easily lead to abuse of a dominant position. This is why the Commission often obliges companies to grant rivals access to these assets. Previously, there was much concern that this practice would diminish incentives to innovate for the companies owning the network. However, nowadays the Commission's approach has improved: companies buying access to a network now have more incentives to make optimal use of it as well. 93

Would information have been made public anyway?
It is true that obligations to give access to assets can diminish incentives to innovate. However, in various cases, information has or would have been made publicly available anyway, in order to ease market access for new technologies. 94 We can see this in the ICT sector when there are open source alternatives to a certain technology, but also in medical sectors, where information is shared freely in order to let products be used widely.

Differences between sectors
For the assessment of the effects of a merger, it makes a difference in which sector it takes place. This is the case because very different interests are at stake. Moreover, different types of innovation can be distinguished. In the ICT sector, access to IPRs is often needed for follow-on innovation, whereas in the pharmaceutical sector this is mostly not possible, since every product stands on its own. 95 Especially in the medical sectors, the Commission seems to be afraid that certain research lines will be stopped that could have led to a new or better product. In this sector, it thus seems to focus on maintaining diversity between firms and their activities. 96 Outside the medical sectors, it should, however, become even more aware of the possibilities and incentives for firms to innovate even without licencing obligations and similar measures. In this way, it will take even more proportionate measures.
Ireland and Telefónica Deutschland/E-Plus cases. The Commission now seems to better understand when a measure is proportionate or not. In the pharmaceutical cases, one often sees divestiture remedies that are very intrusivebut this is understandable, given the interests at stake. The Commission thus seems able to differentiate between various sectors and their specificities.
Mergers in innovative markets are often difficult to assess because the structure of a market can change quickly, and it is mostly not possible to tell exactly what the next innovative step will be. In recent years, the Commission has developed a more sound methodology for assessing the effects of mergers in these markets. The Commission tries to assess the extent of market power that the firm will have after the merger. Are the merging firms the most innovative ones, are they likely to still face sufficient competition? 97 6.2. The application of remedies in practice 6.2.1. Access to networksan improved approach In network sectors such as telecom, merging firms often commit to give access to the network to new companies, in order for them to be able to access the market. The Commission's approach here is more refined than in earlier years; it tries to confer responsibilities upon the acquirers of access as well. 98 However, it remains questionable whether the current approach gives each party the most incentives to invest and innovate (both merged entities owning a network and new entrants). The remedies applied are more proportionate than before, though. 6.2.2. Access to informationthe need for a balanced approach Also in other sectors, such as ICT, such a strict approach can be seen. Remedies often entail that access to information, even if protected by IPRs, has to be given in order to enable a rival to effectively compete. The Commission seems to do this very often when follow-on innovation is at stakethereby applying Lemley's advice to grant access to information in the ICT sector where followon innovation is frequent. 99 Access to information can thus lead to a new or improved product of a competitorand more effective competition. However, as has been extensively debated, the practice might also reduce incentives to invest and innovate for a merged firm. The Commission should thus look carefully 97 Laskowska (n 19). 98 Cook, Frisch and Novak (n 57). 99 Lemley (n 5). whether such measures are indeed necessary to achieve their goals. There are signs that the Commission is indeed starting to analyse these cases more carefully. For example, it also takes into account whether there are open source alternatives available. 100 It is then not necessary to restrict the dominant firm's contract freedom, since competitors can innovate without having access to the merged firm's technology.

Pharmaceutical marketsoften divestiture is necessary
In the pharmaceutical sector, divestiture of a product business is often necessary. Again, the Commission seems to fear Type II errors more than Type I errors. In this sector, it is mostly not possible to conduct follow-on innovation. Therefore, access to information is less frequently needed in this sector, the Commission follows this insight described by Lemley. 101 So, a new firm will have to come up with a completely new product. This type of competition is something the Commission aims for, because it allows for new medicines to be found the quickest. 102 Therefore, its approach is understandable.
6.3. Conclusion: a differentiated approach, but still room for improvement We can thus see that the Commission is able to differentiate between various sectors, but in each sector, it is trying to keep much discretion. This has remained so over the course of the years, but economic developments have pressed further, allowing the Commission to apply more proportionate measures. Nevertheless, there is still scope for improvement.

Disclosure statement
No potential conflict of interest was reported by the author.