Merger Remedies in the EU: An Overview Massimo Motta, European University Institute (Florence), Universitat Pompeu Fabra (Barcelona), and CEPR (London) Michele Polo, Univ. di Sassari and IGIER (Milano) Helder Vasconcelos, European University Institute (Florence). Very Preliminary! Paper Prepared for the Symposium “Guidelines for Merger Remedies – Prospects and Principles”, Ecole des Mines, Paris, January 17-18, 2002 1. MERGER REMEDIES IN THE EU
Under the Merger Regulation 4064/89, the European Commission (henceforth, EC)assesses proposed concentrations (the legal term which comprehends mergers andacquisitions, as well as full-function joint ventures) on the basis of whether or not they“would create or strengthen a dominant position as a result of which effectivecompetition would be significantly impeded in the common market or a substantial partof it”.1 In case it raises competition concerns, the EC might block the merger proposal. If however the parties modify the merger operation in a suitable way, i.e. they offer“commitments” (or “remedies”), the EC might clear the merger. A considerable andincreasing proportion of the mergers reviewed by the EC is actually approved afterremedies have been offered, as the table below shows. Table 1: Merger cases Years 1997 1998 1999 2000 Cases submitted Remedies (phase I) Remedies (phase II) Source: European Commission, Competition Policy Annual Reports
On 21st December 2000, the EC adopted a Notice on merger remedies (EC Notice fromnow on).2 This Notice sets both the substantial and the procedural requirements thatmerging parties must fulfil when proposing remedies to address competition concernsraised by the EC and, therefore, to win regulatory clearance in the European EconomicArea.3 It also summarises the main lines of intervention that were offered in the recentexperience,4 offering a coherent picture for the future implementation of the policy. Assuch, the Notice can be seen as the EC Guidelines on merger remedies.
1 The fact that the EC uses a dominance test rather than the “substantial lessening of competition” test asin the US is obviously of paramount importance when discussing merger remedies in specific cases. Thesame is true for the consideration of efficiency gains, that are usually disregarded in the EC mergerpolicy. However, we do not discuss them here. Note that the recent Green Paper on the Review of theMerger Regulation adopted by the EC on 11 December 2001 opens a discussion on both issues. 2 EC “Notice on remedies acceptable under Council Regulation No 4064/89 and under CommissionRegulation No 447/98”. Official Journal 2.3.2001, C 68/3. 3 The Notice stresses that it is the responsibility of the notifying parties to propose ways to eliminate thecompetition concerns raised by the EC. 4 For more detailed discussions on single cases, see the Competition Policy Newsletters published by theEC Competition DG and available at their web site. 5 It will be interesting to observe how the newly created Merger Task Force Unit on the enforcement ofremedies will develop.
It also gives an overview of the main types of remedies that have been accepted inmerger cases up to date (such as divestitures, termination of exclusive agreements andlicensing agreements to provide access to infrastructure and key technology). Inaddition, it confirms a clear preference of the EC for structural remedies rather thanbehavioural remedies which would absorb scarce resources since they require intensivemonitoring by the EC.
In this article, we briefly review the EC policy on merger remedies. In doing so, westress the possible problems and risks associated with the different types of remedies. Although overall we believe the EC has taken a very sensible approach (its recent policyon merger remedies clearly makes treasure of the FTC experience), we are still not toooptimistic that remedies will be able to restore competition in the majority of the cases. Although we acknowledge that the EC has taken a number of steps to guarantee thatremedies will be successful in restoring competition, we argue that the EC should makeeven more attention to the possibility that the divestiture favours collusion, and wesuggest that a more widespread use of the practice of finding an ‘upfront buyer’ mightalso help in this respect.
We shall also emphasise that remedies – including structural ones – modify the task ofthe EC Merger Task Force (MTF), making it closer to a regulator than a CompetitionAuthority (CA). This change is inherently linked to the nature of remedies, that by theirvery nature aim at changing the structure of the industry, and it occurs despite the MTFis sensibly trying to avoid becoming a regulator.5 We shall also argue that thesedifferent tasks objectively raise challenges for the EC, and that economic theory cansofar offer little help to it (or other CAs): more work is needed in this field. 2. TYPES OF REMEDIES AND THEIR POTENTIAL PROBLEMS Merger remedies as they have been used in competition policy in the European Union and the US seem to follow a relatively similar pattern. There are no guidelines for the US experience, but the paper by Parker and Balto (2000)6 nicely reviews the evolution in the US policy and the recent changes after the 1999 Divestiture Report. From the EC Notice and other publications, and from the US documents one can group the different merger remedies that can be given as an alternative to blocking the project in three categories. Structural remedies: they include divestiture of an entire ongoing business, orpartial divestiture (possibly a mix and match of assets and activities of thedifferent firms involved in the merger project). Behavioural remedies: they consist of engagements by the merging parties not toabuse of certain assets available to them. Very often, they are aimed atguaranteeing access to rivals to certain crucial inputs. Non-discriminationprovisions are an example of such remedies. Vertical firewalls (requirements ofnon-disclosure of confidential information within different vertical units of thesame firm) are another.
6 Parker R., Balto D. (2000), “The Evolving Approach to Merger Remedies”, Antitrust Report, alsoavailable at www.ftc.gov/speeches
Quasi-structural remedies: they are contractual arrangements or clauses such ascompulsory licensing or access to intellectual property.
As one can see, the classification is fuzzy, the third category being a hybrid between thefirst two.7 Divestments are clearly structural remedies, as they modify the allocation ofproperty rights and create new firms. Certain remedies, going from the promise not toabuse of a dominant position to the acceptance of a non-discrimination in the provisionof an input to rivals, are clearly behavioural, as they set constraints to some specificchoice of the merged firms while they maintain the full use of their property rights. Butwhen a firm is obliged (or agrees with the CA to win clearance of the merger) to licenseanother firm or to confer to it the use of a certain brand, this amounts to depriving it ofan asset (even if formally it keeps the property right of it), and it resembles more to astructural than to a behavioural measure.
Of course, not all different remedies are applicable to the same merger, that is, they arenot necessarily substitute to each other. Also, it is in principle possible to resolvecompetition concerns in a particular merger with a package of different remedies, thatis, they might be complementary measures in certain cases.
When choosing a remedy over others, a CA has in mind the main objective, which is tomake sure that the merger does not have anticompetitive effects. However, a remedythat in theory solves a certain problem might not be effective in practice. This is becausethere exist information asymmetries among the merger parties, third parties and the CA;because certain remedies might be difficult to implement; or because they involveparties that have different incentives than the CA. Furthermore, remedies differ in theengagement required to the CA. Behavioural remedies and contractual arrangementsentail continuous monitoring by the authorities, whereas structural remedies do not. Onthe other hand, structural remedies might be more risky, as they are not reversible: if thewrong buyer is chosen for a certain asset divested by the merging parties, for instancebecause the acquiring firm is not viable or not competitive enough, or because it endsup colluding with the merged firm, the competitive damage is there, and cannot beundone.
In what follows, we briefly review the use made of these different remedies in theEuropean merger policy and underline the possible problems associated with each typeof remedy. 2.1 DIVESTITURES As already mentioned the EC will try to obtain divestments of overlapping assets where possible. Indeed, the Notice (§13) says that “the most effective way to restore effective competition, apart from prohibition, is to create the conditions for the emergence of a new competitive entity or for the strengthening of existing competitors via divestiture”.
As the quotation indicates, divested assets can either create a new firm or be acquired by an existing competitor. In the first case, the EC Notice (§14) stresses that “(t)he divested activities must consist of a viable business that, if operated by a suitable purchaser, can
7 The EC Notice talks of structural and behavioural remedies, but does not define them. Parker and Balto(2000) distinguish between behavioural measures and contractual arrangements. 8 The emphasis on certain words is in the original text.
compete effectively with the merged entity on a lasting basis. Normally a viable business is an existing one that can operate on a stand-alone basis, which means independently of the merging parties as regards the supply of input materials or other forms of cooperation other than during a transitory period.”8
This implies that the acquirer will have the possibility to purchase “all the elements ofthe business that are necessary for the business to act as a viable competitor in themarket: tangible (such as R&D, production, distribution, sales and marketing activities)and intangible (such as intellectual property rights, goodwill) assets, personnel, supplyand sales agreements (with appropriate guarantees about the transferability of these),customer lists, third party service agreements, technical assistance (scope, duration,cost, quality), and so forth.” (EC Notice, §46)
The EC is aware that the viability of a firm is sometimes determined by the possessionof complementary assets, and that economies of scope or (hardware-software) networkeffects make it profitable to produce a certain good or service only if there is jointproduction of other goods or services.9 Accordingly, “(i)n order to assure a viablebusiness, it might be necessary to include in a divestiture those activities which arerelated to markets where the Commission did not raise competition concerns becausethis would be the only possible way to create an effective competitor in the affectedmarkets.” (EC Notice, §17)
An example of a case which illustrates both these points is the Unilever/Bestfoods10case. To remove the competition concerns raised by the EC, the parties undertook todivest a significant number of brands (such as Lesieur, Royoco and Oxo). First, toensure the viability of the divested businesses, the divestiture package also includedelements such as appropriate supply arrangements, manufacturing facilities, sales forcesand intellectual property rights associated with the individual businesses. Second, inorder to assure that the acquirer would be able to fully compete with the merging entity,the merging parties had to divest a full range of products, including products for whichthe EC had not raised competition concerns.
One other case which is related to this second remark is the Total Fina/Elf Aquitainecase.11 There, the parties had first proposed to sell several assets to eliminatecompetition concerns in the LPG (liquefied petroleum gases) industry. However, due tothe negative feedback obtained through the EC market test about the viability of theproposed remedy, the merging parties had to divest a full subsidiary, a remedy that wentclearly beyond the elimination of the identified overlap.
It is conceivable that the acquirer of divested assets is a firm already active in themarket. If this is the case, then it would not need all the assets, resources and contractslisted above, but the divestiture can be limited to particular production plants, or retailoutlets, or brands, or more generally assets that would be integrated in the business ofthe acquirer. However, the EC does not look favourably at this “mix-and-match”9 This principle is present in the US practice as well. The FTC often asks for divestiture of a greater set ofassets than those that participate in the market overlapping, if ancillary assets are required to replicateeconomies of scale or economies of scope without which competition could not be restored. See Parkerand Balto (2000). 10 Case No. Comp/M. 1990 – Unilever/Bestfoods; Article 6(2). Decision of 28/09/2000. 11 Case No. Comp/M. 1628 - Total Fina/Elf Aquitaine, Article 8. Decision of 9/02/2000.
approach: “a divestiture consisting of a combination of certain assets from both thepurchaser and the target may create additional risks as to the viability and efficiency ofthe resulting business. It will, therefore, be assessed with great care.” (EC Notice, §18)This is certainly a sensible approach, also in the light of the FTC divestiture study, thatreveals that the likelihood of successful entry was much higher when an entire ongoingbusiness was divested, whereas entry was significantly more problematic in the caseselected assets were selected.12
A case which is related to this approach is the one involving the world’s leadingprovider of Internet connectivity (MCI WorldCom) and one of its main competitors,Sprint (MCI WorldCom / Sprint case).13 The EC concluded that this merger would haveresulted in the creation of a dominant position in the market for top-level universalInternet connectivity. To try and remove the EC competition concerns, the partiesproposed to divest Sprint’s Internet business. However, the EC decided to prohibit themerger since its investigation showed that Sprint’s Internet business was completelyintertwined with the rest of Sprint’s telecom business. In other words, the divestedbusiness would have never constituted a strong and viable competitor of the mergedentity.
Of course, the viability of the business might also depend on the identity of thepurchaser. If the latter does not have any experience in the market, or does not haveappropriate know-how or financial standing, there might be a problem. In normalcircumstances, a Competition Authority is not a consulting firm and should not carewhether a firm is viable or not. However, when it comes to merger remedies, theviability of the acquirer is crucial because the degree of competition of the marketdepends on the competitiveness of the acquirer! Therefore, “(i)n order to ensure theeffectiveness of the commitment, the sale to a proposed purchaser is subject to priorapproval by the Commission. The purchaser is normally required to be a viable existingor potential competitor, independent of, and unconnected to the parties, possessing thefinancial resources, proven expertise and having the incentive to maintain and developthe divested business as an active competitive force in competition with the parties.”(EC Notice, §49).
For these reasons, the EC Notice states that in some cases the merger will not beauthorised unless “the parties undertake not to complete the notified operation beforehaving entered into a binding agreement with a purchaser for the divested business(known as ‘upfront buyer’), approved by the Commission”. (EC Notice, §20)
The first case in which the EC imposed this condition was the Bosch/Rexroth case.14The EC investigations revealed that the merged entity would have a dominant positionon the market for hydraulic piston pumps. Rexroth produces only axial piston pumpsand Bosch radial piston pumps. However, the EC’s review showed that there was a highdegree of substitutability between the two types of products. To address the ECconcerns regarding the potential creation of a dominant position, Bosch proposed to sellits radial piston pumps business to a competitor. None the less, the investigation showedthat to restore effective competition, it was not sufficient to sell. The EC had to makesure that the acquirer was a strong competitor. Otherwise, over time, Bosch would have
12 (Parker and Balto (2000, 6/19). 13 Case No. Comp/M. 1741 - MCI WorldCom / Sprint, Article 8(3). Decision of 28/06/2000. 14 Case No. Comp/M. 2060 – Bosch/Rexroth; Article 8(2). Decision of 4/12/2000.
been able to win back the market shares lost through the sale. This is so because Boschbenefits from strong costumer’s relations in the industrial hydraulics field and this couldbe used to persuade its former consumers to switch from radial to axial piston pumps. POSSIBLE PROBLEMS WITH DIVESTITURES
Structural remedies are, in general, the best corrective measures for potentiallyanticompetitive mergers, and the Commission is right in emphasising it, as well as inpreferring divestment of entire businesses to a mix-and-match approach. Structuralremedies, contrary to the behavioral or quasi-structural measures we shall analysebelow, have also the additional advantage that they do not occupy further the scarceresources of a CA after they have been implemented. Once the buyer has been identifiedand the transaction relative to the divested assets finalised, the EC will not have tomonitor further the deal (unless of course suspected infringements of articles 81 or 82arise).
None the less, this measure is probably trickier than one would think at first sight. Parker and Balto (2000) and FTC (1999) unveal that structural remedies can go wrongin a number of respects, due to a combination of informational asymmetries andincentives of the parties not in line with the objective of restoring competition.
First of all, it is clear that the merging parties have all the incentive to make sure that thepurchaser of the divested assets will not be a competitive firm. This might result inseveral problems. For instance, in the period the assets are for sale and it still managesthem, the seller might have an incentive to decrease their value, by transferring valuablepersonnel, disposing of certain brands, patents and activities, or not maintainingproperly the production plants or the shop premises.15 The divesting firm has also littleincentive to find a proper buyer (not so say to sell at all), and it would probably use verydifferent criteria than the CA to select the buyer. The EC is aware of these problems,and to this end the EC Notice establishes the figures of the ‘hold-separate trustee’ and ofthe ‘divestiture trustee’, that replace the Commission in ensuring that the seller does notengage in activities that could reduce the value of the assets or hinder the sales.16
Second, as already mentioned above, the FTC ex-post study of merger remedies revealsthat the mix-and-match approach is not very successful in fostering entry. One of the
15 Parker and Balto mention a case where the seller purposedly acted so as to decrease significantly thevalue of the assets to be divested. Although it was later sued and fined, they argue that this strategy wasmore profitable than having a dangerous competitor in the industry. 16 The trustees, as well as their mandate, has to be approved by the Commission. See EC Notice, §50-58,for details. 17 The difficult task of assuring viability of the new firm is usually run by assigning to the merging firmsthe burden of action: if competitive concerns are raised by the enforcer, the merging firms have to presenta remedy plan that will be analysed, discussed, corrected and possibly approved by the authority. Although this sequencing might be efficient, it is evident that the asymmetry in information between theproposing firms and the Authority is not solved by leaving the firms the first strike.
reasons why this occurs lies in the significant informational asymmetries between theseller and the buyer, and the problem also concerns sales of ongoing businesses.17 Thestudy reveals that when the latter is not already operative in the industry, it often doesnot know what are the crucial assets to be an effective competitor in the industry, and itmight end up with a package of assets that falls short of what is needed to be successful. The problem is made more serious by the fact that the seller has all the incentive todesign a package that does not include the right (from the point of view of thecompetitor) assets, and that a competition authority is not an industry regulator and hasthus limited expertise in any given sector.18
Third, the study underlines that whenever some relationships were needed between theseller and the buyer of the divested assets (for instance, if the buyer needs supply ofcertain inputs or technical assistance) the remedy did not manage to restore competition. In the FTC study, in thirteen out of the nineteen cases reviewed where there existedsuch a relationship, either the buyer did not manage to operate effectively, or there wascollusion between the two firms. (Parker and Balto (2000, 6/19)). The same difficultiesarise when technology transfers are an integral part of the divestiture: the combinationof the informational disadvantage of the buyer, who does not know the technology, andthe seller’s lack of incentives to provide the buyer with assistance and know-how, implythat technology transfers often do not achieve the desired results.
To the list of problems with divestitures emphasised in Parker and Balto (2000), we addanother, namely the possibility that the remedy might exacerbate the risk that collusionwill occur in the industry.
Fourth, it is obvious that the merging parties have all the incentives to select a buyerthat does not jeopardise its market position, but – perhaps less obvious – it is far fromclear that the an ‘aggressive’ buyer will be the one who will secure the divested assets. Suppose that there are two potential buyers, identical in other respects but who differ intheir market attitude. If it secures the assets, one expects that it will use a soft pricingpolicy, share the markets, or (tacitly or overtly) collude with the seller (perhaps this is afirm that – like the seller – is present in other markets: multimarket contacts helpcollusion, see below). The other instead is a firm that is planning an aggressive pricestrategy (for instance because it knows that the incentive constraints for collusion withthe merging firms would not be satisfied). It is likely that the expected profit of the
18 The EC tries to gather as much information as possible from competitors, buyers and consumers thatare regularly consulted about the likely effectiveness of the remedy (the so-called ‘market testing’ of theremedies). This might somehow alleviate its informational disadvantages, but we doubt that all the abovesources have the incentives to truthfully disclose their information to the EC. As for buyers-consumers,they are often not organised and it might accordingly be difficult to perceive the impact on them of acertain operation. 19 True enough, the Commission has to approve the final sale of the assets, but it is probably very difficultfor it to veto a buyer who fulfills the basic requirements set out in the commitment. If the identity of thebuyer was known upfront, instead, it would be easier for the Commission to assess the likelihood it wouldbe a serious competitor in the industry. 20 See Compte, O., Jenny, F. and Rey, P. (2000), “Capacity Constraints, Mergers and Collusion”,European Economic Review (forthcoming), Kuhn, K.-U. and Motta, M. (1999), “The Economics of JointDominance”, mimeo, University of Michigan, and Vasconcelos, H. (2001), “Tacit Collusion, CostAsymmetries and Mergers”, mimeo, European University Institute. For empirical evidence, see Barla, P. (2000), “Firm Size Inequality and Market Power”, International Journal of Industrial Organization,18(5), 693-722. 21 Case No. Comp/M. 190 – Nestlè/Perrier; Article 8(2) (b). Decision of 22/07/1992.
former is higher than the latter, and it will accordingly be willing to pay more to obtainthe assets. An auction will therefore not guarantee the best possible outcome fromwelfare’s point of view. Again, the identity of the buyer is therefore crucial, not only forthe viability of the business, but also to make sure that the purchaser will be an effectivecompetitor. In order to evaluate these aspects, it seems to us that the resorting to anupfront buyer should be systematic: the CA should lead a full assessment on whetherthe buyer is more or less likely to engage in effective competition, whereas a trustee isnot in the position to decide on such aspects.19
Fifth, the use of structural remedies, especially when the divested assets are used tostrengthen an existing competitor, might increase the risk of collusion in the industrydue to two problems: symmetry and multimarket contacts, two features that facilitatecollusion. To understand better this point, recall that to ensure the viability of thebusiness to be formed, a CA would give preference to an existing competitor or to apotential entrant, the latter probably consisting of a firm active in a related productmarket or another geographic market.
Consider first the case where the buyer is a firm already active in the market. Bypurchasing the assets divested by the merging parties, the risk of single-firm dominancedecreases, as a competitor is made more powerful. However, to the extent thatcapacities, market shares and other assets become more symmetrically distributed, therisk of a collusive outcome (the so-called joint or collective dominance) increases. Indeed, that symmetry helps collusion is not unknown to competition authorities andcourts, and it has been stressed in a series of recent papers.20
One well-known case of a merger involving asset transfers amongst rivals is theNestlè/Perrier case,21 in the French mineral water industry. The EC authorised (subjectto a set of commitments) the purchase of Perrier by Nestlè and the contemporaneoustransfer of ownership of one of the major Perrier brands (Volvic) to the main rival ofNestlè, BSN. Surprisingly, the remedies accepted by the EC to clear the concentrationdid not involve the Volvic parallel sell-off deal, even though it helped Nestlè and Perrierto restore the symmetry in the industry which would have been lost had Volvic not beentransferred to BSN. The problem is that the market structure induced by the merger wasmore symmetric, which, as stressed above, tends to enhance the likelihood of collusionbetween the remaining firms in the market. In a detailed analysis, Compte, Jenny andRey (2000) argue convincingly that symmetry was indeed an important issue in thiscase and conclude that “the proposed takeover of Perrier by Nestlè with the resale ofVolvic to Nestlè was the worst possible solution from the point of view ofcompetition.”22
Consider next the case where the buyer is a firm active in a neighbouring productmarket or in the same product market but in another geographic area. Again, such a firmwill probably be a viable market participant if given the appropriate set of assets. Relative to a new entrant, it should have more expertise and suffers less frominformational disadvantages. However, it is possible that entry into this particularmarket will make the buyer and the seller operate in the same markets. If this is so, thereexists the danger that a collusive outcome will arise. Indeed, economic theory has
22 Compte, Jenny and Rey (2000, p. 28).
showed that multi-market contacts facilitate collusion,23 and there exists some empiricalevidence that this has been the case in some markets.24 This is perhaps a more seriousproblem, to the extent that the EC shows to be aware that symmetry helps collusion inits joint dominance decisions, whereas multi-market considerations appear rarely, if atall, in the EC decisions.
In our view, a case in which a collusive outcome might arise after the merger due tomultimarket contacts is the recent EDF/EnBW case.25 The case concerns the acquisitionby Electricité de France (EDF) of a stake of 34 percent in EnBW, therefore taking jointcontrol with OEW in Germany’s fourth largest electricity firm. Before the merger, EDFenjoyed a dominant position for the supply of eligible customers (i.e., large customers)in France. Two different factors justify why the merger would strengthen the dominantposition of EDF in the market for eligible consumers in France. First, EnBW, due to itslocation, is one of the most likely potential entrants in the French market for eligiblecustomers. Its supply area is in the Southwest of Germany, therefore having a longcommon border with France. Second, by acquiring EnBW, EDF raises its potential forretaliation in Germany, thus becoming less exposed to competition in France.26 To solvecompetition concerns raised by the EC, EDF undertook to make available tocompetitors 6,000 MW of generation capacity located in France.27 Access to thiscapacity will be granted via auctions prepared and operated by EDF under thesupervision of a trustee and will enable foreign suppliers to have access to a large shareof the French market. Notice, however, that if this capacity was bought by a strongGerman competitor there would be the risk of multimarket contacts that might favourcollusive outcomes.28
Therefore, the solutions that seem more easily implementable to solve the problem ofthe viability of the firm created or augmented by the divested assets are often likely tobe conducive to more collusion in the sector. Further, note that the new entity receivesassets, including human capital, that previously were in one of the merging firms. Theinformal linkages with the old firm are therefore very strong, something that mightallow to implement subtle schemes of tacit collusion quite easily.
Moreover, finding the buyer among the existing competitors can give the direction ofthis ‘new’ entity to one of those “good old boys” that have been in the market for a long
23 See Bernheim, B. and Whinston, M. (1990), “Multimarket Contact and Collusive Behavior”, RandJournal of Economics, 21(1), 1-26. 24 See for instance Parker, P. and Roller, L.-H. (1997), “Collusive Conduct in Duopolies: MultimarketContact and Cross-Ownership in the Mobile Telephone Industry”, Rand Journal of Economics, 28(2),304-2225 Case No. Comp/M. 1853 – EDF/EnBW; Article 8(2). Decision of 07/02/2001. 26 EDF could use its presence in the German market via EnBW to deter other actual German competitorsfrom pursuing aggressive competition for the supply of eligible customers in France. 27 This capacity amounts to around 30% of the market for eligible customers in France. 28 As underlined by the EC itself, by accessing generation capacity in France, German suppliers will “beable to gain a foothold in France and thus become sufficiently strong in France to cope with EDF’spotential for retaliation resulting from its presence in Germany” (EC decision, § 108).
time. Although this is not equivalent to reinforcing the attitude to collude, thedestabilizing role of mavericks is rarely found among the existing long run competitors.
All this points to a tension between two problems. On the one hand, CAs have toguarantee the reinforcement or the creation of a viable firm to avoid problems ofunilateral effects (single firm dominance by the merging firm). On the other hand, theyalso have to avoid pro-collusive effects after the merger (joint dominance). We arguethat the implementable rules to solve unilateral effects emphasise the problem of pro-collusive effects. The EC is probably aware of this danger, when it recommends(Notice, §24-25), among the ancillary clauses of a remedy, divestiture of shareholdingin joint ventures and minority cross-ownership and the removal of interlockingdirectories.29 But unfortunately cutting these structural linkages among competitors isonly part of the story: divestiture might create a fertile environment for collusion, forthe reasons we have just explained.
An example of a case in which the EC cleared a merger after the merging partiescomplied with the commitment of divesting their shareholdings in a Joint Venture wasthe Kali&Salz MdK/Treuhand case.30 Incidentally, this specific case gave rise to aninteresting debate on what constitutes structural links and their significance for findingjoint dominance. The EC argued that the proposed concentration would create asituation of joint dominance on the part of the merged entity and the French (state-owned) producer SCPA. The EC decision was based on three criteria: the degree ofpost-merger concentration; the structural factors regarding the nature of the market andcharacteristics of the product; and the existence of “structural links” between the twoleading firms in the industry.31 As a result, the EC required the merged entity toeliminate its structural links with SCPA to clear the proposed concentration. In responseto appeals against the EC decision, the European Court of Justice (ECJ) found,however, that the EC had not proved, using a detailed and prospective economicanalysis, that an oligopolistic dominant position would be created or strengthened by thelinks and, consequently, annulled the EC decision.32
To conclude, we think that the evaluation of merger remedies should follow the sametwofold approach used in merger analysis, that is the evaluation of unilateral effects (orsingle firm dominance: does the merger reduces the degree of competition in a non-cooperative equilibrium of the market) and of pro-collusive effects (or joint dominance:does the merger facilitates the condition for a (tacit or overt) collusive outcome toarise?). Remedies should be accepted, and the merger proposal cleared, only if bothtests are satisfied.
29 As will be expalined later on in this paper, in the Vivendi/Canal+/Seagram case the commitmentspackage included the elimination of the notifying party shareholding on the British pay-TV companyBSkyB. This example confirms the fact that the EC has insited on eliminating minority shareholdings orlinks amongst competitors which could prevent them from effectively competing in certain markets. 30 Case No. Comp/M. 308 – Kali&Salz MdK/Treuhand; Article 6(1) (b). Decision of 09/07/1998. 31 These links were the following: (1) the control of a Joint-Venture in Canada (Potacna), in whichKali&Salz and SCPA each had 50% of the shares; (2) cooperation in the export cartel Kali-Export GmbH,wich coordinated its members’ sales of potash in non-member countries and in which Kali&Salz MdKand EMC/SCPA and the Spanish potash producer Coposa each had a 25% interest; and (3) longestablished links on the basis of which SCPA distributed almost all of Kali&Salz’s supplies in France. 32 France and Others vs. Commission (Joined Cases C-68/94 and C-30/95):  E.C.R. I-1375  4C.M.L.R. 829-953.
It is also our opinion that the Commission should use its bargaining power - consistingof the possibility to reject the merger proposal - to order to cease practices that mightfoster collusion. For instance, if information exchange agreements or other businesspractices (basing point pricing, best price clauses, or other) that facilitate collusionwithout proved efficiency reasons are in use in the industry, the EC should require themerging firm to discontinue its participation in them.
Summing up, whereas it is clear that structural remedies, if available, are the easiestsolution to competitive concerns created by a proposed merger, there exist severalreasons why such measures might have more difficulties in restoring competition thanone would think at first sight. We have argued that, despite the EC shows awarenesswith some of these difficulties and it appears to have taken appropriate safeguards toface them, these measures might not be enough. In particular, information disadvantagesand lack of incentives on the seller’s side to collaborate might result in widespreaddifficulties for new firms to successfully enter the industry.
Further, successful entry by the acquirer of the divested assets is not synonymous ofrestored competition: first, both the buyer and the seller of the assets have all theincentives not to fiercely compete to each other; second, the new configuration of theindustry assets after divestiture might structurally favour a collusive outcome because ofmore symmetric distribution of the assets or the creation of multimarket contacts. Therefore, the EC should take extra care not only that the assets go into the hand of aviable firm – as it is rightly emphasised in the Notice – but also that the conditions for acollusive outcome after divestiture are eliminated or alleviated – an aspect that thecurrent EC practice and the Notice does not in our opinion stress enough.
We also argue that a proper assessment of the likelihood that divestiture restorescompetition can be done only if the identity of the buyer is known to the authority. Tothis purpose, the requirement of an upfront buyer should be systematic rather thanoccasional.33 Otherwise, we are pessimistic that the remedies would be an effectivemeasure to restore competition. 2.2 NON-STRUCTURAL REMEDIES
Although the EC explicitly declares that divestiture is its preferred remedy, there aresituations where divestiture is not feasible, for instance because a buyer for the divestedassets cannot be found (this was the case for instance in Boeing/McDonnell Douglas),cannot solve the problem (the Notice mentions the existence of exclusive agreements,network effects and the combination of key patents), or would entail inefficiencies (forinstance when in a high-technology market where R&D is carried out on a number ofprojects that are related but would involve separate markets: in this case a divestituremight disrupt R&D efforts and licensing might be preferred).34 It is also possible thatdivestiture must be complemented by additional measures to ensure competition will berestored. In these circumstances, behavioural or quasi-structural remedies might beused.
33 According to Parker and Balto (2000, 7/19), upfront buyers are currently used by the FTC in the US inover 60 per cent of the cases where the remedy is non-behavioural, whereas we know only of two caseswhere the MTF used it in the EU. 34 See Parker and Balto (2000, 8/19) who mention the Ciba-Sandoz case to this purpose.
Behavioural remedies consist mainly of commitments aimed at guaranteeing thatcompetitors enjoy level playing field in the purchase or use of some key assets, inputsor technologies that are owned by the merging parties. Therefore, this situation mainlyarises when the merged entity is vertically integrated, and might foreclose eitherdownstream or upstream rivals.35
Typical remedies might then be purely behavioural, as when the parties ‘commit’ togive access to rivals and/or accept non-discrimination provisions, that is they agree notto make offers to competitors that are less attractive in quality and price than thosemade to the own subsidiary. The EC has showed some scepticism towards such type of‘essentially behavioural’ remedies in some cases (e.g., Vivendi/Canal+/Seagram, whereit refused a commitment by the parties not to discriminate against rivals), whereas it hasaccepted them in others (e.g., Vodafone Airtouch/Mannesmann).
They might also be of a contractual type, and therefore ‘quasi-structural’. For instance,the merging parties might be obliged to license a given technology to a rival. Or, in casethe merging parties’ key assets are not owned but were secured by exclusive long-runcontracts, the remedy might involve giving up or shortening part or the totality of suchcontracts.
Similar provisions might be taken in horizontal cases: if following the merger most ofthe buyers-distributors were linked to the merged entity by exclusive contracts, aremedy might consist of asking to shorten the length of a proportion of such contracts,or simply to cancel them.36
Another category of behavioural remedies might consist of the so-called ‘verticalfirewalls’. When the merger creates a vertically integrated firm, say one where theupstream unit supplies not only the downstream unit but also rivals, it is possible thatcompetitively sensitive information about downstream rivals be passed from theupstream to the downstream unit of the merged entity, thereby distorting thecompetitive process.37 It might then be required by the CA that no such information iscirculated within the different units of the firm (non-disclosure provisions).38 We are notaware of any case where the EC has used this measure (and the Notice does not mentionit), but this has been repeatedly used in the US and belongs to the set of remedies theEC might resort to.39
35 We should emphasise that foreclosure is not as likely to happen as one would think when reading someEC decisions. Once said so, it might be appropriate that a CA wants to avoid the risk that a relativelyunlikely event such a foreclosure of assets might occur, to the detriment of welfare. 36 We are not aware of any remedy of this type sofar. 37 The sensitive information might be of different type, depending on the industry involved. It is naturalthat in a long-standing relationship between a supplier and a buyer these parties exchange businessinformation (of financial and commercial order, relative to the specification of some products orprocesses, or the units/volumes bought and sold, and so forth) that they wish rivals would not share. 38 Obviously, the same might happen when it is the downstream unit that distributes products not only ofthe upstream unit of the merged firm but also of rivals. 39 U.S. antitrust authorities approved several vertical mergers subject to the imposition of non-disclosureand/or nondiscrimination requirements upon the post-merger vetically integrated entity. For a verydetailed analysis of U.S. merger cases in which these types of commitments have been used see Klass, M. W. and Salinger, M. A. (1995), “Do New Theories of Vertical Foreclosure Provide Sound Guidance forConsent Agreements in Vertical Merger Cases?”, The Antitrust Bulletin, Vol. XL, No.3, 667-698., andWillcox, T. C. (1995), “Behavioral Remedies in a Post-Chicago World: It's Time to Revise the VerticalMerger Guidelines”, The Antitrust Bulletin, Vol. XL, No.1, 227-256. PROBLEMS WITH NON-STRUCTURAL REMEDIES
Most of these remedies by their nature require some type of ongoing regulation ormonitoring, and they are therefore likely to engage the resources of a CA long after themerger has been cleared and carried out. Some of these measures are relatively easy toevade unless there is a careful monitoring and the regulator knows the industry verywell, which is not likely to be the case for a CA.
When the CA identifies the risk of foreclosure, for instance, short of divestiture (thatmight be unfeasible, as the very reason behind the merger might precisely be tointegrate vertically related or complementary activities) behavioural remedies aredifficult to administer and not likely to be successful unless there is heavy monitoring. Foreclosure or discriminated access might take different forms, from obvious (say,bluntly refusing to supply an input) to more subtle ones (increasing prices, reducingquality, blaming insufficient capacity to justify missed shipments, delaying supplies,reduce accessory services and so on). Therefore, a remedy that calls for an obligation tosupply is tantamount to an empty promise, but even a seemingly more sensibleobligation to non-discrimination might not be easily enforceable. As just mentioned,discrimination might often occur at different levels and with different features, and it isprobably rare the case where one can just look at transaction prices to determinewhether discrimination has occurred or not.
Furthermore, even when prices were the only relevant variable, it cannot be excludedthat transfer prices, allocation of common costs, or other compensatory measures mightoccur between vertically units of the same firm, so as to hide a different treatmentbetween a subsidiary and a rival. Of course, industry regulators – whose main job isoften to guarantee that access is granted to all competitors – are aware of theseproblems, but these are often difficult problems to cope with for a regulator, and theywill a fortiori be for a CA whose expertise lies elsewhere and whose knowledge of theindustry is not like that of a regulator.
Vertical firewalls are another instance of behavioural remedy, aimed this time atpreventing the disclosure of confidential information between the two levels of avertically integrated firm. This might even be a reasonable remedy to solve thecompetitive problems involved,40 but there are some doubts on specific aspects of theirimplementation. In particular, it is not clear to us how one can guarantee that no suchcommunication will take place between different units of the same firm (what aboutemployees meeting in the cafeteria?), and – if it does – that it will not be misused.
At the very least, such behavioural remedies need continuous monitoring, either by theCA itself, or by an industry regulator. The former option being excluded for the verynature of CAs and for their lack of resources, the latter option remains available: non-discriminatory access or firewalls should be implemented only when the firms involved
40 However, notice that firewalls might also have efficiency-destroying effects, as pointed out by Klassand Salinger (1995): “since sensitive information does not necessarily come in neatly labeled packages,firewalls could prevent the sort of productive information traditionally thought to be a benefit of verticalmergers” (p. 690). For a formalised paper that offers a critical view of firewalls, see also Milliou (2002),“Vertical Integration and R&D Spillovers: Is There a Need for Firewalls?”, mimeo, European UniversityInstitute.
are subject to the scrutiny of a regulator. It follows that, to guarantee success of suchmeasures, the industry regulator should be involved in the discussions leading toremedies as soon as possible by the CA. SOME RECENT EXAMPLES OF NON-STRUCTURAL REMEDIES
As already mentioned, there are situations in which a divestiture either is not a feasibleremedy or has to be complemented by additional measures. In such cases, the EC mayaccept non-structural remedies on a case-by case basis. An investigation of a sample ofrecent merger cases reveals that the EC, when accepting non-structural commitments,has been specially concerned with two main problems: avoiding foreclosure effects onthe market and facilitating market entry (the two issues are obviously often linkedtogether). Avoiding foreclosure effects on the market
Cases of vertical integration constitute the natural example where foreclosure practicescan constitute a problem, especially when one of the merging parties enjoys significantmarket power in an upstream or downstream market. The Vivendi/Canal+/Seagramcase is one of such cases.41 This case regards the acquisition by French communicationsand media company Vivendi and its subsidiary Canal+ of the Canadian group Seagram,whose activities range form drinks to films to music (in the films and in the musicsegments Seagram is active through its subsidiary Universal, one of the six majorHollywood studios). One of the markets where the EC raised competition concerns wasthe European TV market, where Canal+ is the largest pay-TV operator. Canal+ is thefirst acquirer of premium films for pay-TV signed with the US major studios and inparticular with Universal in France, Spain, Italy, Belgium and The Netherlands and theNordic countries.42 After such a merger, the parties would gain access to all stages inthe vertical chain. Therefore, chances were that upstream content providers could denyor limit the access to premium films to some downstream active users or potentialentrants. In addition, two different factors justify why the merger would strengthenCanal+ position as a pay-TV supplier. First, Canal+ would obviously secure the renewalof its contracts (output-deals) with Universal on a permanent basis. Second, theproposed acquisition would enhance its ability to renew existing contracts with themajors and also enter in new output-deals with the other US studios.43
In order to eliminate the doubts raised by the EC, the parties, in a first round, proposedthe following remedy. In France and Spain, Canal+ had an ongoing output deal withUniversal Studios for broadcasting of Universal films in the first pay-TV window. Upon
41 Case No. Comp/M. 2050 - Vivendi/Canal+/Seagram; Article 6(2) Decision of 13/10/2000. 42 Premium films constitute a key quality input to increase the attractiveness of pay-TV and the level ofsubscriptions. They are acquired through the so-called output deals. These output deals include “first-window” agreements signed on an exclusive basis, where first-window is the first period of premiumfilms available on pay-TV. A pay-TV operator with no access to first window films can only offer “old”premium films, the so-called “second-window” films. However, since “second-window” films areregarded as “second quality” by the subscribers, the pay-TV operator may have to reduce its subscriptionprice to differentiate itself accordingly. 43 Due to the higher budgets required by films production, US studios resort more and more frequently toco-financing with other US studios or financial partners. Following the merger, Canal+/Universal wouldcertainly be in a stronger and attractive potential partner in terms of co-financing of films (subject,obviously, to exclusive output deals for the involved films).
the expiry of these agreements, the parties committed that Universal should give allinterested pay-TV operators in France and Spain the opportunity to bid for a new outputdeal for Universal films for the first pay-TV window, and should conduct thenegotiations in a “fair and non-discriminatory manner”, attributing the contract to thebest bidder. The EC, however, considered these undertakings unsatisfactory based ontwo arguments. First, they were “essentially behavioural”, clearly stressing again itspreference for structural remedies. Second, the mechanism proposed by the parties tosingle out the winner of an output deal was the one usually used in a normal negotiationprocedure, where no safeguard was provided with regards to the independent nature ofthe bidding process, and no real evaluation of the best bid was proposed. The parties, ina second round, submitted another set of undertakings, committing not to grant Canal+“first-window'' rights covering more than 50% of Universal production and co-production. This commitment covers the territories where Canal+ is active (not onlyFrance and Spain), for a period of 5 years after the expiration of the current output deals(the EC considered 5 years the necessary period rivals need to adapt to the new marketstructure).44 45
Another type of non-structural remedy which aims at avoiding foreclosure effects on themarket is the termination of existing exclusive agreements. This type of commitmentwas used both in the Astra/Zeneca case 46 and in the Lufthansa/SAS case.47 The mergerbetween Astra and Zeneca is part of the ongoing consolidation process which is takingplace in the pharmaceutical industry. The EC investigations regarding the market forplain betablockers showed that Zeneca is Astra's main competitor in Sweden andNorway. In particular, Zeneca has been very actively promoting its plain betablockers(Tenormin) as a competitive alternative to Astra's largest selling betablocker in thosecountries. In order to address this specific competition concern, the parties committedthemselves to “grant a viable independent third party exclusive distribution rights forTenormin in Sweden and Norway for a period of at least 10 years (italics added).”
The Lufthansa/SAS case, on the other hand, is a cooperation agreement to create a long-term alliance between the two airlines, establishing an operationally and commerciallyintegrated air transport system. The agreement provides a setting up of a joint venture toact on behalf of the two airlines as their exclusive vehicle for offering integrated airtransport services between Scandinavia and Germany.48 The EC decided to authorizethe cooperation agreement for a period of 10 years subject to certain conditions. One
44 As already mentioned, the notifying party also undertook to divest its stake on the British pay-TVcompany BSkyB, which has links with Fox, another major US film studio. 45 The AOL/Time Warner case (Comp/M. 1845, Decision of 11/10/2000) is another interesting example ofvertical integration. The merger would create the first Internet vertically integrated content providerdistributing Time Warner’s branded content (music, news, films,.) through AOL’s Internet distributionnetwork. Because of the structural links and some existing contracts with Bertelsmann, the merged entitywould have had access to Bertelsmann content (especially to its large music library). As a result, themerged entity would have controlled the leading source of music publishing rights in Europe and wouldbecome the “gate-keeper” of the emerging market for Internet music delivery on-line. In order to ease thecompetition concerns raised by the EC, the parties offered a package of commitments whose ultimategoal was to break the links between AOL and Bertelsmann. 46 Case No. Comp/M. 1403 - Astra/Zeneca; Article 6(1)(b). Decision of 26/02/1999. 47 IV/35.545 LH/SAS. Decision of 16/01/1996. This decision was not taken under the Merger Regulationbut under article 81. However, since the last revision of the Regulation full-function joint ventures arenow reviewed as mergers. 48 This integrated transport system involves a joint network planning, a joint pricing policy and theharmonization of product and service levels, without creating however a new common entity.
such condition was that the involved airlines should give up slots at saturated airports incase there were potential entrants. This commitment intends to diminish the risk offoreclosure by the incumbents.49
Facilitating other firms to enter the markets and compete with the incumbents is anotherconcern which has been influencing the EC decisions on which commitments to accept. In the Lufthansa/SAS case, for instance, the EC also imposed that the parties shouldconclude interlining agreements with new entrants and freeze on the number offrequencies operated to facilitate entry. An interesting feature of this market is that theroutes between Scandinavia and Germany are essentially used by businessmen who arelocked in through frequent-flyer programs. The EC was aware of this problem andimposed that Lufthansa and SAS should allow the entrants to participate in theirfrequent-flyer programmes, in case they do not have their own international frequent-flyer program.50
The EC has also considered appropriate to facilitate entry a commitment to give accessto a network/infrastructure, or to license a key technology. The VodafoneAirtouch/Mannesmann51 case and the Astra/Zeneca case constitute examples ofdecisions where commitments of this type were offered by parties to clear theconcentration.
The Vodafone Airtouch/Mannesmann merger gave rise to the creation of the first singleEurope-wide mobile network. The effect was to enable the merged entity to provide aseamless pan-European service, especially oriented to serve multinational enterpriseswith substantial amounts of European cross-border business. The Commissionrecognized that there exists a demand for these “advanced pan-European mobileservices'' which is distinct from national mobile telecommunications services. Inaddition, the EC acknowledged that since after the merger, the new entity would havesole control of mobile operators in eight Member States and joint control in three, itwould be in a unique position to build an integrated network which would enable aquick implementation of seamless pan-European services. Other operators, on the otherhand, would not be able, in the short to medium term, to replicate the merged entitynetwork footprint through mergers and/or agreements. This raised doubts as to thecompatibility of the merger with the common market. To address this issue, the mergingparties proposed a remedy which consisted in granting other mobile operators thepossibility to provide pan-European seamless services to their costumers by using theintegrated network of the merged entity, for a period of three years.52
The idea was that by granting access to its network on a non-discriminatory orfavourable terms, the merged entity would not be able to make third party offerings ofadvanced seamless services across Europe unattractive or simply not competitive. We49 As far as we have understood the decision does not specify who chooses the slots to be attributed, andthe details of how this measure was to be implemented and monitored. 50 Again, it is not clear to us how this measure has been practically implemented. If it has indeed been leftvague, we wonder to which extent such a measure would be effective. 51 Case No. Comp/M. 1795 - Vodafone Airtouch/Mannesmann; Article 6(1)(b). Decision of 12/04/2000. 52 The Commission also identified horizontal geographic overlaps regarding the national mobilenetworks (namely, in Belgium and in the U.K.). In order to remove these competition concerns, theparties undertook to de-merge Orange Plc and all its subsidiaries.
have doubts, however, that this remedy will be effective since, even though the mergedentity has committed to the provision of a roaming tariff and/or wholesale services on anon-discriminatory basis between operators of the merged entity's group and othermobile operators, it is entitled not to be obliged to provide such services in cases of“unavailability of adequate network capacity” and/or “technical unfeasibility”.53 Inparticular, the first exception (unavailability of adequate network capacity) might, bestrategically used by the merged entity in periods of very high demand (booms). Inaddition, this undertaking will certainly turn out to be extremely demanding in terms ofex-post monitoring by the EC.
In the Astra/Zeneca case described above, the EC also raised doubts with respect to themarket for local anaesthetics. Astra's Bupivacaine is the most widely used longer actinglocal anaesthetic and is already long off patent. In addition, although until 1998 Zenecawas not present in this market, in March 1998 it concluded an exclusive world-wide(except for Japan) agreement to license-in Chirocaine, a longer acting local anaesthetic. Since there are no other strong competitors in the market, the EC was concerned withthe fact that the exclusive license for Chirocaine constituted the only potential source ofcompetition in this market segment. Therefore, it would not be likely that the mergerparties would be willing to support and invest in the launching of a new product whichwould compete harshly with other products in the merged entity portfolio of existingproducts. To address this concern, Astra committed to reverse all the agreementsrelating to Chirocaine (surrender of license, trademark, etc.). Astra will also support athird party, during a transitional period, in the process of launching Chirocaine. Threeremarks are in order regarding this commitment. First, since the EC investigationshowed that this specific market has no other strong competitors, it is not clear to uswhether there exists a suitable third party with assets and competence to launchChirocaine and strongly compete Bupivacaine. Second, notice that the subset of Astra'semployees involved in the process of supporting the third party to launch Chirocaine,will obtain sensitive non-public information about the product they support to launch,which will be the main rival of the merged entity's local anaesthetic Bupivacaine. Lastly and perhaps more importantly, there is certainly lack of incentives for themerging partners to make the buyer of Chirocaine successful. Hence, given that acollaboration between them is necessary during a transitory launching period, problemsare likely to arise. Interestingly, this last point is related to the Abbot/ALZA case in thepharmaceutical markets discussed in Parker and Balto (2000, 14-14/19). There, partiesproposed to sell several assets to another pharmaceutical company and, because of therisks involved in a necessary ongoing relationship between the merged entity and thepotential third party, the FTC refused the commitment, leading to the parties’withdrawal of the merger proposal…
3. MERGER REMEDIES: A CHALLENGE FOR COMPETITION AUTHORITIES (AND FOR ECONOMIC THEORY)
We have seen that there are different types of remedies that a CA can resort to. Weargue in this section that, in each type or remedies, CAs face new questions andproblems that differ substantially from the approach followed and the tool kit needed inthe traditional antitrust enforcement activity. Moreover, we claim that in many
situations they cannot even rely on a strong and consolidated guidance from economictheory, that has hardly devoted effort to these issues.
3.1 Competition authorities or regulators?
It is difficult to distinguish in a clearcut way the differences between competition policyand regulation, and between competition authorities and industry regulators. There areseveral criteria that one can use to make such a distinction, but for each criterion thereare probably exceptions and qualifications to be made that make the line of demarcationbetween the two blurred. We believe that merger remedies contribute to make such lineeven fuzzier.
Rey54 (2000) classifies competition policy and regulation along the followingdimensions (among others). 1) Procedures and control rights: whereas CAs generallylimit themselves to checking the lawfulness of firms’ activities, industry regulators havemore extensive powers, as they can constrain firms’ conduct in several ways, forinstance by capping or fixing their prices, checking their investment decisions,restricting their product choices. They can also modify the structure of the industry byestablishing when new entry in the sector is allowed, and fixing the criteria that decidemarket entry. 2) Timing of oversight. Generally, CAs intervene ex-post whereasregulators act ex-ante. Also, CAs have usually more time available for investigations,whereas regulators have to come up with rapid decisions, as the firms’ normal business(such as pricing, investments, launch of new products) might need the preventiveauthorisation of the regulators. Regulators’ involvement with a particular case is long-run and continuous, whereas CAs’ interventions tend to be occasional. 3) Informationintensiveness. Regulators usually have an industry-specific expertise whereas CAs havenot. In part, this is also determined by the fact that the regulators’ relationship with anindustry is continuous and of a long-term nature.
The distinction between competition policy and regulation along the lines describedabove only captures part of the story: antitrust intervention is correctly describedaccording to the features suggested mainly when it acts as a law enforcement activity,that is in the areas of agreements and abuse of a dominant position: in these cases thelaw identifies certain conducts which are unlawful, and the authority verifies ex-post ifsuch behaviour has occurred.
Merger control by itself is already a mixed area. In fact, points 1) and 2) above arecertainly not satisfied in concentration cases: the authority intervenes ex-ante and withshort deadlines, and it has implicitly the power to decide on the structure of the industry. What still is different with respect to the role of a regulator is that the antitrust authoritydeals with concentration projects in any industry of the economy, and it is thereforelacking the specific knowledge of the industries investigated that a sectoral regulatoraccumulates over time.
The regulatory component of antitrust intervention is even more pronounced when welook at merger remedies. To the extent that non-structural remedies are followed, theCAs must monitor the behaviour of the firms long after the merger has been authorised
54 Patrick Rey (2000), “Towards a Theory of Competition Policy”, presented at the WorldMeeting of the Econometric Society. See page 44 and ff.
and carried out. It is still true that merger remedies do not change the specialisation ofthe CAs, but to the extent that behavioural remedies imply that a continuousrelationship is engaged, a better knowledge of the industry monitored will probably bebuilt over time.
When we consider structural remedies, other regulatory features come in. In atraditional merger case framework, in fact, the authority has the power to decidebetween two industry structures: the status quo and the market that would arise if themerger project is realized. With merger remedies, in a sense, we fill the gap betweenthese two extremes, and many intermediate structures can be implemented as a result ofthe bargaining process between the firms and the authority. In other words, the authorityis in the position to fine tune its intervention on industry structures much more than intraditional merger cases.
Moreover, we have stressed that the key point in the evaluation of structural remediesis to assess if the divested assets can create a viable competitor. Guaranteeing theviability of a firm is an objective specific to the merger remedy area, as in no other areaof competition policy the enforcer promotes the creation of a new firm (perhaps with theUS exception, rare indeed, of breaking up firms that have been found guilty ofmonopolisation). Moreover, the theoretical arguments and tools that are needed toassess the competitive viability of a remedy are in a large part different from those thatare used by competition authorities in general, and in the evaluation of the originalmerger project in particular.
Mergers always require a prospective exercise on the equilibria that will arise after thestructural modification of the market; for this reason, merger analysis is moretheoretically based and less dependent on facts findings than other areas of interventionas agreements or abuse of a dominant position. And the theoretical tools that are neededto perform such analysis are those of oligopoly theory, as well as, to the extent that theycan forecast possible impact upon prices, of econometrics. Whether the merged firm hasthe ability to make profits and survive in the market is never a relevant issue in theanalysis of the original project, for the proposing firms are assumed to have addressedand solved the problem.
Structural merger remedies, on the contrary, require to identify and evaluate which arethe competences, assets, know how, personnel and other common resources that mustbe packaged in the new entity to create a competitive enterprise. Hence, the tool kit formerger remedies is pretty different from that needed for merger analysis.
More importantly, we think that there is no piece of economic theory that offers a robustand consolidated background to such analysis (which are the features needed to create anew competitor), as instead oligopoly theory does for the evaluation of the originalmerger project. In a general sense, creating new enterprises is exactly what the market isrequired to do, and under this respect the formidable task of designing viable structuralremedies is something very far from the ‘light’ approach of competition policy.
While structural remedy design is an exercise quite different from the traditional mergeranalysis, we can find some examples in which regulatory policy at large has beenrecently challanged with relatively similar problems. One case is that of market designin public utilities reforms, notably the electricity industry, and the other is the design of
auction mechanisms in telecommunications, and specifically the UMTS licencesauctions in the European countries.
Since the pioneering reform of the early Nineties in UK, the electricity industry hasbeen a very interesting workshop for market design in most industrialized countries. The European Commission Directive 1996/92 has pushed all the member countries toset up within February 1999 the domestic liberalization plans according to some verygeneral common principles. While the national experiences are very different, we canfind some unifying themes across countries.
One of them is the possibility to create a competitive market in the generation segmentof the industry, by divesting assets and power plants of the incumbent monopolist. Thestarting point is the recognition that economies of scale and scope do not justify a veryconcentrated market, in particular in countries with a sufficiently high demand. Theissue has been addressed focussing on the number of operators and the kind ofgeneration technologies (plants) to be packaged into the new firms. The theoreticalexcercise behind this discussion was that of creating an industry structure, defined bothin terms of number of operators and their cost structures, competitively viable.
The theoretical tools implicit in the debate have been those of oligopoly theory, as thetraditional Cournotian link between concentration and equilibrium prices and the morerecent developments of supply function equilibria55. From this point of view, we mightpoint out that the market design exercise performed in the liberalization plans sharesmany features with the merger remedy problem we are discussing: the status quo ofquasi-monopolies has the same role of the original merger project, i.e. an industrystructure that must be evaluated from a competitive perspective; the divestiture planshas the same competition enhancing role that is required to merger remedies. In bothcases the policy maker has to identify which structural modifications of the industry canlead to a more competitive environment.
However, this excercise has been performed, in the case of liberalization plans, in amuch more abstract framework than in a merger case, since a very large set of relevantelements (including the rules of the bulk energy market, the identity of the competitors,etc.) were not usually defined at that stage. This explains why the analysis focussed onthe desirable structural properties of the reformed industry, but did not address thesubsequent step of assessing whether the divested assets were able to become viablecompetitive firms. As we claimed above, in a merger remedy analysis, instead, themarket environment is much more defined and the viability of the new competitorsbecomes a more relevant issue.
Evaluating if a potential competitor is viable has been recently an issue also in thedesign of auction rules for UMTS mobile telephone licences in many Europeancountries.56 The more relevant issues in that case have been those of preventing
55 See Green R. and Newbery R. (1992), Competition in the British Electricity Spot Market, Journal ofEconomic Theory, v. 100, pp. 929-53, and Klemperer P. and Meyer M. (1989), Supply FunctionEquilibria in Oligopoly Under Uncertainty, Econometrica, v.57, pp.1243-1277. 56 See Klemperer P. (2000), What Really Matters in Auction Design, mimeo, for a discussion of the mainissues involved and the comparison of the different auctions used in European countries. See alsoKlemperer, P. (2002), “How not to Run Auctions: The European 3G Telecom Auctions.”, European
collusion among participants during the auction and ensuring a sufficiently competitivestructure of the new markets. The two key goals were strictly linked, since auction rulesunable to attract many participants would make collusive bidding easier to sustain andwould force the Government to limit the number of licences (and future operators)issued, as for instance the Netherlands case shows.57
Once the auction rules have been set, deciding the number of licences to be issuedrequires an evaluation of the incumbent operators as well as of the new entrants, inorder to forecast how many bidders will participate. This is a particularly difficultexercise when, as in the UMTS case, it is applied to markets that are not yet beenopened. While existing mobile telephone operators have been considered as technicallyequipped to run the next generation of mobile services, assessing if new entrants withno previous experience were viable competitors turned out to be a much more difficulttask.
In order to avoid such a negative outcome, some countries that organized in a later stagetheir national auction, as Italy, adopted a rule that set initially n licences, to be reducedif less than n+1 participants were admitted to the auction. That seemed a simple andgood idea to maintain competition in the auction, although a possible adverse effect, incase of an insufficient number of participants,58 might be an excessively concentratedmarket. However, a key point in the implementation of the rule was the ability to assessif the firms, in particular those with no previous experience in the market, were reallycompetitive rivals and serious and reliable participants.
In the pre-auction phase candidates were examined and some conditions on financialstability (as bank warrantees) and industrial experience were set. Fixing a positive,although relatively small, entry cost allowed in principle to eliminate frivolousparticipants, but not to prevent the participation of (and side payments to) weak firmscolluding with the larger firms just to keep high the number of (formal participants and)licenses.59
But even in this relatively similar case the selection of viable competitors was notperformed in such an artificial way as in merger remedies, as the applicants were ingeneral existing firms or consortia of existing firms self proposing for the auction, andEconomic Review (forthcoming), and Van Damme, E. (2002), “The Europena UMTS-Auctions. andNext”, European Economic Review (forthcoming).
57 The UMTS auction in Netherland was organized just after the great UK success, following the samelines of an ascending auction. However, five licences were issued in a market with five incumbents. Moreover, it was permitted to some foreign important operators to partner with the domestic incumbents,rather than participate on their own. There was a single very weak entrant that quitted after few rounds,with revenues less than _ of those expected. 58 It must be reminded that the number of participant depends on the characteristics of the auction, e.g. ascending vs. sealed bid, , that leave more or less possibility of success to new entrants. See Klemperer P. (2000) What Really Matters in Auction Design, mimeo. 59 The final outcome of the Italian auction was open to many interpretations: five licenses were auctionedwith six participants, the three major mobile operators, two new groups including important internationaloperators and the fourth mobile telephone operator, still in a start-up phase, a joint venture of Italianinvestors and British Telecom. After few rounds this latter exited the auction, that ended up with revenuesmuch lower than expected. Whether the firm that quitted was a frivolous (or even a colluding) participant,or simply an operator whose shareholders were unable to agree on a coherent bidding strategy cannot beverified. But we can say that the formal conditions needed to participate were not sufficient to identifythe firms with a long lasting incentive to bid.
not entities created by the regulator picking up assets of existing corporations andlooking then for a buyer. 4. CONCLUSIONS
It has recently been observed that the Merger Task Force of the European Commission(EC) has been adopting a tougher stance in merger control, and the one year-old Noticeon merger remedies, as well as the recent decisions in this area, would be further proofof its stricter approach.63 We argue in this article that this (alleged) strictness might stillnot be enough to restore competition in industries where mergers take place.
If it is well known that behavioural remedies might be problematic, and the EC hasrightly expressed its preference for divestitures, we stress that structural remedies aremade difficult by information asymmetries between the merging partners on the onehand and both the competition authorities and the prospective buyers of the divestedassets on the other. Further, given the lack of incentives for the merging partners tomake the buyer of such assets successful, whenever a collaboration between them isnecessary problems are likely to arise.
Perhaps more importantly, we argue that several reasons exist to believe thatdivestitures might exacerbate pro-collusive effects. In other words, we believe that theEC has paid attention mostly to the issue of the viability of the new firm created withthe divested assets, whereas it has not attached enough importance to the possible pro-collusive effects of divestiture. We suggest therefore that the EC should follow forstructural remedies the same double test it uses to assess mergers in the first place: (1)Single firm dominance will likely not arise after divestiture (unilateral effects) (2) Jointdominance will likely not arise after divestiture (pro-collusive effects).
We point out that steps taken to guarantee that the first problem is solved - that is, thatthe new firm is viable – might often worsen the second problem. This is the case, forinstance when, in order to guarantee that the buyer has the ability to run the divestedassets, it is chosen among the firms already active in the same market (risk thatsymmetry increases, favouring collusive outcomes) or in a neighbouring market (risk ofmultimarket contacts, again favouring collusion).
We also argue that to increase the likelihood of success of structural remedies the ECshould make more systematic use of the upfront buyer requirement.
60 We should also add that viability is not a sufficient condition for restoration of competition: we haveunderlined above that the risk of collusion should also be avoided. 61 For an overview of the recent Third-Generation (UMTS) auctions in Europe, see Klemperer, P. (2002),“How not to Run Auctions: The European 3G Telecom Auctions.”, European Economic Review(forthcoming), and Van Damme, E. (2002), “The Europena UMTS-Auctions. and Next”, EuropeanEconomic Review (forthcoming). 62 In the US language, competition agencies have to assess the possible ‘coordinated effects’ of themerger. 63 See Ersboll, N. C. (2001), “Commitments under the Merger Regulation”, European Competition LawReview, Issue 9, 357-364; De Matteis, A. (2001), “The Commission Develops its Policy on Merger Cases:The New Notice on Remedies and its Recent Application”.
We also acknowledge that competition authorities do not have an easy job in dealingwith merger remedies. The task of devising, implementing and monitoring mergerremedies belongs more to the realm of regulation than that of competition policy. Accordingly, competition authorities have to act with tools and deal with problems theyare not used to. Admittedly, they often have to do that with little guidance fromeconomic theory, that has not devoted much attention to the study of issues such as thefeatures which make a firm competitively viable.
To understand whether our preoccupations are only theoretical or real, a more accurateanalysis of the recent EC practice in merger remedies is needed. After some years ofpolicy in this area, the time is now ripe for the EC to carry out an ex-post study on thelines of the recent FTC report.
As drug-resistant HIV strains become What would need to change more common it will increasingly if PEP for sexual exposure 1. HIV Post-Exposure Prophylaxis: Guidance from impact on PEP’s effectiveness and was to become more widely the UK Chief Medical Officers’ Expert Advisory Group on AIDS (Dept of Health guidelines on PEP the need for resistance-testing of availab