Recent Developments at DG Competition: 2014

The Directorate General for Competition at the European Commission enforces competition law in the areas of antitrust, merger control, and state aids. In this year’s review of the recent economic work undertaken at DG Competition we focus on two themes: the modernisation of state aid (with a specific focus on the Aviation Guidelines); and the application of economic techiniques to remedies evaluation in mergers (with a number of case studies).


Introduction: Main Developments in 2014
As has been true in past years, we briefly describe in this article some highlights from competition policy and enforcement that was done at DG Competition: the Directorate General of the European Commission that is in charge of competition issues, during the year 2014. 1 This year we focus on two main topics: First, we discuss what has been arguably the main policy initiative completed during 2014: the State Aid Modernization (SAM) reform: 2 a program that was started in 2012 by Joaquín Almunia, Vice-President of the European Commission responsible for Competition Policy. Second, we discuss some issues that are related to merger remedies. Most of the resources of the Chief Economist Team are employed in merger enforcement, and we economists have been increasingly involved in all stages of merger cases, including the evaluation of remedies.
The SAM reform had two principal objectives: First, modernization aimed to focus the enforcement activity of the Commission on the largest and most distortive types of subsidies, by simplifying the rules and rendering the decision-making process more efficient and quicker. Second, it intended to ensure a coherent and economically sound approach to all types of state aid. In Sect. 2 we briefly deal with the main pillars of the reform-the extension of the scope of the General Block Exemption Regulation and the adoption of several enforcement guidelines of different types of state aids-before discussing more in depth the Aviation Guidelines; the latter are a good example of how economic principles can be used, in a particular context, to assess whether state intervention should be considered illegal.
Prohibitions of mergers are relatively rare, and therefore most of the burden of merger intervention falls upon remedies. An appropriate design and implementation of the remedy is therefore crucial to ensure that a merger does not impact negatively upon market competition. In the last few years we have been increasingly aware of the complexity that some remedies may involve, especially when the purchaser of the divested assets depends on the merged entity for some inputs (for instance, in shared facilities) or in access remedies. Unless the remedy is designed in such a way that the buyer is viable and has both the ability and the incentive to compete, the merger will have anticompetitive effects.
Remedy issues of this nature have arisen in a number of merger cases, and in very different sectors. Although we also refer to mergers that did not take place in 2014, we thought it would be useful to report our experience in this respect, as it is likely to be of interest for other agencies and for competition scholars in general.
Before discussing the two above-mentioned topics in detail, let us also briefly review the main activities that were undertaken by DG Competition in 2014.

Policy and Enforcement at DG Comp in 2014, in a Nutshell
In terms of competition policy, the two main initiatives that were completed in 2014 were: the SAM reform (which is the subject of Sect. 2), and the adoption of the Directive on antitrust damage actions. 3 This piece of legislation completes an initiative that the European Commission undertook a few years ago to support private actions for damages (which, until recently, were virtually inexistent in European competition law). The Directive has the objective of ensuring that damages actions before national courts (private damages are not a competence of the European Community Courts) for antitrust violations are both effective and implemented in a coherent way across the Member States. It also aims at clarifying the relationship between private actions and public enforcement of competition law (as it was feared in particular that damages actions may jeopardize leniency programs, which have turned out to be a crucial tool to fight cartels).
In terms of competition enforcement, the European Commission's activities can be divided into antitrust (cartels, other non-cooperative agreements, and abuse of dominance), mergers, and state aid. 4 In 2014, the Commission issued ten cartel decisions (against an annual average of slightly above five cases in the period 2009-2013) for a total of EUR 1.69 billion in fines. The decisions concerned very different sectors: from automotive parts to chemicals, from financial derivatives to smart card chip producers, from power cables to steel abrasives. Notably, eight of the ten decisions involved a cartel settlement procedure: a relatively recent change (the first settlement decision was adopted in 2010) that allows the Commission to save resources and speed the process.
As for the remaining antitrust cases, the Commission also took six decisions (a number that is more-or-less in line with recent years) with total fines of EUR 499 million. Four of them were prohibition decisions: Romanian Power Exchange (OPCOM): the exchange, was found to discriminate against non-Romanian EU electricity traders; Motorola: a Standard Essential Patent case that we discussed in last year's issue; Perindopril: an anticompetitive agreement between Servier and five generic pharmaceutical companies to delay or block generic entry; and Slovak Telekom: a margin squeeze case. Moreover, two commitment decisions were adopted: one against Visa Europe for multilaterally-agreed interchange fees; and the other against Samsung, which was another Standard Essential Patent case.
In the field of mergers, the Commission took 300 decisions, which was indicative of an increasing M&A trend relative to the five previous years (with an annual average of 273), but still far from the peak of the pre-crisis years (the record was the 398 merger decisions in 2007).
In 2014, no prohibition was issued; five decisions ended in the second phase (that is, an in-depth merger review) with remedies; and 13 mergers were approved after 3 European Parliament and Council Directive 2014/104/EU of 26 November 2014 on certain rules governing actions for damages under national law for infringements of competition law provisions of the Members States and of the European Union, OJ L 349, pp. 1-19 available at: http://eur-lex.europa.eu/ legalcontent/EN/TXT/PDF/?uri=OJ: JOL_2014_349_R_001&from=EN. 4 The enforcement data described below is reported in DG Competition's 2014 Annual Activity Report (available at: http://ec.europa.eu/atwork/synthesis/aar/doc/comp_aar_2014.pdf). remedies were offered in the first phase of the investigation. Note that in two cases the merger was approved unconditionally during the second phase.
If we define as interventions the sum of cases of prohibition and of approvals subject to remedies (as well as transactions abandoned during second phase), the 6 % intervention rate (18/300) is more-or-less in line with previous years.
As for state aids, the Commission adopted 866 decisions in 2014. 5 But-SAM reform apart-the most notable development in state aid is probably given by the fact that DG Competition started to investigate possible fiscal state aid that some Member States may have given to selected multinational companies. In 2014 an indepth investigation was opened on tax rulings (i.e., decisions by the national fiscal authorities) of Ireland, Luxemburg, and the Netherlands, to see whether they have given an illegal favorable tax treatment to certain multinational firms (Apple in the Irish case; Fiat and Amazon in Luxemburg; and Starbucks in the Netherlands).
The extent to which economic analysis is used in competition cases varies to some degrees across instruments and type of cases, and this is reflected in the involvement of the Chief Economist Team (CET). For instance, the CET is typically not involved in hard-core cartel cases or in antitrust cases where the infringement is considered obvious, since in such cases there usually is little need for a formulation of a theory of harm and for an assessment of the effects of the practice at hand. On the contrary, our involvement is typically the highest in the most significant merger cases, where economic analysis plays an important role throughout the investigation and one or more members of the CET are working together with the case teams (it is rare for the CET not to be involved in a second phase case). Finally, in state aid the CET is typically involved in a few cases every year, when economic issues are more relevant.
In general, about 60 % of the CET's labor resources are dedicated to mergers, with the remaining time split more-or-less equally between state aid and antitrust. Under the new State aid architecture, in order to achieve the first objective of focusing on the most distortive subsidies, a large number of state measures will no longer be subject to prior review by the Commission. In particular, under the new General Block Exemption Regulation (GBER) that was adopted in 2014, a wider range of state measures and higher aid amounts can now be granted without prior notification. Maximum allowed aid intensities (i.e., the ratio of aid over costs defined as eligible for state support) have been raised as well.
The widening of the scope of the GBER and the increased aid ceilings relied on evidence from case practice regarding the least distortive and problematic types of state measures. For example, under the new GBER, the ceiling for risk finance aid was raised up to €15 million per small-and medium-size enterprise (SME) and as a one-off aid limit covering the entire development cycle (as opposed to €2.5 million annual tranches per SME under the previous regime). These changes were based on case practice and evidence that activity in venture capital markets remained below the pre-crisis levels and that the SME equity gap is wider than previously thought.
Alongside the broadening of the scope of the GBER, which would effectively remove from the Commission's scrutiny significant amounts of state aid, the Commission updated the main guidelines for specific types of aid. These guidelines provide guidance on how to assess aid measures that are subject to prior authorization by the Commission, and they describe the conditions under which state aid can be authorized. A common framework of analysis has been introduced across all guidelines, with the aim of bringing a consistent approach across the different types of aid and sectors covered by the guidelines. This common framework relies more directly on economic arguments and empirical evidence on the effectiveness of subsidies.
Specifically, in 2014, the Commission adopted new guidelines for Rescue and Restructuring aid (R&R), for R&D and Innovation aid (R&D&I), for Risk Finance aid, for aid to airports and airlines (Aviation guidelines), and for Energy and the Environment aid. In 2012-2013, the Commission had already adopted new guidelines for Regional Aid and guidelines for aid to broadband networks.
The widening of the GBER was associated with greater obligations for Member States to implement ex-post evaluation studies to assess empirically the effectiveness of their exempted aid schemes. The 2013 guidance paper on ex-post evaluation-which describes in detail appropriate methodologies, processes, and data requirements for implementing a robust evaluation plan-was discussed in detail in last year's review of economics at DG Competition. 7 In what follows, we briefly discuss the common principles adopted across the new state aid guidelines, with a particular focus on the Aviation Guidelines which advocate a new approach for determining when subsidies constitute compatible state aid in the case of infrastructure.

The Common Principles to Assess When State Aid is Compatible
Across the newly adopted guidelines for the assessment of individual aid or schemes, a set of common principles has now been systematically included. Note that these principles also underpin the reasoning of the GBER, though the requirements for fulfilling these principles may differ between the GBER and the guidelines. 8 These principles guide the Commission's analysis of whether a particular aid measure or scheme can be declared compatible under the Treaty on the Functioning of the European Union (i.e., authorized). In particular, the Commission may approve state aid if seven criteria are satisfied. Most of these criteria can be seen as ''filters'': necessary conditions to be fulfilled. Once these conditions are fulfilled, in a second stage a balancing test is undertaken to evaluate whether the expected positive effects of the measure would outweigh any potential distortions of competition. Finally, the aid measure should also be transparent.
The criteria for assessing the compatibility of an aid measure are detailed below: The first and second conditions relate to the clear identification of the objective and rationale for the state intervention. First, the aid measure (either individual aid or a scheme) should contribute to a well-defined objective of common interest (i.e., in line with the Treaty on the Functioning of the European Union). Second, it must be shown to be necessary to pursue its objective because the market is not able to deliver the same objective. From an economic perspective, there are two main rationales to justify public intervention into economic activities: (1) public intervention may increase welfare (efficiency rationale) when markets do not deliver socially optimal outcomes due to market failures (such as externalities, coordination issues, or information asymmetries); (2) it may contribute to redistribute market outcomes more equitably when markets lead to outcomes that are considered not to be equitable (equity rationale).
As an illustration, in the 2014 Aviation Guidelines, avoiding air traffic congestion at major European Union hubs and promoting regional development are considered well-defined objectives of common interest. In the 2014 Risk Finance Aid Guidelines, the provision of finance to viable SMEs from their early development stage up to their growth stage is considered a (general) well-defined objective of common interest. However, in the case of the 2014 Risk Finance Aid Guidelines, Member States must also provide an ex-ante assessment of the existence of a funding gap that affects the targeted undertakings, resulting from market failures (e.g., such as information asymmetries that lead to sub-optimal provision of finance to SMEs in their early development stages). The 2014 Energy and Environmental Aid Guidelines consider that the increase of environmental protection beyond what would be achieved without aid constitutes a well-defined objective of common interest.
The third and fourth conditions ensure that public intervention is effective in achieving its goal (say, addressing a market failure) as well as kept to the minimum for the achievement of its objective. In particular, Member States are required to show an ''incentive effect'' (third condition), which indicates that the state intervention has the expected causal impact on the beneficiary's activity. Indeed, it should lead the beneficiary to engage in activities that it would otherwise not undertake, or not in a similar manner (in terms of scope, timing, etc.).
Establishing that aid has an incentive effect requires analyzing the counterfactual: What would the beneficiary have done without the aid? The difference between the counterfactual and the intended activity under aid represents the incentive effect of the aid. Recent empirical studies have shown, for example, that the provision of aid may have very different effects on small and large firms. Looking into the impact of a public scheme to support manufacturing jobs (''Regional Selective Assistance''), the authors of a study found that the aid only had an effect on small firms; the aid did not cause the observed changes in employment levels at larger firms. 9 In other words, in this study, the incentive effect was confined only to small firms, as the same behavior would have been observed in large firms without the aid.
In addition to having an incentive effect, the aid must be proportional or kept to the minimum (fourth condition). This means that it should not exceed what is required to ensure that the beneficiary engages in the expected activity (or locates in a particular area). For the analysis of the incentive effect and proportionality, a usual method includes the calculation of a project's or investment's net present value (NPV); and when such NPV is negative (which would justify the need for state intervention), the funding gap (i.e., the difference between the expected positive and negative cash flows over the life of the project/investment) serves as a basis for evaluating the proportionality of the aid. Internal documents and internal investment models are also useful to document precisely the decision-making process in the recipient company and the reasons why aid is needed to carry out specific projects/ investments.
Note that under the new R&R Guidelines, the proportionality conditions require a significant contribution to the restructuring costs from the beneficiary's own resources as well as burden-sharing between the State and existing investors. These conditions aim at limiting the excessive risk-taking and moral hazard that may arise when firms (and their investors) anticipate state support in case of financial difficulties.
A fifth condition is that providing state aid constitutes an appropriate policy instrument to achieve the objective. This condition requires Member States to explain why public funding is a better instrument than say, regulation; also, among state-aid instruments, Member States are expected to explain why the chosen measure is appropriate (e.g., tax exemption versus direct grant).
If the aid does not fulfill all of these conditions, it cannot be deemed compatible. Indeed, if (for example) state aid does not provide incentives for a firm to engage in activities that fulfill the objective of the aid or if the company would have engaged in such activities anyhow, then the public funding is wasteful: public funds simply constitute windfall profits for the beneficiary. The same reasoning applies if the aid is not proportionate: if the same effect could have been achieved with less public funding.
When the aid is found to pursue a clear objective, to fulfill the incentive effect, to be appropriate and kept to the minimum, the Commission is still required to assess the sixth condition: the balancing of the positive effects of the aid (identified under the previous conditions) and the potential negative effects of the aid on competition and trade. The overall effect of the aid must be positive.
A number of potential theories of harm can be evaluated, depending on the context, and these are described throughout the guidelines in terms of the distortions that state aid could entail. The main theories of harm include the prevention of exit, the distortion of dynamic incentives, the creation or maintenance of market power, and location decision distortions.
For example, a potentially harmful effect of R&R aid derives from the prevention of exit, thus interfering with the normal competitive process by keeping inefficient firms active. Under the R&R guidelines, the minimization of such negative effects is achieved by ensuring that R&R aid is provided on an exceptional basis: the ''One time, last time'' principle, under which ailing companies are allowed R&R only once in a period of 10 years. Further, R&R aid must be accompanied by so-called ''compensatory measures'', which is aimed at limiting the distortions of competition by often requesting beneficiaries to divest certain assets or to reduce their production capacity. Devising appropriate compensatory measures is admittedly a delicate exercise given that such measures can affect the viability that the restructuring plan is aimed at re-establishing in the first place.
Under the R&D&I framework, the dynamic incentives to innovate of the beneficiary's competitors could be negatively affected (a crowding-out effect), and the beneficiary itself could be more prone to risk seeking. The R&D&I guidelines describe the various elements that the Commission would consider in identifying the scope for distortions of dynamic incentives such as market growth, exit barriers, product differentiation, closeness to the market of the activity being aided, etc.
In the Aviation Guidelines, investment aid or operating aid will be expected to be distortive if the beneficiary airport is located in the catchment area of an existing airport, especially if there is spare capacity in that existing airport. Indeed, in such case, aid would likely lead to the inefficient duplication of unprofitable infrastructures.
Finally, as a seventh (final) condition for compatibility, Member States are expected to give easy access to all of the information that is relevant to the aid that is being granted: the transparency requirement.
The new Guidelines that were adopted under the SAM are therefore more firmly grounded in economics in terms of determining a clear objective, filtering out aid that is wasteful (incentive effect and proportionality conditions) and introducing a structure for the balancing of positive effects (e.g., remedy market failures) with negative effects (e.g., distort dynamic incentives to invest or prevent the exit of inefficient firms at the expense of efficient rivals).
In the next section, the New Aviation Guidelines will be discussed in greater detail. In particular, the Guidelines introduced two novelties in the way that state aid is now assessed in the aviation sector.

Background
The aviation sector in Europe has experienced significant developments in the last 20 years. With the development of low-cost carriers, route liberalization, and airport privatization, European airports have mushroomed, and competition amongst airports has intensified to attract low-cost carriers as well as develop hubs. Large airports compete both for point-to-point and transfer traffic, whilst regional airports have focused on low-cost, point-to-point traffic. These developments have led to lower prices (on competitive routes), greater access to regions, and an increased number of served city-pairs, and most likely have promoted regional development.
However, the public funding of airport infrastructure has also led to the duplication of unprofitable competing infrastructure. Competition to build airports and attract low-cost carriers has led to the multiplication of ''ghost airports'' and unused capacity in small, regional airports. A 2014 report by the European Court of Auditors 10 analyzed in detail the situation of 20 publicly funded airports in five Member States (Estonia, Greece, Spain, Italy, and Poland). The report concluded that too many airports in close proximity to each other were oversized and unprofitable and could survive only with continuous public support.
Generally, data on airport profitability shows that the share of unprofitable airports decreases with airport size. 11 This can be explained by the fact that there is arguably a structural difficulty in achieving full cost coverage for smaller airports as these have high fixed capital costs and operating costs that cannot be spread across a large traffic base, and a more limited capacity to generate non-aeronautical revenues such as revenues from parking fees, shopping, restaurants, hotels, and other services (which represented about a third of all revenues at European airports in 2013). The empirical observation that airport profitability and size are closely linked was reflected in the 2014 Aviation Guidelines.
Furthermore, EU funding for airport infrastructure is not coordinated across countries and across Member States, and the multiplication of unprofitable airports that rely on public support is also often decried by the local or the international press. 12 Against this backdrop, the 2014 Aviation Guidelines focused on identifying the cases in which investment aid would be targeted at increasing capacity where it is effectively needed, gradually phasing out operating aid, allowing start-up aid to new destinations and untested airports, and clarifying the conditions under which airlines (indirectly) benefit from state aid through (net) airport charges (also considering rebates, incentives, or any other financial arrangement between airports and airlines) that are too low to be commercially justifiable by the subsidized airports. In the remainder of this section, we focus on the ''lump-sum'' approach for operating aid and on the ''incremental cost'' approach for assessing aid to airlines (as these 10 Special Report No 21/2014: EU-funded airport infrastructures: poor value for money. 11 The 2012 ACI Europe Economics Report shows that whilst 51 % of Europe's airports with more than 5 million passengers were loss-making (including non-operating income), this share reached 73 % for airports with fewer than 1 million passengers. The latest 2014 ACI Europe Economics Report shows that with respect to net profits, 77 % of small European airports (below 1 million passengers) were loss making in 2014 compared to 17 % of airports with 5-10 million passengers, 24 % of airports with 10-25 million passengers, and 0 % for the largest airports (above 25 million passengers). incorporate economic concepts more directly), as well as a brief description of the approach to investment aid.

Operating Aid: The Lump-Sum Approach
Given the existing situation of over-capacity and duplication at numerous airports (mostly small regional airports), the 2014 Aviation Guidelines take a forwardlooking approach, since investments in infrastructure are sunk. Based on the available empirical evidence relating to airport profitability and size, operating aid to airports with more than 3 million passengers a year would generally not be found compatible.
For a transitional period of 10 years, airports with fewer than 3 million passengers will be allowed to receive operating aid under certain conditions: the amount of compatible aid should not exceed 50 % of the initial operating funding gap of the airport for a period of 10 years and the airport must present a business plan that demonstrates that the airport will become viable by the end of the transitional period. 13 Furthermore, in order to provide proper incentives for the efficient management of the airport, the amount of operating aid will be determined ex-ante, as a lump sum (that is payable at once or in installments). The important feature is that the lump sum is determined ex-ante and it is thereafter the airport's responsibility to achieve full operating cost coverage before the end of the transition period: for example, through cost-reducing measures, pricing adaptations, and any other appropriate business restructuring measures. The main purpose of the new approach is to reverse the situation of excessive competition between airports to attract lowcost carriers that arose from the airports' soft budget constraints and continuous (expost) subsidisation. Public funding of operating losses led to airports' being structurally unprofitable and to airlines' facing airport charges that do not reflect the actual costs that are generated by those airlines.
By introducing an efficiency-inducing approach, leading to the tightening of the airports' budget constraint, it is expected that full operating cost coverage would be achieved at least at some existing airports at the end of the transition period. Given the existing excess capacity relative to the actual traffic generated, the 2014 Aviation Guidelines aim at a smooth transition allowing airports to adjust their capacity and cost base in order to return to profitability. After the transitional period, operating aid will not be allowed.
There are exceptions for start-up aid and traffic considered as services of general economic interest. Indeed, irrespective of their size, airports that are located on islands or remote areas can benefit from the compatibility conditions relative to Services of General Economic Interest and are not subject to the lump-sum approach or the transition period. Such airports can be inherently unprofitable but their contribution to regional connectivity justifies a more lenient treatment towards operating aid.
As part of the compatibility analysis, the competitive situation around the airport under consideration will be assessed and should be taken into consideration in the submitted business plan. In particular, there will be doubts about the compatibility of the aid if there are already other airports with unused capacity operating in the same catchment area and Member States will have to demonstrate that full cost coverage could be achieved by all competing airports in the catchment area.
To this date, the Commission has taken decisions that relate only to past operating aid (i.e., aid that was granted before the adoption of the 2014 Aviation Guidelines), and therefore the ex ante lump-sum approach has not yet been applied in practice.

Aid to Airlines: The Incremental Cost Approach
The 2014 Aviation Guidelines also introduced a new approach to assess the existence of state aid in contracts between airports and airlines. In particular, as a large number of existing airports have received state aid, one important question relates to whether aid was also indirectly granted downstream to the airlines that use the airport.
When airports receive subsidies, they are able to charge lower airport charges in order to attract carriers-low-cost carriers in particular, since they focus on pointto-point traffic and are able to switch more easily between airports. In order to determine whether the airport charges paid by airlines constitute state aid (i.e., whether a private operator maximizing profits would have charged the same price), the Guidelines propose two approaches: first, a benchmarking approach that consists of a comparison of the price that is charged to the airline with a ''market price'': the prices that are charged by comparable airports for similar services; and second, the long-run incremental cost approach that consists of an identification of whether the airport charges that are paid by the airline cover at least the long-run incremental costs caused by the presence of the airline at the airport.
Given that many airports have been subsidized in Europe, the first approach may not be meaningful until the market normalizes and prices adjust to a level that is not affected by the granting of state aid. The possibility to use benchmarking would therefore depend on the specific situation (and extent of subsidization) of the airports that are used for benchmarking.
Under the second approach, the Commission is required to check whether the revenues that are generated by the presence of an airline at the airport (both aeronautical and non-aeronautical revenues) cover at least the long-run incremental costs (that is both operating and any investment costs) that are caused by the airline's activity at the airport. This approach allows for differential pricing. 14 Under the long-run incremental test approach, no state aid is deemed to be indirectly transferred to airlines if fixed costs that would have been incurred anyway by the airport are not being covered by the airport charges. Given that airports are infrastructures with a high proportion of fixed costs, the Guidelines recognize that 14 The overall strategy of the airport to achieve profitability over the long run when setting prices is also taken into account (Paragraph 66 of the Aviation Guidelines).
pricing and common cost recovery would tend to lead to differential pricing across users of the infrastructure.
Hence, if the airport is able to show that ex-ante, the airport charges were set at a level such that the contract with the airline is profitable on an incremental basis (i.e., contributing to the airport's profitability), then the contract would be considered free of state aid, and no state aid would have been transferred to the downstream user of the (subsidized) infrastructure.
Since April 2014, the Commission has found that some airlines have received illegal state support given that the pricing arrangements in their contracts with the airport were such that expected revenues from the activity of the airlines did not even cover the incremental costs of their presence in the airport. For example, this was the case for agreements between the airport of Pau-Pyrénées and Ryanair and Transavia, or between the Alghero airport and Meridiana and Germanwings. By contrast, pricing arrangements between Ryanair and the Frankfurt-Hahn airport, or between Air Berlin and the Saarbrücken airport have been found not to constitute state aid as the pricing schedules were found to contribute incrementally to the airports' profitability.
It is also noteworthy that a similar approach to evaluating the profitability of a contract for the use of infrastructure was applied in a case that related to a waste water treatment plant. 15 In that case, the Commission examined whether the fees paid by a paper mill to a waste-water treatment plant covered the long-run incremental cost of the use that the paper mill made of the treatment plant. In particular, it found that the fee paid covered operating costs and maintenance and repair costs, as well as personnel and management costs. Imputed depreciation and an additional imputed interest on 20 % of the investment capital were equally covered. Costs that the plant would have to incur anyway, without the use by Propapier, did not have to be covered by the pricing arrangement.

Airports: Investment Aid
Finally, the 2014 Aviation Guidelines introduced new rules for the assessment of compatibility for investment aid. First, the amount of state aid that is allowed for investment will depend on the size of the airport. Above 5 million passengers a year, unless there are exceptional circumstances, airports cannot benefit from investment aid as evidence suggests that the larger airports are sufficiently profitable to cover 15 The case had started in 2007 with the notification of regional aid to be granted to Propapier, a paper manufacturer, for the construction of a paper mill. The aid was found compatible; but, following an appeal by Smurfit Kapa (a competitor), the Court annulled the decision on the grounds that the Commission had failed to substantiate sufficiently its analysis of the balancing between the negative and positive effects of the aid, and therefore, it ''did not put itself in a position to overcome all doubts as to the compatibility of the aid in question with the common market.'' [Case T-304/08]. The Commission reassessed the case; and in October 2014 it re-adopted a compatibility decision with regard to the regional aid. In addition, it adopted a second decision that found that the fees that were paid by Propapier for the use of a waste-water plant (and surrounding infrastructure such as a parking lot) that was built by the German state did not involve any state aid as the fees incrementally contributed, from an ex-ante perspective, to the profitability of the waste-water plant operator. The waste-water plant had not been built specifically for Propapier, and its use was open to all. their capital expenditures. Below 5 million passengers, the share of investment costs that can be subsidized increases as the size of the airport decreases (e.g., subsidization as great as 75 % for airports with fewer than 1 million passengers compared with 25 % for airports with 3-5 million passengers).
For investment aid, the Commission will have doubts with regard to subsidies that lead to the duplication of unprofitable capacity or the creation of additional unused capacity when there are existing airports with spare capacity. The Commission has already adopted two decisions that found that investment aid to airports would lead to competition distortions and therefore constituted illegal State aid.
First, in February 2014, the Commission adopted a decision that found that investment and operating aid that had been granted to the Gdynia-Kosakowo airport, in Poland, had led to competition distortions and had to be recovered. The Gdynia-Kosakowo airport is a military airport which would have been transformed into a civilian airport, targeting mostly low-cost carriers, charters, and general aviation 16 activities. It is located 25 km. from the Gdansk airport, which is uncongested, already serves the Pomeranian region and also focuses primarily on low-cost traffic.
Furthermore, Gdansk had already an approved investment plan to increase capacity to 5 million passengers by building a second passenger terminal. The business plan for the Gdynia airport presented during the investigation contained unrealistic traffic and revenue projections given the activity at the existing Gdansk airport.
Following the Commission decision, the Gdynia-Kosakowo airport was declared bankrupt.
Similarly, in October 2014 the Commission found that investment aid to the airport of Zweibrücken, which is located about 40 km by road from the Saarbrücken airport (which has existing and ample capacity to serve the Saarland region), merely duplicated capacity. Hence, Germany was ordered to recover the grants and capital injections paid into the airport in the preceding years. In anticipation of the negative decision, the Zweibrücken airport had filed for bankruptcy in July 2014; and by November 2014, all scheduled flights had ceased. Some of the flights were relocated to Saarbrücken (e.g., TUIFly had a base at Zweibrücken and relocated to Saarbrücken).
Following these two negative decisions by Commission, airport capacity was thereby reduced (in the case of Zweibrücken) or not increased (in the case of Gdynia-Kosakowo) in two regions of Europe. Both decisions are currently being challenged in Court (though the Zweibrücken decision has been challenged by an airline, Germanwings, which was ordered in the same decision to repay illegal subsidies that relate to its pricing arrangement with the airport).
Whilst the actual market impact of the 2014 Aviation Guidelines will not be known for some time (depending on the Commission's enforcement action and how airports adjust to the new rules), these guidelines show how empirical evidence and economic theory can feed into the design of policies that are aimed at tackling subsidies in a particular sector.

Some Economic Issues in Remedy Evaluation
Over the past few years economists at DG Competition have been increasingly involved in the evaluation of remedies that have been proposed by merging parties in order to address the potential competition concerns that have been raised by their transactions. The appropriate design of remedies, and their assessment by the competition authority, often raises complex legal and economic issues. These are typically related to the economic theory of harm that is associated with the merger at hand, but can also extend to broader concerns (for example, as is discussed in-depth below, a divestiture ''carve-out'' that is designed to address a standard horizontal concern, can in turn raise vertical foreclosure issues).
The appropriate design of a remedy is essential for effective merger enforcement. Most of the merger interventions by DG Competition (defined as prohibitions, merger withdrawals in Phase II, and remedies offered in Phase I or Phase II) consist of remedies. For example, if one considers the four calendar years that ended in December 2014, there have been 56 interventions by DG Competition; 51 have been remedies [the majority in Phase I (37), but also a significant number in Phase II (14)].
In this section of the article we focus on some of the more complex design issues that can be raised by remedies, based on our recent experience. These include three related issues: (a) the impact of any dependence of the purchaser of the divested assets on the merging parties (be it temporary or of a more lasting nature); (b) the determination of the cost structure of the divested entity, relative to the cost structure of the merging parties pre-merger (and the related, but broader, issue of ''future-proofness''); and (c) the relevance of the level of interest expressed by potential buyers in a given remedy package in the overall evaluation of the effectiveness of that package.
The other over-arching issue in remedy assessment, which often involves detailed economic arguments, is one of ''scope'': For example, how large does divested capacity need to be and which assets need to be included in order to offset a given competition concern?. We do not deal with this issue in this article in order to focus more in-depth on the other issues that were listed above. Four of the recent Phase II intervention cases that we have reviewed in past issues of our annual contribution to the Review of Industrial Organization 17 (the Universal/EMI, Outokumpu/Inoxum, UPS/TNT and Ineos/Solvay transactions) all raised issues that are related to the scope of the remedies offered by the merging parties. These aspects of the cases are in part reviewed in the corresponding articles in the Review of Industrial Organization.

Complex Divestitures
Divestiture remedies are the most prevalent form of remedies that are accepted by the European Commission when clearing mergers subject to commitments. The benefits of divestiture remedies relative to access or conduct remedies are well established. They potentially allow for the creation on a lasting basis of a new competitor in a market, or a stronger competitor (when the divested assets are purchased by an existing player, without raising in turn competition concerns).
A suitably-designed divestiture can also allow a new player to have access to the same or similar production technology as the one available to the merging parties, thus benefitting from any future improvements in that technology (e.g., in the form of lower costs, additional capacity, or additional innovation). A divestiture remedy can also allow the new firm to be independent of the merging parties, avoiding in particular the risk of foreclosure.
Most of the remedies that have been accepted by the European Commission over the past four calendar years (ending in December 2014) have been relatively clearcut and traditional asset divestments: designed to reduce a horizontal overlap, or to mitigate barriers to entry.
Depending on the nature of the assets involved and of the transaction, it is however not always feasible (or indeed necessary) to design an asset divestment that eliminates all links between the merged entity and the buyer of the divested assets, both in the short-and the long-term. 18 For example, transitory agreements for some inputs may be required, in case the new buyer needs time to set up its own supply arrangements for some key input. The nature of such transitory arrangements will depend on the complexity of the industry and of the input at stake. 19 Similarly, in the case of a merger that involves multi-product assets where only a sub-set of the relevant products raises competition concerns, it may not be possible to divest only the problematic assets without allowing for some structural links to remain in place between the merged entity and the divested units (e.g., in the form of joint access to shared facilities, or input supply contracts). For reasons of proportionality, the Commission may accept a ''carve-out'' solution where only the assets that are directly related to the problematic product are divested, and not the entire production facilities where the assets are located. However, for reasons that are discussed below, specific safeguards are necessary to ensure that a carve-out 18 In order to avoid a reduction in competition following a merger, a competition authority should accept a complex divestiture that, for example, allows for some links to remain in place between the divested assets and the merged entity, only if it is satisfied that such a complex remedy is sufficient to remove the competition concerns. Therefore, the fact that a ''clear-cut'' divestment may not be feasible in a given case is not in itself a reason to accept a complex divestment in order to approve the transaction. 19 If the required access arrangements involve inputs that account for a substantial part of the value of the divested business, and if they last for a significant amount of time, then the divestiture remedy should in practice be regarded as similar to an access remedy. The line between a complex divestiture and an access remedy is in practice blurred, and to some extent a matter of definitions. Moreover, even an access remedy can be designed in a way to make it quasi-structural in terms of its impact on the market; and therefore, depending on the design, it can be looked at as a complex divestiture (on the basis of terminology used in this article). This is how the remedies that were accepted by the Commission in the mobile mergers that were cleared in 2014 are dealt with in this article. remains effective in terms of resolving a competition concern, and a ''reverse carveout'' (or even a full divestiture of the multi-product assets) may ultimately be required to address a competition concern. Figure 1 below illustrates the principle of a carve-out remedy. It takes as an example a multi-product merger that involves two products (A and B), only one of which raises competition issues as the parties compete for sales of that product premerger (through the sale of products A1 and A2, respectively). Each plant has dedicated access to a set of self-produced inputs (illustrated as U1 and U2, respectively, in Fig. 1).
A carve-out remedy would consist of the sale of the activities that are linked to product A1 to an independent firm, carving it out from the existing multi-product plant that produces products A1 and B. 20 Depending on the nature of the carve-out arrangement, the new owner of A1 may need to be supplied with a key input from the multi-product plant retained by the merging parties (shown as U1 in Fig. 1).
Over the past 4 years, the European Commission in some cases has accepted complex divestitures that involved either access to shared assets between the merged entity and the divested buyer, carve-outs of existing businesses, or some form of access arrangements for some of the inputs that are required by the buyer (typically, but not always, on a temporary basis). 21 The issues that are addressed in B A2 A1

U1 U2
Input contract Fig. 1 Multi-product merger, and a carve-out remedy 20 For simplicity, in this illustration product A2 is assumed to be produced in a single-product plant. In practice, product A2 too may be produced in a multi-product plant, and therefore any divestment remedy would raise the issue of how to deal with multi-product assets. 21 Recent examples of recent complex divestitures in Phase II include Munksjo/Ahlstrom, Syniverse/ Mach, Ineos/Solvay, Hutchison/Telefonica Ireland, Telefonica Germany/E-plus, and Huntsman/ Rockwood. the remainder of this section are typically raised in cases where such complex divestitures are carried out.

Foreclosure Issues Due to the Dependence of the Divested Assets on Inputs that are Provided by the Merging Parties
A typical competition concern in remedies where the divested business depends on the merged entity for some key input is one of foreclosure. 22 The concern is that the input would be supplied at worse competitive conditions than those that are available to each of the merging parties pre-merger, thus reducing competition relative to the counterfactual without the transaction. Input foreclosure can take the form of higher prices and/or lower quality relative to the pre-merger situation. Economic theory predicts that there may be incentives to engage in input foreclosure (be it partial or complete) in a post-remedy scenario such as the one depicted in Fig. 1. Suppose that the owners of products A1 and A2 are close competitors prior to the merger, and that they are both vertically integrated and receive any necessary input for production at (marginal) cost. Under these assumptions, it is likely that the un-remedied merger would generate a significant upward pressure on prices, due to the loss of direct competition between A1 and A2.
Under a carve-out remedy, the assets producing A1 would be carved out from the merged entity, but continue to be supplied with some key input U1 by the merged entity. This can be thought as a situation where an upstream input monopolist that owns both U1 and U2 is integrated downstream (i.e., it owns A2), and also supplies a downstream competitor with the input (i.e., it supplies A1 with input U1). The supply contract for U1 consists of a two-part tariff, with a fixed fee equivalent to the upfront purchase price for asset A1, and a wholesale price for each unit of the input U1 that is supplied to A1. 23 Under plausible assumptions, the merged entity would face an incentive to raise the price of asset A1 in order to reduce the competitive constraint that A1 exercises on asset A2 (thus enabling A2 to raise its prices too). 24 This conduct would reduce competition downstream, and thus allow the merged entity to increase profits compared to the pre-merger situation. The profits that are earned by A1 in the downstream market would then be extracted by the merged entity through the fixed 22 Another possible concern is that the supply relationship between the divested business and the merged entity may increase transparency in the market, and thus lead to a risk of coordinated effects. We do not address this possible concern in this article. 23 The counterfactual in the absence of the merger would be instead a situation where two upstream suppliers of U1 and U2 are each vertically integrated downstream (with A1 and A2 respectively), and compete in the downstream market. 24 See for example Reisinger and Tarantino (2015). An important assumption to generate foreclosure incentives in this setting is that the merged entity can only contract on wholesale terms with the new owner of A2, but cannot pre-commit to its own retail output and prices. This assumption is realistic in a remedies setting since an agreement between the merged entity and A1 on retail prices or quantities would be effectively equivalent to collusion in the downstream market, and it would therefore not be legal. Under the assumption that the merged entity cannot pre-commit to retail outcomes, once the wholesale contract between the merged entity and A1 has been agreed upon, A1 and A2 compete in the downstream market on the basis of the respective upstream costs. fee paid by A1 (e.g., this is the case if one assumes a competitive divestiture process that determines the upfront purchase price for assets A1).
The precise form of foreclosure in the remedy setting described above depends on the economic framework that is most relevant to the case at hand: • If the downstream products are homogenous and A1 and A2 have the same downstream costs, then the merged entity would face incentives to completely foreclose A1 in order to obtain monopoly profits downstream through its affiliate A2. In such a setting, a carve-out remedy without any additional safeguards on the terms of supply would be ineffective, and it would potentially not result in any pro-competitive effect relative to the merger scenario. • In an alternative setting where A1 and A2 have different costs (say, A1 is more efficient than A2), then the merged entity would face incentives to engage in partial foreclosure of A1 (i.e., set a wholesale price that is above the upstream cost) if the cost difference between A1 and A2 is not too large. 25 Under this setting, partial foreclosure of A1 would still reduce the effectiveness of the remedy, and lead to consumer harm relative to the counterfactual absent the merger (where both downstream firms have access to the input at cost). • Finally, in a setting where A1 and A2 offer differentiated products, there will be typically incentives for partial foreclosure but not for complete foreclosure (since the merged entity would profit from maintaining the competing downstream firm active, due to the presence of product differentiation). 26 The partial foreclosure incentives in this setting can be usefully characterized as a ''vertical Gross Upward Pricing Pressure Index (GUPPI)'', 27 whereby the wholesale price that is charged to A1 is raised above cost in order to benefit the owner of A2 (via the diversion of sales from A1 to A2, and the resulting higher margins on A2). 28 The reduction in competition from A1, and the wholesale margin that is earned by the merged entity if sales are diverted from A1 to A2, would in turn provide incentives for the merged entity to increase its retail prices. Also in this setting therefore partial foreclosure of the divested assets can reduce downstream competition compared to the pre-merger counterfactual where two vertically integrated firms compete downstream.
A competition authority would typically seek to address the risk of foreclosure that is present in carve-out remedies by constraining the price at which the divested assets receive inputs that are supplied by the merged entity. For example, the remedies could envisage that the input should be supplied at cost, and at nondiscriminatory conditions. However, setting the correct supply terms to avoid any type of price or non-price foreclosure can be challenging, in particular in dynamic environments where market conditions are expected to change over time. This is discussed in more detail in the next sub-section of the article. Moreover, whilst the determination of cost-competitive terms for the input might be feasible as part of a remedy (and may be contractually enforceable), it may be more difficult to mitigate the risk of quality degradation, especially for complex and specialized inputs. That is, the remedies may effectively guarantee that the input is to be provided at a given price, but may not be able to properly define other nonprice conditions of supply (e.g., quality, delivery conditions, etc.).
The economic literature suggests that if wholesale prices are effectively constrained by regulation (where regulation in this case would be the result of conditions that are contained in the remedies to avoid the risk of price-based foreclosure), then there may be incentives for the merged entity to engage in nonprice foreclosure (i.e., quality degradation or ''sabotage''). 29 To understand the nature of these foreclosure incentives, consider the case where the remedies envisage that access to the input to A1 is to be provided at variable cost, in order to replicate the pre-merger counterfactual. If this were the case, absent quality-degradation, downstream competition would remain as intense as it was in the counterfactual absent the merger, and each downstream firm would earn a share of the industry duopoly profits (with the profits earned by A1 then extracted by the merged entity via the fixed upfront fee). In this scenario, the merged entity may be better off by completely foreclosing the independent downstream firm if possible (e.g., by significantly degrading the quality of the input) and then extracting monopoly profits from its downstream affiliate.
The foreclosure incentives are the stronger when the upstream price regulation is tighter (i.e., the closer is the wholesale price to cost) and when the cost difference between the two downstream competitors is smaller. In extreme situations (e.g., where the remedy is completely unable to constrain the quality of the input to be offered to the divested asset), quality degradation can result in an outcome that is even worse than the merger scenario, as it may lead to foreclosure of the more efficient downstream firm. In this case, the carve-out remedy would therefore be completely ineffective. Under more realistic assumptions whereby the remedy constrains the wholesale input price to cost and is only partially effective in constraining the level of quality, the incentives to engage in quality degradation would persist, since this form of foreclosure would still allow the merged entity to increase its downstream profits.
An alternative, and potentially more effective way to address the risk of foreclosure is to ''reverse'' the carve-out, in a way which makes the merged entity depend on the divested assets for any key input, rather than the other way around (see Fig. 2 below). The idea of a ''reverse carve-out'' is to give the control of any common assets or key input to a party that does not face foreclosure incentives, and that instead should have incentives to supply the input in the most efficient way.
If we use the example of Figs. 1 and 2, under a reverse carve-out arrangement the new owner of assets A1 would also take over ownership of the input U1. Under this structure, the owner of the divested assets would not face incentives to foreclose product B, by worsening the term of supply of the key input U1 that it supplies to it, since it does not compete with product B. The new owner of A1 should instead have the incentive to offer input U1 to product B for a variable fee that is set at variable cost, and also charge a fixed fee that captures the profits that are earned by B. Such an arrangement would maximize the profits that B can capture in its market, and preserve efficiency. 30 An additional benefit of a reverse carve-out solution is that it may help to address any asymmetries of information that are faced by the competition authority during the remedy evaluation stage. In the context of a complex carve-out remedy, a competition authority may not be fully aware of all of the inputs (or access arrangements to shared facilities) that are required by the divested business in order to be a competitive business. A reverse carve-out can ensure that the party that has access to the best available information on what input arrangements need to be made in order to ensure an efficient and workable carve-out (that is, the merged entity), also has the incentives to reveal all of the relevant information in a timely fashion.
A recent example of a case where the Commission accepted a reverse carve-out solution is a merger in the European specialty paper industry that was approved after an in-depth investigation in 2013 (Munksjö/Ahlstrom). 31 The proposed transaction combined Munksjö and some assets of Ahlstrom in ''NewCo'', leading to significant overlaps in two specialty paper markets. As part of the transaction, Ahlstrom B A1 A2

Customers of Products B
Reverse Carveout Remedy

U2
Input contract Divested business U1 Fig. 2 Multi-product merger, and a reverse carve-out remedy 30 In the context of a competitive asset divestment process, the profits that are earned by B as a result of having access to the input U1 at cost would be effectively captured by the new owner of the divested assets A1 as part of the negotiation of the purchase price for the divestment (on the assumption that the terms of the input contract between A1 and B and of the purchase price for asset A1 are jointly negotiated with the merged entity). 31 COMP M. 6576, Munksjö/Ahlstrom, Decision of 24.05.2013. retained a 15 % stake in NewCo, and its shareholders owned a further 50 % of NewCo (on a pro-rata basis). The two overlap products were both manufactured at one Ahlstrom plant (Osnabrück), in addition to one non-overlap product.
The initial remedy that was proposed by the merging parties to address the Commission's competition concern consisted of the sale of the Osnabruck plant back to Ahlstrom, and then the carve out of the overlap products from the plant. The Commission was concerned that Ahlstrom's structural links to NewCo (i.e., the combined impact of its corporate stake and common shareholding structure) would give rise to input foreclosure. The concern was that after the divestment, Ahlstrom may have the ability and incentives to render the divested business less competitive, in order to benefit NewCo.
The remedy that the Commission ultimately accepted to clear the transaction consisted of a ''reverse carve-out'', whereby the entire Osnabrück plant was divested to a new owner, with Ahlstrom retaining ownership of the non-overlap product, whilst receiving services from the new owner for some of the functions that are shared between the overlap and the non-overlap product, and also entering into a joint venture with the new owner for certain essential infrastructure and utilities located at the plant.

Implementation Issues: How to Replicate the Competitive Constraint that is Exercised by a Vertically Integrated Player by Granting Access at Cost
In the case of a complex divestment, even if an effective input contract is put in place to address the risk of foreclosure, the practical question remains of which costs to use as the appropriate benchmark for the access terms. This concern has both a static and dynamic dimension. In terms of the first dimension, a typical concern is that a ''cost-plus'' basis may be relied upon in order to set the access price to be paid to the merged entity, in order to compensate the merged entity for any fixed costs that are involved in the supply of the input, and/or for its cost of capital. Such a cost-plus arrangement may reduce the competitive constraint that is created by the divested business, relative to a counterfactual where each of the merging parties had access to the same input at variable cost. The divestment would therefore introduce a ''double-marginalization'' issue that was not present pre-merger, thus leading to higher downstream prices (in an analogous way to the partial foreclosure scenarios described above).
The solution to this concern is to ensure that the input supply contract with the divested business replicates as closely as possible the conditions that were enjoyed by the vertically integrated firm pre-merger. This means in practice that the input should be provided at (variable) cost. Any rents that are earned by the divested business as a result of the possibility of obtaining the input at cost should be reflected in the fixed fee or purchase price that is paid by the acquirer of the divested assets (as long as the divestment process is reasonably competitive). 32 32 Determining that access should be provided at cost does not of course address the risk of non-price discrimination (e.g. in the form of quality degradation), as discussed in the previous sub-section.
A potentially more challenging issue is how to deal with non-price considerations, relating in particular to future technological developments. The concern in this case is that whilst an input supply contract may be reasonably competitive in the short-term, it may become uncompetitive over time, as a result of rapid technology developments (e.g., the introduction of new technologies that result in lower cost and/or significantly higher capacity). A vertically integrated firm with access to the underlying technology would be able to benefit from any ongoing technological development, and thus remain competitive over time. In the context of a complex carve-out, this is more difficult to achieve, since it is typically difficult to craft remedies that are ''future-proof'' and that allow a divested business to remain competitive over time.
The Commission has confronted some of the issues that relate to cost structure and ''future-proofness'' in some recent significant mergers.
In the INEOS/Solvay merger (cleared with remedies in 2014), 33 the Commission rejected one of the plants that were offered by the merging parties as a divestment because of its uncompetitive cost structure. The merger concerned the market for S-PVC: a product that is part of a complex vertical value chain that includes the production of an upstream product (VCM), and the production of a by-product (caustic soda). The plant that was offered for divestment by the merging parties was ''virtually'' integrated into the upstream business via a contractual relationship with a 3rd party, which included a price formula for the determination of the price for VCM.
The Commission undertook a detailed economic analysis of the functioning of the VCM price formula to assess the level and nature of the variable cost of the plant. The merging parties had argued that the price formula fully replicated the cost structure of a vertically integrated chain and thus the pricing incentives of a vertically integrated producer. 34 The Commission undertook a detailed analysis of the price formula that showed that this was not the case, and therefore that the divestment of the plant could not replicate the constraint exercised by the smaller of the two merging parties (Solvay) pre-merger.
The Commission also analyzed an amendment to the price formula that was proposed by the merging parties as part of an alternative remedy package. Also for this case, the Commission found that the dependence on the VCM price formula undermined the competitiveness of the divested assets, also in a forward-looking perspective.
The remedy that was finally accepted in Ineos/Solvay included the divestment of three vertically integrated S-PVC plants, and some of the related upstream assets. However, some complex carve-out arrangements were also necessary for these plants. For example, the divestment of some of the S-PVC assets included the reverse carve-out of a non-overlap product that was also produced by these assets (caustic potash).
The Commission concluded that whilst such a carve-out did not undermine the competiveness of the divested business, it reduced the financial attractiveness of the package (and hence potentially its viability), and also increased its complexity. This consideration was taken into account in the determination of some of the additional safeguards that were introduced in the remedy package (in particular the need for an upfront buyer with proven industrial expertise). Moreover, in the remedies implementation phase, a portion of the caustic potash business was added back to the divested business, to address the Commission's concerns about the implementation of the remedy (thus increasing the financial strength of the package).
Another case where the Commission has had to assess in detail whether and how a complex divestment would have affected the competitiveness of the divested package is the Irish mobile telephony merger (H3G/Telefonica Ireland) 35 that was cleared in 2014 (similar issues were raised in the German mobile telephony merger, Telefonica Germany/E-plus, also approved in 2014). This case concerned the merger between the second and fourth largest mobile operator in Ireland, reducing the number of firms from four to three. The Commission concluded that the transaction would have led to a significant reduction in competition.
The remedy that was accepted to clear the transaction consisted primarily of a long-term ''capacity-based'' Mobile Virtual Network Operator (MVNO) contract to be divested by the merged entity to two new entrants. The contract was not structured as a traditional ''pay-as-you-go'' MVNO contract given that the Commission was concerned that a contract of this type would not replicate the competitive constraint exercised by a Mobile Network Operator (MNO) that has access to network assets at (marginal) cost. Under the capacity-based model, the MVNO purchases access to a given level of network capacity in exchange for a predetermined lump-sum payment, without incurring any additional per-unit cost as function of usage. This model approaches the cost structure of an MNO, at least in the short term.
The Commission did not determine the level of the annual lump-sum payment to be made by the MVNO entrant, and allowed this to be set as part of commercial negotiations between the merged entity and the new entrant (as in any traditional asset divestment). However the remedy stipulated that the merged entity had to enter into two capacity-based MVNO agreements (each with a given specified minimum level of capacity), thus ensuring that the pre-defined level of network capacity would need to be sold by the merged entity, for a fixed price that was acceptable by the two new entrants (based on market conditions).
This design has the benefit that the competition authority does not need to determine, as part of the remedy evaluation, the level of the access price that is required to make an entrant cost-competitive. Under the capacity-based MNVO, the entrants essentially obtain access to capacity at no marginal cost, but instead pay an upfront fixed fee that does not distort their pricing incentives.
Whilst a well-designed capacity-based MVNO model may succeed in replicating competitive conditions in the short-run, its main challenges related to dynamic issues (''future-proofness''). In the context of the mobile telephony industry, some of these issues relate in particular to the setting of the appropriate level of capacity to be made available over time, in a context where capacity is growing rapidly with the shift to more intensive data usage and the adoption of new technologies. 36 A related challenge is how to make a capacity-based MVNO contract also effective at the time of contract renewal, given the difficulties with inducing a would-be MVNO entrant to sign a long-term capacity-based agreement in a situation where the market is expected to experience significant change, and where the remedy does not envisage the transfer of the underlying assets. 37 Whilst there are some contractual provisions that were included in the Irish MVNO remedy that are aimed at addressing some of the ''future-proofness'' issues discussed above, it is also clear that these concerns are particularly difficult to remedy in industries that are expected to experience significant technological change. 38 These issues are also significantly more difficult to address than the static pricing concerns that are associated with traditional access remedies.

Relevance of Information on the Level of Buyer Interest
A third issue that in our experience is often raised in the context of complex divestment packages is the relevance that should be attributed to the level of interest that is expressed by potential buyers of a given divestment package. 39 Consider a situation where a competition authority may have concerns that a given divestment may not be cost-competitive and therefore may not replicate the competitive constraint exercised on each other by the merging parties. In such a case, merging parties often make the argument that such concerns are not warranted if there are buyers who have expressed an interest for the divestment package, and therefore consider that the divestment can be competitive in the market place.
From an economic perspective, the presence of interested buyers should be seen as a necessary condition for an effective divestment, but not as a sufficient one. It is a necessary condition because clearly if there are no potential purchasers interested in the package, it is unlikely that an effective divestment can be successfully implemented. The lack of buyer interest is also a strong indication that the package may not be competitive.
However, the fact that some buyers may be interested in purchasing a given package does not imply that the divestment is sufficient to compensate for the loss of competition from the merger. Consider again the carve-out example shown in 36 The concern here is therefore not about setting the right price for the capacity that is made available, but rather to ensure that the level of capacity that is offered is sufficient over time, in order to avoid binding capacity constraints (which would lower the incentives to compete of the capacity-based MVNO). 37 A more traditional concern that is often associated with quasi-structural access remedies is how to ensure that the new entrant can effectively differentiate its offers from the merged entity, in particular in terms of non-price dimensions. 38 It is also relevant to note that the structural elements that have been included in the remedies for some of the recent mobile mergers (e.g. H3G/Orange Austria, and Telefonica Germany/E-plus) have not resulted to-date in structural entry. 39 This issue was encountered in a number of recent Phase II merger assessments, including the mobile telephony mergers that were cleared in 2014 and Ineos/Solvay. Fig. 1, and assume that due to the presence of foreclosure incentives the input that is provided to the divested business A1 is priced above variable cost (in order to reduce the competitive pressure that is exercised by A1 on the merged entity). The higher cost that is faced by divested A1 reduces the value of the business relative to a counterfactual where the input is priced at cost. However, this does not mean that the value of the business is zero or negative. Therefore, as long as potential buyers are compensated for the prospect of partial foreclosure through a lower payment for the divested business (as one would expect in a competitive sale process), then some level of buyer interest may still be present. 40 This would be entirely compatible with a finding that the remedy does not replicate the competitive constraint lost through the merger. This reasoning implies that the observation that some buyers may be interested in a divestment package is not sufficient to show that the remedy addresses a competition concern.