Jump to Content Jump to Main Navigation
Big Data and Competition Policy by Stucke, Maurice; Grunes, Allen (1st June 2016)

Part IV What are the Risks if Competition Authorities Ignore or Downplay Big Data?, 15 Risk of Inadequate Merger Enforcement

Maurice E. Stucke, Allen P. Grunes

From: Big Data and Competition Policy

Maurice Stucke, Allen Grunes

From: Oxford Competition Law (http://oxcat.ouplaw.com). (c) Oxford University Press, 2015. All Rights Reserved. Subscriber: null; date: 23 October 2019

Subject(s):
Market power — Rights — Internet — Technology — National merger control

(p. 217) 15  Risk of Inadequate Merger Enforcement

15.01  We saw in Chapter 6 several shortcomings in the way the competition agencies have analysed data-driven mergers, and how they have not fully considered the implications of a data-driven economy. Some officials, we saw, believe that ‘the antitrust principles are the same’ for data-driven mergers and other conventional mergers.1 This chapter identifies four significant risks if the agencies continue along the same path. First, it is questionable whether the competition agencies are accurately predicting many mergers’ competitive effects in concentrated markets. The second risk is when the competition agencies consider only the merger’s impact on the ‘paid’ side of multi-sided platforms and ignore the free side. To illustrate, the chapter uses the merger wave of US commercial radio stations in the 1990s–2000s, where the Department of Justice (DOJ) considered in its consent decrees only the mergers’ effects on advertising, and not on listenership. The third risk is that even if the agency reviews the merger’s likely effect on each side of the multi-sided platform, their price-centric tools are ill suited for evaluating the free side. Finally, the network effects, we discussed in the preceding chapters, can magnify the likely consumer harm.

A.  The Prediction Business

15.02  For mergers, competition agencies are often in the prediction business. In many jurisdictions, like the US and EU, the merger, if it exceeds certain thresholds, cannot occur unless the merging parties notify the competition authority, and the waiting period passes (or is terminated early by the competition authority).2 The pre-merger (p. 218) notification regime is designed to strengthen antitrust enforcement. It gives the agency time to investigate large mergers before they are consummated and to preserve an effective remedy.3

15.03  With the pre-merger notification regime, the competition agency predicts the merger’s likely effects. In assessing whether the merger may substantially lessen competition or tend to create a monopoly, the competition authority does not have the benefit of economic realities—namely, to allow the merger to occur, assess its competitive impact, and challenge the merger if it proves to be anticompetitive. The competition agency can always challenge a consummated merger. But as the Hart–Scott–Rodino Antitrust Improvements Act’s (HSR) legislative history reflects, the remedy will often be ineffectual. By the time the agency assesses the merger’s competitive effects several years later, among other things, production facilities will likely be closed, employees will likely be fired (or have left), and brands will likely be discontinued or repositioned. Moreover, the company, post-merger, knowing that its behaviour is under review may wait a couple of years longer before exercising its market power.

15.04  So the competition authority does not have the luxury of a wait-and-see approach. The closest thing pre-merger is an analogous historical event or ‘natural experiment’. Here agencies ‘examine the impact of recent mergers, entry, expansion, or exit in the relevant market’ or ‘similar’ markets to assess the merger’s likely impact.4

15.05  Absent natural experiments and consummated mergers, the agencies typically rely on economic theory to predict a merger’s likely competitive effects. This entails predicting several likelihoods, including the ‘likelihood that a merger would have anti-competitive effects in the relevant markets, in the absence of countervailing factors’, the ‘likelihood that buyer power would act as a countervailing factor to an increase in market power resulting from the merger’, the ‘likelihood that entry would maintain effective competition in the relevant markets’, and the ‘likelihood that efficiencies would act as a factor counteracting the harmful effects on competition which might otherwise result from the merger’.5(p. 219)

B.  Most Mergers are Cleared

15.06  The agencies do not have the luxury to review mergers at a leisurely pace. They often are under time constraints. So they must predict quickly (in antitrust time)—typically within six to nine months for the mergers they review.

15.07  The US and EU clear most mergers. For example, between 1990, when its Merger Regulation came into force, and 2012, the European Commission cleared over 4,600 deals and blocked only 22 mergers.6 Or, as Commissioner Joaquín Almunia exclaimed, ‘Fewer than five in every thousand cases!’7

15.08  In 2011, the Commission received 309 notifications, and blocked only one merger.8 Over 90 per cent of its merger cases are typically resolved early on (in its Phase I), generally without remedies.9

15.09  Likewise, the DOJ investigates few mergers. As Figure 15.1 reflects, between 1990 and 2010 the DOJ investigated, on average, 4.4 per cent of all HSR filings. It issued (p. 220)

Figure 15.1  Requests for HSR filings, 1990–2010

Source: DOJ, Antitrust Division Workload Statistics FY 1990–2010, http://www.justice.gov/atr/division-operations.

on average a Second Request for 1.74 per cent of all HSR filings, and challenged 0.37 per cent of all HSR filings.

15.10  The Federal Trade Commission (FTC) challenges roughly the same percentage of mergers.10 Consequently, of the many pre-merger notifications they receive annually, the EU and US competition agencies investigate a small percentage of mergers; an even smaller percentage reach the Second Request or Phase II stage. Even fewer mergers are challenged. And of those few mergers challenged, rarely do the merging parties go to trial; most settle with divestitures.

C.  The Big Mystery: How Often Do the Competition Agencies Accurately Predict the Mergers’ Competitive Effects?

15.11  One cannot fault the competition agencies for allowing so many mergers to sail through unchallenged (often without any scrutiny) if their analytical tools accurately screen anticompetitive mergers from the benign and procompetitive mergers. After all, in assessing the merger’s likely competitive effects, the competition agencies, one prior European Commissioner noted, rely on ‘sophisticated qualitative and quantitative analyses’.11

15.12  So how often do the agencies predict the mergers’ competitive effects accurately? No one really knows. Ask any competition authority, and they cannot provide their batting average, namely the percentage of mergers where they accurately predicted the likely competitive effects. Competition agencies are often evaluated on how quickly they assess mergers, the predictability of their review, and the cost imposed on firms. The agencies are not assessed on how often they accurately predict the mergers’ likely competitive effects.

15.13  What we do know is that over the past 45 years the US competition agencies have increased the concentration thresholds, permitting more and more mergers in concentrated industries.12 Originally, the DOJ consistent with the Clayton Act, applied ‘an additional, stricter standard in determining whether to challenge mergers occurring in any market, not wholly unconcentrated, in which there is a significant (p. 221)

Table 15.1  Concentration levels of mergers challenged by the US competition agencies, 1999–2003

Change in the HHI

0–99

100–199

200–299

300–499

500–799

800–1,199

1,200–2,499

2,500+

total

Post-Merger HHI

0–1,799

0

17

18

19

3

0

0

0

57

1,799– 1,999

0

7

5

14

14

0

0

0

40

2,000–2,399

1

1

7

32

35

2

0

0

78

2,400–2,999

1

5

6

18

132

34

1

0

197

3,000–3,999

0

3

4

16

37

63

53

0

176

4,000–4,999

0

1

3

16

34

30

79

0

163

5,000–6,999

0

2

4

16

9

14

173

52

270

7,000+

0

0

0

2

3

10

44

223

282

Total

2

36

47

133

267

153

350

275

1,263

Source: FTC and DOJ, Merger Challenges Data, Fiscal Years 1999–2003, 18 December 2003, http://www.justice.gov/atr/merger-challenges-data-fiscal-years-1999-2003.

trend towards increased concentration’.13 The Reagan administration dropped that, mentioning only ‘the congressional intent that merger enforcement should interdict competitive problems in their incipiency’.14 The US competition agencies, in the ensuing years, challenged mergers mostly in highly concentrated industries, as reflected in the Herfindahl–Hirschman Index (HHI).15 Table 15.1 gives a snapshot showing the concentration levels of mergers that the agencies challenged between fiscal years 1999–2003.16

By 2010, to reflect their enforcement activity, the US competition agencies pushed the permissible concentration levels higher.17

15.14  The agencies’ increasing tolerance for concentrated industries post-merger is not due to empirical analysis. Although the HHI figures appear precise, there is no (p. 222) scientific basis for the agencies’ cut-offs. Neither the FTC nor DOJ calibrated the likelihood of anticompetitive effects with actual post-merger risks. They do not know whether, and the extent to which, mergers with post-merger HHIs below 2,000 are less likely to cause harm than those below 3,000.

15.15  How can we state this so confidently? Because the competition agencies in the US and elsewhere generally do not revisit the mergers in either highly or moderately concentrated industries to see whether their predictions were accurate.18 In one survey, 43 per cent of the competition agencies said that they had not done in the prior five years any ex-post evaluations for any competition interventions.19 Outside hospital and petroleum mergers, the US competition agencies typically do not revisit the industries post-merger to assess whether they predicted correctly.20 The European Commission does ex-post reviews for only ‘a small number of cases’.21 Most notable was its 2005 merger remedies study.22

15.16  So why don’t the agencies revisit mergers to see if they predicted correctly (especially when they permit many more mergers in moderately or highly concentrated industries)? Cost is one reason. The European Commission said ‘ex post evaluations can require substantial econometric expertise and detailed datasets, which render them costly’.23 The FTC said it was ‘limited by agency resource constraints and the costly, time consuming nature of such research’.24 It is interesting that it is easier for the agencies to confidently predict the merger’s likely competitive effects (and to increase the concentration thresholds) than to see if their predictions are accurate. Granted it may be hard to compare the post-merger world with a competitive counterfactual,25 but it is not impossible.

(p. 223) D.  The Ex-Post Merger Reviews Paint a Bleak Picture

15.17  In 2013 Professor John Kwoka collected all the recent post-merger reviews. The available post-merger reviews, he concluded, suggest that the US competition agencies are inadequately enforcing the competition laws. Of the 53 post-merger reviews with price estimates in 16 different industries, ‘40 or 75.5 per cent report postmerger price increases’.26 Of the mergers for which data on the agency’s actions were available, the agency opposed five mergers, cleared eight mergers, and obtained remedies in ten mergers.27 As Kwoka concluded, ‘[c]‌ollectively, these results suggest that merger control in these studied cases may overall be too permissive, that the remedies chosen may be inadequate to the task of preserving competition, and that conduct and conditions remedies may be especially ineffective.’28

15.18  Granted, the finding of a price increase in one case does not necessarily mean that the agency’s predictive tools are inadequate. There might have been intervening, unforeseen events (such as a competitor’s plant closing due to weather, changes in consumers’ tastes) that caused the price increase.29 But it is questionable that 75 per cent of the industries studied involved natural disasters or other unforeseen events. Or perhaps the old management did not price its goods to maximize profits, negotiated poorly, or otherwise left money on the table, while the new management understood the market better and was more capable. Again, it is unlikely that this explains the price increases Kwoka found. It is absurd to think that all the acquired firms are amateurish while all acquiring firms are professional. Still, as Kwoka cautioned, ‘the number of observations is not especially large, classifications are sometimes difficult, the data have other limitations, and selection issues abound’.30 Perhaps as more post-merger reviews are undertaken, we will obtain a clearer picture of markets or mergers where the agencies’ tools work well and where they do not.

15.19  Based on what we do know, three out of four mergers for which we have post-merger analysis led to price increases. A 0.250 batting average may suffice for a mediocre (p. 224) minor league baseball team, but not for mergers that affect trillions of dollars of commerce. It may be that selection bias abounds—namely the 53 post-merger studies involved the more contentious, problematic mergers. Perhaps the agencies predicted accurately for most other mergers. But as the next chapter explores, that is unlikely. If Kwoka’s survey of post-merger reviews even approximates the agencies’ track record (a big if, we concede), then the agencies and courts are simply not enforcing the competition laws well. If the agencies’ price-centric analytical tools are often wrong (at least for the industries covered in the merger retrospectives), then there is no reason to assume that these price-centric tools somehow work well in assessing data-driven mergers in multi-sided markets, where the products on one side are free.

E.  The High Error Costs When the Agencies Examine Only One Side of a Multi-Sided Platform

15.20  The risk of inadequate merger enforcement increases when the competition agencies focus only on one side (eg, advertising) of a multi-sided platform. For example, the European Commission in broadcast TV cases ‘considered only the market for advertising’ and ‘did not consider the “free” TV side to be a market’.31 Likewise, as we shall see, the DOJ considers only the advertising side in commercial radio mergers.

15.21  It is unclear why any competition agency today would consider only one side of a multi-sided platform. Perhaps the assumption is that the anticompetitive effects are linked: namely, the merger, in lessening competition on the paid side, must also lessen competition on the free side. This, at times, is true. A newsweekly that monopolizes the advertising market may also lack meaningful competition on the free side. A search engine may dominate both the free and advertising sides.

15.22  The fallacy is to assume that if anticompetitive effects are unlikely on the paid side (eg, unlikely that the merged entity could raise advertising rates), then anticompetitive effects are also unlikely on the free side. One assumption is that given the spill-over network effects, the platform to attract advertisers may ‘subsidize’ the other side, by offering the product for free. Thus, the platform has little incentive to exert market power on the free side. This may be true if one platform competes against other platforms offering similar products, which users view as substitutes, and users frequently switch among the platforms. But this is not always true.

(p. 225) 15.23  We can see this with newspapers, magazines, commercial television, and commercial radio. They compete for an audience to read their articles, and listen and watch their programmes, which often are offered for free or a nominal subscriber fee, and for advertisers that seek to target this audience. But the advertising competition and advertising rates do not necessarily capture the editorial competition. Today a free newspaper in the US likely competes against Google and Craigslist for advertising revenue. But Google, while aggregating the news, does not produce in-depth local journalism. Craigslist does not gather the news. Google and Craigslist, while formidable competitors to the traditional media for certain advertisers, are not competitive threats for newsgathering. If Craigslist entered a local geographic market, the increase in classified advertising competition will not translate into better local journalism. Thus, competition on the paid advertising side may lower the newspapers’ advertising rates, but it will not necessarily improve the news content’s quality on the free side. Likewise, if two competing newsweeklies merge, readers will likely be worse off with fewer choices for local investigative journalism, while classified advertisers, given online alternatives, may be less affected.

15.24  Thus a merger can lessen non-price competition on the free side (eg, editorial competition and diversity of viewpoints needed for an effective democracy), without lessening competition on the paid side, and vice versa.

15.25  The emerging consensus is that competition authorities must consider the merger’s impact on each side of a multi-sided platform. For mergers in data-driven industries, the Organisation for Economic Co-operation and Development (OECD) warned, ‘focussing on one side of market will rarely lead to a proper market definition’.32 It also cautioned that ‘[m]‌echanical market definition exercises that exclude one side usually lead to errors.’33 At a minimum, the competition authority must understand the linkages and interrelationships among the sides.34 The fact that assessing the merger’s impact on one side of the platform is more difficult, the OECD counselled, does not warrant ignoring it:

The difficulties inherent in measuring diversity and quality, and in predicting how a merger might affect them is not a good justification for ignoring them in merger review. This is especially true if there is little or no sector regulation dealing with these matters.

The non-price effects of media mergers can have just as significant an impact on economic welfare in media markets as they have in other markets. It is (p. 226) quality-adjusted changes in prices that are pertinent for determining effects on economic welfare.

There could be jurisdictions where diversity and quality are so tightly regulated, through detailed content rules for instance, that competition authorities can justifiably assume that media mergers will leave diversity and quality unchanged or even improved. Absent such regulation, there is no a priori reason why the diversity and quality effects of media mergers should not be examined by competition authorities when diversity and quality changes have a direct effect on economic welfare.35

15.26  In evaluating commercial radio mergers, however, the DOJ erred in considering the merger’s impact solely on the paid side (namely advertisers and the rates they paid), and did not account the mergers’ impact on the free side (namely non-price competition regarding programming quality, listener choice, or the likely impact of these mergers on the marketplace of ideas).36 As a result, both consumers and advertisers have paid the price.

15.27  Nothing in the Clayton Act restricts the DOJ from considering solely advertising competition. For other media markets, like free or paid newspapers, the DOJ for decades has examined both sides of the platform.37 Nor can one claim that the listenership side is best left to the Federal Communications Commission (FCC). This is contrary to Congress’s intent in enacting the Telecommunications Act of 1996.38 And, we discuss elsewhere, the FCC’s ‘public interest’ standard and restrictions on common ownership do not adequately address a media merger’s impact on the marketplace of ideas.39

(p. 227) 15.28  So what happened after the 1996 Telecommunications Act, which weakened ownership limits on radio stations nationally and locally,40 and the DOJ’s examining only the advertising side of the two-sided platform? There was not surprisingly a merger wave, and radio ownership became significantly more concentrated. Between March 1996 and March 2007, the number of commercial radio stations increased 6.8 per cent, but the number of radio owners declined by 39 per cent.41 This trend was already apparent in 2001, by which time the number of radio owners had already declined 25 per cent from when the 1996 Act commenced.42 Over the same period and until 2007, the nation’s largest radio group owners grew even bigger: ‘In 1996, the two largest radio group owners controlled 62 and 53 stations, respectively. By March 2007, the leading radio group, Clear Channel Communications, owned over 1,100 radio stations.’43

15.29  The ensuing merger wave, not surprisingly, had an adverse impact on non-price competition, including on programming quality and programming choices for listeners.44 One complaint, the Project for Excellence in Journalism reported, was ‘that Clear Channel’s domination was diminishing the quality of the AM/FM radio dial by monopolizing key markets and homogenizing content’.45 A frequent complaint was that the deregulation ‘allowed for unprecedented consolidation in commercial radio, which has resulted in a homogeneity that is often out-of-step with artists, entrepreneurs, media professionals and educators—not to mention listeners’.46 An FCC Commissioner observed how ‘[r]‌espected media watchers (p. 228) argue that this concentration has led to far less coverage of news and public interest programming’, how one multiyear study found a homogenization of music that got air play, and how radio served more ‘to advertise the products of vertically integrated conglomerates than to entertain Americans with the best and most original programming’.47 Mel Karmazin, the former head of commercial radio for Infinity Broadcasting and CBS and former CEO of Sirius XM, recognized that commercial radio after the 1996 Act became ‘totally homogenized’.48 Karmazin advocated for radio consolidation ‘[s]trictly for business reasons. No one asked [him] if it was good for consumers’.49

15.30  Sometimes, prices increases and quality degradation on one side of the platform are offset with discounts and quality improvements on the platform’s other side. So to ‘evaluate the market power of a platform one has to look at the markups on both sides’.50 Here consolidation, brought by ineffective antitrust enforcement, harmed both radio listeners and advertisers.

15.31  As a consequence of the radio merger wave, the largest firms often dominated the market in terms of audience and revenue share. By 2012, the largest commercial radio firms, as the FCC found, ‘enjoy[ed] substantial advantages in revenue share—on average, the largest firm in each Arbitron Metro market ha[d]‌ a 45 per cent share of the market’s total radio advertising revenue, with the largest two firms accounting for 73 per cent of the revenue’.51 In over a third of all Arbitron Metro markets, ‘the top two commercial station owners control[led] at least 80 per cent of the radio advertising revenue’.52 Not surprisingly, radio advertising rates nearly doubled during this period of consolidation53—suggesting that even on this advertising dimension, the DOJ’s antitrust review may have been inadequate. So anticompetitive effects were seen on both sides of many commercial radio markets.

(p. 229) F.  How Data-Driven Mergers Increase the Risks of False Negatives

15.32  If the competition agencies ignore or downplay data-driven mergers, then they will allow mergers that could significantly harm consumers.

15.33  First, as Kwoka and others found, it is questionable whether the competition agencies are generally reaching the right outcome in assessing the mergers’ likely effect on prices in concentrated markets.54 The belief is that the competition authorities have more sophisticated tools for assessing mergers’ price effects (in particular unilateral price effects). If the competition authorities’ batting average in predicting price effects is a mediocre 0.250, then their batting average in predicting the data-driven mergers’ effects on non-price competition is not likely to be any better.

15.34  Second, even when the agencies accurately predict the merger’s price effects on the advertising side of a multi-sided platform, there remains the significant risk of false negatives when they ignore the platform’s other sides (such as the harm US radio listeners suffered when the merger wave degraded quality).

15.35  Third, even if the agencies review the merger’s likely effect on each side of the multi-sided platform, their price-centric tools are ill suited for free services. The agencies, like the European Commission, already have a hard time assessing a merger’s net effect in multi-sided platforms where each side has a positive price:

The complexity primarily arises from the presence of two (or more) unique, but interdependent, classes of agents or customers. The analysis needs to account for (1) the responses of two (or more) distinct sets of agents to platform owners (2) platform owners responses to two sets of agents, and (3) the responses of one set of agents to changes in the others’ behaviour and vice versa—particularly as demand conditions change on each side. This pattern of cross-responses will generally affect each step of standard antitrust analysis, from product market definition, the competitive assessment, entry, efficiencies, etc. However, as argued in this contribution, this does not imply a need to abandon the typical tools that one applies in the analysis of single-sided markets, only to adapt them.55

15.36  It will be even harder for the agency with its price-centric tools to assess (i) the effect of a price increase (or decrease) on the paid side of the platform on the quality of (and (p. 230) demand for) the ‘free’ side; (ii) the effect of quality degradation on the free side on the demand for the paid side; and (iii) the effect either has on the platform’s overall profitability.56

15.37  Say, for example, the Facebook/WhatsApp merger benefits advertisers in enabling Facebook to amass even more data on us to better target us with behavioural ads. But many of us who do not like being tracked are worse off, when there is no viable alternative social platform/texting app. Suppose the merger increases advertisers’ welfare by USD 100 million (in lowering the costs to target us with behavioural ads), but significantly reduces our privacy protections. Suppose the agency cannot quantify the consumer harm. Whose interest should prevail: the advertisers’ or ours? How does the competition authority assess the trade-off? One concern is that when it is easier to assess and quantify the benefits or harm to advertisers, that trumps any privacy harms to individuals that are more difficult to quantify.

15.38  With the Facebook/WhatsApp example, the agency can at least inquire how users of texting apps would react to the degradation in quality of the texting app. The problem in the Alliance/Conversant merger is that individuals do not always interact with the merging parties. We may not even know how, and the extent to which, they are tracking us and using our personal data. So do we count? The competition authorities often proclaim how antitrust should promote consumer welfare. Thus the competition agency should be especially vigilant regarding mergers like Alliance/Conversant, as we may not even know of the merging companies, the data they collect on us, and how the merger will affect our welfare.

15.39  Finally, many single-sided markets are not characterized with network effects. But with data-driven mergers, network effects can arise with a vengeance. One may see the traditional direct and spill-over network effects, amplified by the data-driven network effects (learning-by-doing and scope of data). Where such substantial economies of scale exist, the European Commission noted, ‘the typical market structure is likely to consist of a few large firms each with significant market power’.57 So the data-driven network effects increase the cost of false negatives. With these market realities, firms will strive to tip the market to their advantage, including through mergers. Thus the harm when the agencies get it wrong is not just market power, but monopoly power, which market forces will not quickly correct.

15.40  Consequently if the competition authorities continue to assess a merger’s price effects (where, if the post-merger reviews are a guide, their batting average is 0.250) (p. 231) because they lack a solid analytical framework for evaluating ‘free’ products and services in multi-sided platforms, and if data-driven mergers often fall outside the agencies’ pre-existing horizontal, vertical, and conglomerate merger categories, many data-driven mergers will continue to pass through without any significant scrutiny. For the anticompetitive ones, consumers will bear the brunt, which, as the next chapter discusses, can be severe.

Footnotes:

1  Lisa Kimmel and Janis Kestenbaum, ‘What’s Up with WhatsApp?: A Transatlantic View on Privacy and Merger Enforcement in Digital Markets’, 29(1) Antitrust (Fall 2014): p 55 n 63.

2  See eg European Commission, Merger Control Procedures, 13 August 2013, http://ec.europa.eu/competition/mergers/procedures_en.html (must receive notice of any merger ‘with an EU dimension’, where the merging firms ‘reach certain turnover thresholds’); Hart–Scott–Rodino Antitrust Improvements Act of 1976, 15 USC s 18a.

3  United States v Computer Associates Int’l, Inc, Case No 01-02062(GK), 2002 WL 31961456 (US Dist Ct (D DC), 20 November 2002).

4  US Department of Justice (DOJ) and Federal Trade Commission (FTC), Horizontal Merger Guidelines, 19 August 2010, s 2.1.2 (‘US Horizontal Merger Guidelines’), http://www.justice.gov/atr/public/guidelines/hmg-2010.html.

5  European Commission, Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations Between Undertakings [2004] OJ C 31/03, para 11 (‘EC Horizontal Merger Guidelines’), http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52004XC0205%2802%29.

6  Joaquín Almunia, Vice President of the European Commission responsible for Competition Policy, ‘Merger Review: Past Evolution and Future Prospects’, Speech given at Conference on Competition Policy, Law and Economics, Cernobbio, Italy, 2 November 2012, http://europa.eu/rapid/press-release_SPEECH-12-773_en.htm.

7  Ibid.

8  Ibid.

9  European Commission, Merger Control Procedures, above note 2. A Phase I review may involve ‘[r]‌equests for information from the merging companies or third parties’ and ‘[q]uestionnaires to competitors or customers seeking their views on the merger, as well as other contacts with market participants, aimed at clarifying the conditions for competition in a given market or the role of the merged companies in that market.’ Ibid.

10  FTC Competition Enforcement Database, Merger Enforcement Actions, http://www.ftc.gov/competition-enforcement-database (reporting 351 merger actions (which includes both HSR and non-HSR transactions, and mergers that were abandoned during the review process) between 1996 and 2011).

11  Almunia, above note 6.

12  Compare DOJ, 1968 Merger Guidelines (1968), s 6 (‘US 1968 Merger Guidelines’), http://www.justice.gov/sites/default/files/atr/legacy/2007/07/11/11247.pdf with DOJ, 1982 Merger Guidelines (1982) (‘US 1982 Merger Guidelines’), http://www.justice.gov/sites/default/files/atr/legacy/2007/07/11/11248.pdf.

13  US 1968 Merger Guidelines, above note 12, s 7.

14  US 1982 Merger Guidelines, above note 12, s 1.

15  The HHI is calculated by summing the squares of the individual market shares of every firm in the relevant market. So the HHI of a market equally divided among ten firms would be 1,000.

16  FTC and DOJ, Merger Challenges Data, Fiscal Years 1999–2003, 18 December 2003, http://www.justice.gov/atr/merger-challenges-data-fiscal-years-1999-2003.

17  US Horizontal Merger Guidelines, above note 4, s 5.3.

18  Tomaso Duso and Peter Ormosi, ‘Capacity Building Workshop on the Ex-Post Evaluation Of Competition Authorities’ Enforcement Decisions: A Critical Discussion’, Prepared for the OECD, Directorate for Financial and Enterprise Affairs Competition Committee, Working Party No 2 on Competition and Regulation, 26 October 2015, p 2, http://www.oecd.org/officialdocuments/ publicdisplaydocumentpdf/?cote=DAF/COMP/WP2%282015%298&doclanguage=en (‘quite a few of the over 110 competition authorities around the world are yet making use’ of ex-post reviews).

19  International Competition Network, Advocacy Working Group, Interim Report on the Explaining the Benefits of Competition Project, April 2012, p 56, http://www.internationalcompetitionnetwork.org/uploads/2011-2012/interim_benefits%20project%20report%20-%20final.pdf.

20  Daniel Hosken, Deputy Assistant Director, Bureau of Economics, US Federal Trade Commission, ‘Ex-Post Evaluation in the US: Lessons Learned’, http://www.oecd.org/daf/competition/workshop-expost-evaluation-competition-enforcement-decisions.htm (‘Since 1983, the FTC Economists have conducted more than 20 ex-post evaluations.’).

21  Ibid, pp 58 and 64 (‘owing to the costs involved, ex post evaluations cannot be systematic; they should only cover a select number of cases’).

22  DG Competition, European Commission, Merger Remedies Study, October 2005, http://ec.europa.eu/competition/mergers/legislation/remedies_study.pdf.

23  Hosken, above note 20, p 58.

24  Ibid.

25  The most widely used method has been difference-in-differences, where one compares ‘the change in prices in a market affected by the analysed practice (cartel, merger or abuse of dominance) with the change in prices for the same period in a different market unaffected by the analysed practice.’ Duso and Ormosi, above note 18, p 10.

26  John E Kwoka, Jr, ‘Does Merger Control Work? A Retrospective on US Enforcement Actions and Merger Outcomes’, 78 Antitrust LJ (2013): pp 619, 631–2 (the average increase was 9.40%, ranging from a 0.06% up to a high of 28.4%; with 13 transactions (24.5% of the total) found to result in price decreases, which average 4.29% and range from 0.04% to 16.3% in absolute value; the survey included three joint ventures and four airline code-shares; of the three joint ventures ‘(all in petroleum), two reported price increases while one reported a decrease. The magnitudes are all small, and average a positive 0.43 per cent. On the other hand, all four of the studied airline code-sharing arrangements are found to have resulted in price decreases.’).

27  Ibid, p 638.

28  Ibid, p 641.

29  Duso and Ormosi, above note 18, p 15.

30  Kwoka, above note 26, p 644.

31  Teresa Vecchi, Jerome Vidal, and Viveca Fallenius, ‘The Microsoft/Yahoo! Search Business Case’, 2 Competition Policy Newsletter (2010): p 45, http://ec.europa.eu/competition/publications/cpn/2010_2_8.pdf.

32  OECD, Data-Driven Innovation for Growth and Well-being: Interim Synthesis Report, October 2014, p 59, http://www.oecd.org/sti/inno/data-driven-innovation-interim-synthesis.pdf (noting that in multi-sided markets, as enabled by data, one cannot focus only on one side of the market, since one side will often exert an externality on the other side).

33  OECD, Policy Roundtables: Two-Sided Markets, 17 December 2009, p 11, http://www.oecd.org/daf/competition/44445730.pdf.

34  Ibid.

35  OECD, Policy Roundtables: Media Mergers, 19 September 2003, p 9, http://www.oecd.org/competition/mergers/17372985.pdf.

36  See, eg, United States v Entercom Communications Corp, Case No 1:15-cv-01119 (US Dist Ct (D DC), 14 July 2015), Complaint (alleging that acquisition likely would substantially lessen competition for the sale of radio advertising to advertisers targeting English-language listeners); United States v Cumulus Media Inc, Case No 1:11CV01619 (US Dist Ct (D DC), 8 September 2011), Complaint, para 9; United States v CBS Corp, Case No 98CV00819, 1998 US Dist LEXIS 10292, pp *1–2 (US Dist Ct (D DC), 30 June 1998); United States v Westinghouse Elec Corp, Case No 96 2563, 1997 US Dist LEXIS 3263, pp *1–2 (US Dist Ct (D DC), 10 March 1997); United States v Bain Capital, LLC, Case No 1:08-cv-00245 (US Dist Ct (D DC), 13 February 2008), Competitive Impact Statement, pp 5–7; United States v Clear Channel Communications, Inc & AMFM Inc, Case No 00-2063 (US Dist Ct (D DC), 29 August 2000), Complaint, pp 2–3 (complaint filed with consent decree); Joel I Klein, Acting Assistant Attorney General, DOJ, ‘DOJ Analysis of Radio Mergers’, Speech, Washington DC, 19 February 1997, http://www.justice.gov/atr/public/speeches/1055.pdf.

37  OECD, Policy Roundtables: Two-Sided Markets, above note 33, p 149 (United States) (‘In some markets, the network interactions between the two sides are so significant that both sides of the market are important for economic analysis. In newspaper markets, methods that account for network interactions between newspaper readers and advertisers have been used in economic analysis for decades.’); United States v Village Voice Media LLC & NT Media LLC, Case No 1:03CV0164, 2003 WL 22019516 (US Dist Ct (ND Ohio), 2003).

38  Maurice E Stucke and Allen P Grunes, ‘Antitrust and the Marketplace of Ideas’, 69 Antitrust LJ (2001): pp 249, 288–91.

39  Ibid, pp 291–5.

40  Telecommunications Act of 1996, Pub L No 104-104, 110 Stat 56 (codified in scattered sections of 47 USC). Section 202 of the 1996 Act abolished the Federal Communications Commission’s (FCC) limits on the number of radio stations a single entity could own nationally. In 1996, the FCC in revising s 73.3555 of its rules (47 CFR s 73.3555) eliminated the national multiple radio ownership rule and relaxed the local ownership rule. FCC Order, In re Implementation of Sections 202(a) and 202(b)(1) of the Telecommunications Act of 1996 (Broadcast Radio Ownership), 47 CFR s 73.3555 (7 March 1996), http://www.fcc.gov/Bureaus/Mass_Media/Orders/1996/fcc96090.txt.

41  George Williams, FCC, Review of the Radio Industry 2007, p 1, http://hraunfoss.fcc.gov/edocs_public/attachmatch/DA-07-3470A11.pdf.

42  FCC, Review of the Radio Industry 2001, September 2001, p 3, http://www.fcc.gov/mb/policy/docs/radio01.pdf.

43  Williams, above note 41, p 1. In 2008, Clear Channel controlled 833 US radio stations, ‘508 of which were located’ in the largest 100 Arbitron markets. Bain Capital, Competitive Impact Statement, above note 36, p 4.

44  See Maurice E Stucke and Allen P Grunes, ‘Toward a Better Competition Policy for the Media: The Challenge of Developing Antitrust Policies That Support the Media Sector’s Unique Role in Our Democracy’, 42 Connecticut L Rev (2009): pp 101, 111 n 43, 123 (discussing decline in the amount of local news by radio stations and noting how increased concentration has not increased the average number of formats across markets).

45  Project for Excellence in Journalism, State of the News Media 2007, Radio Ownership, http://www.stateofthemedia.org/2007/radio-intro/ownership/.

46  On the ‘Future of Radio’: Hearing Before the Senate Committee on Commerce, Science, & Transportation, 110th Cong 3, 24 October 2007, Testimony of Mac McCaughan, co-founder of Merge Records, http://commerce.senate.gov/public/?a=Files.Serve&File_id=f83d2199-70b0-4633-9fc6-1db93c1d71bd.

47  Michael J Copps, FCC Commissioner, ‘Remarks to the NATPE 2003 Family Programming Forum’, New Orleans, La, 22 January 2003, http://www.fcc.gov/Speeches/Copps/2003/spmjc301.pdf.

48  Phil Rosenthal, ‘Homogenized Radio Stations Bottle Up Growth’, Chicago Tribune, 11 November 2007, Business, p 3.

49  Ibid.

50  OECD, Policy Roundtables: Two-Sided Markets, above note 33, p 220.

51  FCC, Further Notice of Proposed Rulemaking and Report and Order, adopted 31 March 2014; released 15 April 2014, para 92 (‘2014 Quadrennial Regulatory Review’) in In re 2014 Quadrennial Regulatory Review—Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 202 of the Telecommunications Act of 1996, MB Docket No 14-50.

52  Ibid, para 92; Williams, above note 41, p 2.

53  2014 Quadrennial Regulatory Review, above note 51, para 92 n 237 (also noting that while the Consumer Price Index increased approximately 3% per year between the 1996 Act and June 2011, the annual growth rate in radio advertising rates was approximately 6.5%); Williams, above note 41, p 16.

54  Orley Ashenfelter et al, ‘Did Robert Bork Understate the Competitive Impact of Mergers? Evidence from Consummated Mergers’, 57(S3) J L & Econ (2014): pp S67, S69 (‘While we agree with some of Bork’s critique of merger enforcement circa 1970, we believe that he went too far in dismissing likely competitive harm resulting from mergers that fell short of creating a monopoly or dominant firm. Subsequent empirical studies examining the price effects of consummated mergers have shown that, contrary to what Bork believed, mergers in oligopolistic markets can increase prices and harm consumers.’).

55  OECD, Policy Roundtables: Two-Sided Markets, above note 33, p 158.

56  Ibid, p 168.

57  Ibid, p 169.