Exclusionary Conduct: Structuring an Approach to Monopolization Cases

1. Exclusionary Conduct: Structuring an Approach to Monopolization Cases Tim Brennan 2006: T.D. MacDonald Chair in Industrial EconomicsCompetition Bureau, Industry Canada Permanent: Professor, Public Policy and EconomicsUniversity of Maryland, Baltimore County Senior Fellow, Resources for the Future, Washington, DC brennan.tim@cb-bc.gc.ca, brennan@umbc.edu Office of Fair TradingLondon, England 19 June 2006

2. 9/13/2012 Brennan: Exclusionary Conduct 2 By way of introduction Personal background Most experience US rather than Canada, references often to US law 8 yrs full-time with US Antitrust Division; �on call� for about 13 more Staff consultant to FTC, 2003-2005 Disclaimer: views not necessarily those of Competition Bureau Views on how law should work, not necessarily how it does work Both US and CA; not as familiar with EU But these ideas do show up in the strongest, clearest cases Special features of US law Common law vs. (relatively) detailed statute Many cases; long history of jurisprudence Extensive rights of private action (not necessarily helpful) Integrated legal staff vs. separate investigative units

3. 9/13/2012 Brennan: Exclusionary Conduct 3 Important UK contributions on abuse, exclusion From legal formalism to economic effect Classification of three categories of antitrust Collusion, mergers, abuse of dominance Recognizing problems with Article 82 Not just �market distortion�, but �consumer welfare� focus Problems implementing cost-based tests Difficulties of establishing dominance Market share not enough Questioning �collective dominance� Distinguishing abuse categories (crucial here) �Predation-type� vs. �RRC abuses� Impropriety of profit sacrifice for the latter

4. 9/13/2012 Brennan: Exclusionary Conduct 4 Still, issues on the table Is there more than distinguishing between �exclusionary� and pro-competitive� behavior? �Rivals� focus, distinguishing only between revenues and costs? Role of prior market power, dominance? Quest for a uniform standard, �consistency of approach�? Do we want an equally efficient competitor test in all cases? Role of intent, willfulness? Are rebates, bundles bad because they are predatory? What cost-based tests should be used for bundles? Should there always �be a gap between the thresholds � relating to abuse of dominance and to mergers�? [Vickers, 2003]

5. 9/13/2012 Brennan: Exclusionary Conduct 5 With that, what to take away Why monopolization, abuse controversial, difficult: Rival focus Recognize exclusion fundamentally different from predation Ask, �Are initial effects bad (exclusion) or good (predation)?� Predation should be hard, but exclusion more like mergers Exclusion, barriers to entry, RRC entail acquiring control over a complement market to raise effective price Apply empirical merger techniques to complement markets Reject predation-based profit sacrifice, equally efficient competitor, prior dominance screens Case examples Applications to bundling, e.g., Canada Pipe New options for remedies Focus on creating new monopoly, not maintaining old ones

6. 9/13/2012 Brennan: Exclusionary Conduct 6 Two competition areas relatively uncontroversial Horizontal collusion; US Sherman �1, CA S.45-49, UK CA Ch. I Fact-based, like criminal investigations Rule of reason vs. per se boundary Entry as mitigating factor; need for policy (libertarian critique) Horizontal mergers; US Clayton �7, CA S.91-103, UK EA02 Market delineation empirical core of cases Entry as mitigating factor, even with large share Critical loss/elasticities debate Use of differentiated product empirical models, simulations Necessity of market definition under unilateral effects Efficiency defenses, consumer vs. total vs. weighted welfare However, little dispute on theory, conceptual foundations

7. 9/13/2012 Brennan: Exclusionary Conduct 7 Troublesome #3: Monopolization in its many forms Sherman �2 in U.S. law, S. 75, 77-79 Canada Competition Act Foreclosure Abuse of dominance Exclusive dealing Tying/bundling �Raising rivals� costs� Predatory pricing Chapter II of 1998 Competition Act, Article 82 in EU Imposing �unfair� prices or �trading conditions� �Limiting production, markets or technical development� �Applying dissimilar conditions to equivalent transactions� Conditioning contracts on obligations �with no connection�

8. 9/13/2012 Brennan: Exclusionary Conduct 8 Persistently controversial Market power presumably horizontal, not vertical Practices reduce demand for monopolist�s product Too much competition: Low prices, more products Does monopolization law protect competitors, not competition? Recent observation on US, EU: This branch [of antitrust] has been confused since its inception, and has become a vehicle of suppressing, not promoting, competition in the marketplace�an unfortunate fact of U.S. antitrust law and even more so of European competition law. Ronald Cass, former US ITC vice chair Comment during FTC presentation: �Why would you want to save Section 2?�

9. 9/13/2012 Brennan: Exclusionary Conduct 9 Questions Why is monopolization different from collusion, merger? Is monopolization about economic welfare or something else (competitor protection, intent)? Is there any way to harmonize monopolization law with principles underlying collusion, merger law? What are the practical implications of the lack of harmony? Do predation �screens� make sense for exclusion case? Would other screens better sort good conduct from bad? What can vertical aspects usefully contribute to exclusionary conduct analysis? Is there a simple way to change monopolization doctrine to reconcile it with its antitrust cousins? Can we make exclusionary conduct cases simpler, easier, clearer?

10. 9/13/2012 Brennan: Exclusionary Conduct 10 Why?: The fallacious Section 2 �rivalry� syllogism Premise 1: Section 2 cases about hurting rivals Dominant firm abusing the little guys still animating vision Premise 2: Because competition hurts rivals, the burden in cases based on rival injury should be very high (insurmountable) Conclusion: Section 2 cases should bear a very high (insurmountable) burden Debate so far about Premise 2, not Premise 1 �Chicago school� sceptics accepts both premises, conclusion Post-Chicago/populist activist side Reject conclusion by rejecting second premise Protect competition by protecting competitors But both sides accept Premise 1!

11. 9/13/2012 Brennan: Exclusionary Conduct 11 Implications of the present standpoint Consensus has proven difficult Familiar Type I error of false positives: Inconsequential or beneficial practices attacked But don�t forget Type II error: Harmful practices may be ignored Using inappropriate screens Monopolization ignored because firm didn�t look big enough Misidentifying the relevant upstream market Encourage �anything can happen� non-robust theory Game theory can make evaluating practices more difficult Canceling theorists making litigation harder, especially for new, smaller agencies and courts Can courts handle hard theory? (Microsoft example) Benefits of possible precision vs. costs of error, process abuse

12. 9/13/2012 Brennan: Exclusionary Conduct 12 Why not reject Premise 1? Emphasis on rivalry has made monopolization cases harder to bring and justify Focus not on rivalry in primary market but on cornering previously competitive complement market Complement market monopolization (CMM) hallmark of exclusion Use horizontal, not predatory screens to assess that monopolization Possible approach to bundling cases Rescue good vertical cases in regulated sectors U.S v. AT&T (1982) contrasted with Verizon v. Trinko (2004) Regulation as evidence of risk, not excuse to walk away Bring much of monopolization law in line with less problematic collusion, merger law

13. 9/13/2012 Brennan: Exclusionary Conduct 13 Not all monopolization cases are the same Exclusion different from predation cases Unnecessary, spurious screens invoked because of anti-rival focus Exclusion cases may be too hard to bring Exclusion cases involve monopolizing a new market Raising entry barriers means the cost of something needed to enter goes up That�s where the relevant market is! Market power not the ability to exclude from X That requires market power over something outside X needed to enter, supply, and compete in X! Don�t try to find common standard with predation cases � the effects are not the same But how to tell the easy cases from the controversial cases?

14. 9/13/2012 Brennan: Exclusionary Conduct 14 A solution? First, classifying business practices

15. 9/13/2012 Brennan: Exclusionary Conduct 15 Second, antitrust-related examples

16. 9/13/2012 Brennan: Exclusionary Conduct 16 Third, subdividing monopolization cases

17. 9/13/2012 Brennan: Exclusionary Conduct 17 Failure to distinguish, and its costs US: Treat PBDB like PGDB, focus on minimizing Type I error Cases far too difficult The fallacious syllogism Erroneous screens EU: Treat PGDB like PBDB, focus on minimizing Type II error Presumption of bad behavior even with proximate goods Dominance => abuse? Willingness to consider �efficiency offenses� Quest for the �holy grail� of a single monopolization standard Treating all cases alike fails to minimize error costs Take distinction between exclusion and predation seriously Customize screens to the two different situations

18. 9/13/2012 Brennan: Exclusionary Conduct 18 Complementary market monopolization (CMM) To exclude, must raise complement prices => Acquiring market power over an complement, e.g., an input �Exclusion,� �foreclosure� as tying up the market for the complement Retailing, distribution, shelf space, typical examples Input example: Beer company signs exclusives with networks for TV ads Barriers to entry raised outside the market, not within the market Complementary market monopolization like mergers What would happen if contracted complement providers merged? E.g., those who had signed exclusive dealing contracts? The key to reconciling exclusion with merger, collusion law

19. 9/13/2012 Brennan: Exclusionary Conduct 19 Using merger-based effects tests Allows application of merger frameworks, empirical techniques!! E.g., US Horizontal Merger Guidelines (HMGs) Relevant market is the complement, not the upstream product Apply �SSNIP� standards for market delineation Treat conduct �as if� merger among complements suppliers Share of complement mkt. controlled, e.g., via exclusionary contracts Is that market easy to enter? Carlton�s idea of the �shadow price� CMM comes up indirectly in cases�make it the prime focus Use instead of predation based tests!! Profit sacrifice, equally efficient competitor, prior dominance Appropriate for high burdens, but not for PBDB exclusion

20. 9/13/2012 Brennan: Exclusionary Conduct 20 Finesse (in part) market delineation for monopoly Monopolization, abuse of dominance currently requires finding of a monopoly (dominance) that is then abused No good answer on how to do that Different from mergers, with �make matters worse� counterfactual Everybody has a little market power in a differentiated product world Cellophane fallacy: observed substitution doesn�t answer question �Monopoly� profits could come from inframarginal rent Overall cost coverage requires lifetime data So complicated that much of U.S. v Microsoft was about whether MS had a monopoly in which operating system market CMM approach avoids problem Use HMGs approach: Does practice make complement market worse? Problem still needs to be solved, but for regulation, not antitrust [Aside for later?: A �hypothetical regulator� test?]

21. 9/13/2012 Brennan: Exclusionary Conduct 21 Vertical irrelevance and relevance Irrelevance Complement, input market power necessary, sufficient for concern Why should the identity of the monopolizer of that market matter? Don�t dismiss because the monopolizer too small in primary market Vertical relevance Market delineation: Like using buyer information in a merger Derived demand: Are affected buyers a relevant market? Flesh out �unilateral� story: Can X% firm raise input price? How much does price have to go up to have significant effect? A small margin can create a huge competitive advantage Still need model to assess effects But predation-based tests still should not be used!

22. 9/13/2012 Brennan: Exclusionary Conduct 22 Questionable test #1: Profit sacrifice, �but for� test Sometimes called �no business sense�; essentially the same idea Any investment has positive opportunity cost, so need something relevant to competition concerns Why give primacy to welfare of alleged perpetrator? If complement market monopolized, why does price paid matter? Irrelevance of ability of complement suppliers to extract profits Like ascertaining crime by looking at criminal rather than victim �But for� test need doesn�t imply economic harm or benefit �Free� innovations can reduce welfare; even though would�ve been undertaken even without exit Induce exit by less efficient competitor; price goes up Some beneficial innovations may not be profitable unless rivals exit

23. 9/13/2012 Brennan: Exclusionary Conduct 23 Additional �profit sacrifice� considerations Wrong answer in regulatory evasion, network externality cases Refusal to interconnect immediately profitable Maintain network monopoly to evade regulation Exclusion via bundling isn�t predation Absolute exclusion shouldn�t be the test �Raising rivals� cost� contribution: The lack of need for a profit sacrifice Tantamount to absolute efficiency defense No profit sacrifice implies practice has a penny of efficiency Makes conduct immune if even trivially efficient

24. 9/13/2012 Brennan: Exclusionary Conduct 24 The �? sacrifice� irony � indicator of efficiency! Recall the Chicago argument regarding vertical restraints E.g., Telser, �Free trade� with RPM Vertical practices involve profit sacrifice Tying reduces demand for the monopolist�s product Upstream monopolist wants lower retail price, given wholesale price Avoid cost of dealing with regional monopolist Pay retailers more to be exclusive dealers Dominant firm ostensibly better off with complement competition �Chicago� inference: Profit sacrifice => unrecognized efficiencies Solving principle-agent problems with incomplete contracting How do we know if � We aren�t repeating 100 year old US AT mistakes? When our accounting of profits is complete?

25. 9/13/2012 Brennan: Exclusionary Conduct 25 Questionable test #2: Equally efficient competitor Inefficient competitors are competitively relevant Bertrand homogeneous products Dominant firm, competitive fringe Cournot oligopoly All have prices below the monopoly price Dubious standard even for predation Justification as test for competing too hard Would allow above costs tests to exclude Predating against less efficient competitors could reduce welfare Elhauge: Equally efficient where? Everywhere? At the margin? Lesson: Type I, Type II errors different for exclusion, predation cases

26. 9/13/2012 Brennan: Exclusionary Conduct 26 Questionable test #3: Prior dominance Focus is on the wrong market Dominance neither sufficient (EU) nor necessary (US) Excuses actions by firms that don�t have prior monopoly, yet tie up complement Inherent contradiction in litigation To establish prior monopoly, show high barriers, etc. But if entry that hard, does practice at issue matter? U.S. v. Microsoft example Microsoft dominance: Lock-in, scale economies, network effects But then what�s the threat to application platform monopoly? Posner�s �fragile monopoly� a useful concept: If monopoly secure, efficiency likely motive The �but for� test: would monopoly exist if practice banned?

27. 9/13/2012 Brennan: Exclusionary Conduct 27 Dominance a defense, not condition of guilt? Large size in one market suggests large involvement upstream Restraints commensurate with scale Improve coordination with similar size of input market Disproportionality would warrant concern Marginal harm from upstream monopolization not great Even if �single monopoly profit� special case, more profit taken at one stage, less to be taken at others Exceptions may enhance welfare (price discrimination, more efficient input mix) Place burden of showing no effect from CMM on defendant Eliminating double marginalization not quite on point Applies only if market power inevitable at both stages CMM assumes (more or less) competition in complement

28. 9/13/2012 Brennan: Exclusionary Conduct 28 Reality check: Should screens apply to mergers? Profit sacrifice? Block only mergers unprofitable but for the exercise of market power? Not just total welfare, but absolute efficiency defense Gut enforcement Equally efficient competitor? Defend any merger if one party less efficient than another Allow most efficient firm to buy up all less efficient competitors Even prior dominance Some market share matters, of course But with very large market share, harder to tell that the merger will make matters worse Relevance: View CMM as �as if� merger in complement market Same points for collusion: Screens irrelevant, even if effects matter (e.g., CA �undue� SLC)

29. 9/13/2012 Brennan: Exclusionary Conduct 29 Applications to some exclusion cases Heinz (Canada) Alleged monopolization: Baby food companies sign contracts for carriages, prime display at retail Real monopolization: Market for retail display Focus: Ease of entry into complement market, e.g., �shelf space� Dentsply Alleged: Artificial teeth via exclusive contract with dealers Real monopolization: Market for distributing teeth to dental labs Focus: Entry into distribution market, including self-provision U.S. v Microsoft Alleged monopolization: �Intel-based PC operating systems� Real monop.: Browser distribution via PC inclusion, ISP promotion Focus: Could Netscape get its browser out effectively? Remedy: DOJ settled primarily on eliminating exclusive contracts

30. 9/13/2012 Brennan: Exclusionary Conduct 30 Some lessons from those cases Core issues: Share, entry into relevant complement market Monopolization of that complement market (e.g. distribution) If distribution not affected, competitors can�t be excluded Look at how exclusive contracts tie up market, �as if� merged Profit sacrifice, equally efficient competitor not relevant Monopolization paradox: Having monopoly in order to monopolize Microsoft OS dominance (lock-in, scale economies, network effects) Denstply in teeth, Heinz in baby food But then what�s the problem? Dominance as defense? Have dental distribution share match upstream market Incentive to develop �middleware� by capturing rents Avoid double marginalization (e.g., browser + OS monopolies)

31. 9/13/2012 Brennan: Exclusionary Conduct 31 Bundling, discounts, rebates, etc. Dominant topic in antitrust debate Is the relevant test a predation-based standard? Does discount lead to pricing below cost? Does/would it exclude equally efficient competitor? State of U.S. law SmithKline v. Eli Lilly: Bundle price need not be below variable cost LePages v. 3M won at divided Third Circuit Court of Appeal; Supreme Court denied review DOJ, FTC: Review would be �premature�; issues �novel and difficult� Focus of recent models Nalebuff: Bundling exclusionary entry barrier Greenlee, Reitman, Sibley: Bundling reduces welfare if stand-alone prices rise

32. 9/13/2012 Brennan: Exclusionary Conduct 32 Limitations of models Nalebuff Bundling increases consumer and total welfare in all equilibria Would allow competitors to agree not to enter each other�s markets Limits of other bundling models Rent extraction, but welfare often up (GRS like 2 part-tariff) Like price disc., results depend on unobservable correlations Whinston � tie hardens output commitment, but ever observed? Consumers buy in models, but cases about intermediate markets Concord Boat v. Brunswick (engine discounts to manufacturers) LePages v. 3M (discounts for carrying 3M�s house brand tape) Ortho Diagnostic Systems v. Abbott Laboratories (Abbott discounted virus screening tests but legal as above Ortho�s costs) SmithKline v. Eli Lilly (Lilly bundled antibiotics in sales to hospitals) CA Canada Pipe (full line discounts to cast iron pipe distributors) Use CMM exclusion approach, not predation

33. 9/13/2012 Brennan: Exclusionary Conduct 33 Canadian pioneering: Canada Pipe Allegations (from initial Bureau notice of application) Canada Pipe holds monopoly in relevant market in cast iron drain pipe Instituted �Stocking Distributor Program� provides rebates (up to 20%) for distributors who purchase exclusively from Canada Pipe Equivalent to exclusive dealing, forecloses distribution Lost at trial (from Tribunal decision) SDP exclusive, but with no anticompetitive effects; not �predatory, exclusionary, or disciplinary� Found some entry from imports and a new manufacturer No switching costs; SDP not contractual, distributors could leave at any time Key argument control over distribution of pipe Currently under appeal � should total exclusion be necessary?

34. 9/13/2012 Brennan: Exclusionary Conduct 34 Bundle discounts via CMM: A first screen Issue fundamentally static upstream monopolization Not tying for rent extraction (Nalebuff, Sibley et al.) Nor commitment (Whinston) Intermediate good (CM) focus First screen: Whether bundling ties up complement market Reach of the bundle Necessary and sufficient condition HMGs approach to market definition, informed by �buyer� market Examples Outlets needed to sell unbranded tape Distributors of false teeth to denture labs Distributors of cast-iron drain pipe

35. 9/13/2012 Brennan: Exclusionary Conduct 35 Second tentative screen Resemblance to exclusive dealing: Is loss of discount from carrying competitor equivalent to payment for breach of exclusive dealing contract? How much does exclusive dealing raise complement price? Exclusive contracts not absolute Entrant has to compensate distributor for the breach payments. These will be based on the incumbent�s lost profits (Price � short run marginal cost) from breach of contract, implying � Entry possible only if it can undercut the incumbent�s marginal cost Is exclusive dealing breach the right quantitative standard? Courts may find that short-run marginal cost is very low Less efficient competitors could knock down price For intermediate goods, a small price increase can create great competitive disadvantage

36. 9/13/2012 Brennan: Exclusionary Conduct 36 Observations on remedies Exclusive dealing, rebates, bundling not good or bad per se The harm arises from CMM, but OK short of monopolization Ideal remedy not �all or nothing�, but share based, like mergers Identify level of exclusive contracting that is non-threatening OK to exclude, but only if one does not go beyond X% of the relevant complement market (distribution, retailers) Effective remedies assume competition in complement market Prior retail, distribution monopoly perhaps a DM defense Can an upstream monopolist leave money on the table? Share of market with money left? Share-limited remedies may reveal anticompetitive effect, if dropped if they can�t cover the whole market

37. 9/13/2012 Brennan: Exclusionary Conduct 37 Other issues that may arise in exclusion cases Tying, bundling can exclude but need not be anticompetitive Hurts rival, but consumers may be better off Add product features, quality control, shopping convenience Consider �aftermarket� cases (locked-in consumers) as contract breach, consumer protection matters Predation, dynamic cases may not be reducible to the direct exclusion �complementary market� format Microsoft as example Controlling browser, media player distribution vs. preempting future competition in application platforms Sector regulation need not reduce the need for antitrust Regulation can make anticompetitive conduct more likely, e.g., U.S. v. AT&T (contrast 2004 USSC Trinko decision) When can antitrust apply in regulated sectors? [CA �regulated conduct�, US �state action� doctrines, UK CA98 S.10(1)(a)]

38. 9/13/2012 Brennan: Exclusionary Conduct 38 Elhauge�s monopolization test Does a practice by a monopolist discriminate against rivals? Allows refusals to deal if uniform, but not targeted Fundamental flaw: Does it work when most needed? Some inputs are bought only by rivals Network elements, operations support in telecom markets Bridge access in Terminal Railroad Either no discrimination, since only rivals are buyers Or discrimination an almost certain byproduct of vertical integration, making VI by a monopolist abusive per se Also, is anything equivalent to price discrimination per se OK? (Carlton, IIOC)

39. 9/13/2012 Brennan: Exclusionary Conduct 39 Exclusion: monopoly creation, not maintenance �Create� focus appropriate for exclusion cases Has a monopoly/market power been created in a complement market that was not there before? Look at using familiar merger, collusion horizontal terms, e.g., MEGs �Maintain� leads to the problems Sets up entrants, small competitors as targets to be protected Focuses on inhibiting competition, leading to high if not impossible tests Looks at the wrong market, introducing error in both directions Forces establishment of monopoly that�s being maintained Fits predation cases, perhaps, but not exclusion Dominance neither necessary nor sufficient Profit sacrifice as absolute efficiency excuse May lead to focus on intent rather than effect Promote monopolization law by reducing its scope?

40. 9/13/2012 Brennan: Exclusionary Conduct 40 Summary Not all monopolization cases are the same: Exclusion distinct from predation Fallacious syllogism: monopolization led awry by focus on rivals Those with direct bads, like exclusion, can be addressed as we do mergers, collusion These involve complementary market monopolization approach which can clarify analysis, remedy in exclusion cases Profit sacrifice, equally efficient competitor, prior dominance, problematic screens for exclusion cases Vertical matters to get story straight, as in merger cases Applies to exclusion cases, perhaps bundling Focus on �create� monopoly in complement market rather than �maintain� monopoly in primary market