第三届CRESSE-JUFE竞争政策国际研讨会【专题】.pdf

Incomplete information approaches to competition policy Simon Loertscher & Leslie M. Marx CRESSE-JUFE virtually @ Nanchang 20 October 2022 Loertscher & Marx IIIO competition policy 20 Oct 2022 1 / 61 Introduction Digital age tremendous increasing returns to scale in data rather than an age of disintermediation, digital age has brought forth giant platforms and firms no end in sight Renewed importance of competition policy regulation Larger markets are better, all else equal, because they can execute the same trades as smaller, standalone markets, and sometimes execute more valuable trades. Loertscher & Marx IIIO competition policy 20 Oct 2022 2 / 61 Motivation for incomplete information approach With incomplete information that uniform pricing by a monopoly is optimal is a conclusion under some conditions rather than an assumption (e.g. Loertscher-Muir JPE 2022) likewise, that perfect price discrimination is not possible is a conclusion, not an assumption the tradeoff between social surplus and profit exists without any restrictions of the contracting space Loertscher & Marx IIIO competition policy 20 Oct 2022 3 / 61 Motivation for incomplete information approach Bargaining in the field Backus-Blake-Larsen-Tadelis (QJE 2020): Bargaining breaks down in 55% of over 25 million observations on eBay Incomplete information (mechanism design) approach emerging as preferred conceptual framework, e.g.: Larsen (RES 2021) Backus-Blake-Pettus-Tadelis (2022) Larsen-Lu-Zhang (2022) Bargaining at heart of numerous contemporaneous merger cases Loertscher & Marx IIIO competition policy 20 Oct 2022 4 / 61 Today 1 Buyer power in procurement markets simple setup, empirically relevant clear-cut definition, results analysis of coordinated effects 2 Countervailing power can bargaining power on one side be offset by power on the other? Galbraith vs Stigler 3 Investment with incomplete information, efficient bargaining implies efficient investments 4 Merger policy and remedies, digital monopolies, and regulation Loertscher & Marx IIIO competition policy 20 Oct 2022 5 / 61 Background literature Published papers: Loertscher-Marx, Journal of Political Economy 2019 [LM-JPE] Loertscher-Marx, International Journal of Industrial Organization 2020 [LM-IJIO] Loertscher-Marx, Journal of Law & Economics 2021 [LM-JLE] Loertscher-Marx, American Economic Review 2022 [LM-AER] Working paper: Loertscher-Marx: “Vertical integration with incomplete information” 2022 [LM-VI] Loertscher & Marx IIIO competition policy 20 Oct 2022 6 / 61 Buyer power in procurement markets (based on LM-JPE) Loertscher & Marx IIIO competition policy 20 Oct 2022 7 / 61 Setup and motivation Look at procurement setup Buyer with single-unit demand procures a good from a set of suppliers Simple setup, permitting clear-cut definitions and results Extends naturally to analysis of coordinated effects Procurement is economically relevant Loertscher & Marx IIIO competition policy 20 Oct 2022 8 / 61 Public procurement as a share of GDP Loertscher & Marx IIIO competition policy 20 Oct 2022 9 / 61 Procurement of substantial economic importance Government procurement: double-digit share of GDP in most developed economies China’s public procurement market grew 10-fold over 2001-11 to $180 billion Japanese govt construction procurement of $10 billion/year in 2003–06 37% of projects found to be affected by collusive bidding (projects pre-allocated to a bidding ring member) Barkley (2022): collusion kills (insulin procurement in Mexico) Business-to-business procurement: $8-trillion in B2B wholesale and distribution in the U.S. in 2018 technology enabling the use of competitive procurement in ever more settings “Government Procurement: Data, Trends and Protectionist Tendencies,” EC Chief Economist Note, 2018. “Amazon’s next mountain: B2B procurement,” Globe and Mail, 2018. Kawai & Nakabayashi, “Detecting Large-Scale Collusion in Procurement Auctions,” Journal of Political Economy 130, 2022, 1364–1411. US China Business Council, “Navigating China’s Public Procurement Market: Background, Challenges, and Best Practices,” 2012. Loertscher & Marx IIIO competition policy 20 Oct 2022 10 / 61 Buyer power Perceived wisdom: Powerful buyers limit the unilateral (and coordinated) effects of mergers Powerful buyers can withstand upward price pressure and are thus immune to merger harm Buyer power features prominently in antitrust and merger analysis Frequent merger defense Appeals to judges and juries – embraced by courts as having “talismanic power that cured all doubts about a merger” (Steptoe ALJ 1993) Loertscher & Marx IIIO competition policy 20 Oct 2022 11 / 61 What is buyer power? The notion of “buyer power” is widely invoked but not always precisely defined Powerful buyer: “Often able to negotiate favorable terms with their suppliers” (US HMG) Has “bargaining strength ... due to its size, its commercial significance ... and its ability to switch to alternative suppliers” (EC HMG) “Exists when the specific characteristics of a buyer—such as its size, its commercial significance to suppliers or the manner in which it purchases from suppliers—provide the buyer with additional negotiating leverage” (AUS HMG) Loertscher & Marx IIIO competition policy 20 Oct 2022 12 / 61 Buyer power: a bit of a conundrum Yet, real plausibility... Theoretical foundation? Standard models (Cournot, Bertrand) do not permit buyer power Ad hoc assumptions, e.g. powerful buyers withstand collusion Loertscher & Marx IIIO competition policy Proposed merger of oilfield services firms 20 Oct 2022 13 / 61 Buyer power and mergers in a procurement setup Buyer purchases using a procurement mechanism and suppliers bid for the buyer’s business Buyer power is the ability of the buyer to use an “optimal” procurement Merger creates a supplier whose cost is the minimum of the two merging suppliers’ costs Loertscher & Marx IIIO competition policy 20 Oct 2022 14 / 61 Summary of results for unilateral merger effects Mergers always harm the buyer Without buyer power, mergers: are always profitable (and equivalent to perfect collusion) are neutral for rivals increase the merged entity’s incentives for investment With buyer power and pre-merger symmetry, mergers are not always profitable – even to monopoly (while perfect collusion is) benefit rivals (and entrants) increase rivals’ incentives to invest (and may increase the merged entity’s) Loertscher & Marx IIIO competition policy 20 Oct 2022 15 / 61 Setup Pre-merger n ≥ 2 suppliers: supplier i draws its cost ci independently from a distribution with support [c, c] and continuous density that is positive on the interior of the support One buyer with value v > c Post-merger (merger occurs before cost realizations) Merged entity combining 1 and 2 has cost c = min{c1 , c2 } Rivals still draw their costs from the pre-merger distribution Buyer knows the distributions, but not the realized costs Loertscher & Marx IIIO competition policy 20 Oct 2022 16 / 61 IIIO approach: Stigler meets Myerson Incomplete information about firms’ costs affects: firm behavior price formation Tractable methodology and disciplining framework: risk neutral buyers and sellers independent private values one-dimensional private information incentive compatibility, individual rationality explicit price formation process efficiency-profit tradeoff Loertscher & Marx IIIO competition policy George Stigler Nobel Prize 1982: “studies of industrial structures, functioning of markets” “the true father of modern oligopoly theory” Roger Myerson Nobel Prize 2007: “laid the foundations of mechanism design theory” 20 Oct 2022 17 / 61 Model of buyers with and without power Powerful buyer: holds an optimal procurement As in Bulow-Klemperer (AER 1996) Buyer without power: holds an efficient procurement (second-price or English auction with reserve p = min{v , c}) Loertscher & Marx IIIO competition policy 20 Oct 2022 18 / 61 Harmful mergers Theorem 1 LM-JPE A merger (without cost synergies) harms the buyer. Loertscher & Marx IIIO competition policy 20 Oct 2022 19 / 61 Proof of Theorem 1 LM-JPE Fundamentally: Merger eliminates a bid, which is bad Without buyer power, that’s it With buyer power, buyer optimally uses different mechanism (allocation rule) post merger However: revealed preference argument – buyer could use the post-merger optimal allocation rule pre merger, and have an additional bidder Loertscher & Marx IIIO competition policy 20 Oct 2022 20 / 61 Buyer power and unilateral merger effects Proposition 4 LM-JPE Without buyer power With buyer power quantity: ∆Q = 0 ∆Q ≤ 0 social surplus: ∆SS = 0 ∆SS < 0 buyer surplus: ∆BS < 0 ∆BS < 0 comparison: ∆BSwithout < ∆BSwith < 0 Note: A buyer-power based merger defense is self-defeating when authorities take a social surplus perspective Assumes: identical distributions with increasing virtual costs pre merger (regularity) Loertscher & Marx IIIO competition policy 20 Oct 2022 21 / 61 Coordinated effects of mergers (based on LM-JLE) Coordinated effects Merger changes “the nature of competition in such a way that firms that previously were not coordinating their behavior, are now significantly more likely to coordinate” (EC HMG) “An acquisition eliminating a maverick firm in a market vulnerable to coordinated conduct is likely to cause adverse coordinated effects” (US HMG) Loertscher & Marx IIIO competition policy 20 Oct 2022 22 / 61 How to model coordinated effects? Important, central concern “Long been at the core of U.S. merger policy” (Kolasky 2002) “The ultimate issue” (Judge Richard Posner 1986) Challenging Unclear how to model imperfect coordination Miller-Sheu-Weinberg (AER 2021) Interaction with buyer power Key role (but vague/elusive definition) of mavericks Baker (NLR 2002) advocates a maverick-centered approach on grounds that in “many settings, regulators reliably can identify an industry maverick that prevents or limits coordination” Loertscher & Marx IIIO competition policy 20 Oct 2022 23 / 61 Results Provide a model of coordination that allows us to: define what it means for a market to be at risk for coordination quantify the risk define and test for maverick status acquisition of a maverick does not imply that the market is put at risk Provide a measure of coordinated effects and definition of a maverick that are operational based only on pre-merger observables Loertscher & Marx IIIO competition policy 20 Oct 2022 24 / 61 Defining at-risk markets and mavericks Coordination is modelled as participation in an allocation scheme each coordinating supplier is chosen with some probability to be the only supplier from the group to participate for coordination to be individually rational, each coordinating suppliers must be selected with sufficiently high probability Market is at risk if coordination is individually rational for each coordinating supplier Firm is a maverick if the market is not at risk when the firm is present but is at risk if the firm is eliminated Acquisition of a firm is different from eliminating its productive assets Loertscher & Marx IIIO competition policy 20 Oct 2022 25 / 61 Approach Allocation schemes are appealing: no transfers; no communication of private, production-relevant information commonly used by cartels (phases of the moon) accommodate non-all-inclusive coordination More fundamentally: Spell out the coordinating conduct of concern (here: allocation scheme) Derive testable predictions for when market is or is not at risk for coordination operation definition of a maverick based on that conduct This basic approach is, in principle, applicable to any alternative coordination of concern Loertscher & Marx IIIO competition policy 20 Oct 2022 26 / 61 Key findings Coordination reduces expected buyer/consumer surplus and social surplus Some, but not all, markets are at risk Some, but not all, mergers put markets at risk Risk varies with: degree of outside competition symmetry and strength of coordinating firms buyer power vertical integration of buyers Tradeoffs between unilateral and coordinated effects Loertscher & Marx IIIO competition policy 20 Oct 2022 27 / 61 Setup Suppliers N Consider possibility that K ⊆ N coordinate K contains k ≥ 2 suppliers Allocation scheme i ∈ K is active (others inactive) with probability si Πi is i’s expected payoff with no coordination Πi (K ) is i ∈ K ’s expected payoff when it is the active supplier Assume that failure by any supplier in K to participate results in no coordination conservative approach: if market is not at risk, then it is also not at risk if we allowed for unilateral defections Loertscher & Marx IIIO competition policy 20 Oct 2022 28 / 61 Critical shares and coordinated effects index Coordination is profitable for i if and only if it is selected with probability greater than its critical share: si (K ) ≡ Πi Πi (K ) Coordinated effects index I(K ) ≡ 1 − X si (K ) i∈K I(K ) can e.g. be estimated based on market shares and one markup, with parameterized distributions Loertscher & Marx IIIO competition policy 20 Oct 2022 29 / 61 Interpretation I(K ) ≤ 0 (not at risk): no selection probabilities satisfying IR coordination as a allocation scheme is not profitable I(K ) > 0 (at risk): exist selection probabilities satisfying IR coordination as a allocation scheme may be profitable larger I implies more leeway for coordination to remain profitable in the face of additional costs of coordination as the index approaches 1, selection probabilities of 1/k suffice for I close to zero, small positive costs or frictions could impede coordination Loertscher & Marx IIIO competition policy 20 Oct 2022 30 / 61 Mavericks A firm outside K is a maverick with respect to K if I(K ; N) ≤ 0 and I(K ; N\{m}) > 0 Possibility of: no maverick more than one maverick with respect to K Loertscher & Marx IIIO competition policy 20 Oct 2022 31 / 61 Reason for concern about coordination Assuming efficient procurement markets: Proposition 1 LM-JLE Coordination reduces expected buyer surplus and expected social surplus. With positive probability, the participation of the lowest-cost supplier is suppressed Coordination increases the expected cost conditional on trade and reduces the probability of trade if trade does not occur with probability 1 under efficiency Elimination of bids increases the expected price—harms buyer Loertscher & Marx IIIO competition policy 20 Oct 2022 32 / 61 Implication for mergers Proposition 3 LM-JLE A merger of suppliers in K can, but need not, cause a market not at risk for coordination by suppliers in K to become at risk for coordination by the corresponding post-merger suppliers. A merger of suppliers outside K does not affect the risk for coordination by suppliers in K . Merging parties cannot reduce concerns by strategically varying the timing of their merger so that it falls either before or after a merger of outsiders Can’t reduce risk of CE among set of firms by first approving “balancing” mergers among outsiders Loertscher & Marx IIIO competition policy 20 Oct 2022 33 / 61 Countervailing power effects (based on LM-AER and LM-VI) Loertscher & Marx IIIO competition policy 20 Oct 2022 34 / 61 Countervailing power: basic idea Preceding analysis as most parts of LM-JPE take absence or presence of buyer power as given It is, of course, conceivable that bargaining power varies with, say, policies or mergers In short, that is the basic idea behind John Kenneth Galbraith’s idea of countervailing power Now expand the setup and analysis to account for this possibility Loertscher & Marx IIIO competition policy 20 Oct 2022 35 / 61 Countervailing power controversial since beginning “The neutralization of one position of power by another” is a mitigant of economic power of “substantial, and perhaps central, importance” John Kenneth Galbraith, AER 1954 “The theory of bilateral oligopoly can hardly be said to exist ... it simply is romantic to believe that a competitive solution will emerge... I have most serious doubts that anyone can devise a rational explanation for Galbraith’s proxy-minded oligopolists.” George Stigler, AER 1954 Snyder (Palgrave 2005): “Formalizing the concept [of countervaling power] is difficult because it is difficult to model bilateral monopoly or oligopoly, and there exists no single canonical model” Loertscher & Marx IIIO competition policy 20 Oct 2022 36 / 61 Yet, concept remains influential Yet, popular appeal, and influence in antitrust OECD (2007, 2011): role for collective negotiation and group boycotts by merchants to counterbalance market power by providers of payment card services DOJ/FTC (1996): potential benefits of physician network joint ventures EC (2020): retail alliances “generate benefits for their members by improving their bargaining position and leading to better purchasing terms” through increased “bargaining ability” and “bargaining leverage” Now amend model to allow for bargaining powers that can vary continuously Loertscher & Marx IIIO competition policy 20 Oct 2022 37 / 61 Setup (simplified version of LM-AER) Buyers and sellers with single-unit demands and supplies Each agent draws its value or cost independently from a continuous distribution Its value or cost is the agent’s private information Bargaining is modelled as a mechanism (“game”) that is incentive compatible, individually rational mechanism requires no subsidy, and maximizes the weighted sum of the agents’ expected surpluses Generalization of Williams (JET 1987) Private information on both sides means first-best need not be possible (Myerson-Satterthwaite JET 1983) Loertscher & Marx IIIO competition policy 20 Oct 2022 38 / 61 Incomplete information Pareto frontier Proposition 4 LM-AER The incomplete information Pareto (Williams) frontier is concave to the origin, and strictly concave if and only if its intersection with the first-best frontier contains at most one point. (a) Overlapping supports (b) Nonoverlapping supports uB uB FB FB SB Δ=1 Δ=1/2 Δ=1 Δ=1/2 Δ=0 Δ=0 SB Loertscher & Marx uS FB IIIO competition policy uS FB 20 Oct 2022 39 / 61 Countervailing power effects 1 buyer, 2 pre-merger suppliers, w B = ∆, wjS = 1 − ∆ uB pre-merger post-merger ∆=1 ∆=1/2 ∆=0 uS Loertscher & Marx IIIO competition policy 20 Oct 2022 40 / 61 Bargaining weights affecting merger 1 buyer, 2 pre-merger suppliers uB pre-merger post-merger ∆=1 ∆=1 ∆=1/2 ∆=1/2 ∆=0 ∆=0 uS Loertscher & Marx IIIO competition policy 20 Oct 2022 41 / 61 Bargaining weights affecting merger 1 buyer, 2 pre-merger suppliers, k B = kjS = 1 uB pre-merger post-merger ∆=1 ∆=1 ∆=1/2 ∆=1/2 ∆=0 ∆=0 uS Loertscher & Marx IIIO competition policy 20 Oct 2022 42 / 61 Bargaining weights affecting merger 1 buyer, 2 pre-merger suppliers, k B = kjS = 1 uB pre-merger post-merger ∆=1 ∆=1 ∆=1/2 ∆=1/2 ∆=0 ∆=0 uS Loertscher & Marx IIIO competition policy 20 Oct 2022 43 / 61 Requirements for a countervailing power defense For a merger of suppliers, a countervailing power defense requires: Policy objective that goes beyond just buyer surplus a merger with countervailing power is doubly bad for the buyers—competition among suppliers is reduced and the remaining suppliers have increased bargaining power Pre-merger buyer power buyers must have greater bargaining power than the suppliers, so that increased supplier power is movement towards equalization Retention of some post-merger buyer power buyer power would need to diminish, but not vanish—so that society is not simply trading dominant buyers for dominant suppliers EC merger guidelines: “a merger of two suppliers may reduce buyer power if it thereby removes a credible alternative” “it is not sufficient that buyer power exists prior to the merger, it must also exist and remain effective following the merger” Loertscher & Marx IIIO competition policy 20 Oct 2022 44 / 61 Desirability of equalizing bargaining powers LM-VI distinguish between (i) pure bargaining power, which merely represents an agent’s ability to tilt the market mechanism in its favor (ii) bargaining power that derives from economically relevant factors (e.g. firms who serve more consumers may have more bargaining power) In case (i), equalization of bargaining powers is a first-order issue since efficiency is not possible without equal bargaining powers In contrast, in case (ii) a benevolent social planner may want bargaining powers to differ across agents Loertscher & Marx IIIO competition policy 20 Oct 2022 45 / 61 Investment (based on LM-AER and LM-VI) Loertscher & Marx IIIO competition policy 20 Oct 2022 46 / 61 Investment incentives Investment incentives feature prominently in the Theory of the Firm and in recent antitrust cases (Dow-Dupont, GE-Almstrom mergers) How does market structure affect incentives to invest? Assume that each firm i ex ante chooses investment yi at cost Ci (yi ) with yi affecting the density of its value θi but not its support The social planner maximizes over y h i Eθ|y SS(θ) − X Ci (yi ) i Let y be a solution to the planner’s problem Loertscher & Marx IIIO competition policy 20 Oct 2022 47 / 61 First-best market implies first-best investments Proposition 10 LM-VI If the market is efficient given investments y, the investment game has a Nash equilibrium in which each firm i invests y i . Proof: Vickrey-Clarke-Groves (VCG) mechanism aligns each firm’s incentives given private information with the planner’s; carries over ex ante; by payoff equivalence theorem, focus on VCG is without loss of generality Make markets work efficiently is correct guidance with or without investments Simple argument, but as pointed out by Waehrer (2021), far from universally understood (see e.g. Miller IJIO 2014) Loertscher & Marx IIIO competition policy 20 Oct 2022 48 / 61 Merger policy and remedies, digital monopolies and regulation (partly based on LM-VI and LM-IJIO) Loertscher & Marx IIIO competition policy 20 Oct 2022 49 / 61 Merger policy and remedies (based on LM-VI) Model of vertical integration (LM-VI) Firms hold resources that they can use internally or trade to others Vector or resources held by firms is r = (r1 , . . . , rn ) Once firms’ private information is realized, the firms can trade (market modelled as in LM-AER) Some vertical integration is necessary and sufficient for the first-best to be possible Proposition 4 LM-VI Social surplus SS(r) is concave in r. Vertical mergers can increase or decrease social surplus Model provides rationale and guidance for remedies (divestitures) Loertscher & Marx IIIO competition policy 20 Oct 2022 50 / 61 Merger policy and role for divestitures rj SS(ri , r j ; r-i,- j ) = SS0 ri Loertscher & Marx IIIO competition policy 20 Oct 2022 51 / 61 Merger policy and role for divestitures rj SS(ri , r j ; r-i,- j ) = SS0 ri Loertscher & Marx IIIO competition policy 20 Oct 2022 51 / 61 Merger policy and role for divestitures rj SS(ri , r j ; r-i,- j ) = SS0 ri Loertscher & Marx IIIO competition policy 20 Oct 2022 51 / 61 Merger policy and role for divestitures rj SS(ri , r j ; r-i,- j ) = SS0 1 2 3 ri Loertscher & Marx IIIO competition policy 20 Oct 2022 51 / 61 Digital monopolies (based on LM-IJIO) Increasing returns to scale in data-driven matchmaking Digital monopolies (DMs) Like natural monopolies, DMs call for antitrust scrutiny and possibly regulation Sharply contrasting views: distopian future (or present) vs “Golden Age of Music, Movies, Books, and Television” (Waldfogel, JEP 2017) Loertscher & Marx IIIO competition policy 20 Oct 2022 52 / 61 Key questions 1 Does DM use data to improve matching only or matching and pricing? 2 Should an antitrust/competition authority be concerned with privacy protection or with price regulation? We consider only private values environments whoever likes Elton John (R.E.M.) is likely to enjoy being referred to John Lennon (The National) no fraud or criminal behaviour do not discuss freedom of speech vs misinformation today Loertscher & Marx IIIO competition policy 20 Oct 2022 53 / 61 Digital monopoly: setup Monopoly firm One consumer Consumer’s value is a random variable With more data, the monopoly can better match the product to the consumer ng asi Waldfogel’s Golden Age of Pop ta da Privacy protection can be disastrous re inc v Loertscher & Marx Individuals (like me!) who used to watch CSI can now watch The Bridge v IIIO competition policy Netflix could not recommend The Bridge to me without information about me and individuals like me 20 Oct 2022 54 / 61 Data and personalized pricing Of course, the above assumes the monopoly does not also use its improved information to “improve” its pricing If monopoly uses data to match and price, then CS goes to 0 as data or information becomes perfect Monopoly can perfectly target product and perfectly price discriminate That’s the distopian future (or present) for consumers That is why the question what data is used for —matching only, or matching and pricing—is key Loertscher & Marx IIIO competition policy 20 Oct 2022 55 / 61 Regulation New and interesting issues arise within the incomplete information approach: 1 With digital monopolies, Ramsey pricing may not be sufficient to protect consumers 2 When a monopoly uses an optimal selling mechanism rather than just uniform pricing, Ramsey pricing may involve price floors as well as price ceilings Loertscher & Marx IIIO competition policy 20 Oct 2022 56 / 61 Regulating a digital monopoly Consider DM that uses data to match and price Social surplus (SS) is maximized in the perfect information limit and equal to producer surplus (PS) Ramsey pricing will not do since maximizing αPS(p) + (1 − α)SS(p) is equivalent to maximizing PS(p) Alternative: set the price p so that consumer’s “fair” share of social surplus is a constant α ∈ (0, 1): CS(p) = αSS(p) Given α, this yields a unique price This price decreases in α and is bounded away from buyer’s maximum willingness to pay Loertscher & Marx IIIO competition policy 20 Oct 2022 57 / 61 Price regulation in the digital age In the digital age, may be able to implement price regulation by inspecting algorithms Allow personal data to be used for matching, but restrict use of personal data for pricing Loertscher & Marx IIIO competition policy 20 Oct 2022 58 / 61 Ramsey pricing revisited Loertscher-Muir (JPE 2022): When revenue R(Q) = P(Q)Q under market-clearing prices is concave in Q, monopoly optimally sets a uniform price Otherwise, it optimally sets two prices, inducing rationing at the low price Revenue under optimal selling mechanism R 0.100 0.095 0.090 0.085 0.080 0.075 0.070 0.065 0.2 0.3 0.4 0.5 Q Loertscher-Muir 2022 (working paper) show that Ramsey pricing for a monopoly that uses an optimal selling mechanism may involve price floors as well as price ceilings and minimal quotas of production Loertscher & Marx IIIO competition policy 20 Oct 2022 59 / 61 Conclusions Loertscher & Marx IIIO competition policy 20 Oct 2022 60 / 61 Conclusions Provide a framework in which one can rigorously define and analyse: Mergers in a procurement setup Buyer power Coordinated effects Mavericks Countervailing power Corroborate some perceived wisdom while challenging other Key refinements, e.g., to: Buyer power as a merger defense Maverick-based analysis Future research: Incomplete info. models for other aspects of merger review Extend concepts to standard oligopoly models Loertscher & Marx IIIO competition policy 20 Oct 2022 61 / 61