Econ 510. Monopolization I

Transcription

1 con 510 Monopolization 1

2 xclusionary practices practices carried out by a dominant firm with the aim of deterring entry or inducing exit in the same market or in related markets. Common time pattern: nitial aggressive phase in order to reduce (actual or expected) profitability of (actual or expected) competitors. Short-run sacrifice of profits Recoupment period: once (actual or potential) competition has been eliminated, exploitation of market power Higher profits in the long-run 2

3 Fresh attention: Liberalisation and privatisation processes have created asymmetric market structures with strong incumbents. Growing importance of sectors which exhibit network effects Very difficult topic in anti-trust: To distinguish exclusionary practices from tough competition is not an easy task 3

4 Plan of the lectures Predatory pricing Non-price monopolization practices Commitment to aggressive pricing Over-investment in capacity Product Proliferation Bundling and tying ncreasing rivals costs xclusive Dealing Refusal to supply Overinvestment in advertising. 4

5 Search for a theory McGee (1958): we should not expect predation to occur: Criticism to deep pocket arguments: why should the prey not be able to obtain further funds? Predation is inefficient (destroys industry profits): merging with rivals would be more profitable Yamey (1972) s counter-objections: 1. Predation discourages further entry (merging with an entrant would invite further entry) 2. Predation allows to buy rivals at lower prices (see also Saloner, 1987) But: lack of rigorous foundation to predation theory until the 80s. 5

6 Predatory pricing A dominant firm sets low prices in order to exclude rivals. A convincing economic theory of predation has appeared only recently: Reputation models Signal jamming models Financial models All these stories rely on asymmetric information. 6

7 Reputation models A price war today creates a reputation of being a strong and aggressive incumbent thereby discouraging future entry. Simple model: ncumbent monopolist active in a number of identical markets. Potential entrant in each market. f entry, the incumbent can adopt predatory pricing or not. Sequential entry. f information is complete, there is no effect of reputation building and predation will not occur. (Selten s chain store paradox- Finite Prisoner s Dilemma) 7

8 One-shot game: Complete information Mkt 1 time 1 ( M out,0) Fight in Accommodate M A 0 A P P P P A A (, ) (, ) 8

9 Complete information out in Mkt 1 time 1 ( M,0) Fight Accommodate P P A A (, ) (, ) out in Mkt 2 time 2 ( M,0) Fight Accommodate P P A A (, ) (, ) 9

10 ncomplete information The entrant does not know if the incumbent is weak or strong. A Strong incumbent always benefits from fighting. A weak incumbent fights entry at the beginning of the game to establish a reputation of being strong and discourage future entry. (Kreps and Wilson, 1982) x.: The aspartame market. 10

11 Signal jamming models The entrant has imperfect info on market profitability. t tests the market. The incumbent sets prices which are lower than optimal short-run prices: Why? To disturb the signal that the entrant receives from the test. To make the entrant believe that market is barely profitable, thereby inducing it to abandon entry or to reduce its scale of activity. (Scharfstein, 1984 Fudenberg and Tirole, 1986) 11

12 Financial predation Traditional story: The dominant firm is endowed with more abundant financial resources ( long purse ) than a small firm (recent entrant). The dominant firm can afford losses for a longer period. The large firm engages in a price war until the small firm is driven out of the market. Challenge to this view: The prey can obtain external financing during the predation phase. Anticipating this, the dominant firm will not start the price war. 12

13 Financial predation nformation asymmetries give rise to financial market imperfections: The financial contract must be designed in such a way that provides incentives to repay the loan. This may generate inefficiencies. xamples: loan offered if own resources are large enough (collateral); loan extension denied in case of poor performance Credit constraints: profitable projects are not financed 13

14 Financial predation But, financial contracts designed to alleviate agency problems create scope for predation. Price war small firm s profits the small firm does not accumulate enough own resources to obtain the loan exit. Attempt to reduce exposure to predation (ex. Loan offered even if scarce collateral) exacerbates agency problem. (Bolton and Scharfstein, 1990) 14

15 Modelling credit constraints Holmstrom-Tirole (1997) A risk-neutral entrepreneur needs to pay a fixed cost F to enter the industry (or to develop a project). Own assets are A: it needs to borrow F-A from a risk-neutral bank. ntrepreneur protected by limited liability. f financed, entrepreneur can: work diligently (high effort) on the project or shirk (low effort). f diligent, project succeeds with prob. p (revenue=r ) fails with prob. 1-p (revenue=0 ). f shirking, project fails with prob. 1, but private benefit B. ffort is not observable (or not verifiable): impossible to write a contract on it information asymmetry (with moral hazard) between bank and entrepreneur (capital market imperfection). Ass.: if no information asymmetry, investment would be made: pr>f 15

16 The optimal contract The entrepreneur s C (if he obtains S in case of success): ps> B The bank s P: p(r-b/p)> F-A The bank s lending decision depends on firm s assets: it offers the loan iff A A pr f A A a project with positive NPV is not financed (firm is credit constrained). F B 16

17 Financial predation Two firms: (incumbent), and (a recent entrant). They differ only in assets: has a long purse, has limited assets. Assume that both have incurred fixed cost for period 1, but not yet for period 2. Prey or accommodate 1st period payoff Pay F or out (financial contract) effort 2nd period payoff Stage 1: preys or accommodates entry. f preys, both get P ; if not, they get A > P >0. Stage 2: each firm either pays F or goes out of business. Stage 3: effort decisions. f high effort (and both paid F), both earn A with prob. p; if only one paid F, M 2 A with prob p. 17

18 Financial predation Assumptions: p A > F has own assets A > F (always able to finance the investment) s assets in the first period are 0 : its second period assets equal first period retained earnings. F A p A B p F P Since always invests and makes high effort, from stage 2 on, the game is as the financing model above: will be financed only if does not prey, A replaces R. But, does firm have an incentive to predate? Yes, if: M P A A p p Therefore, predation will occur if the future prospect of higher profits outweighs the current losses from predation. 18

19 Financial predation ncumbent s financial advantage assumed. f the recent entrant is financially strong, exclusion does not occur strong entrants can exert a pro-competitive effect!!!! 19

20 nnocent low prices With switching costs, network effects, learning effects, product complementarity, low prices may be consistent with healthy competition. x.: firm launching a new product sets low prices (even below costs) to form a critical mass of consumers. However, this argument is less likely to apply to a firm which is dominant in the industry. Dominance test alleviates the risk of misclassification. 20

21 Two-tier approach Proposed rule: 1. s there enough market power for recoupment? f predator is dominant, go to 2. lse, dismiss the case. 2. s there sacrifice of profits? P>AverageTotalCost (ATC): always lawful P<AverageVariableCost (AVC): presumed unlawful (burden of proof on defendant) AVC<P<ATC: presumed lawful (burden of proof on plaintiff) 21

22 Non-price exclusionary strategies Commitment to aggressive pricing in case of entry (stay): Over-investment in capacity Product proliferation Bundling.. ncrease of the rival s costs xclusive dealing Over-investment in advertising Refusal to supply.. 22

23 Over-investment in capacity (Dixit, 1980) The incumbent installs a larger capacity than a monopolist would do in the absence of threat of entry. Why? f capacity is scarce, incentives to undercut rivals are weak (demand attracted from rivals cannot be satisfied). Large capacity makes it credible that competition will be very tough in case of entry ntry discouraged. x.: DuPont and the Titanium Dioxide ndustry. nvestment choices must be irreversible! Difficult to distinguish from innocent investment. 23

24 Product proliferation ncumbent produces a high number of varieties (larger than short-run optimal). Why? f entry, the entrant forced to produce varieties close to the incumbent s ones. Low differentiation tough price competition entry unprofitable. Variety choice must be irreversible! xample: US market of ready-to-eat breakfast cereals. 24

25 Tying and Bundling Tying the purchase of one good (tying good) is conditional upon the purchase of another good (tied good). the tied product may be bought alone. x.: computer and OS. Requirement tying: goods sold in variable proportions Bundling x.: printer and toner; photocopy machine and after-sale services; mobile telephone and calls. two goods are sold together Pure bundling: goods available only together x.: newspaper and supplement; Mixed bundling: goods available also separately x.: fixed menu and order à la carte. 25

26 xclusionary bundling Bundling can be used by a dominant firm to extend its monopoly into an adjacent market or to protect its position in the home market. ndependent products 26

27 ndependent products Bundling the two products, the incumbent commits to aggressive pricing if entry, thereby discouraging entry. (Whinston, 1990) Market A Market B Why? n order to earn monopoly profits on market A, must sell also in market B. 27

28 Simple model Assumptions 1. Demand: D A 1 0 if p A v otherwise D B 1 0 if p B w otherwise 2. Costs: v>c A, w>c B >c B. 3. Sunk cost F to enter the market B. f entry, Bertrand comp. 4. c B c B >F 5. v c A -F B > c B - c B. 6. Bundling is irreversible and requires a fixed cost F B. 28

29 Simple model bundling no - bundling out in out in v c 0 A w c B F B v c v c F After entry decision, active firms name prices: f bundling and out: f no bundling and out: f no bundling and in: f bundling and in: B A ca cb F B v ca w cb v ca ( c c ) F 0 ( c c ) F 0 B B ~ p B ~ p v w p, A v p, p, A v p, v c B B p B c w B c B p c B B B 29

30 Cont. eq, after bundling and in ~ ~ ~ f w p v w p p p v, consumers buy good B from. f p p v, consumers buy both goods from. quilibrium: ~ p v c B p c B.. Consumers buy from An eq. where not exist: ~ p v c B has incentive to deviate and consumers buy from does p' ~ p v c B An eq. where exist: p c B has incentive to deviate and consumers buy from does not ~ p ' ' p p v v c B c A F B 0 by ass. 5 30

31 Sub-game perfect equilibrium bundling no - bundling out in out in v c 0 A w c B F B v c v c F B A c A c F ( c c ) F 0 0 ( c c ) F 0 B B B B B v c A w c B v c B A B rreversibility is crucial for exclusion. Given that in, s profits lower in case of bundling. f bundling reversed without costs and in a short time, in case of entry would reverse bundling. Anticipating this, would enter. 31

32 Bundling and Portfolio ffects Consumers benefit from joint supply. Bundling provides a competitive advantage over competitors. 32

33 fficiency reasons for tying Reduction of transaction costs for consumers. Quality improvement. Solution to information problems. 33

34 Price discrimination device Simple example: 1 s willingness to pay 2 s willingness to pay Good A 7 4 Good B 5 8 Bundle Without bundling: p A =4, p B =5 =18 With bundling: p=12 =24 34

35 Price discrimination device Requirement tying allows to sort consumers according to their intensity of use and make them pay accordingly. xample: copy machine tied to toner. Demand for toner measures intensity of use. Low price for the machine, high price for toner allows to price discriminate. Welfare effects of price discrimination in general ambiguous: More likely to increase welfare when it attracts new consumers. 35

36 Conclusion Per se prohibition not justified. Rule of reason approach: valuate whether exclusionary effect exists: rreversibility? less likely if complementary products. ntry in one market affects success of entry in another market? valuate existence and importance of (potential) efficiency gains. 36

37 xclusive dealing Contract that commits a firm to deal exclusively with some vertically related firms but not with others. xample: a dominant seller prohibits buyers from dealing with competitors A dominant seller commits to deal exclusively with one (or some specific) vertical related firm(s) xclusive dealing may allow a dominant firm to deter efficient entry: Traditional argument. Chicago school critique. Recent theories. 37

38 Traditional argument Anti-competitive D c c B f D, no entry: firm has nobody to sell to. 38

39 Chicago school critique Why does the buyer agree on exclusivity? f B rejects exclusivity: p m A entry; p*=c ; B =CS(c ); =0 c m C D f B accepts exclusivity: no entry; p*= p m ; B =CS(p m ); = m B s loss = CS(c )- CS(p m )=x*> m = s gain 39

40 Chicago school critique The incumbent cannot (profitably) compensate the buyer to elicit acceptance. The incumbent cannot (profitably) use D to deter entry. D will be signed only if mutually beneficial fficiency considerations explain the use of D. 40

41 Rent extraction (Aghion and Bolton non-stochastic version) An incumbent can use exclusive deals to extract rents from entrants. A simple example: nelastic demand, q =1. B c c, 41

42 Game: 1. offers an exclusive deal with (x, d, w ), where: x = compensation; d = penalty (liquidated damages ) if deal terminated w = price commitment. 2. Buyer B accepts or rejects. 3. decides on entry. 4. f entry, decides p (and if no deal, chooses p.) 5. B decides on termination (if had signed), or on supplier (if free ). Note. Here the buyer is final consumer with willingness to pay v and unit demand. 42

43 P v f buyer rejects, enters and buyer buys at p c. C Any contract should leave buyer with at least: cs v B c C 1 9 f buyer accepts (x, d, w ), it switches to only if: p d p w w d (or: ). ntry occurs only if p. c 43

44 ncumbent maximises its profits, by offering: x*=0, d*=c -c, w *=c. Buyer makes cs B =v -c ; entrant makes zero profit; incumbent makes c c ( d *). The incumbent finds it optimal to allow entry and use the C and the penalty to extract the efficiency rent associated with entry. n this model, entry is pre-empted only if s cost is stochastic and makes mistakes in predicting s costs. 44

45 Contracts as a barrier to entry (Aghion and Bolton, AR 1986) Assumptions: ncumbent has cost c =1/2 B s valuation: v=1 (unit demand) Potential entrant : c unif. distr. in [0,1]. xclusive deal (p,p o ): B will buy from at price p, but: it can buy from if pays "liquidated damages" p o. The game: t₁: firm offers (p,p o ) to B, who accepts or rejects t₂: firm decides on entry and sets price p. (f no contract, chooses its price p) t₃: payoff realisation. 45

46 No exclusive contract f c <1/2, enters, sets p =1/2 and gets all f c 1/2, no entry, sets p=1 Prob. of entry: f=pr(c 1/2)=1/2 Buyer's surplus if entry: v-p =1-1/2=1/2. With probability (1-f), B has surplus v-p =0. B's expected surplus: (1/2)f+(1-f)0=1/4. 's expected payoff: (0)f+(1-f)(1-1/2)=1/4. 46

47 xclusive contract B buys from if: p +p o p: if it enters, sets p =p-p o. Prob. of entry with contract: f =Pr(c p-p o )=p-p o. ncumbent's problem: max p,po =f p o +(1-f )(p-1/2) s.to:1-p 1/4. [B accepts only if better off than no contract (=1/4)] max po s.to p 3/4, (p*,p o *)=(3/4,1/2). Hence, firm enters with prob. f = p*-p o *= 1/4. 47

48 ffects of exclusivity ntry efficient if c 1/2, but occurs under the contract only if c 1/4 welfare loss for 1/4<c 1/2 Does offer this contract in equilibrium? Yes: =(1/4)(1/2)+(3/4)(1/4)=5/16>1/4. When very efficient, prefers not to deter entry (it extracts some of 's rent via t). 48

49 fficiency reasons for D Protect and encourage relation-specific investments x.: stimulate producer s investments into retailers services (free riding). Promote retailer loyalty x.: encourage the retailer to tailor its promotional efforts towards the manufacturer s product. Maintain the value of the product x.: producer of a luxury good commits not to sell it through supermarkets not to damage the image of the good. 49

50 Conclusion Per se prohibition not justified. Rule of reason approach: valuate whether exclusionary effect exists: Buyer s fragmentation? less likely if upstream competition intense. Less likely if downstream competition intense. mportant scale economies upstream? valuate existence and importance of (potential) efficiency gains. 50

51 ncreasing rivals costs A vertically integrate firm may refuse to supply a key input to a downstream rivals. Deny inter-operability to a rival network. Over-investment in advertising. 51