1. 1 Industrial Organization or Imperfect Competition Entry deterrence I Univ. Prof. dr. Maarten Janssen University of Vienna Summer semester 2012 Week 6 (April 26)

2. 2 Definition of entry deterrence Incumbent’s choice of business strategy such that it can only be rationalized in face of threat of entry Two different mechanisms often contemplated:  Building up capacity Studied in both Cournot and Stackelberg context  Choice of prices to signal (low) cost structure Context of game with asymmetric information First, discuss briefly Cournot and Stackelberg models  Then extensions to entry deterrence Later, pricing choices

3. 3 Capacity Expansion and Entry Deterrence Central point: For predation to be successful—and therefore rational—the incumbent must somehow convince the entrant that the market environment after the entrant comes in will not be a profitable one. How this credibility? One possibility: install capacity  Installed capacity is a commitment to a minimum level of output

4. 4 Cournot Model 2 (or more) firms Market demand is P(Q) Firm i cost is C(q) Firm i acts in the belief that all other firms will put some amount Q -i in the market. Then firm i maximizes profits obtained from serving residual demand: P’ = P(Q) - Q -i

5. 5 Demand and Residual Demand Market demand P(Q)=P(q 1,Q -i =0) q1q1 P(q 1 ) P(q 1, Q -i =10) P(q 1, Q –i =20)

6. 6 Cournot Reaction Functions Firm 1’s reaction (or best-response) function is a schedule summarizing the quantity q 1 firm 1 should produce in order to maximize its profits for each quantity Q -i produced by all other firms. Since the products are (perfect) substitutes, an increase in competitors’ output leads to a decrease in the profit-maximizing amount of firm 1’s product (  reaction functions are downward sloping).

7. 7 Cournot Model The problem Max{(P(q i +Q -i ) q i – C(q i )} defines de best-response (or reaction) function of firm i to a conjecture Q -i as follows: P’(q i +Q -i )q i + P(q i +Q -i ) – C’(q i ) = 0 Q -i qiqi qiMqiM qjqj r1r1 qi*(qj)qi*(qj) Firm i’s reaction Function Q -i =0

8. 8 Cournot Equilibrium Situation where each firm produces the output that maximizes its profits, given the the output of rival firms Conjectures about what the others produce are correct. No firm can gain by unilaterally changing its own output

9. 9 Cournot Equilibrium q2q2 q1q1 q 1 M =30 r1r1 r2r2 q 2 M =30 Cournot equilibrium q 1 * maximizes firm 1’s profits, given that firm 2 produces q 2 * q 2 * maximizes firm 2’s profits, given firm 1’s output q 1 * No firm wants to change its output, given the rival’s Beliefs are consistent: each firm “thinks” rivals will stick to their current output, and they do so!

10. 10 Properties of Cournot equilibrium The pricing rule of a Cournot oligopolist satisifes: Cournot oligopolists exercise market power:  Cournot mark-ups are lower than monopoly markups  Market power is limited by the elasticity of demand More efficient firms will have a larger market share. The more firms, the lower will be each firm’s individual market share and monopoly power.

11. 11 Concentration measures Different industries have very different structures and also different behaviours SCP paradigm  Structure (cost, entry conditions, number of firms)  Conduct (prices, product differentiation, advertising,etc.)  Performance (Lerner index (P-MC)/P, profit, welfare, etc.) Concentration measures try to provide indication of conduct and/or performance on the basis of structural features  Preference for one number representation  Use this for regression analysis (e.g. Lerner index on concentration measure)

12. 12 Different concentration measures C4 is sum of four largest market shares  Can’t be used in highly concentrated sectors such as in mobile telephony No difference between four firms with 25% market share and monopolist  Why 4? Market shares of 5 th, 6 th etc. largest firm has no effect HHI uses all information: sum of all squared market shares  Larger market shares get more weight

13. 13 “Justifying” HHI

14. 14 Changes in marginal costs

15. 15 Another look at Cournot decisions Firm 1’s Isoprofit Curve: combinations of outputs of the two firms that yield firm 1 the same level of profit A BC Increasing Profits for Firm 1 D Q1Q1 Q1MQ1M r1r1 Q2Q2

16. 16 Q2Q2 Q1Q1 Q1MQ1M r1r1 Q2*Q2* Q1*Q1* Firm 1’s Profits Firm 2’s Profits r2r2 Q2MQ2M Cournot Equilibrium Profits at Cournot equilibrium

17. 17 Cournot versus Bertrand I Predictions from Cournot and Bertrand homogeneous product oligopoly models are strikingly different. Which model of competition is “correct”? Kreps and Scheinkman model two stages  firms invest in capacity installation  then choose prices.  Solution: firms invest exactly the Cournot equilibrium quantities. In the second stage they price to sell up to capacity.  We discussed this implicitly when discussing capacity constraint Bertrand competition

18. 18 Cournot versus Bertrand II Cournot model is more appropriate in environments where firms are capacity constrained and investments in capacity are slow. Bertrand model is more appropriate in situations where there are constant returns to scale and firms are not capacity constrained

19. 19 Stackelberg Model 2 (or more) firms  Producing a homogeneous (or differentiated) product Barriers to entry One firm is the leader  The leader commits to an output before all other firms Remaining firms are followers.  They choose their outputs so as to maximize profits, given the leader’s output.

20. 20 Stackelberg Equilibrium Q1SQ1S Q2SQ2S Follower’s Profits Decline Leader’s Profits Rise Stackelberg Equilibrium r2r2 Q1Q1 Q1MQ1M r1r1 Q2*Q2* Q1*Q1* Q2Q2 Cournot Equilibrium

21. 21 Stackelberg summary Stackelberg model illustrates how commitment can enhance profits in strategic environments Leader produces more than the Cournot equilibrium output  Larger market share, higher profits  First-mover advantage Follower produces less than the Cournot equilibrium output  Smaller market share, lower profits

22. 22 Stackelberg Mathematics I Linear Demand and No production cost Stackelberg Follower’s Profit Stackelberg Follower’s Reaction Curve:

23. 23 Stackelberg Mathematics II Stackelberg Leader’s Profit Or, Optimal Output Leader: Is credibility used somewhere?

24. 24 Stackelberg with Fixed Entry Cost: Follower Q1Q1 Q2Q2 Follower’s Profits are High Follower’s Profits are Low With Entry Cost: follower’s profits in the market can be too low to recover entry cost Reaction Curve with Entry cost

25. 25 Follower’s decision with entry cost f Stackelberg Follower’s Profit (with α=β=1) Stackelberg Follower’s Reaction Curve: If π F ≥0, i.e., if (1-q L ) 2/ 4 ≥ f or q L ≤ 1 - 2√fq F = (1-q L )/2 Otherwise q F = 0

26. 26 Stackelberg with Entry Cost: Leader Q1Q1 r1r1 r2r2 Q2Q2 Q1SQ1S Stackelberg Equilibrium Optimal output

27. 27 Stackelberg with Low Entry Cost: Leader Q1Q1 r1r1 r2r2 Q2Q2 Q1SQ1S Stackelberg Equilibrium Entry deterrence is not optimal (accommodated entry)

28. 28 Stackelberg with High Entry Cost: Leader Q1Q1 r1r1 r2r2 Q2Q2 Q1SQ1S Stackelberg Equilibrium Monopoly Output is enough for entry deterrence

29. 29 When do the different cases occur? Leader’s profit of entry accommodation is 1/8 (as p = ¼ and its output is ½); follower’s profit is 1/16 – f. Leader’s profit of entry deterrence is 2√f(1-2√f) (as p = 2√f and [total] output is 1- 2√f);  choosing minimal output level to deter Entry deterrence profitable if 2√f(1-2√f) > 1/8, i.e., iff √f > ¼(1- ½√2)  0 < √f < ¼(1- ½√2) is too costly  ¼(1- ½√2) < √f < ¼ entry deterrence in proper sense (distort output decisions compared to monopoly decision)  √f > ¼ monopoly output to deter entry

30. 30 Is entry deterrence in Stackelberg context always bad? Welfare (TS) if entry takes place is ½ - 1/32 – f  Total output is ¾; price is ¼ Welfare (TS) if entry is deterred is ½ - 2f  Total output is 1-2√f; price is 2√f Thus, TS is higher under entry deterrence if f < 1/32 Entry deterrence is individually optimal for incumbent and takes place if (1- ½√2) 2 /16 < f < 1/32 Thus, entry deterrence is sometimes optimal from a TS point of view (entry can be excessive)