THE IN-BETWEEN MARKETS MONOPOLISTIC COMPETITION AND OLIGOPOLY

1

In-Between Markets • Perfect competition and monopoly are idealized paradigms • Market power refers to a firm’s ability to set the price of the good • In Perfect Competition firms are price takers. We say they have no market power • In Monopoly the firm sets the price, constrained only by the market demand curve. In this market the firm has the most market power

• Most markets do not fit exactly into perfect competition or monopoly • There is a continuum of markets between these two paradigms classified by how much market power firms in the market have • Two additional paradigms are oligopoly, which refers to a market with only a few firms, and monopolistic competition, which refers to a market with many firms, but with products that differ somewhat • Firms in these markets also have some market power 2

Market Structure depends on four key factors • The number of sellers • The number of buyers • Entry conditions • The degree to which consumers see the products of different firms as substitutes (called product differentiation) • Product differentiation exists when products possess attributes that make consumers think the product from one firm is not a perfect substitute for the product of another firm • Most products are differentiated • • • •

Coke and Pepsi taste different Cars have different styles, power and accessories Grocery stores carry different items or cater to different clientele Restaurants cater to different tastes and budgets 3

Classifying Market Structure Degree of Product Differentiation Firms produce identical products Firms produce differentiated products

Number of Firms Many

Few

One

Perfect Competition

Oligopoly with Monopoly homogeneous products Monopolistic Oligopoly with Competition differentiated -----------products

• Mostly focus on monopolistic competition and oligopoly with homogeneous products. We may touch on oligopoly with differentiated products • The key difference between market structures is what the demand curve facing the firm looks like 4

Monopolistic Competition • Product differentiation: consumers do not see the products of different firms as perfect substitutes • Differences can be based on real attributes, like taste, quality, location or service • Differences can be more perceived that based on attributes (bleach)

• Product differentiation creates brand loyalty, a willingness of customers to pay a higher price to purchase a specific good in the product class • A measure is the cross price elasticity of demand for products in the same group. Strong brand loyalty means the cross price elasticity is small; a big difference in the prices of goods in the same product group does not causes customers to go to the cheaper brand, weak brand loyalty means the cross price elasticity is large • The stronger brand loyalty, the less two goods in the product group are seen as substitutes • For most customers, there will be less brand loyalty between a Toyota Rav-4 and a Subaru Forrester than between a Chevy Impala and a Toyota Rav-4 5

Brand loyalty and the firm’s demand curve • Result of brand loyalty is that firms face downward sloping demand curves • If a firm raises its price, the customers “most loyal” to the brand will stay with it, while those “least loyal” move to a cheaper brand

P

• The stronger brand loyalty is over-all, the steeper the demand curve facing the firm • With more alternative brands, the demand curve facing the firm is less steep – brand loyalty erodes • D1 shows weak brand loyalty, D2 shows stronger brand loyalty

D1

D2

Q

6

The Firm under Monopolistic Competition • Once we have downward sloping demand, we have a situation “like a monopoly” • Firm has MR0 as shown, what will happen? Why?

Price MC AC

P1 C1

d q1

MR

Quantity

7

The Firm under Monopolistic Competition- LR • Firm in the are making positive profit (red demand curve)

Price

• P1>C1 at q1

• Firms enter the market • Market demand is shared among more firms P1 • Demand facing the firm shifts in (it P2=C2 could also become flatter) C1 • Entry stops when profits=0

MC

MR1

• P2=C2 at q2

• LR equilibrium • MR=MC for each firm • Profit=0 for each firm

AC

MR2

q2

D1

D2

q1

Quantity 8

Market Dynamics in Monopolistic Competition • Increase in demand  Firms make profit • Entry occurs, until profits are dissipated • More firms, larger firms (go back to previous graph)

• Increase in demand  Firms make loses • Exit occurs, until profits are back to zero • Fewer firms, smaller firms (go back to previous graph)

• Cost Decreases  Firms make profits • Entry occurs, until profits are back to zero • Fewer firms, smaller firms (go back to previous graph)

• Cost increases  Firms make losses • Exit occurs, until profits are back to zero • Fewer firms, smaller firms (go back to previous graph) 9

Product Differentiation and non-price competition • Brand loyalty is measured by a willingness of customers to pay a price premium to get a particular brand • Promotional activities (advertising) are designed to increase brand loyalty, and thus shift out the demand curve facing the specific firm • But advertising and promotion of a good is expensive, hence it raises costs • Firms will advertise if the (expected) gain in market power – by a shift out in the demand curve it faces, and hopefully it getting steeper – enables it to raise price by more than the per unit cost of advertising • Enables profit in the short run • But the profit attracts entry or counter advertising by competitors • Long run result is zero profit 10

Optimal Product Differentiation • A group of closely related products that are substitutable • There are n firms competing in the market, each with its own attributes, ai • The product’s attributes affect its demand qi(pi, Pn, ai, An) here pi is the firms price, ai are its attributes, and the capital letters with subscript n indicate the prices and attributes of other firms in the

• Attributes also affect its costs

Ci(qi, ai)

11

First order conditions • Firm i’s profit: i = piqi – Ci(qi, ai) • First-order conditions for a maximum:

(1)

 i pi Ci  pi  qi  0 qi qi qi

(2)

 i qi Ci  pi  0 ai ai ai

• First condition (1) is to set output where MR=MC. Since firm is a pricesetter, choosing quantity is the same as choosing price • Second condition (2) is to set attributes so the marginal revenue of the attributes (the first term) equals the MC of the attribute 12

Hotelling’s Beach: Location as an attribute • Ice cream stands located on a beach • Demanders are located uniformly along the beach • One at each unit of beach

• Ice cream cones are costless to produce • But carrying them back to one’s place on the beach results in a cost of td 2 • t = temperature • d = distance

• Applicable to fast food, gas stations, lots of things, including non location attributes, which is why products of different producers tend to be similar

13

15.5 A’s demand

0

B’s demand

A

x

B

L

• Ice cream stands A and B are located at points a and b along a beach of length L. The consumer who is indifferent between buying from the two stands is located at x. Consumers to the left of x buy from A and to the right buy from B. • A person located at point x will be indifferent between stands A and B if pA + t(x – a)2 = pB + t(b – x)2 where pA and pB are the prices charged by each stand • Solving for x we get

b  a pB  p A x  2 2t (b  a)

• If the two stands charge an equal price (pB=pA), the indifferent consumer is located midway between a and b 14

Firm’s problem: Choose location to make the highest profit • • •

Since there are no cost to production, getting the highest price maximizes profit Hence, want to locate where they get the biggest demand Suppose firm A is the first to locate – wherever it locates, it gets the entire beach A’s demand demand.

0

A

L

• Where should B open? A’s demand

0

B’s demand

AB

L 15

Where is the long run equilibrium? • If firms can move location, they will end up right next to each other, in the center • This is why you tend to see gasoline stations located near each other • And car styles look similar • And fast food looks similar

• Competition doesn’t just push prices of different firms together, it also pushes product attributes together A’s demand

0

B’s demand

AB

L 16

Comparison to Perfect Competition • Perfect competition and monopolistic competition, long run profit is zero. No entry because there is no incentive • Customers pay more for the differentiated products (Pm) than they would for the identical products in perfect competition (Pc), even if the costs of production are the same. • LR equilibrium in perfect competition price equals minimum long run average cost, but not in monopolistic competition, so monopolistic competition is not fully efficient • QmMC, so there is dead weight loss

Pc

Qc

17

Oligopoly • Oligopoly is competition between a few firms • Sellers are large relative to the market • But no single seller can meet the entire market demand • It means firms are price-setters

• In pure theory, products are identical or nearly identical, so consumer buys from whomever has the lowest price • Buyers go to the lowest price, until it is no longer available from that firm • Then they buy from the next lowest price • If a firm raises it price, it loses a lot, but not all of its sales

• Expect that there are barriers to entry, so firms in the market can make positive profit without causing entry • Usually analyzed with game theory, but we will do some simpler models 18

The Kinked Demand Curve model of Oligopoly • A model of individual firm • Market “starts” with a firm charging P0 • If it lowers price, competing firms follow the price drop, so it attracts only a few new customers (new to the market) – demand is steep for a price cut • If it raises price, competitors don’t follow, and it loses a lot of customers to competitors – demand is flat for a price increase • Causes a “kink” in the demand curve, and a discontinuity (vertical portion) to the MR curve

19

Cost changes with kinked demand • Since the firm is producing Q0 it must mean that MR=MC at that output • So it must MC must cross the vertical portion of the MR curve • This means that costs can change without changing output or price • Give price stability in the market • Only large cost changes (say to MC3) leads to instability and a new price • This model does not explain how the market gets to stable price

MC3

MC2 MC1

20

Price Wars: Competition for market share • Alternative to kinked-demand curve model, is that firms choose price simultaneously • Bertrand model of price competition • Two identical firms producing identical products (often simplify by assuming a constant MC=c) • Each chooses a price for its product • Sales go first to the firm with the lowest price • Sales are split evenly if the price is the same

• If my price is higher than my competitors, I will lower price to meet or beat its price • What will the equilibrium be? Hint: think of price as a product attribute and the Hotelling model 21

Equilibrium in the Bertrand model • P0, get entry or advertising (raising costs) until profit=0 • Notice, MC>MR, so cannot be profit maximizing

22

Cournot Duopoly Model: A start of game theory • Each firm chooses its output assuming other firms will not change market price, but knows its choice changes market price • P(Q) is the market inverse demand curve • Q   qi is the market output, where i indicates specific firms i • So for the firm we have  i  P( qi )  c(qi )

• To maximize profit

i

 i  P  Q   P '  Q  qi  c 'i (qi )  0 qi

• So for the firm, MR=MC • Price>MC, so unlike perfect competition, there is dead weight loss 23

Market dynamics Cournot model • As each firm makes an output decision, it changes market price, and the other firms react by then changing their output • We can understand this better in a two-firm model • Firm A goes first, and starts as a monopoly, so produces where MR=MC (a constant to keep it easy)

24

Now add firm B • Because A is making a profit, B enters the market, maximizing its profit off the residual demand – the part of market demand A is not meeting • Now A can only charge P2 instead of the Pa it was charging • So A gets to react to firm B’s output • Equilibrium is where

Pa

• When B produces Qb, A maximizes its profit by producing Qa • When A produces Qa, B maximizes its profit by producing Qb 25

An algebraic model of reaction functions • P=300-Q where Q=Qa+Qb • Firms assume other firm’s output is fixed, let MC be constant =30 • Residual inverse demand for Qa is thus given by P=(300-Qb)-Qa so MRa =(300Qb)-2Qa • Profit is maximized where MC=30= MRa =(300-Qb)-2Qa • Solving for Qa we get Qa=(270-Qb)/2 • This is firm A’s reaction function to firm B’s output

• By symmetry, the reaction function for firm B to A’s output is Qb=(270-Qa)/2 • Cournet Equilibrium solves the two reaction functions simultaneously • Solution is Qa=Qb=90, P=120, each firm makes a profit of 8100. 26

Stackelberg Model: Trying to lead the market • Since firm’s know they react to other firms output, they can try to game the market by using the other firm’s reaction function • Leader chooses its output first, know the other firm will react • Substitute B’s reaction function into A’s problem and solve for A’s best strategy • Result is A makes profit of 9112.5, while B makes profit of 4556.5

• But if A tries to lead, why doesn’t B – result would be the Bertrand equilibrium, both make zero profit

P  300  Qb  Qa so 270  Qa P  300  Q a 2  165  0.5Qa  MRa  165  Qa MR  MC  30  165  Qa  Qa  135 and Qb  67.5

27

15.1

Comparison of Oligopoly to Perfect Competition and Monopoly Price PM PA MC=AC

PC

D

MR QM

QA

QC

Quantity

Market outcomes under imperfect competition depend on how firms behave, and of course on market demand. In the graph we assume MC is constant, to facilitate comparison. The Bertrand equilibrium is at point C, which is the same as what we get in a perfect competition. A monopoly would be at point M. Almost any point in-between is possible depending on how individual firms behave, and how they play “the game of competition.” A Cournot equilibrium could be at a point such as A. The deadweight loss given by the shaded triangle increases as one moves from point C to M. 28

Some Brief Game Theory: Prisoner’s Dilemma

Suspect 1

• Game theory studies strategic behavior between competitors • A classic game is the Prisoner’s dilemma. Two crooks get caught, but there is not sufficient evidence to convict them • If one confesses and the other doesn’t, the one who confesses gets a off, the other a bad deal • If both confess, they get a lesser deal • If neither confesses, they get convicted of a lesser crime

Suspect 2 confess

don’t confess

confess

U1=1 U2=1

U1=3 U2=0

don’t confess

U1=0 U2=3

U1=2 U2=2

The Payoff Matrix shows years of freedom over the next four years of different outcomes

29

Prisoner’s Dilemma

Suspect 1

• Looking at the payoffs, we might think both would be to not confess, we each get 2 years of freedom • Suppose I am suspect 1. If my partner in crime isn’t going to confess, my better move is to confess, then I get 3 years of freedom • If he is going to confess, I will also confess, giving me 1 year of freedom rather than 0 • Symmetry means suspect 2 has the same choices • Confessing is the dominant strategy (it is always the best choice)

Suspect 2 confess

don’t confess

confess

U1=1 U2=1

U1=3 U2=0

don’t confess

U1=0 U2=3

U1=2 U2=2

The Payoff Matrix shows years of freedom over the next four years of different outcomes

30

Golden Balls split

steal

split

U1=1/2 U2=1/2

U1=0 U2=1

steal

U1=1 U2=0

U1=0 U2=0

Player 1

• Pot of money • Players choose to split or steal • Payoff matric shows the share of each decision • If player 1 splits, player 2 should steal • If player 2 splits, player 1 should steal • But both better off if split than if both steal • What would you do?

Player 2

https://www.youtube.com/watch?v=yM38mRHY150 https://www.youtube.com/watch?v=S0qjK3TWZE8 31

Nash Equilibrium

• Given the other person’s choice, neither suspect should change • Not true for the other cells

• Whenever there is a dominant strategy for both players, it will be a Nash equilibrium • Equilibrium strategy results in longer jail terms

Suspect 1

• A Nash Equilibrium involves strategic choices that, once made, provide no incentive for either player to change • It shows the best choice for each player, given the other player’s choice • Confess, confess is a Nash equilibrium

Suspect 2 confess

don’t confess

confess

U1=1 U2=1

U1=3 U2=0

don’t confess

U1=0 U2=3

U1=2 U2=2

The Payoff Matrix shows years of freedom over the next four years of different outcomes

32

Battle of the Sexes Husband (2)

• A couple needs to make a choice between activities • Both prefer time together to doing nothing • Now there are two Nash Equilibria

• Mixed cells are not Nash equilibria. If they are in a mixed cell, both would have an incentive to change

Baseball

Ballet

U1=2 U2=1

U1=0 U2=0

Baseball

U1=0 U2=0

U1=1 U2=2

Wife (1)

• If the wife chooses ballet, the husband will also choose ballet • If the wife chooses baseball, the husband will also choose baseball • Neither has an incentive to change, given the other’s choice, if they are in top left or bottom right cell

Ballet

The Payoff Matrix shows enjoyment of different choices. If not the same, they do nothing.

33

Prisoner’s Dilemma in the Duopoly • We can use game theory to understand the Cournot duopoly model • If firm A behaves Cournot, firm B should behave Stackelberg • If firm B behaves Cournot, firm A should behave Stackelberg • If my opponent behaves Stackelberg, I should behave Cournot • Two off diagonals are Nash equilibria 34

Measuring Market Power • Perfect competition are two ends of a continuum of market power • Monopolistic competition and oligopoly fall in-between • Like perfect competition and monopoly, they are idealized paradigms of markets • Most markets have characteristics of both, with “not too many” firms, and with some degree of product differentiation • Thus, understanding how much market power (price setting ability) individual firms have is incomplete

• Economists and policy makers have come up with several measures of market power. The three most popular are: • n-firm concentration ratio ( a measure of market share) • Herfindahl Index (another measure of market share) • Lerner’s Index (a measure of market inefficiency based on MC and price)

• Market share indices assume large market share implies price setting power 35

n-firm Concentration Ratio • Usually denoted Cn, the n-firm Concentration Ratio adds up the shares of the n largest firms in the market. Shares can be defined over • • • • •

Output Profit Revenue Sales (units) Other

• The more concentrated the industry, the larger will be Cn • Small values indicate more concentration • The most common measures used are C2 and C4 36

n-firm Concentration Ratio, continued • Some example values of C2

• Wheat – 0 (in fact, the value for this industry to large n is 0) • Breakfast Cereal – 0.66 (1992) (Kellogg had about 38%, General Mills 28%) • Aluminum – 0.97 (1945) (Alcoa had 92% share, Reynolds a 5% share)

• It does not account for the entire market • One very large firm with a “competitive” fringe may not be as concentrated, or may be more concentrated, than an industry with several large, but not very large firms • • • •

Example 1: Firm 1 has 80%, 20 firms each have 1%. C4=0.84 Example 2: Firm 1 has 40%, firm 2 has 38%, firms 3 and 4 have 3% each. C4=0.84 Which market is more concentrated? What firm has the most market power? 37

Herfindahl Index • Compensates for the n-firm concentration ratio by using the entire market • Let si be the market share of firm i. The Herfindahl index, H, is give 2 2 2 2 H  s  s  s  ...  s by i 1 2 n

 i

where there are n firms in the market • A monopoly has a value of H=1 • A perfect competition, with many firms having miniscule shares, has a value of H0 • Values in-between 0 and 1 indicate different measures of market power derived by skewed market shares 38

Herfindahl Index, continued • An example, breakfast cereal: Kellogg had 0.38 so 0.382=0.1444. Similarly for General Mills 0.282=0.0784. • If two additional firms each have 17% share (covering the remaining 34%) we have H=0.1444+0.0784+0.0289+0.0289=0.2806 • If the remaining 34% is split so one firm has 14% and 10 firms have 2% each, H=0.1444+0.0784+0.0196+10*0.0004=0.2464. The additional fringe increases the competitiveness of the market

• A problem is comparing across values • Is 0.2464 significantly more competitive than 0.2806?

39

Comparing Measures -- equal sized firms • One way to compare different values is to compute the number of equalsized firms that give the same value • If H=x for a firm, then we can find m equal-sized firms that give the same value by the formula H=ms2 constraining for the fact that ms=1 because the sum of all shares must be the full market • Notice since ms=1we have that s=1/m so H=mm-2=m-1 which means m=1/H • Example, if H=0.2806 m=1/0.2806=3.5634 firms. • if H=0.2464 m=1/0.2464=4.058 firms.

• Can do the equal-sized firms for Cn as well. • m=1/[Cn/n]=n/Cn • For cereal, C2=0.66 so m=2/0.66=3.03

• These tell us about concentration of market share, but not about market power – the ability to set price above MC (or MR) since  max  MR=MC 40

Lerner’s Index – a direct measure of market power P  MC MC  1 • Lerner’s Index has the form LI  P P

• Since for  max firms MR=MC

 1 MR  P 1    e  p    1 P 1   e p  LI  1  P

    11 1  1 ep ep

because e p  0 • This is the Lerner’s index for each firm, and it is equivalent to the markup over MC that we talked about earlier 41

Lerner’s Index, continued • To the extent that more competitors makes the demand for a specific firm more elastic, Lerner’s Index tells us how much market power each firm has • To understand a market, we need to aggregate Lerner’s index. Let a subscript i be the Lerner’s index for firm i. Then

qi P qi P si Q P si si 1 LI i       e pi P qi P Q P Q ep ep because qi  si Q and qi  Q and e p is the market elasticity of demand

42

Market Lerner’s index • The firm’s Lerner’s Index tells the market power of individual firms • It does not tell about the market power of all the firms in the market • The weighted average of the firm Lerner’s Indices tells the average market power of firms in the market 2

si si H LI   si LI i   si   ep ep i i i ep • So market LI equals the Herfindahl index divided by the market elasticity of deman 43

Department of Justice use of Indices • The DOJ generally uses a measure of the HHI which is what we computed x 10000.

• Unconcentrated Markets: HHI < 1500 (our 0.15) • Moderately Concentrated Markets: 1500 < 2500 (our 0.25) • Highly Concentrated Markets: 2500 < HHI • Transactions that increase the HHI by more than 200 points in highly concentrated markets are presumed likely to enhance market power under the Horizontal Merger Guidelinesissued by the Department of Justice and the Federal Trade Commission • http://www.justice.gov/atr/public/guidelines/hhi.html

44

DOJ standards about mergers • Small Change in Concentration: Mergers that increase the HHI less than 100 points are unlikely to have adverse competitive effects • Unconcentrated Markets: Mergers resulting in unconcentrated markets are unlikely to have adverse competitive effects • Moderately Concentrated Markets: Mergers resulting in moderately concentrated markets that increase the HHI by more than 100 points potentially raise significant competitive concerns and often warrant scrutiny. • Highly Concentrated Markets: Mergers resulting in highly concentrated markets that increase the HHI between 100 points and 200 points potentially raise significant competitive concerns and often warrant scrutiny. Mergers resulting in highly concentrated markets that increase the HHI by more than 200 points will be presumed to increase market power 45