L6 - Collusion

Question 1

How is collusion defined?

Answer 1

Hard to define…

• “firm conduct intended to coordinate the actions of firms” Church & Ware (p.308)
• “firms … avoid competing with one another” Belleflamme & Peitz (p.336)
• “firms’ prices are higher than some competitive benchmark” Motta (p.138)

It may entail: agreeing on price levels, allocating sales quotas, mutual forbearance (not entering each other’s markets) …

How is an agreement reached?

• Explicit collusion: secret communication – meetings, phone calls, emails
• Tacit collusion: no communication – an understanding develops over time
  
  • one firm raises their price to signal that if they work together prices could be higher.

Question 2

Current Policy surrounding collusion?

Answer 2

• Article 101: prohibits cartels (prohibits explicit collusion)
  
  • easier to prove explicit collusion in a court of law (usually have some hard evidence)
    
    • Harder to do for tacit collusion and could get in the way of a market functioning properly
• Merger control: prevents coordinated effects (increased likelihood of tacit collusion)

Question 3

History of Cartels in Europe?

Answer 3

• How common is the collusion? –> from top graph
  
  • jump up in prosecution from the end of the 90s into the 2000s and beyond.
    
    • Optimistic view –> European commission got better at detecting cartels
      
      • Why?
        
        • In 1996 the EC created their leniency program - where the commission offers leniency to those who come forward and tell them they are part of a cartel - creating a prisoners dilemma for companies to tell on each other
    • Pessimistic view –> There are more cartels out there and the EC is just detecting the same proportion of cartels from a larger batch
• Fines have increased enormously overtime
  
  • The bottom graph is the total fine given out to all cartels in those 5 years
  • The left side is the average fine - which has increased by over tenfold over the period.

Question 4

What are the incentives that lead firms to collude?

Answer 4

• Two firms in the part
  
  • Where the lines represent their best response function given the price the other firm’s set
• Firm’s have the joint incentive to collude at the monopoly price
  
  • Both have individual incentives now to undercutting each other (to the price on their best response function) to increase their profits

Question 5

Example of Tacit collusion?

Answer 5

• a concrete example (Albæk et al, 1997)
• The Danish Competition Council (DCC): wanted to promote market transparencyWhy? Increase price sensitivity of consumers to intensify competition
• Danish ready-mixed concrete market:
  
  • Sellers offered individualised and confidential discounts to buyers
  • In 1993, DCC collected firm-specific prices for 18 production sites in 3 regions
  • Published list prices, average prices, average of 5 lowest prices (3 month delay) –> hoped this will lead prices to fall
    
    • Within a year: prices ↑ by 15-20 per cent & price differences had disappeared
• Was it simply a result of competition?
  
  • ● No substantial increase in demand or input prices
  • ● Inflation was only by 1-2 per cent over the period
  • ● No increase in prices in other regions where they hadn’t increased the prices
• The policy provided firms with the means of coordinating prices & detecting deviations –> tried to make prices transparent to consumers but also made them transparent to the firms themselves too
• Emphasis on promoting market transparency abandoned in 1997

Question 6

How can we create apply game theory to the collusion of firms on price?

Answer 6

• Have to assume infinite –> If its a finite game, the Nash equilibrium is to both deviate (and as this is the subgame perfect nash equilibrium) it will be the equilibrium for any repeated game of length less than infinity

Question 7

infinite repeated game approach: Present Discounted Payoffs?

Answer 7

Question 8

infinite repeated game approach: Nash equilibrium of the trigger strategy?

Answer 8

• The started discount factor is the ‘critical’ discount factor

Question 9

How does the baseline model change if the number of firms changes?

Answer 9

• as long as the discount factor is greater than the critical the firms would collide as a monopolist
• As n increases the critical discount factor tends to 1
  
  • This means the more firms in a market the harder it is to collude
  • Why do we get these results???

Question 10

Why does the number of firms affect their ability to collude?

Answer 10

i) the amount of monopoly profit each firm gets for colluding gets smaller –> price war is less of a threat as the firms would lose less profit in the end
ii) the short term benefit of deviating increases with more firms in the market

So there is always a greater incentive to defect when their are more firms

–> assuming firms are symmetric

Question 11

How will asymmetric firms affect collusion?

Answer 11

• as homogenous products –> consumers will buy from the cheapest so the deviation profits are the same
  
  • The only difference in the profits are from collusion
• As firms A’s share increases –> wants to collude more:
  
  • Its punishment becomes harsher and its short-term benefit is lessened –> less likely to deviation
• Firm B share decreases –> more likely to deviate:
  
  • Long term punishment falls and deviation benefit rises

B’s discount factor is the binding one –> firm’s more likely to collude the more symmetric they are

• Cartels should arise in markets with few firms that are relatively symmetrical

Question 12

What is the evidence for Cartels?

Answer 12

• graph of detected cartels in Europe (1990-2010)
  
  • Left axis (market share of the largest firm) against the market share of the second-largest firm
  • The diagonal line downwards is the duopoly line
  • Diagonal line upwards –> symmetric share between the two firms
• There are some firms that align with theory (few symmetric firms) –> yet many don’t
  
  • Dominants firms
  • many firms with large fringes
• Could be the case there are more firms that align with theory but they have just not been detected –> collusion is so easy that it is hard to detect
  
  • Only looking at explicit collusion –> maybe for a few symmetric firms tacit collusion occurs more often –> don’t need to form a cartel

Question 13

How does product differentiation affect collusion?

Answer 13

• Two effect work in opposite directions - so there is no rule on which will be larger
• Policymakers however believe that collusion is more likely when there is little product differentiation
  
  • Easier to coordinate on a collusive outcome when everyone is selling the same product –> not considered in the theories

Question 14

How does collusion and product differentiation fit into the hotelling framework?

Answer 14