Administrative Issues

1
Administrative Issues
2
Response to Feedback

• More cases
• Cases coming up
• GE vs. Westinghouse
• Bitter Competition
• African Communications Group
• Raytheon
• Lead time for ordering HBS cases such that we
  cant add HBS cases at this point.
• Will try to add some mini-cases.
• Also some philosophy here about value of
  deductive vs. inductive learning for business.
• Will post questions about HBS other cases (as
  did with GE vs. Westinghouse).

3
Feedback continued

• More student participation
• Okay, but two-way street you have to talk up,
  ask questions, point out connections, etc.
• Give more explanation and guidance about CSG
• Will talk about in a moment
• Give some guidance about the final exam
• Final will consist of
• Short problems to test understanding of
  theoretical ideas (1/3 roughly)
• Short answer questions about aspects of strategic
  situations (1/3 roughly)
• A case analysis (1/3 roughly)
• I have posted some examples of the type of
  questions Ive asked previously on the course web
  site (Exam questions from old exam ).

4
Some Things that are Tricky

• Be more clear when assignments are due
• but best laid plans
• Confusion over when case reader would be ready
• Technical problem with CSG
• Stochastic elements in pacing of course
• Switch to Catalyst
• Never used it
• Will look into it this weekend but no promises.

5
One Thing I Wont Do

• Distribute printed handouts lecture notes
• Not clear theres sufficient demand and not clear
  demand for what.
• Expensive (even at 6 slides/page, double sided,
  talking 720 pages per lecture on average).
• Logistically a pain
• Dear Haas Community,
  Unfortunately, BOTH Xerox digital copier /
  printers (rooms F-580 S-545) are presently
  DOWN. Please do not send print jobs to either
  machine until further notice. Service is pending.
  I apologize for the inconvenience.
• Yours truly,
• Kurt, the Copy Guy
• Fails the market test

6
CSG

• Grading
• Dependent on how you do given your situation.
• Based on your strategy memorandum
• Its intended to be experiential learning
• I want mistakes to happen because through their
  analysis we learn
• Remember Carter Racing you dont learn a lot
  looking at just your successes
• Clarification
• You produce to order

7
Avoiding the Bertrand Trap II Cooperation
8
How do Coke Pepsi Make Money?

• Coke and Pepsi sell essentially undifferentiated
  products
• Prices are widely known, often advertised
• There are no consumer switching costs
• No evidence of serious limits on capacity
• No evidence of cost advantages

9
Coke and Pepsi Recognize Repeated Interaction

• Suppose Coke forbears cutting price today because
  it knows Pepsi will follow suit tomorrow.
• Suppose Pepsi forbears cutting price today
  because it knows Coke will follow suit tomorrow.
• Tradeoff for Coke or Pepsi is forgoing a larger
  market share today in order to avoid the Bertrand
  trap tomorrow.

10
Method 6 Exploit Repeated Play

• If firms play repeatedly, then can use repeated
  play to sustain a form of cooperation on price
  known as tacit collusion
• No firm cheats (undercuts rivals) because this
  will trigger a price war in the future (e.g.,
  reversion to Bertrand competition).

11
To Cheat or Not to CheatThat is the Question
Looking just at today
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
1
time
time
12
To Cheat or Not to CheatThat is the Question
Now take into account the future!
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
1
time
time
13
To Cheat or Not to CheatThat is the Question
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
Smaller benefits today (because split market).
But positive benefits in future.
Benefit today
But Bertrand trap forever after.
1
time
time
14
To Cheat or Not to CheatMore Firms
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
1
time
time
15
To Cheat or Not to CheatHigher Interest Rate
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
1
time
time
16
Tacit Collusion

• Tacit collusion is easier to sustain when
• fewer firms (four or fewer if excess capacity)
• interest rate low

17
To Cheat or Not to CheatDying Industry
Cheat (undercut)
Cooperate (tacitly collude)
Expected PDV of profits
Expected PDV of profits
1
time
time
18
Dying Industries

• In fact, if death date known with certainty,
  then cooperation generally not sustainable at
  all.
• Backwards induction
• In last period there is no future period, so no
  punishment to deter cheating in last period.
  Hence cheating (Bertrand) in last period
• But then same is true of penultimate period and
  so on back to first period.

19
General Phenomenon

• Firm going bankrupt not paid by other firms that
  owe it money.
• Management problems when boss announces shes
  leaving.
• Basically dont let others know the end is coming.

20
The Issue with Detection
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
Detection occurs
1
2
time
time
21
The Issue with DetectionStochastic Discovery
(Demand Fluctuation)
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
Detection occurs
Possibly lost to mistaken price war
1
time
time
22
When Demand Fluctuates

• Play trigger strategies
• Sometimes to avoid temptation, firms dont charge
  maximum price during high-demand periods
• Evidence that gasoline refiners dont charge
  maximum price during summer, the high-demand
  season.
• If too much (unpredictable) variability in
  demand, then would have price wars too often.
• Hence, value of tacitly colluding is reduced.
• Relative cost of cheating today is reduced.
• So difficult or impossible to sustain tacit
  collusion.

23
Making Tacit Collusion Work
no
Tacit collusion not an issue
Incentive to cut price?
yes
no
Easy to detect price cuts?
Tacit collusion will fail the firms risk
finding them-selves in the Bertrand trap
yes
no
Can serious punishments be inflicted?
yes
no
Firms willing to punish?
Tacit collusion is sustainable in equilibrium
yes
24
Electronic Components Distribution Industry

• How do we assess the potential for tacit
  collusion in the electronic components
  distribution industry?

25
Making Tacit Collusion WorkElectronic Components
Distribution Industry
no
Tacit collusion not an issue
Incentive to cut price?
yes
no
Easy to detect price cuts?
Tacit collusion will fail the firms risk
finding them-selves in the Bertrand trap
yes
no
Can serious punishments be inflicted?
yes
no
Firms willing to punish?
Tacit collusion is sustainable in equilibrium
yes
26
Making Tacit Collusion WorkAirline Industry
no
Tacit collusion not an issue
Incentive to cut price?
yes
no
Easy to detect price cuts?
Tacit collusion will fail the firms risk
finding them-selves in the Bertrand trap
yes
no
Can serious punishments be inflicted?
yes
no
Firms willing to punish?
Tacit collusion is sustainable in equilibrium
yes
27
VS.
28
Exiting a Price War

• Need to signal that price war at end without
  engaging in illegal explicit collusion.
• American Airlines and the NYT
• Price leaders
• Traditional leaders are
• GM in automobiles
• American Airlines in airline industry
• Tesco is a price leader with respect to Asda and
  Sainsbury
• Public adoption of means for facilitating tacit
  collusion

29
Facilitating Tacit CollusionImproving Detection

• Firms want to make sure that
• cheating is detected promptly
• cheating is detected accurately
• Numerous devices to make this work
• public posting of prices
• simplified pricing
• e.g., GE and Westinghouse
• airlines per-mile pricing
• collection dissemination of prices (some
  antitrust issuesMaple Flooring Mfrs. Assn v.
  United States)

30
Making Punishments Severe
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
1
time
time
31
Making Punishments Severe
Cheat (undercut)
Cooperate (tacitly collude)
PDV of profits
PDV of profits
Increase the severity of the punishment
1
time
time
32
How to Make Severe

• Most Favored Nation Clauses
• MFN If cut price today, give refund to past
  customers.
• Note the other guy better adopt this too!
• Also contemporaneous MFN All customers get same
  price today (makes detection of price cutting
  easier)

33
How to Make Willing

• Build in doomsday devices
• Dr. Strangelove
• Meeting the Competition Clauses (MCC)
• state that will meet lowest price available
• just advertised policy
• or put into contracts (some antitrust issues)
• if rival cuts price, either honor clause (a
  reputational or contractual obligation) or suffer
  consequences.

We miss you Stanley
34
Tacit Collusion on Non-Price Dimensions to Lessen
Price Competition

• When tacit collusion on price would be difficult,
  firms can tacitly collude to maintain conditions
  that lessen price competition
• Generally, these are conditions that make one of
  the assumptions of the Bertrand model fail.
• Concept of market discipline.

35
Tacit Collusion on Non-Price Competition

• Raising search costs
• tacit agreements not to price advertise
• not locating outlets near each other
• Raising switching costs
• making products incompatible with rivals
• signing customers to long-term contracts
• Note As we will see, these can also serve to
  deter entry.

36
Tacit Collusion on Non-price Competition

• Restrict capacity
• Firms can tacitly agree not to expand capacity
• Note can be difficult to coordinate
• Think about GE vs. Westinghouse
• Industrial capacitor industry
• In reverse lead additive industry

37
Product Differentiation

• Tacitly agree to split market on non-price
  dimensions
• location non-overlapping territories (usually
  invites antitrust scrutiny)
• product space e.g., split market between
  high-end and low-end

38
Other Dimensions of Tacit Collusion

• RD
• (Non-price) advertising
• No poaching

39
Take-aways GE vs. Westinghouse

• Price fixing is illegal you can go to jail for
  it.
• Tacit collusion requires
• An ability to detect deviation
• Sufficient punishments
• Sometimes need to facilitate tacit collusion
• By improving ability to detect (e.g., multiplier,
  audits)
• By increasing punishments, including use
  doomsday devices (e.g., MFN)
• By serving as a price leader

40
Conclusions

• Recognizing repeated play can allow firms to
  avoid the Bertrand trap via tacit collusion.
• Tacit collusion is distinct from explicit
  collusion or pricing fixing, which is illegal.
• Tacit collusion works best when
• The number of firms is relatively small.
• The future is sufficiently important.
• Detection of undercutting is easy and not too
  subject to error.

41
Conclusions (continued )

• Firms can take steps to facilitate tacit
  collusion
• Making prices public
• Using MFN and MCC clauses to increase punishment
• Signal to each other through the press and other
  means
• But be careful Devices intended to facilitate
  tacit collusion can run afoul of the antitrust
  authorities.

42
Conclusions (continued )

• The logic of repeated games has many applications
  in business. Among them
• Payments to bankrupt firms
• Treatment of a leaving supervisor
• Quality assurance (see reading)
• Entry deterrence (upcoming)

43
List of some of the firms and industries mentioned

• American Airlines
• Asda
• Coca-Cola
• GE
• Pepsi
• Sainsbury
• Tesco
• Westinghouse

• Airline industry
• Electronic components distribution industry
• Industrial capacitor industry
• Lead additive industry
• Maple flooring industry