Slides - Competition Policy International

David S. Evans
University of Chicago, Global Economics Group
Elisa Mariscal
CIDE, Global Economics Group
Topic 9 | Part 2
5 September 2013
Your CPI teaching team
Professor David S. Evans
[email protected]
Professor Elisa Mariscal
[email protected]
Teaching assistant
Mr. Alexis Pirchio
[email protected]
Course Support
Ms. Meara Hamidiani
CPI Community Manager
[email protected]
Part 1
Part 2
Replicating the
monopoly outcome
with a Cartel
Explicit collusion:
Detecting cartels
The incentives to cheat
Discouraging cartels:
Leniency programs
Methods to detect and
punish cheaters
Punishment and
Factors that make
collusion easier
Cooperating for
Guest Speaker:
Screening techniques
Explicit Collusion: Detecting
Explicit vs. Tacit Collusion
Explicit collusion is the result of an agreement to fix price, divide
markets, or otherwise achieve a non-competitive outcome. It
requires direct communication between the cartel members.
Tacit collusion refers to market interactions between firms that
encourage them to keep prices high and discourages them from
competing. For example, if all firms believe that a reduction of price
will trigger a price war in which they will lose, none will reduce price.
In the US and most jurisdictions only explicit collusion is unlawful. In
effect, the law doesn’t prohibit “collusion” so much as it prohibits the
use of explicit communication to facilitate that collusion.
We focus on explicit collusion but some of the economic analysis is
equally applicable to tacit collusion.
Caught in the Act
Sales meeting, huh? (Laughs).
Well, it's easy to know the price, so everything's clear.
And one from FBI. (Laughs).
(ui) joke.
And seven from the FTC.
Yeah. FTC. (Laughs).
Seven from FTC. (Laughs).
Well, I also feel cautious in the way I came into this, uh, hotel. I
actually... (Laughs). (ui).
In Europe, we...
You know and, uh, (ui).
Even in Europe now, they're cracking down.
Everyone has to be careful.
Oh, yes.
When we leave, we're better off to leave separately
Detection of Explicit Collusion In Practice
An entrant who hasn’t been part of the cartel is solicited and reports
this to the authorities.
Customers become suspicious about the prices they are being
charged or the bids they are getting.
Disgruntled employees who know about the cartel report it.
The authorities become suspicious because of institutions or contract
features that encourage collusion or punish cheaters.
Members of cartel approach authorities as part of a “leniency
program” which reduces penalties on those who report it.
What Can Competition Authorities Do?
Head them off at the pass
•Forbid/limit practices that facilitate cartels (Information exchange, MFNs, RPM—
but are these practices ever efficient?)
•Prohibit mergers that will facilitate collusion (e.g. ones that eliminate a maverick
player or increase the likelihood of coordinated effects)
Hunt them down
•Focus investigative efforts on industries that have features that facilitate
cartels—but are these cartels the most damaging or the easier ones to detect?
•Screening technologies (our guest speaker will elaborate)*
•Dawn raids when enough suspicion—requires specialized know-how and can
be legally challenging in some jurisdictions
Get them to squeal
•Encourage cartel members to report each other (leniency programs)
•Get employees to be “whistleblowers”
Carry a big stick
Head Them Off At The Pass: Facilitating Practices
A facilitating practice is an activity that makes collusion more likely
or more effective, either by making coordination easier or making
it easier to sustain a collusive agreement.
A facilitating practice may be prosecuted either as an anticompetitive agreement in and of itself or as circumstantial
evidence of price fixing.
Some examples include:
• Information exchange of current individual firm sales or prices
• Announcements of future price changes
• List (or posted) pricing with no discounts
• Most favored consumer clauses, meeting competition clauses
• Delivered or basing point pricing
Information Exchange Agreements Help Collusion
Exchanging future pricing intentions
help coordinate price
Exchanging past prices help police
price agreement
Exchanging demand and cost data so
that firms have:
• A more common set of beliefs (this
makes it more likely that, without express
communication, they can settle on the
same collusive price)
• More precise estimates of demand
Information Exchange Agreements: Cons
They can circumvent cartel prohibition
• In the UK Cartels became illegal with the
Restrictive Trade Practices Act of 1956…
Subsequently, as much as 50% of British industry
adopted information exchange agreements.
They can be a creative way of monitoring and
enforcing a price fixing agreement, e.g. Maple
Flooring Association (US)
• 1916-21: MFA establishes a minimum RPM for the
sale of maple and birch flooring supported by
severe financial penalties for deviators.
• Jan 1921: After a government investigation, MFA
abandons the price floor policy, and institutes an
information sharing system.
• Supreme Court ruled in favor of the defendants–
the DOJ did not prove the companies used
information to collude in prices and production.
Information Exchange Agreements: Pros
Sharing past sales and profits can be used for
benchmarking and establishing relative
performance contracts.
Sharing past prices and sales allow firms to
more accurately estimate current demand.
•Industry profit and welfare effects are ambiguous.
Knowledge of past prices may assist consumers in
bargaining with firms.
•Positive welfare effect as prices are reduced
•Not in the best interest of firms to make this information
Hunt Them Down: Detecting Cartels
Stages in the process: screen, verify,
• Screening: Identifying markets where
collusion is suspected.
• Verification: Systematically distinguish
between collusion and competition.
• Prosecution: Developing economic
evidence to determine guilt.
Two empirical methods for screening:
structural, behavioral
• Structural: Identifying industry traits
conducive to collusion.
• Behavioral: Identifying collusive behavioral
Structural Approaches
Structural approaches identify markets with traits conducive to the
formation of a cartel.
Problem of too many false positives
• Imagine the “ideal” market for collusion: Two firms, homogeneous
products, stable demand, etc.
• In practice, only a small fraction of such markets probably have
• The reason is that there are many omitted (unmeasured) factors that
influence whether a cartel forms or not.
Behavioral Methods
Is behavior inconsistent with competition?
Does a collusive model fit the data better than a competitive
Is there a structural break in behavior –i.e. statistically
distinguishable change?
Empirical methods will vary according to:
• The type of data that is required
• The need for prior information about collusion
• Reduced form or structural estimation methods
Discouraging Cartels: Leniency
Leniency Programs
A leniency program offers reduced penalties to corporations
and/or individuals involved in collusion, in exchange for
cooperating with enforcement authorities.
U.S. Department of Justice
• 1993: Revised corporate and individual leniency program
• Three major revisions:
• Amnesty is automatic if there is no pre-existing investigation
• Amnesty may still be available even after and investigation has
• All officers, directors, and employees who cooperate are protected
from criminal prosecution.
• Annual number of leniency applications increased 20-fold.
European Commission introduce leniency programs in 1996, revised
in 2002.
More than 50 jurisdictions have leniency programs (Feb, 2010)
The Economics of Leniency Programs
A cartel shares common characteristics with organized
•Cooperation within an illegal environment means that
there is a corporate governance problem at its core
•Repeated interaction, monitoring, threats and reactions
against cheaters, means that participants acquire
important information about all their co-conspirators
A viable cartel requires that 2 conditions be fulfilled
•Participation Constraint:
E(Profit from participating in the cartel) – E(Cost of
•Incentive Compatibility Constraint:
Under incomplete information and different types of
agents, the participant will find it in its interest to act in
accordance with the cartel rules.
A successful leniency program weakens a cartel by
maximizing conflict of interest among its members and
reducing the compatibility of interests
A Model for Using Leniency in the Post-Cartel
Scenario: auction houses where collusion has stopped. Should they
apply for leniency?
Model: Each auction house chooses the option that minimizes
expected penalties
• f is the penalty avoided by receiving leniency (government fine).
• d is the penalty not avoided by receiving leniency (customer
• p is the probability of a conviction when neither applies for leniency
Do not Apply
d + f/2 ; d + f/2
Do not Apply
p(d + f) ; p(d + f)
A Model for Using Leniency in the Post-Cartel
Do not Apply
d + f/2 ; d + f/2
Do not Apply
p(d + f) ; p(d + f)
Equilibria when the probability of being convicted is low:
• p (d + f) < d or p < d/(d+f) [so expected cost of being caught is less
than certainty of paying damages]
• Two possible equilibrium: Both apply for leniency; both do not apply
for leniency
• It is a coordination game. (Firms want to coordinate on “do not
Equilibrium when the probability of being convicted is high:
• p (d + f) > d or p > d/(d + f)
• Unique equilibrium: Both apply for leniency.
• It is the Prisoners’ Dilemma game. (The dominant strategy is “Apply”)
Punishment and Damages
Simple Theory of Cartel Fines
Deterrence and Fines
• Theory is that fines should be set so that the expected fine is greater than
the benefits of cartel
• Set properly, fines should eliminate cartels; a higher fine can offset a
lower probability of detection
• M = monopoly profits from cartel (divided among members)
• p = probability of detecting the cartel
• F = fine imposed on cartel (assume one period for simplicity)
Expected fine for cartel = p x F (divided among members)
• If expected fine is greater than expected monopoly profits then firms will
not join cartel. Therefore p x F > M will deter cartels.
• Therefore, firms will choose not to join the cartel when F>M/p
Fine Questions and Puzzles
Corporate versus individual liability
• For publicly traded companies fines reduce shareholder value.
• Will punishing shareholders reduce cartel formation?
What’s the role of jail terms
• Why not just fines?
• Why jail for cartels and not for other antitrust offenses?
• What is more effective, jail versus debarment?
Fines vs. detection
• Why not just increase fines and reduce resources on detection?
Is cartel enforcement optimal?
• Does the continuing uncovering of cartels evidence that penalties are
still too low?
Punishment: the Fine Arts Auction Houses Case
Customer Damages: Purpose
Compensation to harmed consumers
Deterrence and resistance of cartels
• Additional financial penalties to fines levied by the government.
• Create added incentives for customers to monitor, report, and
Customer damages – US
• Treble damages: Multiplier serves deterrence since the probability of
being caught and paying penalties is well below one.
• Indirect purchasers cannot sue for damages except in some states.
• Class action suits used for when many customers each incur a small loss.
The Economics of Damages
Damages inflicted by firm i from colluding in period t are calculated
to be:
• [ Pic(t) – Pibf(t) ] * Qic(t)
Pic(t) is the observed (collusive) price charged by firm i in period t.
Qic(t) is the number of units sold by firm i in period t.
Pibf(t) is the “but for” (or counterfactual) price for firm i in period t.
Pic(t) – Pibf(t) is the overcharge
Customer Damages
Customer damages EU (White Paper, April 2008)
• Single damages.
• Indirect purchaser suits are allowed.
• If an illegal overcharge is passed on to consumers who are not direct
buyers then those consumers are harmed.
• Those harmed consumers can claim compensation.
Comparison of US and EU
• EU is focused on compensation.
• US is primarily concerned with deterrence and resistance.
Allowing indirect purchasers to sue weakens enforcement
• Direct purchasers have the best information when it comes to detecting
collusion. Weakening their incentives will reduce the likelihood that they
report and sue, thereby reducing enforcement.
• Damage calculation becomes more difficult and could effectively
reduce penalties.
Cartels vs. Cooperation
Firms Often Cooperate For Pro-Competitive
Joint facilities to obtain scale economies (e.g. Computer Reservation
Standard setting to obtain network economies and reduce
transactions costs for consumers (see Lecture 4) (DVD standards).
Solving various market imperfections (e.g. Portobello Road Antique
Exchanging information (costs, best practices, etc.) for procompetitive reasons (e.g. all trade associations). Cooperatives, joint
ventures, standard setting organizations, and other horizontal
combinations pose difficult problem for antitrust.
Engage in obvious efficiency enhancing activities.
Efficiencies from Cooperation
Sharing of facilities and reduction in
duplicative costs
Realization of network effects through e.g.
Cooperation may be essential to the creation
of a product (sports leagues, payment cards)
Other efficiencies
Key Economic Issues for Cooperation
Effect of competition law on choice of organizational form—if mergers are
lawful but cooperation isn’t then we will get more mergers. That may lead to
inefficient organizational forms and perhaps less competition than with
cooperation short of mergers.
Is efficiency justification legitimate or a sham?
Is it possible to obtain efficiencies in less restrictive ways (relative to
US courts deal with these issues by treating cooperative agreements for which
there is an efficiency justification under the rule of reason (since the BMI case
which involved a music collecting society).
EC law deals with these issues under Article 81(3). Coordinated behavior may
be exempt if it is objectively justified.
Music Collecting Societies
Composers have copyright protection
• But cost of enforcing is too high for individual composers
• Would need to negotiate with many venues
• Would need to invest in detection
Music collecting societies form to represent composers
• Negotiate with venues
• File enforcement actions
• Note these are two-sided platforms
But cooperation increases market power when compositions are
• Where do we draw the line between efficiencies from cooperation and
increased market power?
• How have courts/competition authorities resolved this?
Guest Speaker:
Rosa M. Abrantes Metz
Collusion, Screening and LIBOR
End of Topic 9, Next Class Topic 10
Part 1
Part 2
Market power
Supply side
Demand side

similar documents