Topic 8:
Collusion and Cartels
EC 3322
Semester I – 2008/2009
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
1
Introduction

What have we learned before?

Cournot competition induces firms to overproduce.

Bertrand competition with homogenous goods induces price war.

Firms would be better off if they can coordinate their activities  e.g.
restricting their market outputs and raising the market price  however
in a one shot interaction no firms are able to commit to do so 
prisoner’s dilemma.

But, firms typically interact repeatedly in the markets  so they may
have incentive to coordinate  look for strategies that will sustain
cooperation  if they are sufficiently care about the future and
reputation.

The analytical tool  repeated game analysis.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
2
Example: One-shot Cournot Game

Two identical Cournot firms making identical products.

MC are the same for both firms  MC=$30.

Inverse market demand  P=150 – (q1+q2).

Profit for firm 1 and firm 2:
 1  150  q1  q2  q1  30q1
 1
q1
 150  2q1  q2  30  0
q2
 150  q1  2q2  30  0
1
q 2  60  q1
2
1
q1  60  q2
2
*
*

 2  150  q1  q2  q2  30q2
 2
At Cournot-Nash equilibrium 
q1*  q2*  40
Yohanes E. Riyanto
P*  $70
 1*   2*   70  30  40  $1600
EC 3322 (Industrial Organization I)
3
Example: One-shot Cournot Game …

If they are able to coordinate, e.g. collude  behaves as a monopoly.
  150  Q  Q  30Q

 150  2Q  30  0
Q
Q*  60

If each firm produces half of the total quantity q1  q2  30 . Then,
we have: P*  90 and  1*   2*   90  30  30  $1800

Unfortunately, there is incentive to cheat  firm 1’s output of 30 units
is not the best response to firm 2’s output of 30 units.

Suppose firm 2 sticks to the agreement and sets q2 = 30 units.
1
 1*   75  30  45  $2025
q1*  60  q2  45
2
*
P*  150  45  30  $75
Yohanes E. Riyanto
 2   75  30  30  $1350
EC 3322 (Industrial Organization I)
4
Example: One-shot Cournot Game …

So indeed firm 1 prefers to cheat  of course firm 2 can anticipate this
 the best for firm 2 is also to cheat  prisoners’ dilemma.
Both firms have the
incentive to cheat on
their agreement
Firm 2
Firm 1
Cooperate
Cooperate
Deviate
Yohanes E. Riyanto
This
is the1800)
Nash
(1800,
equilibrium
(2025, 1350)
EC 3322 (Industrial Organization I)
Deviate
(1350, 2025)
(1600, 1600)
1600)
(1600,
5
Finitely Repeated Game

Suppose now that the game is played repeatedly finite number of times 
common knowledge to firms.

This allows for reward & punishment strategy.



“If you cooperate this period, I will cooperate the next period”.
“If you deviate (renege) then I will also deviate”.
For example: in our previous Cournot game  repeated twice  Strategy of
firm 1  cooperate in period 1  maintain cooperation in period 2 if firm 2
cooperated in period 1, otherwise deviate.
Firm 1
Firm 2
Yohanes E. Riyanto
Cooperate
Deviate
Cooperate
(1800, 1800)
(1350, 2250)
Deviate
(2025, 1350)
(1600, 1600)
EC 3322 (Industrial Organization I)
6
Finitely Repeated Game …

Such a strategy is not credible (not subgame perfect).

Firm 1’s dominant strategy in period 2 is not to keep its promise to
cooperate because it knows that period 2 is the last period.
Firm 2
Firm 1


Cooperate
Deviate
Cooperate
(1800, 1800)
(2025, 1250)
Deviate
(1350, 2025)
(1600, 1600)
Cooperation in period 2 is based on empty promise.
Thinking backwardly  period 1 is effectively ‘the last period’  so
firm 1 will also deviate in period 1.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
7
Finitely Repeated Game …


The same problems arise for finite period more than 2 periods, e.g. T
periods.

in period T any promise to cooperate is worthless

so deviate in period T

but then period T – 1 is effectively the “last” period

so deviate in T – 1

and so on.
Selten’s Theorem (Reinhardt Selten  Nobel Laureate)

“If a game with a unique equilibrium is played finitely many times its
solution is that equilibrium played each and every time. Finitely
repeated play of a unique Nash equilibrium is the equilibrium
of the repeated game.”.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
8
Infinitely Repeated Game

With finitely repeated game the collusive agreement breaks down in the
“last” period.

Suppose we have an infinitely repeated game  players do not know
when the game will end  some probability in each period that the game
will continue.

Good behavior can be credibly rewarded.

Bad behavior can be credibly punished.

Suppose that in each period the net profit is always the same πt.

Discount factor: 0≤δ ≤1, and the probability of continuation to the next
period is p  probability adjusted discount factor = pδ.

The present value (PV) of the infinite sequence of profits:
PV  t    1  p 2  p 2 2 3  ... p t 1 t 1 t with
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
 = 1   2  ...   t
9
Infinitely Repeated Game…
PV   1  p  p 2 2  ... p t 1 t 1 
(1)
p PV    p  p 2 2  p 3 3  ... p t t 
(2)
(1)-(2)  PV 1  p   

PV  
1
1  p 
Consider the following strategy  grim-trigger strategy.


Cooperate as long as the partner cooperates in the previous periods.
Punished forever by deviating to non-cooperative act if the partner
deviates in the previous round.

Its “grim” because of the swift and harsh punishment for deviation.

An alternative punishment strategy (less harsh)  tit-for-tat  start by
cooperating  in every subsequent round, adopt your partner’s
strategy in the previous round  thus if your partner reverts back to
cooperation you reward it by cooperating as well.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
10
Infinitely Repeated Game…


Payoff from honoring the agreement (cooperation):
1
 C  1800
1  p 
Payoff from deviating from the agreement (deviation):
 D  2025  1600
p
1  p 
Deviation gives a one-time payoff, but thereafter the partner will punish
by deviating forever.

Cooperation is better if
C D
1800
1
p
 2025  1600
1  p 
1  p 
2025  1800 

p 
 0.5294
2025

1600


Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
if p  1 then   0.5294
if p  0.6 then   0.882
11
Infinitely Repeated Game…

In a more general setting. Suppose that in each period





profits to a firm from a collusive agreement are πC
There is always a
value
profits from deviating from the agreement are
πDof δ < 1 for which
this equation is
profits in the Nash equilibrium are πN
satisfied
This is the short-run gain
from cheating
Thisnot
is the
loss
Cheating on the cartel does
paylong-run
so long as:
from cheating
 D  C 

p 
 D   N 
The collusion is sustainable (stable), if:

we expect that  D   C   N

Short-term gains from cheating are low relative to long-run losses

Cartel members value future profits (high discount factor).
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
12
Infinitely Repeated Game…

With infinitely repeated games, cooperation is sustainable through
self-interest.

But there are some caveats


We so far assume speedy reaction to deviation, what if there is a
delay in punishment?  Trigger strategies will still work but the
discount factor will have to be higher.
The punishment is harsh and unforgiving  what if demand is
uncertain?

The shrinking in outputs  maybe a result of a “bad condition” rather
than cheating on the agreed quantities.

So sometimes it is better to agree on the variation in outputs that will
not trigger retaliation, or

To agree that punishment lasts for a finite period of time
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
13
The Folk Theorem

We have so far assumed that firms cooperate by choosing the
monopoly outcome  this need not be the case.

There are many potential agreements that can be made and sustained
in a repeated setting – the Folk Theorem (Friedman, 1971).
Suppose that an infinitely repeated game has a set of pay-offs
that exceed the one-shot Nash equilibrium pay-offs for each
and every firm. Then any set of feasible pay-offs that are
preferred by all firms to the Nash equilibrium pay-offs can be
supported as subgame perfect equilibria for the repeated
game for some discount factor sufficiently close to unity.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
14
The Folk Theorem …

In our example, the feasible pay-offs describe the following possibilities
2
$2100
$2000
$1800
The Folk Theorem states
that any point in this
Collusion
triangleon
is a potential
monopoly
for the
Ifequilibrium
thegives
firms collude
eachcompete
firm
$1800
repeated
game
If the firms
perfectly
they
share
they each earn $3,600
$1600
$1600
$1500 $1600 $1800
Yohanes E. Riyanto
$2000 $2100
EC 3322 (Industrial Organization I)
1
15
Cartel in Practice

Cartel: an association of firms that explicitly attempt to enforce market
discipline and reduce competition (e.g. by coordinating pricing or output
activities or market shares)  it does not necessarily include all firms in
an industry.

Cartel is common  but generally hidden, difficult to detect and is
considered illegal.

But some cartels are explicit: OPEC (oil cartel) and De Beers (diamond
cartel).

Example of hidden cartel agreement: price fixing on long-haul passenger
routes by British Airways (BA) and Virgin Atlantic (VA) & Korean
Airline (The Economist, Aug 4th 2007)  many other examples.

Competition policy authority imposed hefty fines if detected  in the
case of BA&VA, the fine is huge: $546 million.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
16
Cartel in Practice
BA and Virgin: Flying in formation
Aug 2nd 2007
From The Economist print edition
It takes two to fix prices
FOR years British Airways (BA) described itself as “the world's favorite airline”. It no longer looks so popular in London and Washington. On
August 1st the firm was hit with a transatlantic double whammy after it was found guilty of colluding with a rival, Virgin Atlantic, to
fix prices on long-haul passenger routes. Britain's Office of Fair Trading (OFT) handed down a record fine of £121.5m ($246m). A few
hours later, America's Department of Justice (DoJ) imposed a $300m penalty of its own. The severity of the American fine also reflected
BA's role in a different international conspiracy involving Korean Air and Lufthansa.
A clearer example of illegal price-fixing than that between BA and Virgin would be hard to imagine. The two firms discussed “fuel
surcharges” at least six times between August 2004 and January 2006, during which time they rose from £5 to £60 on a return ticket.
A transatlantic bust was particularly fitting for the OFT. During Labour's period in office, it has introduced American-style, cartel-busting
sanctions on companies that prefer cozy deals with rivals to the bracing winds of competition. But despite many protracted investigations into
sectors such as banking and supermarkets that attract consumers' ire, the OFT has struggled to find the kind of smoking-gun evidence of
collusion it needed to look as terrifying as it and the government wished. That is partly the nature of the beast. Collusion is difficult to
prove: as Mr Collins observes, the tricky thing about colluders is that they do their business in secret. Indeed, the airlines' price-fixing
came to light only after Virgin's legal department alerted the authorities.
This was no selfless dedication to consumers' welfare. Virgin hoped to benefit from the “leniency policy”, which was introduced in the 1998
Competition Act and copied from similar laws in America, granting immunity to firms that blow the whistle. Virgin was just as complicit as
BA in the price-fixing and has, presumably, benefited from it financially. Not only was the airline saving itself from the risk of prosecution, but it
was also grassing up a rival with whom it has had a bruising relationship in the past. It grates to see one firm get away with something while
another is punished, but leniency policies are, probably, a good thing. The ability to claim immunity gives a powerful incentive for
businesses to police their own industries, which ought to improve things for consumers. After all, half a victory is better than none.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
17
Cartel in Practice
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
18
Cartel in Practice
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
19
Cartel in Practice
Geographic
Scope
Product
Year
Fine
F. Hoffman-LaRoche Ltd.
Vitamins
1999
$500
International
BASF AG (1999)
Vitamins
1999
$225
International
SGL Carbon AG
Graphite Electrodes
1999
$135
International
UCAR International Inc.
Graphite Electrodes
1998
$110
International
Lysine and Citric Acid
1997
$100
International
Citric Acid
1997
$50
International
Marine Construction
1998
$49
International
Sorbates
1998
$36
International
Graphite Electrodes
1998
$32.5
International
Sodium Gluconate
1998
$20
International
Marine Transportation
1998
$15
International
Dyno Nobel
Explosives
1996
$15
Domestic
F. Hoffman-LaRoche Ltd.
Citric Acid
1997
$14
International
Eastman Chemical Co.
Sorbates
1998
$11
International
Jungblunzlauer International
Citric Acid
1997
$11
International
Vitamins
1998
$10.5
International
Sodium Gluconate
1997
$10
International
Defendant
Archer Daniels Midland co.
Haarman & Reimer Corp.
HeereMac v.o.f.
Hoechst AG
Showa Denko Carbon Inc.
Fujisawa Pharmaceuticals Co.
Dockwise N.V.
Lonza AG
Akzo Nobel Chemicals BV & Glucona BV
Source: U.S. Department of Justice, http://www.usdoj.gov/atr/public/press_releases/1999/2456.htm
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
20
Cartel in Practice …

Three major factors that facilitate the formation of cartel:



The ability to Raise the Market Price without inducing
substantial increased competition from non member rivals.

Market demand should be sufficiently inelastic (relatively
vertical).

No threat of entry by non members (given that price is high) and
close substitutes products.
The probability of getting caught and receive severe
punishment is not excessively high.
Low organizational costs of maintaining the cartel  realized
when 1) few firms are involved, 2) the market is highly
concentrated, 3) relatively similar products (homogenous in
qualities or properties), and 4) there exist a trade association.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
21
Enforcing a Cartel Agreement

The success of a cartel agreement depends on whether it is easy to
enforce  this in turn depends on whether it is easy to detect any
deviation by its members.

Factors determining the ease of detection:



The number of firms in the market  fewer firms makes it easier to
monitor any change in one firm’s market share.
Prices do not highly fluctuate and they are widely known  no
frequent shifts in demand, input costs and other factors  otherwise
cheating by a member cannot be distinguished from other factors that
cause price change.
Whether or not cartel members serve the same type of
consumers along the distribution chain, e.g. retailers or end
consumers.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
22
Enforcing a Cartel Agreement

The cartel members have little incentive to cheat if:




Their marginal cost curve is nearly vertical  costly to increase
production.
There is no large unutilized capacity.
There are many small customers who make small purchases  if a
price cut is unannounced, other small consumers are unlikely to know
 if it is announced, other members will not like it and will retaliate.
Types of cartel agreements  to prevent cheating.


Fix more than just price  divide the market based on geography
or buyers’ type.
Fix market share  as long as market shares are observable  no firm
has an incentive to cheat.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
23
Enforcing a Cartel Agreement

Types of cartel agreements  to prevent cheating….


Most-favored-nation clause  a sales contract that guarantees the
buyer that the seller is not selling at a lower price to another buyer 
if it does it has to also offer the same lower price to its initial buyers
 this rebate mechanism acts as a penalty from cheating.
Low-price (price matching) guarantee  if another seller offers a
lower price, the seller will match it  this clause helps sustaining high
cartel price, instead of the low cartel price that they seem to
guarantee.
Courts
Low Price
Low Price
Best Denki
High Price
Yohanes E. Riyanto
(200, 200)
(0, 400)
High Price
(400, 0)
(300, 300)
EC 3322 (Industrial Organization I)
24
Enforcing a Cartel Agreement

With price matching guarantee:

Best Denki can adopt the following strategy:


Set high price (H), and at the same time adopt price matching
guarantee.
By doing so Best Denki can reduce the benefit Courts can achieve
from setting low price when Best Denki sets high price.

When Courts chooses L  Best Denki must match the low
price when consumers find it out  so both ends up charging
low price  equal profit of 200.

When Courts chooses H  Best Denki chooses H (no need to
match)  equal profit 300.

It is better for Courts to stick with H.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
25
Enforcing a Cartel Agreement

With price matching guarantee …

At the same time, promising to meet Courts low price (L) also
hurts Best Denki  because of the lower profit from the low
price.



Hence, no incentive for Best Denki to deviate from H.
Similarly, Courts will have the same thought  so Courts will sets
H and offer price matching guarantee.
Both Best Denki and Courts set (H,H)  lower price-matching
guarantee is a credible tool to sustain collusion  matching is a
punishment device.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
26
Courts
Low Price
Best
Denki
High Price
Match
Low
Price
200, 200
400, 0
200, 200
High
Price
0, 400
300, 300
300, 300
200, 200
300, 300
300, 300
Match
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
27
Cases
De Beers: Diamond Cartel
De Beers Consolidated Mines Ltd. was established in the 1870s. Since then, the firm owned by the Oppenheimer family, has maintained a
remarkable control over the world diamond industry. De Beers owns all of the diamond mines in South Africa and has interests in other
countries as well. However, in terms of mining its share of the world market is relatively small, especially since the discovery of the Russian
mines. The key to De Beers’ control of the market is the Central Selling Organization (CSO), its London-based marketing arm.
The CSO serves as an intermediary between the mines and the diamond cutters and polishers. More than 80% of the world’s diamonds are
processed by the CSO, although only a fraction of them originate from De Beers’ mines. CSO staff classify the diamonds by category (there are
thousands types of diamonds). This is a highly skil-intensive task in which De Beers has unmatched expertise. The CSO also regulates the
market to achieve price stability, building up its stocks during periods of low demand and releasing those same stocks during periods of high
demand.
The high margins earned by De Beers are a constant temptation for mining companies, who figure the same margins might be earned by selling
directly to the diamond cutters. What stops them from doing so? First, many of the diamond producers see the current cartel structure as a
benefit to the whole industry. In addition to stabilizing prices, De Beers plays the crucial role of advertising diamonds. Both price stabilization
and advertising are “public goods” at the industry level: Every producers benefits, although only De Beers pays for it.
A second reason for compliance with De Beers’ dominance is the fear of retaliation if they defect from the cartel. In 1981, President Mobutu of
Zaire, the world’s largest supplier of industrial diamonds, announced that the country would no longer sell diamonds through the CSO. At the
same time, contacts were set up with two Antwerp brokers and one British broker. Two months latter, about one million carats of industrial
diamonds flooded the market, and the price fell from $3 to less than $1.80 per carat. Although the cource of this supply surplus is unkonown,
many believe the move was De Beers’ way of showing who’s in control.
For De Beers, this was a costly operation, but it was a case of “it’s not the money, it’s the principle.” And the point was made: In 1983, Zaire
requested the renewal of its old contract with De Beers. In fact it ended up with was less favorable than the original one.
Source: Quoted from Cabral, Luis (2000), “Introduction to Industrial Organization”, MIT Press, 354 p.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
28
Cases …
OPEC’s mission is to coordinate & unify the petroleum policies of Member
Countries & ensure the stabilization of oil markets in order to secure an efficient,
economic & regular supply of petroleum to consumers, a steady income to
producers & a fair return on capital to those investing in the petroleum industry.
OPEC
From 1974-1985, OPEC’s share of world production fell from 55% to 30%, and non OPEC producers (Including the US) and
producers from the communist and former communist countries experienced an increase in their share from 45% to around 70%.
The biggest loser in the relative decline in OPEC’s market share was Saudi Arabia, since it is the biggest oil producer among OPEC
members.
To stabilize prices, Saudi Arabia was forced to significantly reduce its output from an average of 6.5 million barrels per day in 1982
to 5.1 million barrels per day in 1983, 4.7 million barrels in 1984, and 3.4 million barrels in 1985. In 1985, Saudi Arabia reached the
limit of its restraint on behalf of the cartel. In late 1985, the Saudis decided to assert their power to control price. Once the Saudis
decided to increase output to preserve their market share, instead of attempting to maintain a high price, industry prices fell to below
$10 a barrel in 1986. At $10 a barrel the output of Non-OPEC members began to decline as marginal wells were shut down. When
the price was eventually stabilized at approximately $18 a barrel in 1986-1987, OPEC’s share relative to non OPEC countries once
again began to increase.
To date, OPEC still commands an important role in stabilizing the world’s crude oil price. For instance, in September 2007, OPEC
agreed to boost its crude oil output by 500000 barrels per day from Nov 1st 2007 in an attempt to calm down the market which has
shown an increasing crude oil price to around $76 a barrel. Oil prices have risen by 27 per cent this year after OPEC curbed exports
to drain inventories.
Source: Various sources.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
29
Cases …
P&G and Kraft Food (Phillip Morris)
On Feb 21, 1997, P&G announced its intention to raise the price of its leading coffee brand,
Folgers, by more than previously intended. The company said that it now would raise the
price of a 130 ounce tin of Folgers by 20% as of March 3.
P&G’s announcement drew an immediate response from the Kraft food unit of Phillip
Morris. This rival coffee maker announced that the price increase on its own leading brands,
Maxwell House and Yuban, would exactly match the Folgers price rise as of March 31.
Do such parallel price hikes indicate collusion? It is difficult to know, especially in this case.
The price of raw coffee had been rising for months prior to the P&G announcement
following a frost that destroyed the Brazilian coffee crop. Yet whether there was a conspiracy
or not, coffee drinkers were in for a different kind of jolt from imbibing their favorite
beverage.
Source: “Procter & Gamble, Rival to Increase Coffee Prices,” Boston Globe, February 22, 1997 as quoted in
Peppal, Norman, Richards (2005), “Industrial Organization: Contemporary Theory & Practice”, Thomson-South
Western, 672 p.
Yohanes E. Riyanto
EC 3322 (Industrial Organization I)
30