B. The cons: how do information exchanges “facilitate” collusion? . 90
2. Internal stability of coordination
Nevertheless, even if the companies can reach a consensus through information sharing, it would be still necessary that the industry meets certain conditions that ensure that the collusive equilibrium can be sustained. After all, there is no sense in reaching an agreement, if this cannot be enforced later on.
Stigler was one of the pioneers in analyzing collusion under this perspective.249 The idea behind his work was that the factors that deter collusion are those factors that make cheating (i.e. deviating from the collusive explicit or tacit meeting of the minds) worthwhile.250
a) The economics of cheating in cartels
To identify them, it is necessary to understand the economics of cheating in these collusive equilibriums. Even though a cartel wishes to behave like a single monopolist, there is a crucial element that differentiates them: the diffused effect of their actions.
In plain words, a cartel member obtains gains as a member of the cartel, but his gains are even higher, if he cheats while the cartel is in place.251 Thus, the question is:
when will a cartel member cheat?
248 (Faull & Nikpay, 2014) op. cit.
249 Stigler, G. "The Economic Effects of the Antitrust Laws." In The Organization of Industry, by G.
Stiegler. Chicago: The University of Chicago Press, 1983, page 268. “when the event we wish to study is clandestine, we cannot rely upon direct observation. I believe that my theory of oligopoly is a useful tool for this study precisely because it seeks to isolate (…) the determinants of successful cheating and hence unsuccessful collusion”.
250 As (Motta, 2004) op. cit. page 150 puts it “detection of deviation is a crucial ingredient for collusion”.
251 Note that these incentives to cheat are observed in a perfect cartel with identical cost structures for all firms, intuition tells us that the incentives to cheat may increase if we add differentiation in the firm’s marginal cost and marginal revenue structures. If firms set a collusive equilibrium, their marginal revenue will exceed their marginal cost, thus obtaining, supra-competitive profits. For the monopolist this is a stable situation, however, for each single cartel member there is an incentive to increase his output and decrease his prices (an incentive to cheat), as he will receive all the profits of the cheating behavior
Well, only when there is limited risk that he will be detected and punished. In other words, his behavior is only beneficial to him for as long as the other cartel members do not react. The moment the other members increase output, prices will decrease for all members; losing, therefore, their supra-competitive profits. 252
b) Monitoring adherence to the collusive agreement or practice
As a result, we can see how information sharing can be a very useful mechanism for firms to monitor adherence to the collusive equilibrium.253
Precise individualized information might allow coordinating firms to identify which exact firm has deviated, on which particular product, and for how long.254 This makes it easier to tailor the punishment for any deviating behavior.255
In other words, even if competitors can reach an agreement and develop the tools to monitor compliance, it is undisputed that the incentive to cheat will be, ultimately, affected by the “punishment” that the cheater faces. In particular, there are two elements of the punishment which will normally determine the likelihood of
252 This was Stigler’s intuition in 1964 when he argued that the likelihood of collusion could be determined by the ability of the members of the cartel to detect deviating behavior. Since then, it has been supported and expanded by a wealth of economic literature. Stigler, G. "A Theory of Oligopoly." In The Organization of Industry, by G. Stiegler. Chicago: The University of Chicago Press, 1983. Stigler suggested three empirical predictions: First, that oligopolistic collusion was more likely in markets with small buyers rather in markets with large buyers. The reason being that while the later are price takers, the former can negotiate prices. Second, that collusion will be always more effective against buyers who report correctly and fully the prices tendered to them (such as the Government) as this would allow the cartelist to monitor prices and thus deviation. Third, and following the previous prediction, collusion is severely limited (excluding market sharing) when the significant buyers change identity continuously.
Summarizing it (Motta, 2004) op. cit. pages 150 to 156, fn 32. Particularly relevant are the conclusions of (Green & Porter, 1984) op. cit. that since observation and punishment are key to sustain the stability of a cartel, the observation of some period with low prices is not sufficient to exclude that the industry is at a collusive equilibrium. Rather, price wars simply are the indispensable element of a collusive strategy.
253 (Bennet & Collins, 2010) op. cit. page 322. (OECD, Unilateral Disclosure on Information with Anticompetitive Effects, 2012) op. cit. background paper, (Werden, Economic Evidence on the Existence of Collusion: Reconciling Antitrust Law with Oligopoly Theory, 2004) op. cit.
254 (Bennet & Collins, 2010) op. cit. page 323.
255 (Motta, 2004) op. cit. page 151 “exchange of information on past prices and quantities (or of verifiable information on prices and quantities set in the current period) of individual facilitates collusion, as it allows to identify deviators and better target market punishments, which then become more effective and less costly for the punishing firms”.
cheating: (i) the response time it takes the cartelists to observe and react to the deviation and (ii) the severity and accuracy of the punishment.
The response time is critical, the longer it takes the cartelists to react, the more profits that the deviating firm will make.256 The severity of the punishment is also crucial and raises an interesting aspect of game theory. The other members of the cartel will want to send a clear message that any deviating behavior will be penalized and they want to ensure that the penalty is higher than the benefits obtained.
However, the cartel members will also “suffer” the consequences of any penalty unless they can take targeted punishments. In the absence of clearly targeted punishments, an interesting trade-off will take place between the willingness of the firms to penalize the cheater, on the one hand, and their own interests in reversing to the collusive equilibrium as soon as possible, on the other.257
An example that is often cited by economists is the Fatty Acids case.258 According to the Commission decision, the market leader initiated contacts with the
256In order for the punishment to take effect expeditiously, the cartelist need: (a) to detect the deviating behavior expeditiously; and (b) possess the necessary spare capacity or financial muscle to reduce prices and penalize the cheater quickly. (Bishop & Walker, The Economics of EC Competition Law: Concepts, Applicaction and Measurement, 2010) op. cit. (Nitsche & Von Hinten-Reed, 2004) op. cit. explain that
“Information exchange about future and past prices and quantities may affect the ability and the cost of punishment. For example, if in a scheduled transport service collusion builds on a market sharing agreement, the announcement of a new schedule that would not be in line with the collusive arrangement could trigger a punishment response even before market entry (e.g. a price cut), this can make punishment very effective and therefore help stabilising a collusive arrangement. Information exchange may also help targeting the punishment schemes. In localised markets punishment can be targeted against the cheating party if sufficient information exists to determine who has cheated. This reduces the cost of punishing a “cheater” and so makes the threat of punishment more credible.”
257 (Bishop & Walker, The Economics of EC Competition Law: Concepts, Applicaction and Measurement, 2010) op. cit. (Nitsche & Von Hinten-Reed, 2004) op. cit.
258 Fatty Acids [1987] OJ L3/17. See Stroux, S. US and EC Oligopoly Control. The Hague: Kluwer Law International, 2004, for a summary of the discussion, page 149. Bissocoli, E. "Trade Associations and Information Exchange under US Antitrust Laws and EC Competition Law." 23 World Competition 1, 2000. According to Capobianco, A. "Information Exchange ." Common Market Law Review, 2004 the first Commission decision on information exchanges. See (Pepperkorn, 1996) op. cit. arguing that “a number of elements indicate that the Commission did not base itself on or was not aware of the apparent underlying non-cooperative prisoner's dilemma game setting of the case. The decision does not make clear that the main effect of up to date monitoring is that punishment becomes more effective while the incentive to free ride is reduced. The Commission also indicates that it considers it of importance whether the companies actually communicated on sharing the market or not, making it clear that such would not
other two large producers which led to an information exchange agreement among these producers which covered firm data on yearly sales and future four-monthly reports about total sales.259 This information was used to discriminate the competitive strategies. In fact, customer switches between the three main competitors were labelled “stolen sales”, whereas gains of new customers were considered “legitimate gains”.260