Multi-Market Collusion with Demand Linkages and Antitrust Enforcement

This paper analyzes dynamic cartel formation and antitrust enforcement when firms operate in demand-related markets. We show that cartel prosecution can have a knock-on effect: bringing down a cartel in one market reduces profits and cartel stability and leads to the break-up of the cartel in the adjacent market. Cartel prosecution can also have a waterbed effect: disrupting a cartel increases cartel stability in the adjacent market and induces cartel formation in previously competitive markets. We discuss the impact of dynamic cartel formation on consumer surplus, explore antitrust spillovers, the optimal scope of antitrust interventions and cartel formation with local firms.     

The discontent cartel member and cartel collapse: The case of the German cement cartel

We hypothesize a particular source of cartel instability and explore its relevance to understanding cartel dynamics. The cartel instability is rooted in the observation that, upon cartel formation, the relative positions of firms are often fixed which may lead some growth-conscious members to be discontent. This incongruity between a cartel member's allocated market share and its desired market share may result in systematic deviations and the eventual collapse of the cartel. This hypothesis is then taken to the German cement cartel of 1991–2002. We argue that Readymix was such a discontent cartel member and, using a rich pricing data set, are able to characterize how Readymix deviated, how other firms responded, and how it led to the collapse of the cartel.     

Stable heterogeneous cartels

We define a nation of cartel stability which allows firms to perceive the impact of their actions on overall competitiveness in the market. We demonstrate that in an industry with linear demand and diverse linearly increasing marginal cost functions a stable cartel always exists. Furthermore, we examine uniqueness and size of such cartels as well as the impact of cost and demand conditions on their characteristics. In particular, we establish that the most efficient firms will be in the cartel while the less efficient ones will remain outside.     

Collusion in a price-quantity oligopoly

In the context of an infinitely repeated oligopoly game, we study collusion among firms that simultaneously choose prices and quantities. We compare a price cartel with a price-quota cartel and analyze when and why firms prefer the latter to the former. Output quota may be required to solve coordination and incentive problems when market demand is sufficiently elastic. If market demand is sufficiently inelastic, then the cartel faces a trade-off between increasing prices and the amount of costly overproduction. We find that a price cartel prices consistently below the monopoly price to mitigate excessive production. In this case, a quota arrangement allows firms to avoid overproduction and to sustain the monopoly price. From a policy perspective, our findings suggest that an overall price increase in conjunction with more stable prices and market shares is indicative of collusion in industries where production precedes sales and outputs are imperfectly observable.     

Cartel sustainability in retail markets: Evidence from a health service sector

This paper studies the role of cost asymmetries and product differentiation on cartel sustainability by drawing data from a failed retail cartel. Unlike the extensive theoretical literature, little empirical evidence exists on these relationships. First, we analyze cartel compliance and find that players are more likely to comply when cost is symmetric and own cost is high. Next, based on a structural model and counterfactual experiments, we show that a cartel price that satisfies all cartel members does not exist. This result indicates an inherent difficulty of sustaining collusion in retail markets with heterogeneous players. We also show that firm heterogeneities, especially product differentiation rather than cost asymmetries, hinder collusion more. Finally, we derive the level of patience (or the discount factor) required for cartel sustainability when firms split profits based on the Shapley value.     

An oligopoly model of ocean liner shipping

In this paper I develop a simultaneous equations oligopoly model of the regulated international ocean liner shipping industry. The firms act as a cartel to determine price jointly and then set their own quality levels to maximize individual profits. The cartel does not attain monopoly profits, because each conference member myopically determines quality without regard for overall cartel profits. The results indicate that an increase in the number of firms in the cartel will increase both cartel price and quality level. An increase in price will also lead to an increase in quality level.I would like to thank Professors Alamarin Phillips, Robert Summers, and Bruce Allen for their helpful comments on earlier research for my dissertation in the Economics Department of the University of Pennsylvania on which this work is based. I am especially grateful to Professor Lawrence J. White for his encouragement and valuable suggestions at various stages of my work.     

Post-cartel tacit collusion: Determinants,consequences, and prevention

We experimentally investigate the determinants of post-cartel tacit collusion (PCTC), the effects of PCTC on market outcomes, and potential policy measures aimed at its prevention. PCTC occurs robustly with or without fines or leniency and is determined both by collusive price hysteresis and learning about cartel partners’ characteristics and strategies. As a result, it is also strongly related to the preceding cartel success. PCTC generates a downward bias in the estimated cartel overcharges. This threatens the effectiveness of deterrence induced by private damage litigation and fines imposed on colluding firms based on the overcharge. This bias further increases with preceding cartel stability such that especially more stable sets of colluding firms may be deterred less when PCTC is present. Rematching colluding subjects with strangers within a session prevents PCTC. This indicates that barring colluding managers from their posts could help impede PCTC in the field.     

Competition,Collusion, and Spatial Sales Patterns: Theory and Evidence*

This article studies competition in markets with transport costs and capacity constraints. Using a rich micro-level data set of the cement industry in Germany, we study a cartel breakdown to identify the effect of competition on transport distances. We find that when firms compete, they more often serve more distant customers. Moreover, the transport distance also varies in the ratio of capacity relative to demand, but only if firms compete and not when they coordinate their sales. We provide a theoretical model of spatial competition with capacity constraints that rationalizes the empirical results.     

Fine-Tailored for the Cartel-Favoritism in Procurement

In this paper, we investigate the interaction between two firms, which are involved in a repeated procurement relationship modeled as a multiple criteria auction, and an auctioneer (a government employee) who has discretion in devising the selection criteria. Our main result is that favoritism substantially facilitates collusion. It increases the gains from collusion and contributes to solving basic implementation problems for a cartel of bidders operating in a stochastically changing environment. A most simple allocation rule where firms take turns in winning, independently of stochastic social preferences and firms’ costs, achieves full cartel efficiency (including price, production, and design efficiency). In each period the selection criteria is fine-tailored to the in-turn winner: the “environment” adapts to the cartel. This result holds true when the expected punishment is a fixed cost. When the cost varies with the magnitude of the distortion of the selection criteria (compared to the true social preferences), favoritism only partially shelters the cartel from the environment. We thus find that favoritism generally facilitates collusion at a high cost for society. Our analysis suggests some anti-corruption measures that could be effective in curbing favoritism and collusion in public markets. It also suggests that the much-advocated rotation of officials is likely to be counter-productive.     

Prosecution and Leniency Programs: The Role of Bluffing in Opening Investigations

This paper characterizes the optimal investigation and leniency policies when the Competition Authority is privately informed about the strength of a cartel case. I show that the Competition Authority can then exploit firms’ uncertainty about the risk of conviction to obtain confessions even when the case is weak. More generally, I show that offering full leniency allows the Competition Authority to open more successful investigations (what I refer to as the ‘activism effect’ of leniency), which overall raises both cartel desistance and cartel deterrence. Finally, I discuss the policy implications of the model.     

Optimal cartel trigger strategies and the number of firms

Recent empirical papers have analyzed collusion in the Joint Executive Committee in an attempt to determine which of several theories of cartel behavior is supported by the behavior of this 19th century railroad cartel. Non-parametric tests of whether high and low profit regimes followed a first-order Markov process when one controls for the number of firms support the theory of optimal collusion given by Abreu, Pearce, and Stachetti. Results on whether transition probabilities depend on the number of firms are inconclusive.     

Cartel pricing dynamics with cost variability and endogenous buyer detection

This paper characterizes collusive pricing patterns when buyers may detect the presence of a cartel. Buyers are assumed to become suspicious when observed prices are anomalous. We find that the cartel price path is comprised of two phases. During the transitional phase, price is generally rising and relatively unresponsive to cost shocks. During the stationary phase, price responds to cost but is much less sensitive than under non-collusion or simple monopoly; a low price variance may then be a collusive marker. Compared to when firms do not collude, cost shocks take a longer time to pass-through to price.     

Fines and reputational sanctions: The case of cartels

In this article, we revive an old debate in the law and economics literature: the relative role of public and reputational sanctions in deterring misconduct. We propose an empirical framework, which accounts for public sanctions (in our case cartel fines) and a more direct measure of reputational sanctions, harnessing recent developments in opinion mining. We use the intensity and the sentiment of media exposure of misconduct as a measure of reputational effect and thus approximation of the reputational sanction. As a demonstration, we combine an event study approach, sentiment analysis, and econometric techniques on a sample of 339 listed cartel member firms, prosecuted by the European Commission between 1992 and 2015. Our results offer evidence that in the context of cartels, public and reputational sanctions act as substitutes: where there is a reputational penalty, increasing this penalty reduces the effect of the public sanction. One the other hand, in the absence of a reputational punishment, the effect of the cartel fine steps in.     

POST-CARTEL PRICING DURING LITIGATION

Standard methods in the U.S. for calculating antitrust damages in price-fixing cases are shown to create a strategic incentive for firms to price above the non-collusive price after the cartel has been dissolved. This results in an overestimate of the but for price and an underestimate of the level of damages. The extent of this upward bias in the but for price is greater, the longer the cartel was in place and the more concentrated the industry.     

Collusion at the extensive margin

We augment the multi-market collusion model of Bernheim and Whinston (1990) by allowing for firm entry into, and exit from, individual markets. We show that this gives rise to a new mechanism by which a cartel can sustain a collusive agreement: Collusion at the extensive margin whereby firms collude by avoiding entry into each other's markets or territories. We characterise parameter values that sustain this type of collusion and identify the assumptions where this collusion is more likely to hold than its intensive margin counterpart. Specifically, it is demonstrated that where duopoly competition is fierce collusion at the extensive margin is always sustainable. Finally, we provide a theoretic foundation for the use of a “proportional response” enforcement mechanism.     

OUTPUT‐EXPANDING COLLUSION IN THE PRESENCE OF A COMPETITIVE FRINGE*

Following the structure of many commodity markets, we consider a few large firms and a competitive fringe of many small suppliers choosing quantities in an infinite‐horizon setting subject to demand shocks. We show that a collusive agreement among the large firms may not only bring an output contraction but also an output expansion (relative to the non‐collusive output level). The latter occurs during booms and is due to the strategic substitutability of quantities. We also find that the time at which maximal collusion is most difficult to sustain can be either at booms or recessions. The international copper cartel of 1935–39 is used to illustrate some of our results.     

Flagging cartel participants with deep learning based on convolutional neural networks

Adding to the literature on the data-driven detection of bid-rigging cartels, we propose a novel approach based on deep learning (a subfield of artificial intelligence) that flags cartel participants based on their pairwise bidding interactions with other firms. More concisely, we combine a so-called convolutional neural network for image recognition with graphs that in a pairwise manner plot the normalized bids of some reference firm against the normalized bids of any other firms participating in the same tenders as the reference firm. Based on Japanese and Swiss procurement data, we construct such graphs for both collusive and competitive episodes (i.e when a bid-rigging cartel is or is not active) and we use a subset of graphs to train the neural network such that it learns distinguishing collusive from competitive bidding patterns. With the remaining graphs, we test the neural network’s out-of-sample performance in correctly classifying collusive and competitive bidding interactions. We obtain a very decent average accuracy of around 95% or slightly higher when either applying the method within Japanese, Swiss, or mixed data (in which Swiss and Japanese graphs are pooled). When using data from one country for training to test the trained model’s performance in the other country (i.e. transnationally), predictive performance decreases (likely due to institutional differences in procurement procedures across countries), but often remains satisfactorily high. All in all, the generally quite high accuracy of the convolutional neural network despite being trained in a rather small sample of a few 100 graphs points to a large potential of deep learning approaches for flagging and fighting bid-rigging cartels.     

Examining models of collusion: The market for lysine

We compare the experience with collusion in the market for lysine with the predictions of theory. The lysine market provides an ideal setting following the confessions of cartel participants in antitrust investigations. Data availability allows demand and cost functions to be estimated and observed mark-ups compared with predictions. We find that several integral aspects of collusion in the lysine market are not adequately addressed in the literature: the dynamics associated with entry and investment; persistent asymmetries between firms; the cartel's bargaining problem; and the existence of cheating in equilibrium. These issues are likely to have much wider applicability beyond the lysine market.     

The Effect of Bid Rigging on Prices: A Study of the Highway Construction Industry

This paper examines the effect of multimarket contact in afirst price sealed bid government procurement auction market. It investigates whether bidprices in the highway construction industry are related to conditions that favor the formation of a cartel.Repeated contacts among firms are found to have a significantly positive effect on the winning low bidwhich leads to higher profit. Further, rivalry among few firms tends to exacerbate the multimarket effect.The results in this study additionally support the recent theoretical predictions that collusion isbetter sustainable during economic downturns.     

Cartel birth and death dynamics: Empirical evidence

This paper examines how a gradual tightening of antitrust enforcement impacts cartels’ births and deaths. To avoid the inherent sample selection bias in prosecuted cartel studies, we use a unique dataset of Swedish legal cartels registered between 1946 and 1993. We compare estimates from a count model (considering only registered cartels) and a Hidden Markov Model (allowing for potentially unregistered cartels) to identify observed and hidden cartel dynamics. The count model suggests that strengthening antitrust enforcement has a deterrent effect, but the Hidden Markov Model suggests otherwise. Despite stricter competition laws and a credible threat of cartel prohibition, cartels continue to form, but do so undercover. Additionally, our results suggest that the strengthening of competition law has little impact on destabilizing existing cartels.