There is a growing concern among some scholars and practitioners that penalties for companies participating in a cartel have become excessive in the sense that they are more than sufficient to deter and may be causing social harm. Here I will argue that, even if there is validity to those claims, it is imprudent to begin lightening up on enforcement.
There is no doubt that “cartel fines from public enforcement … are staggering” and that, when a cartel operates in a jurisdiction allowing for private customer damages, the amounts can be vastly higher than staggering, shall we say gargantuan? Of course, the incremental profits earned through collusion may as well be staggering or even gargantuan. Thus, any assessment of whether penalties exceed or fall short of what is necessary to deter cartel formation requires a careful comparison of penalties with those profits while taking account of the likelihood of cartel members ever paying those penalties.
Recently, some studies have pursued such an analysis. Using a traditional approach that compares the profit from colluding with the expected penalty (which equals the penalty multiplied by the probability of discovery and conviction), Connor and Lande (2012) argue that penalties are in the under-deterrence region. However, a different conclusion is reached in Allain, Boyer, and Ponssard (2011) who take issue with the method and estimated overcharges used in the Connor-Lande study and conclude that “the current level of EU cartel fines is adequately dissuasive.” Consistent with this methodological critique, my own analysis suggests that penalties are significantly more deterrent than is generally recognized because they accumulate over time and cartel formation is deterred as long as cartel stability is undermined which is a less stringent criterion than making collusion unprofitable. 
Due to the lack of good data, all of these findings are tentative but my task at present is not to critique these analyses. Indeed, for the sake of argument, let me suppose that current enforcement policy exceeds that necessary to make collusion unprofitable. Even given that, it is not clear that we should be any less aggressive in our enforcement.
Even if penalties are at a level to make collusion unprofitable, are managers aware of this fact?
The global vitamin cartel was one of the big early successes in this new regime of enforcement. Hoffman LaRoche was given the largest fine in U.S. history of $500 million and private customer damages were more than twice as large. In spite of imposing fines and damages in the billions of dollars, these penalties fell short of the incremental profits from colluding in the U.S. market. Collusion was ex post profitable! Of course, those penalties were levied around 2000 and, since that time, some jurisdictions – in particular, the EU – have increased their penalties and, as mentioned above, some studies find them at levels that make cartel formation ex ante unattractive. But this begs a question: What do managers believe? Do managers believe that the vitamins case is representative? Or are their beliefs more in line with the findings in Allain et al (2011)? Managers may witness cartels being discovered and prosecuted in other markets and hear about the fines and damages (as well as possibly prison sentences) but they surely have rather limited information about the illicit profits (and how it compares to those penalties) and especially about the likelihood of having to pay those penalties if they chose to collude. We can calculate the expected profit from forming a cartel – and that is relevant information – but it isn’t clear that is the calculation going on in the head of a manager when he is contemplating forming a cartel. For that reason, it would seem unwise to lighten up on enforcement based on calculations comparing collusive profit with expected penalties. It is prudent to wait until there is convincing evidence that cartels are actually being deterred. While it is admittedly difficult to collect such evidence, it is not impossible and doing so is necessary before stemming the tide of intensified enforcement.
Even if penalties are at a level to make average cartels unprofitable, what about above average cartels?
The most vociferous calls that penalties are excessive relate to the European Commission’s 2006 Fine Guidelines, which is the focus of the analysis in Allain et al (2011). (And, with private customer damages developing in the EU, over-enforcement claims are sure to be accentuated.) Here, it is critically important to recognize that fines have been shown to be sufficient to deter collusion based on the average cartel overcharge. EC fines are tied to revenue in the affected markets and not to incremental profits or customer losses, so the penalty does not scale up with the overcharge. If we take these estimates on face value, the only cartels that will form are those with abnormally high overcharges which are the ones imposing the largest losses on consumers. This is hardly reassuring.
The problem here resides in the penalty formula not being proportional to the additional profits from colluding. While I have not engaged in a systematic analysis of competition authority practices, my sense is that an EC-style revenue-based formula is typical. That is the case in the U.S. as well. Though U.S. Sentencing Guidelines have a maximum of “not more than the greater of twice the gross gain or twice the gross loss,” apparently that sort of calculation is not standard practice when the U.S. Department of Justice sets a fine.
That cartel profits are not taken account of in setting or negotiating fines is a criticism of both the competition authority and the body that sets their budget. One defense of this practice is that it is too costly to calculate those profits. That does not seem credible. There are many plaintiffs who perform exactly that exercise for much smaller markets involving much smaller sums. If a plaintiff can engage in a cost effective calculation of the impact of collusion on profits when hundreds of thousands of dollars of claims are at stake then a competition authority should be able to do so when millions of dollars of fines are at stake. A second defense is that a competition authority has limited resources and it is better for it to use those resources to develop additional cases. That is a valid point but then the argument should be made to increase the competition authority’s budget so they can engage in the proper setting of fines. We must remember that the ultimate goal is not to convict and penalize cartels but rather to deter their formation, and that requires tying penalties to illicit profits. This point is worth emphasizing as competition authorities may attach too much weight to disabling cartels relative to deterring cartels. But let me not dwell on that point as it will be the focus of a future post on this blog.
Even if penalties exceed what is necessary to make collusion unprofitable, where’s the harm?
As all cartels are harmful, on first glance it would seem misplaced to worry that penalties exceed levels sufficient to deter. But, as has been noted by others, there are at least two sources of social harm from excessive enforcement. First, firms may avoid legitimate activities out of fear that their behavior would be misconstrued as collusive. Second, at least in the case of the U.S. where there is an overly active litigation scene, customers may pursue unjustified cases with the hope that the prospect of legal fees, discovery, and the small chance of having to pay large customer damages will induce settlement by innocent suppliers.
I’m skeptical of these concerns, at least for the U.S. The standards for proving guilt for a Section 1 violation have always been high. Furthermore, Twombly has raised the bar as now discovery can be avoided unless the plaintiff can plead “facts that are suggestive enough to render a §1 conspiracy plausible.” At present, it is quite difficult for a plaintiff to get past the pleading stage without some reasonably convincing evidence that there was collusion and it was of the unlawful variety.
Even with that skepticism, the first concern should not be dismissed out of hand. I could imagine a scenario in which companies may choose not to form a trade association or a research joint venture because of the possible misinterpretation that their actions have the intent to coordinate prices and allocate markets. But again, show me the evidence! We have clear-cut evidence of cartels continuing to form and causing harm to consumers. If the claim is that enforcement is harming companies not engaged in collusive practices then evidence is needed of these deterred legitimate activities before enforcement is weakened.
Another potential harm that, to my knowledge, is generally not mentioned is that senior management may put in place costly mechanisms to monitor lower level employees from colluding. Even if enforcement is so severe that collusion is not in shareholders’ interests, employees may still collude in order to improve their perceived performance within the firm. Senior management may then need to spend resources to detect and deter lower level employees from participating in a cartel, just as it might put in place monitoring systems to detect and deter accounting fraud and embezzlement.
On this point, AkzoNobel – a notorious cartelist involved in at least nine cartels in the last 15 years – issued a Competition Law Compliance Manual to its employees in 2008. CEO Hans Wijers states in no uncertain terms: “The Board of Management considers compliance with competition law to be more than a legal requirement; it is core to AkzoNobel’s value of integrity and responsibility. … Disciplinary action will be taken against any employee who violates competition law. … In this area we are a ‘zero tolerance company’.” This is all well and good but are they investing resources to monitor employees and not just warn them? If not then perhaps senior management is not so concerned with their employees colluding which suggests that penalties are not excessive and causing harm.
In conclusion, before making enforcement less aggressive, I would say: Show me the deterrence of cartels. Show me the harm from over-deterrence. Until that is done, competition authorities should continue to “keep the pedal to the metal” and maintain an enforcement policy that is “fast and furious.”
 George L. Paul, “Leniency Programs and Growing Fines: The Risk of Over Deterrence, White & Case, Winter 2009; p. 1.
 John M. Connor and Robert H. Lande, “Cartels as Rational Business Strategy: Crime Pays,” Cardozo Law Review, 34 (2012), 427-90.
 Marie-Laure Allain, Marcel Boyer, and Jean-Pierre Ponssard, “The Determination of Optimal Fines in Cartel Cases: Theory and Practice,”Concurrences – Competition Law Journal 4-2011, 32-40; p. 32. However, for a different view, see Emmanuel Combe and Constance Monnier, “Fines Against Hard Core Cartels in Europe: The Myth of Overenforcement,” Antitrust Bulletin, 56 (2011), 235-76.
 Joseph E. Harrington, Jr. “Penalties and the Deterrence of Unlawful Collusion,” The Wharton School, University of Pennsylvania, November 2013 (Economics Letters, forthcoming).
 John M. Connor, “The Great Global Vitamins Conspiracy: Sanctions and Deterrence,” American Antitrust Institute, Working Paper 06-02, February 2006. The vitamins cartel was even more profitable in the EU and other countries.
 One method for assessing the impact of a new policy on the latent population of cartels is devised in Joseph E. Harrington, Jr. and Myong-Hun Chang, “Modelling the Birth and Death of Cartels with an Application to Evaluating Antitrust Policy,” Journal of the European Economic Association, 7 (2009), 1400-35.
 This statement is based on Gregory J. Werden (U.S. Department of Justice) in an email correspondence with Yannis Katsoulacas, December 4, 2013.
 Bell Atl. Corp. v. Twombly, 550 U.S. 544 (2007) at 556.