This working paper, which is available here, is still rough around the edges, but it contains a number of interesting insights, which I thought might be of interest.
This essay traces how the institutional setting of U.S. cartel enforcement evolved over the years, and assesses these developments from an optimal deterrence framework.
In doing so, the authors also review the outcomes of the various US policy regimes in terms of number of cartels prosecuted, the level of financial penalties imposed per individual and firm, and of jail time for cartel crimes. The authors also offer an analysis of how cartel enforcement has varied with recent US Presidential administrations.
Section 3 describes how cartel enforcement has evolved in the US since 1890.
Cartel enforcement in US began with the passage of the Sherman Act, which imposed a maximum fine for collusion of USD 5,000, raised to USD 50,000 in 1955. Jail time was not actively pursued until the late 1950s, when the DOJ pushed for greater incarceration for collusion. This occurred in tandem with an increase in the robustness of cartel enforcement in 1958-61, reflected in a number of electrical equipment price fixing cases which involved 30 corporations and 45 defendants. These cases received significant public attention due to the size of the overcharges, the high profile of the companies involved in the collusion, and the relatively small financial penalties and jail time imposed on “respectable” people.
As a result, the years between 1960 and 1978 witnessed important structural shifts in the institutions governing cartel enforcement, three of which merit a reference: the imposition of higher financial penalties, greater criminal sanctions, and the implementation of leniency programmes. While criminal enforcement increased during the 1980s in terms of the number of cases and criminal prosecutions, and marginally in terms of penalties imposed, the first iteration of the leniency program was ineffective.
It was only in the 1990s that cartel detection mechanisms, and the leniency program in particular, were successfully reformed. The new leniency program, which provided for full amnesty, proved much more effective and was a major success, transforming cartel detection. One important consequence of this program was a change in the type of cartels investigated by the US authorities. Prior to 1995, less than 1% percent of all DOJ cartel indictments involved non-US firms. Leniency helped uncover an increasing number of international cartels, and started a trend towards the internationalisation of cartel prosecution. This, in turn, triggered an increase in the depth of cooperation of US authorities with foreign enforcers, including coordinated cartel enforcement and information exchange. In more recent years, penalties kept increasing and there has been an enhanced focus on individual liability.
Section 2 describes the intellectual context for cartel enforcement.
Cartels are not very interesting from a conceptual standpoint. The identification of consumer welfare harm caused by a cartel welfare is simple, and it used to be thought that the per se application of liability to hard-core cartels made cartels less effective and destabilised them. This is apparent in how the Chicago School dealt with cartels. Stigler noted that collusion was possible, but believed that cartels would break down easily. Bork treated cartels summarily in the ‘Antitrust Paradox’, reflecting his view that cartels were both conceptually and doctrinally easy situations to rectify. Posner focused on the need for higher penalties for cartels and on the challenges of distinguishing tacit collusion from anticompetitive agreements, but other than that devoted little attention to the topic.
However, this lack of doctrinal attention to cartels does not reflect the practical importance of fighting cartels – or how cartel enforcement has evolved over the years. In effect, the authors make the point that developments regarding cartel enforcement over the years do not reflect critical reflections on its conceptual foundation or developments in economic thinking, but merely the need to reform prosecution mechanisms.
Section 4 reviews intertemporal patterns in cartel enforcement.
This section engages in a statistical analysis of cartel prosecution and enforcement data between 1969 and 2016 in the US. The specific annual data examined include: the total number of antitrust cases; the total number of civil and cartel cases; penalties imposed in cartel cases, such as the number of individuals incarcerated and the number of days served in jail; and monetary fines levied on individuals and corporations.
The authors find that the DOJ’s average won-loss record on cartels is near perfect, ranging from 98% to 100% for most years. This implies that cartel cases filed nearly always lead to prosecution. Secondly, the typical time from initiating a cartel case to prosecution appears to be in the 1.5-year range.
The authors also identify a number of interesting developments over time, and by Presidential administration. Prior to 1979-80, the mean number of US cartel cases was about 20 per year. The mean number rises sharply during the Reagan years to about 60 cases per year. Thereafter, the number falls steadily, reaching a low of 25-30 cases per year during the George W. Bush administration, after which there is a sharp increase in the mean number of cartels prosecuted during the Obama administration. This evolution seems not to be related to the actual prevalence of cartels in the market, but to the attitudes of those responsible for competition enforcement at the time. This conclusion finds support in the ratio of cartel cases to total antitrust cases. The pre-Reagan years saw an average ratio of about 30%, which increased to over 70% during the Reagan years. The ratio then declines steadily after that, reaching a low of 36% during the George W. Bush, before picking up again during the Obama administration. It should be noted, however, that there are significant variations between years of the same administration, however.
As regards sanctions, the number of individuals and corporations fined have steadily declined over time, even if there is no definitive shift in the number of individuals sentenced to incarceration. The total number of individuals incarcerated increases dramatically during the Reagan administration, then falls to relatively low levels before showing an uptick more recently. However, the severity of penalties has increased dramatically over time, with the average number of days of incarceration having risen dramatically from under 100 days to over 700 days recently. Similarly, there has been an increase of an order of magnitude in the number and amount of fines levied on individuals and corporations after the mid-1990s. This is clearly a result of policy changes which have taken effect across multiple Presidential administrations.
Finally, the authors look at the total funding for the Antitrust Division, and the ratio of cartel fines to the Antitrust Division’s funding. Total funding decreased markedly during the Reagan administration, and then increased for several years before becoming relatively flat since 2003-04. At the same time, the “ratio” of total cartel fines to the Antitrust Division’s budget averages around 0.4 before the 1995-1996 period, after which it becomes closer to 5, with highs reaching values as large as 10-12.
Section 5 seeks to identify the various forces driving cartel enforcement, while Section 6 discusses whether there has been a decline in cartel enforcement in the U.S.
Multiple forces likely influence the DOJ’s cartel enforcement. The authors focus on procedural developments that are widely thought to have had an impact on cartel enforcement – namely the adoption of an effective leniency program in 1993 and reforms increasing sanctions and prison terms –, on political variables – such as presidential terms –, on the institutional context – e.g. whether the levels of merger control or funding impact cartel enforcement – and on the macroeconomic context. Their analysis does not find a significant relationship between increases in the severity of sanctions and the number of cartels prosecuted; nor is there any significant relationship between the levels of merger control, the macroeconomic context, the DOJ’s funding and cartel enforcement. It seems that Republican administration charge fewer cartels, but given the high level of enforcement during the Reagan years the authors hypothesise that this is a consequence of the very low levels of enforcement during the George W. Bush years.
The question this then raises is what are the reasons behind the DOJ prosecuting fewer cartels over time. This has been the case for a number of years, aside from the increase under the Obama administration.
One possible reason is that perhaps leniency is working, and that, in combination with stronger cartel enforcement around the world, and increased compliance and audit efforts on the part of companies, maybe fewer international cartels are being formed. The opposite hypothesis is that cartels have gotten smarter and become more adept at avoiding detection. From this perspective, cartels have learned how to avoid stricter antitrust screens and scrutiny, and may even have started to game leniency and use it to punish competitors and/or harden existing cartels.
Another possibility is that the increased difficulty of private plaintiffs in bringing successful cases in situations where (at least) tacit agreement is present, in conjunction with aggressive prosecution of cases where direct evidence becomes available through leniency, has led rational firms to adopt collusion strategies that fall in the murkier legal area of tacit agreement, e.g. via partial and common ownership. Such strategies make it harder to identify instances of anticompetitive agreement, and to distinguish them from mere market interdependence. Additional possibilities are that shifts in market structure and increased market concentration are making it easier for firms to pursue strategies of mere tacit collusion, or that a number of local cartels go undetected by local (state) authorities.
Section 7 briefly seeks to determine whether there any links between U.S. and European Commission cartel investigations.
There are some parallels between the US and the EU as regards cartel enforcement. While the first legal prohibition against collusion was adopted in the 1957 Treaty of Rome, cartel enforcement was rather sparse for decades in the EU. Significant fines and prosecution of cartels did not materialise until after the First Leniency Notice was adopted in 1996.
While empirical literature on European cartel enforcement is relatively small, the evidence to date suggests a similar trend in enforcement to that of the United States. The time-path of the fines-per-cartel case is relatively similar to that of the U.S., with fines rising dramatically after the year 2000. However, the number of cartels prosecuted is very low: the average over the full sample period is only 4.5 cartels per year, and the highest number is 10 cartels prosecuted in 2001 and in 2014. These numbers are even lower than in the U.S.
Based on aggregate data, the authors find that there is no statistically significant interrelationship between U.S. and EU cartel cases. This is not to say there are no dynamic interrelationships; it is more likely that if we study the same issue on a cartel-by-cartel basis, we may well find that specific cartel prosecutions in one jurisdiction lead to sharing of information and prosecutions in the other, with consequent implications for fines and other penalties.
This piece is still clearly a working paper – it was unclear to me how some sections related to others, or whether the structure makes complete sense (you will notice that I described section 3 before section 2); how the analysis of US and EU cartel enforcement trends is relevant for the paper; or whether comparing cartel enforcement in the EU and the US at a (con)federal level is appropriate, given the very active role played by national competition agencies in Europe by comparison to state Attorney Generals in the US.
The paper nonetheless provides very interesting insights into US cartel enforcement over time. Some of the findings are quite striking, such as the US DOJ’s success rate, but the most interesting implication of this piece for me is how important procedural developments have been for effective enforcement. Given that the focus of the paper is on assessing cartel enforcement by reference to optimal deterrence theory, this conclusion is fertile with implications for how relevant it is (or not) that sanctions are set at the optimally deterrent level.