A Rebuttable Presumption of 20% Price Rise for Damages against Proven Cartels?

(by Bruce Lyons)  The UK Government yesterday published a consultation on private actions in competition law.  It includes a suggestion that there should be a rebuttable presumption that the cartel has resulted in higher prices.  20% is offered as a possible presumed increase due to the cartel.  This is an excellent idea and should be widely supported.  The only room for debate should be over the precise presumption to adopt and whether to extend this approach beyond cartels.

Even when a cartel has been successfully prosecuted by the OFT, the current burden of proof is that a customer seeking damages has to prove how much she has been harmed.  It is as if the presumption is that the proven cartel did not raise prices.  If that were true, it would make you wonder why cartels were illegal in the first place.  Of course, it is not true and there is well documented evidence that most cartels raise prices very substantially.  Consequently, it is both equitable and efficient to presume that a cartel raises prices.

Why does the burden of proof matter if all you have to do is look at the evidence on prices?  There are two problems.  First, statistical data needs to be collected, but most of this is in the hands of the cartel.  Second, the data must be processed to understand the economic effects of the cartel, and there is more than one way to do this.  Taken together, there is plenty of room for obfuscation and it makes it very hard work for the (often numerous) customers to prove a precise level of damages.  In other areas of law, judges fully understand the principle that an informational advantage of one party should naturally lead to that party taking on the burden of proof.  There is no reason not to adopt it for cartels.

The next issue is that if we are to presume a price rise, we need to have a suitable number to presume.  Needless to say, no two cartels have exactly the same price increase, but that is not the point because the presumption is rebuttable if either plaintiff or defendant has the evidence to show otherwise.  A casual thinker might claim that this requirement to specify a default number is a fundamental problem with the change in presumption from ‘no harm’.  This is not a serious argument because zero is just another number but, unlike twenty, it is one that we know to be at the lower bound rather than somewhere in the middle of the true range.  For the same reason, it provides a useful focus for a quick, low-cost, out-of-court agreement if 20% is ‘about right’ for a significant number of cases.

Is 20% the best number to use?  Much of the economic evidence points to a somewhat higher average cartel effect, so I would not be averse to a slightly higher figure, but it is a cautious start that might be revised in the light of evolving evidence.  As the consultation document notes, it also has the virtue of being the number I suggested when making the case for a rebuttable presumption in my response to last year’s European Commission consultation on its draft guidance paper on the quantification of damages.

My only serious concern with the BIS proposal is in an important footnote attached to the relevant section (starting #4.40).  This states that references to cartels should be taken to refer to any breach of the Chapter 1/Article 101 prohibition on anticompetitive agreements.  This goes too far.  In the case of cartels, damages in addition to a fine have a positive effect on deterrence.  This is because the probability of detection times the size of fine is likely to be less than the payoff to cartel formation. Even in the unlikely event that it is not, there is no competition downside of excessive deterrence of cartels.  However, where the main offence is an exclusionary practice, as opposed to exploitation of consumers, there is neither economic evidence to support the 20% figure nor a strong presumption of under-deterrence.

In particular, there is a substantial danger of chilling competition in the context of business practices that may result in foreclosure but in other circumstances may be pro-competitive (e.g. quantity discounts, exclusive dealing).  Excessive deterrence is possible if the penalties of a business practice are seen to be large in one case where it is anticompetitive, and consequently other businesses play safe in avoiding the practice in circumstances where it would be pro-competitive. Furthermore, the reward of damages can act as an incentive for a weak competitor to threaten a private action in order to induce a strong competitor to compete less aggressively. This does not mean that there should be no damages actions in foreclosure cases.  However, alongside the informational advantages a competitor is likely to have in relation to the relevant calculation of damages (relative to the information available to customers), it does suggest that courts might reasonably place the burden of proof on the plaintiff (i.e. foreclosed firm) in quantifying damages above zero.

5 Responses to A Rebuttable Presumption of 20% Price Rise for Damages against Proven Cartels?

  1. David Elliott says:

    Bruce
    I share your enthusiasm up to a point but surely we don’t have the evidence to support any particular number (no doubt you will correct me). I very much suspect than any so called calculation of the price hike resulting from a cartel has not been based on an adequate counterfactual. Certainly the surveys by Connor seem to have little foundation. Of course you might argue that it does not matter what number we choose for the rebuttable presumption.
    A key point here is what an economist might call competitive is based upon effect (i.e. at one extreme P=MC). For cartels the law looks at behaviour, thus in terms of simple oligopoly models any conjecture ≠ 0 is collusion, whether tacit or explicit. This I take to suggest the law regards zero conjectures as competitive. Now even between simple oligopoly models zero conjectures will give prices that differ. Thus the competitive counterfactual will vary with the oligopoly model chosen and more importantly its “effect”. I am not sure the lawyers will like that and it clearly influences your choice of rebuttable price rise. (The zero conjecture price cost margin for a “Cournot” cartel with 8 equal sized firms and an elasticity of market demand of 1.0 is 12.5% – I am sure you can correct me on that if wrong!)
    Of course we can use various empirical methods to estimate a “counterfactual” (the OECD has a very good survey at http://www.oecd.org/dataoecd/12/9/48849122.pdf). However often the data do this will not exist and we are forced back to theoretical approaches as in Archer Daniels Midland where the damages issue was ultimately is it Bertrand or Cournot?
    David Elliott

  2. Bruce Lyons says:

    [uploaded by BL on behalf Sebastian Peyer] I agree with the consultation paper’s assumption that cartel victims face considerable difficulties in obtaining redress. However, a rebuttable presumption for a 20 percent price increase in cartel cases is debatable for two reasons. First, even in some cartel cases it is difficult to establish the economic effects. Even on the horizontal level – and Bruce has pointed out the dangers of extending the presumption to all anticompetitive agreements – there are situations where the effects on prices are unclear like, for example, in the OFT’s cover pricing investigation. Second, it is unlikely that shifting the burden of proof to the defendant will save cost as it is hoped by BIS. The defendant will in all likelihood use economic experts to rebut this presumption which, in turn, will prompt the plaintiff to use experts too. Sebastian

  3. Bruce Lyons says:

    David Elliott makes two points. First, damages depend on the counterfactual (i.e. what would have happened if there had not been a cartel) and this is difficult to determine. Second, the literature on cartel price rises is too weak to build a quantified presumption. There is a useful review of typical cartel price rises and a detailed discussion of alternative counterfactuals in the DG Competition draft guidance paper on the quantification of harm in actions for antitrust damages (http://ec.europa.eu/competition/consultations/2011_actions_damages/draft_guidance_paper_en.pdf). I agree that the literature has weaknesses, but overall I think the findings are stronger than you suggest. On the counterfactual, if you adopt a sophisticated structural model there will always be room to argue, much as there is when predicting the economic effects of a merger. However, an advantage over merger models is that there are usually comparative periods when the cartel was, and when it was not, operational. This makes cartel damages a less inexact science. In any case, when we are estimating the consequences of illegal activity, I see no reason or moral justification to say “it is difficult to estimate damages so we should give the proven guilty party the benefit of the doubt”.
    Sebastian Peyer also makes two points. His first is similar to the counterfactual issue. For example, was the cover pricing ‘cartel’ really likely to raise prices? (https://competitionpolicy.wordpress.com/2011/03/15/%E2%80%98cover-pricing-object-or-effect/) While I agree that it is unlikely that in this particular case prices would have increased by 20%, it would be wrong to throw away the baby with the bathwater – isolated exceptions should not drive competition policy. As I said in my blog post, we should restrict the presumption to cartels, but inevitably there will be a grey area. We want to draw the line as to the definition of a cartel as clearly as possible, but the rebuttable nature of the presumption means there should not be an injustice even if a case falls on the wrong side of that line for some reason. Seb’s second point is that there might still be battles between experts so costs will not be saved. I disagree because sometimes 20% will provide a focal point for a quick and low cost agreement. It will also help the judge if both sides get bogged down in an unhelpful destructive battle focused only on the weakness of the other’s evidence.

    • Bruce Lyons says:

      I have been asked to justify my claim that judges understand the principle that an informational advantage of one party should naturally lead to that party taking on the burden of proof. Competition law adopts two clear and central examples in which ‘defendant’ firms have to satisfy the burden of proof based on their informational advantage:
      1. Article 2 of Regulation 1/2003 states: ‘The undertaking or association of undertakings claiming the benefit of Article 81(3) of the Treaty shall bear the burden of proving that the conditions of that paragraph are fulfilled.’ This is repeated in its Article 101(3) guidance. Furthermore, in its Article 102 guidance, the Commission uses wording implying a similar burden in the case of a dominant firm claiming offsetting benefits (even though there is no equivalent Article 102(3)): ‘the Commission will look into claims by dominant undertakings that their [apparently abusive behaviour] may lead to savings in production or distribution that would benefit customers.’

      2. Paragraph 87 of the Horizontal Merger Guidelines states clearly that the burden of proving efficiencies is on the merging parties precisely because they have access to all the relevant information:
      ‘Most of the information, allowing the Commission to assess whether the merger will bring about the sort of efficiencies that would enable it to clear a merger, is solely in the possession of the merging parties. It is, therefore, incumbent upon the notifying parties to provide in due time all the relevant information necessary to demonstrate that the claimed efficiencies are merger-specific and likely to be realised. Similarly, it is for the notifying parties to show to what extent the efficiencies are likely to counteract any adverse effects on competition that might otherwise result from the merger, and therefore benefit consumers.’

      As far as I am aware, these have not been challenged in the Court and are unlikely to be. Beyond this, I am moving far beyond my economics training – I am not a lawyer. However, conversations with lawyers and my reading of legal texts lead to the following observations.
      a) In criminal law, the prosecution must place the accused at the site of the crime, but the defendant has the burden of proof if he claims an alibi that he was elsewhere at the time. This seems natural as the defendant is best placed to produce the evidence of where she was.
      b) In a tort case of contributory negligence, for example a car accident in which a passenger had encouraged a drunk driver to behave recklessly, the passenger’s damages are reduced and a 50% reduction appears to be a standard starting point, even though further evidence may be used to determine a different final figure [e.g. Deakin et al ‘Tort Law’, OUP, 6th ed., pp899-901].
      c) The doctrine of res ipsa loquitur (‘the issue speaks for itself’ because there is no reasonable alternative story) seems consistent with the idea we should start from non-zero damages; e.g. the cartel has been proved so this can be taken to imply higher prices.

      Bruce Lyons

  4. […] reforms proposed in the UK have been discussed in previous blog posts by Lyons and Peyer. They include: allowing the Competition Appeal Tribunal (CAT) to hear more kinds of […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: