Quantification of damages

-    Presumptions and statistical guidelines
-    Economic models for the calculation of damages
-    Econometric methods and data collection
-    Conclusions
-    Further information

Presumptions and statistical guidelines

Any plaintiff seeking damage compensation in court must provide sound evidence of the facts of the case in order to (i) demonstrate and prove the occurrence of a cartel-related damage and (ii) enable the court to estimate the amount of damages. In this respect, plaintiffs can rely on the following legal presumptions and statistical guidelines:

Factual presumption on the occurrence of damages

The Court of Justice of the European Union, the General Court and different national courts across Europe such as the Federal Court of Justice in Germany [Berlin Ready-mixed Concrete (2005)] and the Provincial Court of Valladolid in Spain [Sugar (2009)] acknowledge that anticompetitive ‘hardcore’ agreements between competitors regularly result in illegal profits of the cartel members to the detriment of their customers, and in a deterioration of the customers’ negotiation position. In the Italian judgment of the landmark Manfredi price-fixing case in 2007, the Justice of the Peace of Bitonto concluded that there was a ‘simple presumption’ (presonzione semplice) of an overcharge of 20 %, which the defendant had failed to rebut.  In another case, the Italian Court of Cassation [No. 2305/2007] held that, with respect to causation, a direct link between a cartel and the damages suffered by consumers may be presumed, because downstream contracts between cartel participants and consumers are usually the means by which the cartel is put into effect. Thus, it is generally sufficient for a plaintiff to prove the existence of a cartel agreement and it is for the defendants to rebut the presumption that this agreement has caused a loss.

Statistical and other guidelines on the amount of cartel-based damages

There are several empirical studies aimed at assessing the cartel-related damage, that is, especially the price overcharge. Studies evaluating cartel statistics back to the time of World War I reveal that cartels cause an average overcharge between 21 % and 43 % above the hypothetical competitive price. Some cartels have even raised prices 200 % above the unit cost of production and distribution. In their meta-analysis, Connor and Bolotova [Cartel overcharges: Survey and Meta-analysis (2005)] used 800 different overcharge rates, which were collected from a variety of sources published over the past 125 years, and which were obtained by using  various quantificational methods. The mean and median values of the evaluated overcharges were 29 % and 19 % respectively. Connor and Lande [The Size of Cartel Overcharges: Implications for U.S. and EU fining policies (2006)] used a separate data set for European wide cartels where the overcharges varied between 28 % and 54 %. European cartels which were country-specific exhibited lower overcharges in the range of 16 % to 48 %. In a more recent study, John M Connor investigated 516 private international hardcore cartels from 1990 to 2008, which were subject to either public or private legal action. He concluded that the median cartel-related price overcharge varied between 17 % and 21 % [AAI Working Paper # 09-06, Cartels & Antitrust Portrayed, Private International Cartels from 1990 to 2008]. The findings of the 2010 Oxera report published by the European Commission show that the median overcharge is 18 % and the average (mean) overcharge is around 20 %.

Statistical studies and guidelines may be relevant in terms of judicial damage estimation and may serve the purpose of specifying the average overcharge in order to avoid the time and expense that would be required by the court to determine the actual loss in each case. For instance, Section 88/C of the 2009 Hungarian Competition Act expressly stipulates that in the course of evidencing the effects of price-fixing, market-sharing and other hardcore antitrust infringements on the price level it ‘shall be deemed that the infringement affected the price by 10 % unless the contrary is evidenced’. Similarly, the cartel fining guidelines issued by the US Sentencing Commission are based on the assumption that price-fixing, bid-rigging or market-allocation agreements impose an overcharge of 10 %. 

Economic models for the calculation of damages

The right to damages under European and national laws aims to compensate cartel victims by putting them into the equivalent position that would exist ‘but for’ the damaging illegal conduct. Therefore, most of the economic models on cartel damage estimation are guided by the principle of establishing a hypothetical scenario, commonly referred to as the ‘but for’ world, which would have prevailed on the market in the absence of the cartel. This hypothetical outcome is used as a reference to which the actual (cartelized) market outcome is compared, and the difference between the two serves as the basis for the estimate of damages caused by the cartel. The most common factor used when comparing the two markets is price or overcharge. In order to obtain a value for the damages incurred by a particular victim, one would typically multiply the estimated overcharge by the quantities purchased by that buyer during the period in which the cartel resulted in price effects.

The key challenge in any damage quantification is the determination of the hypothetical ‘competitive price’ that would have been paid by the customers in the absence of the cartel. Clearly, some assumptions have to be made and an estimate of a hypothetical market price absent collusion is, to a large extent, an attempt to construct the most plausible alternative scenario. One should aspire to employ sound modelling techniques and make assumptions and limitations transparent. At the same time, the model should be connected to the facts of the case.

Naturally, there are many different models for damage estimation. The most common approaches can be divided into three subgroups:


•    Comparative approaches; the ‘but for’ market conditions are proxied by the periods before and after the cartel activity, comparable sectors and/or geographic markets or a combination of the two that were in any event not polluted by collusive practices.

•    Cost-based approaches are conceptually based on the estimation of profitable cost margins.

•    Simulation techniques; by calibrating a standard oligopolistic competition model to the facts of the case and specific factors influencing supply and demand in the cartelized market, one tries to simulate the market outcome of various competitive counterfactual determining prices and quantities that would have been able to be charged and sold, had the interactions of the firms been more competitive.

Different methods can, and sometimes should, be used as complements in order to verify sensitivity and plausibility of damage estimates in any particular market.

Comparison Approaches

•    Before-and-After Method
With the before-and-after method, the overcharge is calculated by comparing prices set during the cartel period and the prices set in the period(s) before and/or after the price-fixing occurred. It is assumed that the price level in the cartel-free periods provides a reasonable approximation of the price ‘but for’ the cartel. A simple regression analysis with a dummy variable for the cartel effect can demonstrate how much of the observed price may be attributed to price-fixing. The analysis can be extended to account for external effects not caused by the cartel activity but changing over time. By introducing control variables into the econometric model, various parameters of demand and supply can be accounted for.

However, using post-cartel prices as the sole reference point can lead to an underestimation of the overcharge for two reasons. Firstly, tacit collusion may be easier to sustain once the firms have had explicit cartel agreements in place for a decade or more. This would prevent a gradual price decrease to more competitive levels even after the cartel has been discovered or fined. The second explanation lies in the deliberate attempt of the cartel members to maintain the higher price in order to reduce the damage estimation in the follow-on lawsuits [Harrington, Post-Cartel Pricing during Litigation (2004)]. Unless the pre-cartel prices are available and can be used for the ‘but for’ analysis, one can conclude that the approach of using the post-cartel prices exclusively would yield a lower bound of damages caused.

The before-and-after method has been used to calculate damages in US private antitrust cases since the 1920s. In Europe, it is a standard tool in the fine-setting of the competition authorities and courts [for example, European Commission, Vitamins Cartel (2001); German Federal Court of Justice, Paper Wholesalers (2007)]. In Germany and Austria, for example, the before-and-after method has also been applied by the courts also with regard to private antitrust actions for damages sustained [Germany: Higher Regional Court of Berlin, Berlin Ready-mixed Concrete (2009); Regional Court of Dortmund, Vitamins Cartel (2004) Austria: District Court of Graz-Ost, Driving School Cartel (2007), upheld by the Regional Court of Graz (2007)].

•    Yardstick Approach
The yardstick approach can be used to quantify the damage caused by the cartel by comparing the prices in the collusive market with the prices in a comparable market that was unaffected by any conspiracy. The latter is usually a similar product market and/or a similar geographic market or a combination of the two. In order to attribute the difference in prices (or other reference factors) to the illegal price-fixing, the benchmark market should ideally be characterised by competitive forces similar to the cartelized market (that is, similar cost and demand structures). As in the before-and-after method, it is possible to extend the approach to an econometric model which assesses differences across the markets including control variables, thereby isolating the effect of the illegal conduct.

•    Multidimensional Approaches
The ‘before-and-after’ and the yardstick method can be combined in a multidimensional approach. This method assesses the price developments in the benchmark market before, after and during the cartel activity. In other words, it combines the cross-sectional and time-series data into an econometric model in order to estimate the ‘but for’ price in the relevant market. Supply and demand are estimated by using a considerable amount of information of non-conspiracy periods as well as information on production, costs and capacity utilization. The cartel’s price effect can then be captured by including a qualitative variable for the time periods where the cartel was in place. Certainly, data availability can severely limit the applicability of this approach in damage calculation.

Profit- or Cost-based Approaches

•    Profit-based Estimation
By comparing the cartel profit realised during the conspiracy period with a normal return from oligopolistic competition that would have occurred ‘but for’ the cartel, one can estimate the excess benefits reaped from the conspiracy. The price overcharge per unit sold can additionally be calculated by dividing the ‘but for’ profit by the production volumes during the cartel period. In Germany, Section 33(3)(2) of the 2005 Act against Restraints of Competition suggests, alternatively or supportively to the above-mentioned economic models, that the judicial estimate of the size of the damage ‘may take into account, in particular, the proportion of the profit which the undertaking has derived from the infringement’. According to German law, this profit shall be calculated by subtracting manufacturing costs as well as accrued operating expenses from sales proceeds, whereas overheads and non-cartel-related operating expenses, which would have accrued without anticompetitive behaviour, are not to be subtracted.

•    Cost Plus Margin Method
Another method for estimating the overcharge uses a cost-per-unit plus a predefined mark-up as an estimate of the ‘but for’ price. Various cost measures can be used for this purpose – for example, a short-run incremental cost, long-run incremental cost or average unit production cost – although the question of which is the most appropriate one is case-specific. It is also important to take the temporal horizon into account when deciding on the suitable cost type. For example, the short-run marginal cost does not take long-run changes in fixed costs, overhead costs and investments in cost reduction into account, but measures only the cost of producing an incremental unit. Cost estimates are usually obtained by using accounting data, internal management reports, business plans or financial statements. The cost-based approach has been explicitly suggested as a possible means of damage estimation in antitrust cases by the German government with regard to the 2005 Act against Restraints of Competition.

•    Critical Loss Analysis
Critical loss analysis attempts to determine an upper bound of the overcharge which would still be profitable for the cartel to implement before the decrease in sales causes a decrease in revenue. In other words, by estimating the price elasticity of the cartelised product in question one can forecast the break-even point where the increase in profits obtained via the increase in price would no longer outweigh the decrease in profits from lower demand. This break-even price would serve as an upper bound for the overcharge. To be able to calculate the critical loss, one would need data on the cost structure and contribution margins of cartel members. The profitability limit of cartel overcharges can also be used as a plausibility test for the estimated damage obtained with other economic damage estimation models.

Simulation Models

Theoretical modelling uses the theory of industrial organisation to predict prices in different market outcomes. They mainly involve the use of models of oligopolistic competition (for example, the Cournot or Bertrand model of pricing) in order to quantify the effects of cartel behaviour, that is, to specify an economic equilibrium but for the cartel. The data categories typically employed in simulation are market factors influencing supply and demand under competitive conditions (such as industry concentration, marginal cost of production, the own-price elasticity of demand)  which are used to predict the ‘but for’ price. The method allows the calculation of not only a price which would prevail on the market absent the cartel, but also cartel profit margins as well as non-cartel profit margins. These profit margins can consequently be used as a proxy within the ‘cost plus margin’ method (see above). The size of the cartel-related overcharge is the difference between the cartel prices and the prices determined under the simulation procedure. Unlike other economic damage models, simulation models take into account the harm caused to customers by a quantity reduction. In this approach it is crucial to choose the appropriate competitive model, which would adequately explain competition and the market outcome in the absence of the price-fixing.

Econometric methods and data collection

Econometric Methods

Econometric methods allow for a multitude of non-cartel factors (such as demand and cost factors) to be accounted for in a given model. This is usually done by setting up a so-called ‘reduced form equation’ which describes the price (dependent variable) as a function of the external factors affecting it (the explanatory variables). Within the reduced form equation, the two most commonly used approaches in damage estimation are the dummy variable model and the residual model:

•    The dummy variable model uses data from the conspiracy and non-conspiracy periods to capture the cartel’s effect on prices. A dummy variable is introduced, which represents the presence of the cartel in the model by taking on the value ‘1’ in the conspiracy period. Running the regression will lead to price effects which cannot be explained by control variables (such as change in the input costs) and which are therefore captured in the cartel dummy. The method also allows controlling for different degrees of price-fixing by using values smaller or greater than ‘1’ so as to adjust the dummy variable to periods when price-fixing was weak and high values when price-fixing was strong.

•    The residual model uses data from a non-conspiracy period and predicts the hypothetical competitive price in the cartel period by assuming that the relationship between the explanatory variables and the price remained constant during the cartel activity. This approach is also known as the ‘prediction’ model. The overcharge is calculated as the difference between the actual cartel prices and the ‘but for’ prices estimated by the model.

Employing regression models should be done with both sufficiently large amounts of good quality data and a sound economic reasoning of why certain variables should be included.

Data Collection

The collection and preparation of the relevant information is of crucial importance and is a prerequisite for the sound application of any of the above mentioned damage quantification methods. For this reason, CDC has developed specific IT-solutions like the ‘tradebaCCC’ tool, which are designed for the collection and analysis of large volumes of purchase and market data. CDC’s approach under tradebaCCC gives an answer to two essential practical questions:

•    Significance of Data?
The more relevant transaction data are available, the more precise the assessment of the effects of the cartel in question and hence the damage estimation by use of the above-mentioned economic models and econometric methods. Correspondingly, the tradebaCCC tool is based on the idea of aggregating knowledge by collecting and processing relevant data of a large and diversified group of injured parties. In this way, it overcomes the problem that an individual claimant who can only demonstrate and evidence data on his or her own purchases of the cartelised good is hardly ever in the position to offer sound information on the actual effects of a cartel on the markets concerned (Approach).

•    Availability of Data?
Another essential benefit of the tradebaCCC tool is that it does not depend on any contribution by anybody other than the injured parties. It thus avoids lengthy discovery or disclosure proceedings against the cartel members (or third parties), whose results are uncertain anyway, as each cartel member can provide only data on its individual transactions itself, if at all. The right of an injured party to access the file of the competent competition authority does not provide any help in this respect either, as the competition authorities are not legally obliged to, and thus often do not, collect any evidence on the actual effects of the cartel in the case of a hardcore infringement

Conclusions

The burden of proving cartel damages generally lies with the victims of the cartel. It is up to them to obtain and present hard factual evidence and sound economic analysis which will then serve as the basis for discretionary damage estimation by the court. In the absence of a universal economic model of damage estimation in cartel cases, estimation of cartel-related damages requires a case-by-case approach, which gives due consideration to the specific characteristics of each market and cartel agreement. Indeed, all of the above-mentioned methods may complement each other and can be applied simultaneously to cross-check the plausibility and accuracy of the damage estimate obtained. However, irrespective of the method or combination of methods ultimately used to establish the counterfactual, which is the market outcome including prices that would in all likelihood have occurred ‘but for’ the infringement, such an estimation of a hypothetical market crucially depends on the availability of data to the plaintiffs as well as their financial, technological and expert resources. In this respect, CDC’s tradebaCCC tool, which is adapted to each case in order to fit the individual product and market characteristics, offers indispensable assistance.

Printer friendly version of this page