The England and Wales Court of Appeal has overturned the Decision of the United Kingdom Competition Appeals Tribunal (CAT) in the collective proceedings Walter Merricks v MasterCard, where final consumers are represented by Merricks. This is an important decision because it ensures that financial damages are recoverable when a firm’s anti-competitive practices affect millions of consumers. There are already a few articles published on the legal consequences of the judgment, in this short piece I am going to discuss the reasons for which the decision of the Court is a big step towards the acceptance of economic analyses in class action cases.
Walter Merricks appealed against the CAT’s decision on three grounds. They claimed that (i) the CAT wrongly assessed the level of evidence needed in the certification phase, that (ii) it followed a wrong approach concerning the question of distributing damages between individuals, and that (iii) the CAT did not properly assess how individual claims lead to common issues. The Court decided in the appellant’s favour on all grounds.
All points of the appeal are linked to the expected robustness of the passing on estimate. The CAT reached the conclusion earlier that they must reject giving a collective proceedings order (CPO) as it is not demonstrated that a reasonable estimation can be made for the level of passing on in the relevant retail sectors during the infringement period. The CAT identified two basic issues with the approach of the claimant. First, that the claimant did not demonstrate to a sufficient extent that the relevant data are available for an appropriate assessment. Second, that the methodology provided is not sufficient to carry out an estimation. I argue that both findings are premature and can be addressed at later stages of the trial.
The Court of Appeal stated that the economic experts of Merricks are not obliged to identify all data necessary for the passing on estimation at the certification stage, and that the experts have done enough by setting out the proposed economic approach (which the CAT accepted as a correct methodology) and pointing to the potential sources of data.
It is clear that the more diversified the number of claimants, products and markets affected by the infringement, the more data are needed for the estimation of pass-on. It requires a significant amount of time and financial resources from the claimant even to identify the exact data sources and prepare the data for analysis, let alone to come up with a detailed economic model (which is likely to be data driven anyway). Hence, the Court is right in saying that the CAT required an excessively high level of certainty in data availability and methodology at this stage (i.e. at the stage of certifying a CPO). Had the CAT’s decision been upheld, the chances of customers at downstream levels of the supply chain to open a damages trial could have been significantly diminished, and not only in this specific case but in all class action cases.
In Merricks v MasterCard there are more than 46 million claimants and, potentially, tens of thousands of goods and services that may have been affected by the anti-competitive levels of MasterCard MIFs. Even so, it is not unreasonable to expect that pass-on estimates can be carried out.
First, the Competition Act allows for aggregate damages (Merricks strongly relied on this aspect in the appeal). Moreover, the Court of Appeal has set the appropriate legal framework when they decided that the legislation does not require (i) the aggregate damage estimation to be based on individual losses and (ii) the method of distributing the aggregate award to be compensatory. Second, computational developments in the last decade in big data analytics, data storage and cloud computing make it possible to efficiently analyze multiple hundred billion of transactions. Third, it is not even needed to analyze every transaction-level purchasing data from thousands of retailers to come up with reasonable pass-on estimates. Indeed, the claimant’s economic experts suggest that aggregate data and estimates on pass-on rates in different market segments is sufficient. This latest point is explained below.
Market studies, decisions of competition authorities, evidence from other damages claims against MasterCard can all be considered in the estimation of passing on, but some market sectors may lack existing evidence or that evidence may be obsolete. Therefore, the economic experts of Walter Merricks will need to run their own passing on analyses based on disclosed MasterCard and other market data. How could they go about doing that?
It is widely accepted that the interchange fees are fully passed on to merchants by the merchant’s banks. Hence, there is no dispute that the passing on between the merchant’s bank and the merchant is 100%.
The level of pass-on between the merchant and the consumer, however, can vary on a large scale. In theory, industry-level pass-on (if the overcharge affects all retailers in the market segment) is expected to be between zero and 100% but, in certain circumstances, higher than 100% pass-on is also demonstrable (i.e. the merchant increased its prices by a higher percentage amount than the MIF). In markets with textbook demand and supply characteristics, the passing on varies between 50% and 100%. This is because, under general assumptions, perfect competition in the market implies 100% pass-on (i.e. increased costs are fully passed on to consumers) while monopolies maximize profits if they pass on 50% of the overcharge (assuming linear demand curve).
Passing on is highly dependent on the level of competition which, in turn, depends on the number of competitors and the diversification of products and services in that market. Pass on is larger in markets characterized by a higher number of competitors and less diversified products. These are market characteristics on which data are typically more available and that the expert report of Merricks should focus on.
The level of passing on is also affected by the characteristics of consumer demand in the analyzed markets. Particularly, depending on the price sensitivity of consumers, passing on can be stronger or weaker. The curvature of the demand curve at current price levels also matters. However, the MIF overcharge was always less than 1% of the price of the product or service. Therefore, it is reasonable to assume that the characteristics of consumer demand do not play a significant role (in comparison to pass-on analyses where the relative size of the overcharge is larger). Passing on the MIF is unlikely to have changed the purchasing decision of a substantial number of consumers.
Finally, decreasing or increasing marginal costs in the supply-side of the markets (for larger firms it is cheaper/more expensive to produce an extra unit of the product) also affect passing on. In markets characterized by decreasing marginal costs (which includes the vast majority of downstream markets) the level of passing on is expected to be higher. In Merricks v MasterCard the MIF overcharge could have reduced merchant sales, potentially leading to higher marginal costs, and these costs may have been passed on to consumers in the form of higher prices. Again, it can be argued that the relatively low level of the MIF overcharge did not lead to substantially lower sales, so marginal cost effects may not be considerable.
As can be understood from the above, there are a number of market characteristics that economic experts should consider to come up with a reasonable estimate of the level of pass-on, despite damages being calculated at an aggregate level. Even so, one should avoid drawing conclusions too early that pass-on rates cannot be estimated in a robust enough manner (as the CAT did). Economists can provide the tools to model markets based on the gathered industry knowledge. Applying suitable econometric models on MasterCard and other market data should be sufficient to carry out reliable pass-on estimates.
 Idem, pp 19-20.