Representative Actions in Spain. Questions for Economists.

In January 2023 the Spanish Government published the draft law on representative actions (the “Draft”) ¹, which transposes Directive 2020/1828 on representative actions for the protection of the collective interests of consumers (the “Directive”)² into the Spanish legal system.

The Directive introduces the framework for consumers to seek injunctive and redress measures against unlawful practices that may have been harmed or can harm them. The key concept is to pool together all consumers that have suffered damage from the same conduct and have the same set of circumstances into one single legal proceeding.

This Directive draws the main lines for collective actions to be implemented in Member States. Each Member State has now to transpose this Directive into its own legal system. Some European countries have already done it (e.g., Germany, Netherlands or Portugal).

The focus of the article is the first draft of this transposition into the Spanish legal system that the Government published. While there have been many articles discussing the novelties of the Draft from a legal perspective, this article aims to underline the open questions that the practical application of the Draft would entail for economists and explores potential solutions.

Current representative actions system in Spain

The current Spanish legal system on representative actions is not a systematised and structured body of regulations, but instead it consists of a few rules spread throughout civil procedural regulations.Below, we present a recap of the key aspects of the current system for representative actions in Spain, to provide a reference point to analyse the impact that the transposition of the Directive will have.

• Material scope: Only claims to defend the interests of consumers – meaning that any company (even SMEs) or individuals acting as traders, cannot file a representative action.
• Standing: Only consumer associations that meet certain requirements,³ or consumers themselves when the affected consumers are determined or can be determined, and they represent the majority of affected consumers.
• Commonality requirement: For consumers to be in the same class, they must have been affected by the same damaging conduct. There is no formal process to determine whether the consumers represented meet this requirement.
• Access to evidence: There is no specific regulation of access to the sources to evidence, unlike the disclosure procedure we see for competition claims.
• System: Some version of an opt-out system, meaning that all consumers affected by the same damaging conduct are represented by the claim are bound of the judgment. It also means they cannot bring further individual actions against the same damaging conduct in the future. However, while there is no formal procedure to opt-out, case-law considers that individuals who bring an action even after the judgment in the representative action will not be bound by it.

While representative actions in Spain have peaked up in recent years in the wake of mass-service contracts (the prime example is the litigation in the case of the preferred stock fraud in Spain), it has been very infrequent that a group of consumers affected by the same wrongdoing seek monetary compensation through a representative action. The few exceptions all relate to damages arising from contractual relations, as is the case of mass-service contracts.

In fact, even in those cases, most Courts in Spain have taken the stance that damages should be claimed individually by consumers in separate proceedings if a representative action finds an infringement against consumers. Moreover, there has been close to no representative actions seeking relief or compensation for damages outside a contractual relationship.⁴ This is due to the shortcomings of the current system, which is what the Directive aims at correcting.

Draft Implementation

The Draft improves on many of the shortcomings that the current legal system on representative actions has, namely on expanding the material scope, setting out a certification phase, or regulating a proper opt-out mechanism or the access to evidence. The key improvements are the following:

• Material scope: the extension of representative actions to any infringement of legislation referring to consumer interests, including competition law infringements, (which goes beyond the Directive regime). This means that, for the moment, the draft law appears to be applicable to any infringement affecting consumers.
• Standing: there is no change to the standing rules for representative actions.
• Commonality requirement: For a group of consumers to be eligible to be part of the same class there must be no Claimant-specific factual or legal questions to determine (i) the conduct, (ii) the total damages and (iii) there is a causal link between the two.
• Certification phase: Moreover, the Directive introduces a certification phase, where the court certifies whether certain requirements in terms of legal standing of the claimant, commonality among claimants and on the appropriateness of funding by any third-party.
• Access to evidence: Allows for Claimants to request disclosure from defendants, even at an early stage to determine the identification of the affected consumers. The regulation on disclosure is analogous to the mechanism in the Spanish system set for competition claims.
• System: Preference for the opt-out system, while the opt-in system is limited to specific situations where the Court may impose an opt-in system if it is deemed more efficient, provided that the claimed damages are at least of EUR 5,000 per claimant.
• Settlement agreements are encouraged. While they have to be approved by the Court to be binding, a redress settlement does not imply an acknowledgement of liability or guilt by the Claimant.
• Provision for enforcement: if all the beneficiaries have been identified, the defendant must pay them the corresponding damages as stipulated by the court. If they have not been identified, the defendant must deposit a lump sum in a court deposit account and the Claimant entity will be appointed as a liquidator to distribute this sum among the beneficiaries.

The legal framework set up by the Draft still has many open questions on its practical application. This section explores some of these challenges that are particularly relevant for economists.

As discussed above, the current Draft includes a new commonality requirement for consumers to be part of the same class. The Draft requires that there are no claimant-specific factual or legal questions to determine (i) the conduct, (ii) the total damages and (iii) that there is a causal link between the two.

Commonality Requirement

As discussed above, the current Draft includes a new commonality requirement for consumers to be part of the same class. The Draft requires that there are no claimant-specific factual or legal questions to determine (i) the conduct, (ii) the total damages and (iii) that there is a causal link between the two.

– Do the same legal and factual questions determine the conduct?

First, the Draft requires that there are no claimant-specific legal or factual questions needed to determine the conduct by which consumers were affected. A natural starting point in this assessment is the definition of the market(s) where consumers affected by the infringement interact. In case that all affected consumers are part of the same market, it is likely that they are affected by the conduct in the same way, i.e. no claimant-specific factual or legal question is necessary to determine the conduct.

To determine whether within the consumers of the affected product(s) or service(s) there are any individual factual or legal questions to determine the conduct, it may be useful to turn to the (economic) definition of the affected market(s) by each infringement. All consumers of the affected product or service that are in the same economic market are very likely to be affected by the same conduct, i.e., there would not be any claimant-specific factual or legal question to determine the conduct.

In the cases where there is a decision from a public authority (e.g., the decision of a competition authority), it is likely that the affected market(s) have been defined already in the decision. For the cases where there is no prior decision, it is possible to apply the classical empirical tools used in competition cases to define the affected product and geographic market(s).

For example, let’s assume that there is a cartel decision on a hypothetical product we will name widgets. Widgets can be blue or red, and the cartel decision reached the conclusion that both blue and red widgets are part of the same relevant product market and there is hence only one affected market – the market for widgets. This would suggest that the determination of the conduct does not require any claimant-specific factual question to distinguish consumers of blue and red widgets.

– Do the same legal and factual questions determine total damages?

However, within an affected market, there may be certain differences across (groups of) consumers. In those cases, the question is whether the second requirement – i.e., that no claimant-specific factual or legal questions are required to assess the total damage of the class – is considered to be met.

Relevant differences have to be considered in an economic analysis, for instance when estimating the damages endured. Given good data availability, econometric methods are in general well equipped to control for heterogeneity among class members, even in a single analysis. The question, nonetheless, is whether controlling for heterogeneity – even in the same analysis – means that different factual questions are considered for total damage.

Following the example above, let’s assume that the widget cartelists agreed on a different price overcharge for blue and red widgets, due to e.g., differences in the input price of blue and red paint. Those differences can be controlled for in one analysis using shared evidence.

The question is however whether controlling for the widget colour and the different input costs is considered to be a claimant-specific factual question, in which case consumers of blue and red widgets should be in different classes.

Some differences are however of such an extent that separate economic analyses are in order. In such a case, (even) an economist may be tempted to conclude that the class needs to be divided in subclasses.

– Is there a non-claimant specific causal link?

The concept of causality carries universal import within economic assessments, irrespective of whether they pertain to collective actions. The crux here lies in establishing a clear and direct linkage between the alleged infringement and the subsequent damages incurred. The analytical methodology shoulders the responsibility of substantiating this fundamental causal connection, whereas a theory of harm is needed to establish it. A theory of harm needs to present a mechanism or way in which the claimant has (likely) been damaged. A theory of harm also informs the economic analysis.

It is highly unlikely (if not impossible) that claimant-specific questions need to be considered when assessing and establishing a theory of harm when no claimant-specific questions are needed to assess the conduct and the total damages. If, however, claimant specific facts are needed to establish a sound theory of harm, then they need to be taken into account in the economic analysis.

– Commonality vs. procedural efficiency

Overall, assessing whether or not differences in sub-classes satisfy the commonality test remains a legal question and it is not (yet) clear where courts will draw the line. Factors such as the number of potential sub-classes, the nature of the differences between them, and the required information to control for such differences may be relevant in answering that question.

On this, we also briefly consider the trade-off between the commonality of claims requirement and the procedural efficiency of court proceedings. The complexity of the methodology and the required data collection effort used for the damages estimation tends to increase with the granularity and quantity of individual factors that need to be taken into account. If sub-classes cannot be formed – or not be formed to the desired level of granularity – due to their complexity or lack of commonality, it may be necessary to bring them as separate collective proceedings for the same conduct. These separate class actions would have a higher degree of commonality, while being associated with lower damages per proceeding and a likely higher overall burden on the courts and parties involved.⁵

It therefore remains to be seen how points of commonality of claims and procedural efficiency are weighted against each other. This will likely depend on the nature of the alleged infringement, the market concerned, and the legal interpretation of the collective action regime.

It will be interesting to see how Spanish courts interpret the commonality requirements. From the point of view of a competition economist it will especially be interesting to see how cartel cases and, within cartel cases, umbrella effects⁶ will be treated. Will consumers who were not clients of the members of a cartel, but that may have suffered a damage due to the umbrella effect, be part of the same class as clients of cartelists?

Implications of an Opt-Out System

As discussed above, the Draft currently points to an opt-out system as the general case. The difference between an opt-out and an opt-in system is that, in the former, all consumers that could be part of the representative action (i.e., that meet the commonality requirement described above) are bound by the Judgment unless they explicitly state the opposite. Whereas, in an opt-in system, only consumers who explicitly state their willingness to be part of the claim will be bound by the outcome. There are two implications for an opt-out system, which we discuss in what follows.

First, establishing class-wide harm could be challenging from a methodological point of view in the event that the characteristics of a class are not well defined. If the class is not well defined, certain characteristics of a group of customers that fall within the class may (wrongly) not be accounted for, meaning that the methodology would be flawed. If the characteristics are well defined, then the methodology developed should be able to accommodate any customers that fall within the class – and in the case that it cannot, the class may need to be split up in subclasses.

Second, it implies that the Claimant may not rely on Claimant-specific data, putting them in a disadvantage (in terms of access to evidence and information) at least at the initial stages of a claim – and particularly at the stage of initial assessment. As the Claimant does not have access to private class-wide or market-wide data, such data usually can only be obtained through disclosure requests. If disclosure is rarely granted, this leads to an over-reliance by the Claimant on public data sources. While this may not be an issue for the estimation of total damages, the award of individual damages will become less precise (relying on average damages) and therefore overcompensate some and undercompensate other consumers in the class.

The crux will lie in striking the balance of over- and under-compensation. This underscores the importance of disclosure mechanisms working efficiently. The Draft refers to the disclosure mechanism available in competition law infringement procedures, by which Defendants may be required to provide all relevant and proportional evidence requested. While this mechanism aims at levelling the field between Claimants and Defendant in terms of access to evidence, it remains to be seen to which extent it does so.

The second implication from an opt-out system relates to the difficulty of showing commonality of the class. Mirroring the same argument exposed above, if disclosure is rarely granted also prior to the certification phase, showing commonality of the class may be a challenge, as it relies on factual evidence. However, in opt-out claims there is no claimant-specific evidence and public sources are not likely to be best placed for that purpose.