In this podcast series, Deminor Litigation Funding interviews global professionals to share their perspectives on different aspects of dispute resolution and litigation funding.
Podcast Preface
In this interview, we speak with Thierry Wetzel (TW), Senior Director at Alvarez & Marsal, about calculating damages in antitrust litigation. In conversation with Namita Moorjani (NM), Thierry shares his extensive experience in competition economics, discussing key methodologies for assessing overcharge and pass-on in cartel and abuse of dominance cases.
Podcast Transcript
NM: Good morning. I'm excited to be joined today by Thierry Wetzel, who's a Senior Director at Alvarez & Marsal, a global consultancy. Thierry focuses on damages calculations for antitrust and commercial litigation cases. Thank you for being here, Thierry. Can you briefly talk to us about your role and how you got into damages calculations in the first place?
TW: Hi, Namita. Thanks for having me. As you said, I'm an economist and a Senior Director at A&M. I specialise in economics, the application of economics to provide evidence in competition investigations, litigation, arbitration, specifically collective actions in that space. I also provide advice and economic support on transactions and their regulatory scrutiny.
I studied economics a while ago, and I never really left the field. I specialised further and further. And this is how I got into the role of economic consulting. Learning that there is this world out there where it's a bit of in between academia and consulting, which is quite a nice space to be for an economist. It's a bit more dynamic, but we still get the joy of applying economics to the real world. And of course, we get data as opposed to academia. We actually get the information. So that's how I got into it, it's quite a straight career path from university into this industry. And then I've never left it.
And one of the things we really like to do here is we apply, or I apply economics to the commercial reality of cases and the applicable law, rather than following textbooks in isolation. The importance of which was also highlighted in the recent judgment that I think last Friday [27 June 2025] in the trial 1 umbrella proceeding, which showed again the importance for economists to actually understand the legal setting, the prior records when providing expert services and not simply sit in an ivory tower and apply textbook models.
In terms of case experience, I worked on various cases, for example, the trucks litigation on the defense side or in the lithium battery cartel on the claimant side. We are also the economic experts on the certified commercial interchange fee collective actions in the UK. And I advised on various merger proceedings in the UK, in Switzerland, European Commission, et cetera.
NM: That's incredibly interesting. You've clearly had a lot of experience in a variety of types of cases. Given your extensive experience with antitrust cases, could you maybe walk us through some of the key methodologies you typically apply in those situations?
TW: Sure. To start, whilst this is a legal concept, the counterfactual is why economist services are necessary. We need to not simply know the actual, but define and tell the story of the counterfactual world, the world that should exist absent the infringement in question. That central theme sets the tone for the methods we use and deploy.
Damages are quite a broad concept. They consist of overcharge as well as mitigation. If we take it very broadly, mitigation includes, for example, pass on, but also countervailing benefits. There are also other components to a full damages assessment, such as volume effects. It is also about applying interest to it. It's not just about the overcharge and pass on.
I think for the purposes here, I will focus on overcharge and pass on. These are the two you will come across in almost any case you touch. And with respect to interest, there's not too much to talk about. Statutory interests are awarded automatically. That usually is the way clients end up going. Of course, they could argue for example, that a cost of capital approach is more appropriate and as such that a different interest would have applied, but that then involves, it's not an automatic award. So, then you're back in providing expert evidence, collecting material and so on. But that's just to set the scene because damages is quite wide.
Now, when it comes to overcharge methods, there are, broadly speaking, three categories of methods. You can look at comparator-based approaches. This is most common in cartel cases. It involves a comparison of prices over time: before, during, and after the infringement or across markets. Or the comparison of an affected market versus an unaffected market, or of course, a combination of both. Now, within comparator-based approaches, of course, there are different methods in the detail, but broadly speaking, that's one way to do it.
We then also have the financial / cost-based approaches. These are approaches that examine, for example, profit margins or cost structures. They include assessments of return on capital, return on sales, weighted cost of capital. And then they would compare that to some benchmark, which needs to be defined. And then you can arrive at sort of an overcharge figure. These approaches can also include valuations or event studies, for example.
And then we have the third category, which is very rarely used in practice. Those are more of the structural theoretical approaches. Those involve simulating or basing it on a theoretical model of how markets should work. And we then take it from there. We can simulate it. We can populate it with data. We can run regressions. But these approaches tend to be very assumption-driven. There tends to be, because of that, a lot of uncertainty. And if one were to populate these theoretical models to come up with an actual estimation, they also become very data-intensive without resolving the issues, because it's built around assumptions and so on. So, these models we rarely see. But while they exist, and in academia they're used from time to time, in practice, they're not very well used. It's almost as if we can consider them the third best, if you have absolutely no other way of going about it.
In cartel cases, we see comparator-based approaches predominantly. This is kind of where the market converged. And for abuse of dominance cases, we often see financial cost-based methods. By now, in practice, most methods will involve some combination of the above. One might look at margins as a sense check to a regression, et cetera. It's not in isolation, but those are sort of, broadly speaking, the main categories.
And we didn't really, or I didn't really, touch on the precise nature of how we build these models. I think that's a bit too much for here. But all approaches by now can involve some econometric or statistical modelling. It is increasingly more so in comparator-based approaches. That's really at the heart of it. And with financial-based approaches, it's a bit less so. But you can still deploy regressions. That's for the overcharge.
NM: How would you go about determining whether you need to apply a pass-on discount? And if one is needed, how would you go about working on the relevant percentage?
TW: Re pass-on, I'm going to take a bit of a different approach. I think it's more helpful to understand where the methods are coming from rather than discussing detailed specifications. Pass-on is part of the law. One particular issue is the proximity between the act and the pass-on to the ultimate consumer, and the evidence an economist can provide to answer this legal question.
Now, pass-on can be used as a shield, which is raised by defendants to reduce liability. Or it can be used as a sword raised by indirect purchasers seeking damages. The EU damages directive, I think from 2017, introduced an asymmetric presumption of pass-on. Direct purchasers are presumed not to have passed on the overcharge, if any, thus suffering all damages. While indirect purchasers are presumed to have had pass-on from higher up in the supply chain.
In the UK, there is a useful case. The UK Supreme Court in Sainsbury's has looked at this in great detail. The economic expert and team for Sainsbury's are now at A&M with me. We've moved together. We all worked on these cases. Now, in this case [Sainsbury’s], the UK Supreme Court has given very clear guidance on pass-on. And outlined the four reactions a firm can take in response to a cost change. These are: they can adjust their prices; they can negotiate with suppliers; they can do nothing and simply absorb it; or they can change discretionary spend (and discretionary spend is defined as, for example, marketing or other types of overhead costs).
The Supreme Court also made very clear that only one and two, so adjusting prices and negotiating with suppliers, are considered as pass-on. The other two reactions a firm can have would result in no pass-on. If you change your discretionary spend, for example, you face an increased cost on one side of the business. You may do marketing savings on another side to cancel out the increase. That would not be considered pass-on by the UK Supreme Court as defined in Sainsbury's.
In order to show pass-on, if we follow this framework, it will be necessary or it is necessary to show that a firm did any or some of these mechanisms. And then understand how pass-on could have happened through this. This implies that many standard textbook models do not work. This is because they only allow firms to change prices in light of a cost change. It's almost by assumption. And therefore, they rule out other avenues a firm can take in reality. Standard textbook models wouldn't allow a firm to reduce discretionary spend, but they would allow a firm to change its price. In a sense, these models rule out or they're somewhat inconsistent with the routes identified by the Supreme Court.
What this then in practice means is that documentary evidence, such as business accounts, meeting minutes, budgetary processes, as well as witnesses of fact, have really become indispensable to show pass-on. If one wants to stay consistent with the UK Supreme Court, that is. And when we take it there, the first step to assess pass-on is to establish a causal mechanism. So, was there even a mechanism for pass-on? And this mechanism is one of the four reactions a firm can take. These mechanisms can usually be found in budgetary processes or textbooks. How does a firm price? How does a firm make decisions? How does a firm work? These are historical records in most cases. You can go through them. You can understand how a firm operates, and then you can identify the mechanisms through which it would have, in principle, been possible to pass on a cost.
Once this causal mechanism has been established, which is, in my view, really more in the evidence rather than the theory, then we can, in a second step, go about estimating the level of pass on through that specific route or combinations of routes. And this then has very strong implications for the method used, or should at least guide and inform the method used. That's on a high level. I think a bit later we may get into some challenges this brings about. But roughly, I would say on a high level, these are sort of the methods or the way we go about thinking about overcharges and pass-on.
NM: Have you seen any key trends or an influx of specific types of cases? If yes, how has this changed over your career?
TW: While I would say courts have increasingly eased up to the idea of more technical econometric techniques in the past decade or so, they prefer simplicity and robustness over unnecessary complexity. And perhaps unsurprisingly, judges appear to favour analysis they can actually understand and explain over what they used to refer to as black boxes of the economists.
And this is also highlighted in the UK, you have four cartel cases that went all the way. Of these four cases, only in one did the court accept the econometrics put forward. And the court found issues with the econometrics in the other three cases. Now, this is not to say that they dislike econometrics in any way, they dislike the way it is used. One example is in Stellantis, the CAT critiqued the econometrics of being data mining. So sort of reverse engineering what you're looking for and then finding the data to support it. This is really more about, we can call it bad practice, rather than courts not liking the idea of it.
Now, because of that, I would say documentary evidence and factual witness testimony remain very crucial components, even if economist experts are deployed to answer these questions, because they need to know the facts and the reality of things. Economic findings need to be grounded in the reality of these cases and they need to be cross-validated or informed by documentary evidence or factual witness testimony. Different judges will seek consistency between economic approaches and the commercial reality as it is shown in the factual documents of the case.
And by now in practice, courts tend to rely on more than one method and they triangulate findings. For example, they may consider regression analysis as well as comparative studies and factual evidence all side by side to arrive at the conclusion and understand the consistency between the different approaches. It's unlikely a judge would follow a simple statistic method without explanation, just a regression. That usually is not the way to do it. It needs to be sort of the proper scientific approach. You do the storytelling, you understand why it is what you're attempting to measure and how you're going to measure it. And you do the process in that order and you don't start at where you would like to end up. And then that can give a robust story that is coherent and makes sense and can be followed in logics, not just in statistical terms.
And we have also seen with the introduction of the collective actions, we have seen an introduction of a new set of issues. When it comes to overcharges, it is now about class wide harm, not individual harm. How do you quantify damages for an entire class of consumers, for example, or a class of firms where you may not have access to as much information as you would have in a more traditional individual litigation? And here, this has, of course, implications for the method and is heavily driven by the information that is available.
What we have also seen in this space is an increasing number of abuse of dominance cases. And these have given rise to what I would say, at least in practice, somewhat novel ideas. In academia, they've been used for a while, but I think Meta is a good example where it appears a collective action has been certified on a theoretical bargaining model. This is not something we would have, or I think we would have seen prior to collective actions. This is another trend that I think we're seeing. It makes it a good vehicle to bring a new set of cases. And we have seen that in the type of cases that have come to courts in the past one, two, three years.
And specifically on pass on, collective actions have also given rise to a new issue. And that is consumer collective action may in fact compete, compete may be the wrong word, but they may compete with firm collective action on the same topic. It's now becoming a zero-sum game because two entities are claiming for the same damages pool, and damages in Europe are compensatory. It's now about assessing where the damage has actually taken place and how it has been passed on. And in this case, the lines between who bears the burden of proof have been blurred. It also renders the asymmetric presumption in the guidelines as rather unhelpful and useless. What is clear is that pass-on is becoming a central topic in many of these collective cases.
NM: Thank you for taking the time to go through all of that, and especially the part of courts actually being able to want to understand the logic behind all of the calculations, also just for the consumers as well. I was wondering, do you have a preference on being the economist on the defendant side versus the claimant side? Because to some extent, do you have more data on the defendant side?
TW: It's a good question. In the UK specifically, we are independent experts and our status implies that the assessments we do are objectively done. In that sense, it shouldn't matter that much. We need to assess the question objectively. And whichever side hires us, we need to come up with a credible, robust story rather than aiming at helping a specific side.
But that aside, I think they both have their perks. And for an economist, I would say it's very helpful to have experience on both sides. On the defense side, it again depends on, it really depends on which question, right? If it's an individual case, if you're assessing pass on, the defendants have to prove pass on, but the evidence rests with the claimant. And when it comes to the overcharge, it's quite likely that the defendants have much better price data because they're selling the product. Even in a collective action, the defendants may have the data for the entire class.
Whereas on the claimant side, we have to make do with what we have. There, I would say it's almost a bit balanced because it depends on which part you are in the case. It's somewhat nice to be on one of the two sides because you have access to the evidence. It's easier access to the documents and so on. I wouldn't say I have a clear preference for working on either side. I think they both have their challenges.
Claimant work usually involves more thinking, at least from the outset of a case. In that sense, you can call it a bit more intellectually pleasing because you actually have to come up with a positive case on how to do something rather than just shooting it down all the time. But I think it's very similar. And I'm quite happy that I have case experience on both sides because it does help with the work we do, it does help increase the quality of the models. It helps with anticipating likely critiques from either side.
NM: That makes complete sense. I'm sure you've worked with funders for certain cases as well. Have you seen a difference when a funder is involved in a case versus when they're not involved in the case?
TW: For, based on my experience, it depends a bit on the work, of course, we do for the funder. If we do a risk assessment, we would be very closely working with the funder because it's not necessarily about the case itself. It's about assisting you guys in making a decision about the commercial feasibility of a certain litigation.
From my experience, once cases actually run, all the funders I have worked with, they stayed largely out of it. They didn't, they took it very serious to stay independent, not get involved in the actual case. Sure, you may have a budget discussion here and there, but you're not usually discussing substance with funders once the case is up and running. At least that was my experience and it probably is the right way to go about it.
I think for you guys, it's this, the pre-stage is really important, understanding the case, understanding the facts, understanding the likely quantum, understanding the legal risks, so you can make an informed decision. And then it is really for the legal team to take it forward, for the legal team to collaborate with us, to instruct us in the right and helpful way so that we can assist the CAT or the court in arriving at a conclusion.
NM: Definitely agreed on the funders staying out of it once the case has actually started. You've made calculating damages sound, you know, not too complicated, but I'm sure that there are a few challenges. Could you maybe just talk about them a little?
TW: Yes. I'm trying to stay away from boring you with statistical challenges of very specific technical problems, and there are many in each method, in each case we do, we face a multitude of issues. But in general, it is, of course, very case and jurisdiction specific. In the UK, we have very, or relatively, a very evolved disclosure system compared to the continent. While disclosure is a possibility and courts can award it, it is not something they do as likely as in the UK, so it is much harder. These jurisdictional differences, they naturally set the scene of likely issues we will face. On the continent, I am more likely to struggle with data availability than I am in a UK court.
But now I would say if we stick again to sort of the overcharge pass-on split, the main, what I call non-technical challenges in a sense is definitely the counterfactual. That's the first main challenge we face. Assessing the potential overcharge requires a credible benchmark. And this is the formulation of what would have happened in the market absent the infringement, sometimes called the but-for world. And if econometrics is unable to establish such a credible counterfactual, sometimes it can, we need to revert to other forms of assessing it. This was, for example, with interchange cases, when it came to the assessment of counterfactuals, it's more thinking through using the facts available, how a firm would have likely behaved rather than just putting it into a statistical model.
And then the second big challenge we always face is the data availability, data availability, or information availability is often, at least partially an issue. It's not always an issue that breaks the case, but it is an issue or a challenge we need to address and we need to deal with. If a certain method would be preferred, but the data to deploy it is simply not available, then the method cannot be deployed. You don't have the necessary information to make it work.
This is particularly an issue where infringements are long in the past. A cartel case that concerns 20 years ago, records may not be there anymore. Or in collective actions where much of the evidence rests with a dispersed class of firms or consumers, particularly in opt-out actions, UK courts are very reluctant to ask information of the opt-out class, whether these are firms or consumers, because they didn't actively want this litigation. So courts consider it unfair to put a burden on them, which is also sensible.
In these cases, data availability, at least from the claimant side at early stages, tends to be something that we need to address. And once disclosure is available and is ordered, a lot of these issues may be mitigated. We may suddenly get the beautiful information we have wished to have for two years, or we may not. But in either way, in such cases, economists need to do the best they can, given the information that is available. For example, from public sources. That's the two main sort of overarching challenges that I think any overcharge estimation will face.
At some point, and when we move on to the quantification of pass-on, I again take a slightly different approach than to overcharge. So maybe to take a step back, pass-on, and it is often cited in cases, you may have come across it, economists call it a very well-studied phenomenon. They say there are hundreds of empirical papers. There is plenty of theoretical textbooks that predict the level of pass-on. There is empirical evidence supporting this, but when we look at that pool that is out there, we can draw two distinctions.
I would say that the first one is, let's assume a firm has a policy that they price based on a cost plus method. They have almost an automatic mechanism in place. It takes the cost, the input costs. It applies a 15% margin and it sets the price. If now this input cost has gone up, this policy already tells you that 100% of the cost was passed on because it automatically results in a change in price. But what if it's not cost plus? What if it's beyond cost plus? And I think that's where the interesting bits come in.
Once it is not cost plus, it again becomes much more challenging to understand how a firm operates. Firms are complex entities and often firms that are part of litigation are also bigger. It's not in a class action. Sure, it's different, but in traditional cases, it's unlikely that you would have a small corner shop in a litigation. These are businesses with complex policies, different business units. One unit may be concerned with the gross margin. Another unit may be concerned with the operating cost. And then there are different KPIs for all of these individuals within an organisation. And also economic theory tends to argue that firms in the long run become these perfectly efficient vehicles. But when we look at the reality of big organisations, there is quite a lot of slack here and there.
To really find out how a firm sets prices, if it's not a simple cost plus approach, can be quite challenging. And to do that properly, you need, again, the documentary evidence. You need to understand the business. You need to understand how this business operated in the period you are interested in. And since this is often historical, that information exists, you can find it. It's not that it's not there. You're just not looking for data points. You're looking for policy documents, for meeting minutes. I mean, all of these things tend to exist in larger organisations. And that can really give you insights into understanding the process. Rather than trying to force a textbook onto reality, you try to understand the reality. Or that should be the right approach. And that's one complication. I call it the beyond cost plus.
But then we also have the beyond direct costs. If you would look at them [empirical papers on pass-on], and we have done that at some point. We did actually look at these papers. Out of the few hundred econometric papers on pass-on, it is really the vast, vast majority that concerns large direct input costs. These are costs directly needed for the production of goods and services. If you produce plastic bottles, you need plastic. Plastic is going to make up a vast majority of the costs of the product. And if there is a change or a large change in the cost for the input material, it is fairly likely that this is going to force some change on the price. And this is what the economic theory does. This is what the empirics does. No one looks at the small indirect costs. So there I would call a complication the beyond input cost.
In these cases, again, we need to really understand how a firm operates. Interchange is again a good example. It concerns a small overhead cost that is not needed for the production of goods and services. And within the accounts of a much bigger organisation, it is going to be a fairly small cost, if you compare it to electricity costs or other types of costs that are out there. And a firm simply has a myriad of ways to deal with this. They have a lot of incentives at play, and they have many different ways to achieve a margin. They have different ways to achieve cost savings. They have different pressures that are around.
It's not as simple anymore as saying, yes, economists have studied this for hundreds of years, and we actually don't need to assess it. We can just take a number from a textbook, and we quickly cite 50 empirical studies. But once you look into them, half of them are about oil prices and petrol. And it's probably unsurprising that if the oil price drastically increases, sooner or later the petrol price needs to adjust. Otherwise, the firm will go out of business. I would say the question is more, how much does it adjust? Is it truly 100 percent? Would it be 80? Would it be 70? But on an indirect cost, you may never have the mechanism. It may never be such a clear route to prices. And a simple approach of applying textbooks would simply ignore that.
Here, a big challenge in pass-on is when we're in these, I wouldn't call them niche cases, because that's the wrong word, but sort of not the traditional empirical evidence on large direct input costs. And that's where it becomes much more challenging to understand how a firm behaved and is the reason why we're doing what we do and why you need an economist. I mean, if it was as simple as just applying a textbook, there would be no need for an economic expert in these cases, right?
NM: That makes complete sense. AI has been the topic of discussion the past couple of years. Do you think that that can be used to address some of those challenges?
TW: It's an interesting question, and I think it is quite early stage. My opinion is rather speculative, I would say. When it comes to the actual modelling, and again, it depends how we define AI. Part of AI is standard statistics. It just gets a different name. It's still based on regressions. It's still based on the same principles. Then, of course, AI has many different applications. You can deploy neural network, et cetera. But now we're getting very quickly into the black box territory.
How did your AI arrive at the number? Can it explain it? Can it go line by line through a cross-examination and answer why it made certain decisions and so on? It probably cannot. I think on that spectrum, I don't see it being deployed widely.
Where I do see advantages in sort of technological advancement more generally, is the data collection, the data mining, all the advanced tools we now have available to collect information from a large variety of sources. You could now, with an API, access millions of data rows in a second, which would have taken us years to collect without it. I also understand it is increasingly being used in document reviews or in large-scale document reviews. It's a big hot topic on the lawyer side of things.
Here, I think it strongly depends on the case. If you are looking for at least the way the AI works or my experience with AIs, they're not yet on the level to get all the nitty-gritty details out of these documents. They might find the obvious ones. If you're looking for cost plus in 2 million documents and you're looking for the term cost plus, then arguably you don't need an AI. You just do the search function.
But what if it has to start reading between the lines? What if it needs to start making inferences? Oh, I might not find the answer written down, but I might find the answer in between two documents once they harmonise with one another. There, I think there are still quite some limitations, but it will definitely assist us in the work we do. It can make it a bit easier, but I see very strong limitations, at least based on the current AIs I've had experience in.
Who knows how different things are going to perform in five or 10 years, but at least from the experience I have now, because I use it a lot for smaller things. I ask it to verify things. I ask it to do some research for me, but based on experience, it just tends to be like 70% right? It is also somewhat of a dangerous tool to just deploy blindly.
NM: Definitely agreed on that. Even with my experience, you still have to know what the topic area is a fair amount before you are actually ready to use it for those purposes. But thank you for joining us, Thierry. It's been great talking with you, learning from your experiences and hearing everything that you've been working on.
TW: Thank you. Great to be here.
Podcast Speakers and Further Information:
Thanks for joining Deminor's Litigation Funding Podcast Series as we dive deep into core topics in funding litigation.
Keep a lookout for our upcoming conversations as the Deminor team speaks with several more experts to get their insights into different aspects of litigation funding.
If you would like to connect with either Thierry or Namita on LinkedIn, please click on the links below:
Thierry Wetzel, Senior Director at Alvarez & Marsal – https://www.linkedin.com/in/thierry-wetzel-094853148/
Namita Moorjani, Financial Planning and Investment Manager at Deminor – https://www.linkedin.com/in/namita-moorjani/
***
Click below to view more interviews from Deminor's Litigation Funding Podcast Series:
