A growing number of critics argue that these basic assumptions are both wrong and outdated, as evidenced by the fact that in many industries, particularly technology, companies have been growing to gargantuan proportions and, as anybody who owns a smartphone is painfully aware, they seem free to gobble competitors, hinder innovation, and serve up crappy, overpriced products.
These conglomerations of money and power not only end up widening the inequality gap, but they also threaten democracy itself, as University of Utah antitrust expert Mark Glick and other experts have attested. That’s why tech-focused antitrust voices are sounding the alarm as companies like Google, Amazon, Apple, and Meta expand at a breakneck pace, encroaching into every possible area of our lives, from our cars to our refrigerators to our dreams.
European antitrust and competition expert Cristina Caffarra has been a top advisor and expert before the European Commission and in courts and agencies across Europe, as well as a guide in antitrust efforts in the United States. Her experience includes landmark cases on the economics of platforms and the digital economy for and against Microsoft, Amazon, Apple, Google, Meta, and more. She spoke to the Institute for New Economic Thinking about what she sees as the most pressing areas of antitrust, why regulation and legal actions have largely failed so far, and why, from her perspective, too many economists have been part of the problem rather than the solution.
Lynn Parramore: Let’s start with the overall antitrust landscape in Europe in comparison with what’s happening in the U.S. It has long been said that Americans have lagged behind in taking on Big Tech in competition matters. What’s your view?
Cristina Caffarra: For a good decade, Europe felt itself to be a pioneer in enforcement, particularly against tech. The case against Google Shopping [Google’s shopping comparison service] started in 2010 and the case against Android started in 2015. There was a time when I was visiting the U.S. that we looked at U.S. colleagues and practitioners with some element of smugness, saying, look, we’re enforcing in Europe but you haven’t done anything since Microsoft 2001.
What has happened since is an increasing realization in Europe that the antitrust cases we initiated and pursued, however well-intentioned, have not delivered yet and are not going to.
Take the Google cases. We have led them to a conclusion and there has been a finding of infringement against Google, but it has taken far too long and the remedies have been ineffectual. This is inherent in antitrust assessment. It’s the nature of the beast that you often intervene as a result of complaints. The reason the European Commission pursued the Google Shopping case is that they were inundated with complaints — the case selection in prioritization often follows the complaints you have. What happens is that enforcement bandwidths get tied up. While the Commission spent five years focused on Shopping, Google was moving the monopoly from the desktop to mobile devices. The search monopoly was effectively translated to mobile, and by the time the Commission opened the investigation in 2015, it was far too late.
The same is true in other cases. The Commission is pursuing cases against Amazon, against Apple — all framed in ways that are to my mind not going to deliver anything.
So Europe, having started off as a pioneer, has ultimately not been able to show that you could deliver. To us Europeans, the U.S. was really frozen under the permafrost since Microsoft. But things began to change when you started to see the progressive New Brandeisian discourse [an antitrust movement focused on competition] getting a bit more into the mainstream.
I vividly recall a meeting that I attended back in late 2015 at the Metropolitan Club in Washington, put on by the Jevons Institute, a small group of people who talk about antitrust. Jason Furman and Peter Orszag (at Citibank at the time) came to present what they considered to be an interesting result that was beginning to seep into the conversation. They had been in President Obama’s Council of Economic Advisers, and they had uncovered that over time, concentration was increasing and margins were increasing. This was the start of an antitrust discussion at the time. The gathered antitrust community was totally in disbelief that this was telling us anything important or interesting. Why would concentration be going up? People argued, wait, you’re not measuring market concentration at the right level. You’re not doing market definition properly. Margins are not particularly going up. Why would they be going up?
So the antitrust establishment, such as it was then, was skeptical that there was any degree of under-enforcement or that they were watching anything interesting. At the same time, we started to hear rumblings from other voices like Barry Lynn [journalist and director of the Open Markets Institute] and others who were noticing the phenomena, but it hadn’t percolated through the antitrust establishment. To me, that meeting was the very first time that the antitrust establishment was confronted with the potential for a claim that we have underenforced widely in this country, and you can tell because concentration is on the up, margins are on the up, and the share of GDP that goes to labor is down. We have less creation of companies and less vitality. But most importantly, workers are disadvantaged in this new order.
That was in 2015. Then you had the development of these Brandeisian voices and so on. In the spring of 2019, I went to the Zingales conference in Chicago, as the only European, I think. I was approached by Doug Peterson, who was and still is the attorney general of Nebraska. He heard I had experience working against Google in Europe and asked if I would be interested in assisting a coalition of bipartisan attorneys general bringing cases against Google. I say, whoa, yes! Where do I sign? So I became an advisor to this coalition, initially a single coalition, that later split (I advised both for a period).
I was on the steps of the Supreme Court on the 9th of September 2019 when [Texas Attorney General] Ken Paxton launched, with all the other AGs, that antitrust investigation of Google. Then the initiative split, and it was Texas, on the one hand, that is pursuing Google adtech [accusing Google of monopolizing technology underlying online advertising] and then you had Nebraska and Colorado leading the way in pursuing [Google’s] search [accusing Google of monopolizing the online search market].
Simultaneously, you started to see cases brought by federal agencies. The Federal Trade Commission (FTC) began to look into Facebook and Amazon, and suddenly the Department of Justice (DOJ) began looking into Google and Apple. In December 2019, I had an event in Brussels with a thousand people coming from all over. I had Doug Peterson speaking at that event together with Max Miller [assistant attorney general] from Iowa. The title of the panel was, “Drums Beating From the Hill. Are We Beginning to See Enforcement Revived in the United States?” At that time, the U.S. was way behind, but that impetus made a huge difference and it was at the level of the state AGs. It was the states and it was a bipartisan initiative to pursue Google in particular. Of course, when you are the United States and you start to veer, it carries a lot of force.
The Europeans started thinking, oh, look at them, they woke up and they’re doing stuff! Then you had the election of Biden, and you got Jon Kanter [assistant attorney general for the DOJ], Lina Khan [FTC Chair], and Tim Wu [White House antitrust adviser] and suddenly there was government policy reflecting their views. The current perception in Europe is that in terms of posture, the regulators in the U.S. have overtaken us because Kanter and Khan are very much progressives. They are certainly very much ahead of European regulators. The only exception was the U.K., though now, with the current government and a new CEO to be nominated, it’s not clear what’s happening and they may take us back to where we were.
So the U.S. posture is something we watch with great interest. But will it deliver? In Europe, we are in a world in which antitrust has failed. But we woke up in 2019, too, and said, oh, we’ve got a lot to do, but unlike the U.S., we are a regulatory power — we do regulate a lot. In 2019 there were these political designs to essentially pair up antitrust and regulation, and so we are now in a world where we have digital regulation, the DMA [the EU Digital Markets Act, which addresses perceived unfair business practices by large online platforms designated as important gatekeepers between European businesses and consumers]. This is now law. Of course, what we’ll actually do is a very big question.
LP: You’ve noted a need to scrutinize tech business models and to fundamentally change them. Why are the business models themselves important to address?
CC: This is very much my baby. I came up with this discussion of business models in a 2019 article. It was the first time discussions of business models and their antitrust implications appeared in the antitrust community. The point I made is that when you think about categories of concern in antitrust, which are premised in all cases, like, for example, self-preference, you see that yes, there is a thing called self-preferencing. This is when firms favor their own business. I’m an economist, though, and I also know that not every form of self-preferencing is anti-competitive.
LP: What would be a case in which self-preferencing is benign?
CC: Suppose Amazon is throwing up a particular recommendation in the buy box [the white box beside the product detail page used for customers to purchase items in their cart]. Now (and for clarity, I have advised Amazon), Amazon will say that the algorithm that is selecting that particular product is designed to provide the product that is the lowest price, highest quality, and matches the requirement of the consumer the best. Why? Because, they will say, we care about the consumer coming back to the platform over and over again. They will do so only if they get the sense that they were given a fair recommendation and value. If they get a biased recommendation, the consumer will suss it out and eventually be unhappy. That’s the story.
But is it a form of self-preferencing every time Amazon recommends its own product? It’s not clear, because Amazon makes money either way. If I’m buying a battery, Amazon makes money whether I buy an Amazon battery or a Duracell because they get a commission. Financially – and here’s the monetization point – it’s not clear that Amazon makes more selling Amazon batteries v. Duracell batteries. It depends on the margin. The incentive is not necessarily to sell an overpriced Duracell battery or to sell one of their own if it is crap, because the consumer will say, well, it’s a bit cheaper but it’s crap. I’m not going to go back and buy Amazon again.
On the other hand, I’m saying that there are some business models in which self-preferencing is inherent and most likely bad, and that is typically ad-funded businesses. Why? Because ad-funded businesses monetize in no other way than through directing traffic to themselves, to their own sites. They want to work as the turnstile that directs traffic to themselves. That is inherent when there is no other form of monetization.
If you want to monetize a commission for being the platform – think of Uber or Airbnb, or think of Apple – you monetize mostly on the device. In these cases, the unit economics of their business is fundamentally different from the Amazon example I mentioned, so if you have a general rule that says “thou shall not self-preference,” you get these companies saying, well, “I don’t! My algorithm is completely fair.” It just is designed to optimize what a consumer wants. But then the regulator says: how do I know that what the company says is true? Do I believe you just because you’re Amazon and you tell me you have the right incentives? Maybe. But that is why I would favor some form of algorithmic transparency: a regulator can say, Amazon, come and show me your algorithm and tell me how you constructed it. Obviously, no one really wants to do it, but that would be the way to check.
The point is that to presume that every time you are recommending your own product is a form of self-preferencing and illegal is wrong. We need to think about how companies monetize because companies are driven by incentives. In economics, we talk about incentives, and they are, in turn, designed as a result of the monetization structure. Can I make more money this way or that way? That’s why I was very much a fan of the approach that the U.K. did, which was not a generic “thou shalt not self-preference” kind of thing.
The U.K. approach was intended to be bespoke– that is, tailored to an individual company. The U.K. invented this approach: Amazon, Apple, you will have to do this, and Google, Facebook, you have to do this. Each and every one of them had a different set of problems and a particular bespoke regulation tailored to their business model.
That is the only way we can hope to achieve anything with regulation, because the other model, which is generic and simply says something like “don’t self-preference” is going to be met with such endless resistance that it won’t lead anywhere. I think what we’re going to see in Brussels and with the DMA is absolute and total resistance. Companies are resisting not just the designation but the obligation that is placed on them because they’re going to say this is not relevant to me.
LP: Many people are aware of the problem of “killer acquisitions,” a situation in which a big firm buys a startup in order to neutralize the competitive threat, like Facebook buying Instagram. But you’ve also been critical of the more common “reverse” killer acquisitions – scenarios in which an established company buys features or startups so they don’t have to build something themselves, sort of allowing themselves to throw in the towel on their own efforts. Can you talk about these and why we should be concerned?
CC: It’s quite pervasive, and I can mention half a dozen deals right now that are progressing that have got this feature. It’s much more widespread that a real killer acquisition. A real killer, like some we were looking at in pharma, is when you buy something to suppress it. You suppress it because you’ve got a product that competes with it and you don’t want the competition. That is a killer acquisition. What I see in my practice much more often, the reverse killer acquisition, is based on the idea that you buy something out there that you were actually doing yourself, or could have done if you really put your mind to it, or maybe you were even halfway there. But then you buy this slightly better or more advanced version and thereby you kill your own. You reverse kill. You kill not the target that you’re buying, but your own thing.
LP: Is this, then, a killer more of innovation than of competition?
CC: The two are incredibly close. If we care about dynamic competition, which is really, fundamentally, the creation of new products, and competition in the creation of new products, then we have to care about a situation in which if I didn’t do the deal I would have an incentive, as the buyer, to run as fast as I can to catch up with these other companies that are already in the market. The two become one.
Let me give you an example. Let’s take the FTC case against Meta concerning the [fitness startup] Within acquisition, which is being challenged (for transparency, I have had some involvement with this case). I would argue that the deal has got features of a reverse killer acquisition to the extent that Meta was trying internally to develop capabilities to do immersive reality fitness work. (There’s a further discussion about whether that particular user case is an important building block for them towards establishing themselves in this new form factor which is the metaverse. I would argue it definitely is because they hold a significant number of small BR apps [streaming apps]).
Anyway, fitness is an important user case, and there is evidence in the public domain that’s how Meta was thinking of it. You’ve got your headsets and you’re thinking about things that you can do that are social and fun in the immersive reality world. Exercising is one of them. There is a big focus on exercising in the metaverse. It is a very significant future application and a building block for other things, but this is something that they had their eyes on before. To me they didn’t buy Within to kill Within, but so that they wouldn’t need to make any effort of their own. Within is no longer independent. It’s now owned by Meta, so it can be used and manipulated to do all the other things that we know Meta does when they acquire a company. Meta owns the headset and the store, and they can do various things to make it more difficult for others to establish themselves. The point about the reverse killer is that to the extent that there was an incipient effort inside Meta to develop its own fitness app, that effort is killed.
The company can say, well, the effort was going nowhere, what’s wrong with buying a better version. Well, yes, but the point is that you’re Meta, right? You could get it right. You could make it work. You have the funds and the ability. In the U.K., the exact question was asked when Amazon bought a stake in Deliveroo [an online food delivery company]. The CMA was considering that they could, in principle, have done a delivery service themselves, and by buying Deliveroo they were forgoing their own effort in this space.
The result in reverse killer cases is that we see less innovation. I see this much more obviously as a phenomenon with these big digital companies because they have the cash, they’re so acquisitive, and they constantly acquire complements — and not just those that are totally independent and what have you, but things that they are trying to do themselves. Then they turn and snuff out that internal effort. Think about a world in which Within was attracting other funding and went further into the future, working with all headsets, and then Meta developed their own version. There would be more competition, more choices for people looking for fitness apps. That is the point.
LP: You’ve noted that the best way to avoid many of these antitrust problems is to preempt them by avoiding mergers in the first place. Why is this critical to intervention that works?
CC: It is a very important piece of this conversation. We have established and exposed that antitrust intervention is ineffectual; that we get there far too late; that we are not able to effect change on the ground; and that it will not work because the regulation that is before us is going to attract a series of challenges in which these companies will sue in Luxembourg, every step of the way.
Personally, I do not predict that there will be any sort of movement. So I am in favor of looking at the merger policies and how these companies have grown to the grotesque levels that they have grown today. We know that the regulators, certainly in the U.S., like Jonathan Kanter and Lina Khan, are thinking about this. Will the courts follow? There is no telling. They have to try it out. There is a huge body of the conventional antitrust establishment that says that this is all madness, that it has to end. There are a lot of people waiting for it to fail. But there’s no question that the merger policies have been incredibly permissive.
I’ve been in this business for 25 years and we’ve stopped almost nothing. There’s this view that for companies to merge is a fundamental right. But sorry, where is the evidence that all of these mergers are benign? That efficiencies are being realized? I’ve not yet seen an efficiency study worth a damn and I’ve often advised the parties. The majority of these deals are intended to create market power. The hubris of people who say they are not is extraordinary. I’ve got 25 years in my career to look back on and to say, guys, we’ve left behind a shitshow. A lot of people think that I’ve gone completely mad to say the things that I say, but it’s true. When I came into this business as an advisor 25 to 30 years ago, there was an element of idealism because, certainly in Europe at the time, antitrust was thought to be such a soft science. We thought we needed to bring structure and form and mathematics — and then you had this crazy shift in which the economists understood there was a ton of money to be made by giving lawyers useful narratives and we sold our soul. That’s what happened.
Lynn Parramore – Senior Research Analyst, INET