B School for Public Policy

Professor Herbert Hovenkamp

Antitrust in Labor Markets: How it Relates to Big Tech

Big U.S. tech companies like Apple, Alphabet, Facebook, and others, are coming under fire for being monopolies that should be broken up. This is what we have been hearing from Democratic Presidential candidates like Massachusetts Senator Elizabeth Warren and Minnesota Senator Amy Klobuchar. It is an idea that is gaining steam as these tech giants face accusations of violating privacy rights, squeezing out competitors, and spreading misinformation. A new 150-page report commissioned by the British government includes many of those similar criticisms and say the existing rules governing these companies are outdated and need to be strengthened. And the European Union has repeatedly fined big tech companies. So is it time for the U.S. to look at whether the tech industry is too big and make some changes?

An edited transcript of the conversation follows.

Knowledge@Wharton: Herbert Hovenkamp is a professor with a joint appointment at the University of Pennsylvania Law School and here at the Wharton School, he joins me here in studio. Also joining us on the phone, Hemant Bhargava who is a chair in technology management in the graduate school at the University of California at Davis. And also joining us on the phone, William Kovacic who is a professor of law and policy and also director of the Competition Law Center at George Washington University Law School. He is also a former Federal Trade Commission commissioner. Herb, great to see you again.

Herbert Hovenkamp:  Thank you, great to be here.

Knowledge@Wharton:  Thank you. Great to have you back. Hemant, Bill, great to have you with us today. Thank you both.

Hemant Bhargava:  Good to be here, thank you.

William Kovacic:  Pleasure to be here.

Knowledge@Wharton: So is the assumption of Senator Warren correct, Herb, that these companies are monopolies?

Hovenkamp: Probably not for most products and services that they sell. And we have to disaggregate some of the informational and political concerns from the anti-trust concerns. I mean, I think we do have a monopoly problem in the United States but it relates to price cost margins. And these companies are not notably at fault in that regard.

Amazon has very low margins, in fact that is what people complain about much more. At the consumer level Google most commonly charges a price of zero, so high margins is not the problem. They are however very big companies, and that creates questions about who the anti-trust laws are supposed to protect. Are we supposed to be protecting consumers who have an interest in high output and low prices, or else competitors and other people who deal more laterally with these big companies who might be filling a pinch. Although I have say there is not all that much evidence even of that.

Knowledge@Wharton: Bill, your thoughts?

Kovacic: I agree with Herb. To an anti-trust person the issues are a good deal more complex than set out in some of the public debates so far. And a crucial level of complexity that lies in the background is if you decide in effect to go to the moon, how are you going to get there? And getting there has proven historically to be very difficult. And the suggestion that you might be able to simply step in and start disaggregating these firms in a relatively simple and straightforward way runs up against a great deal of historical experience that says, you can do it but it is hardly an easy venture.

Knowledge@Wharton: Hemant?

Bhargava: Yes Dan, it seems you have three people who pretty much agree on this. I think Senator Warren, what she is trying to do is good. But I think first of all the disease is not that bad. I don’t think these companies — they are certainly big, but they are not monopolies. They are certainly nowhere near as dominant as Microsoft was, for example, in the 1990’s with over 90% of market share in operating systems.

So I brought up the other yard stick, that if you look at it from the perspective of prices, consumer surplus, I mean all four of these companies that she names have created enormous surplus over the years. They may have hurt, like Herb said, some of their smaller competitors but again if you think of Amazon and Google and others, they have enabled tens of thousands of smaller companies to come out of nowhere. So I think I really don’t fear yard stick by which this claim of monopoly dominance leading to bad outcomes seems to be true.

Knowledge@Wharton:  Bill, then one of the conversations we have had on the show, and I would love to get your opinion about it, is where we are headed with some of these companies. And obviously last year we saw a variety of tech CEOs appearing before congressional committees about big data and privacy issues. And the question out there I think for many people is whether or not there needs to be a certain level of regulation with these companies. And I think the expectation is we are headed that way. The bigger question might be is, when would we be getting to that point?

Kovacic: I think Herb is right, that there is a larger, important debate to be had about public policy. In my own mind there are policy areas that need a significant upgrade in the United States. They don’t all necessarily deal with anti-trust law, but to give you one example. I think the U.S. has a badly patchwork system of privacy oversight.

And a basic upgrade of the U.S. regime to create a comprehensive privacy regime that would eliminate a lot of loopholes and improve oversight would be a real improvement. And I would anticipate that one thing that we will see is that within the next few years I expect we will see our Congress adopt comprehensive privacy legislation that creates a new and more effective framework for oversight. I would see that to be one important consequence of the debate we are having now.

Knowledge@Wharton: How much element could there be coming out of something like GDPR with any potential new legislation moving forward in your opinion?

Kovacic: In many respects GDPR is becoming a global standard, at least for western economies. So that if the U.S. does nothing and effects GDPR by default, will increasingly shape U.S. privacy policy. So that is an important stimulus, coupled with the revelations that you referred to before. So the pressure from GDPR, does the U.S. want to set privacy policy or just default, plus the revelations, plus a growing awareness that the U.S. regime has a lot of holes in it, I think is going to give us a new privacy law in a couple of years.

Knowledge@Wharton: Hemant, one of the things that has also been discussed is if you were to break up one of these companies or several of them, the impact that it would have had on the U.S. economy because of the fact that they are a significant piece to that. How do you view a potential break up of one or any of these companies in regards to the impact on the U.S. economy?

Bhargava: Then you can look at the short term and long term aspects of the break up. And everyone in all of these companies, and in different ways, but everyone of them has adopted business models that require making sacrifices. Giving up profit opportunities in one area or at certain points in time in order to get higher profit opportunities elsewhere.

And if you somehow tell them, tech companies, that you cannot do that, and that at some point the spigot will be turned off, then the question is will companies in the future be willing to take on losses for five years, ten years, or give up their most luxury products, like Google search, for free for years without turning a profit. And that will really — I think that to me is the most dangerous part of it, that there is a negative long term impact that people will not see.

Because they might see gains in the short term that companies get broken up. But I think even in the short term, the evidence is not there. And that is where it will be different for every one of these companies that, are you breaking them up vertically, are you breaking it up horizontally, are you breaking them up geographically like the Bell companies are broken up.

So I am really not very optimistic. I will agree with William that regulation is needed, but it should not be on the outcomes you desire, it should really be on fundamental factors that affect or have causal influences on the outcomes that you desire. And privacy is a good one. Some of us have been talking about this for 20 years, how the regulation on privacy is so far behind our technology in general.

And what GDPR has done is some of the most very sensible things on data privacy. But there is sort of the national security aspects of data, there is control of fake news, the formation of standards. I think this is where we should be thinking about policy, and not about mandating size which, as I said, really what do you mean by it is too big? And in fact, if you think of these big tech companies as one industry, they are all fighting with each other. They are in fact competing, and so it is very hard to claim that they are their own monopoly.

Knowledge@Wharton: Herbert?

Hovenkamp:  Just looking at Senator Warren’s two proposals, I think the second one which is a limitation on acquisitions of smaller firms, particularly startups, I think that has some real merit at a particular level. We do have to look a little bit more closely because I think a case is to be made that some of these independent firms that the large platforms have acquired could have developed into rivals.

And acquisitions have foreclosed that from happening. I think it is overstated in her statement, but nevertheless it is a problem we have to consider. The other one I think is affirmatively harmful, and that is the one of disaggregating platforms from markets. Her proposal would say if it is a big platform, $25 million or more in annual revenue, then the platform will be permitted to sell the goods of others, but it will not be permitted to sell that particular firm’s own goods.

For example, looking at Amazon, the biggest target would be Amazon’s home brands, like Amazon Basics. So just think about what that means a second. So I am going to take an ordinary consumer product that has very high volume, and that is household batteries. Alkaline, AA, AAA batteries. The market is dominated by very large firms. Duracell is owned by Berkshire Hathaway, Energizer, Rayovac, and Eveready are all brands controlled by one very, very large battery company.

Amazon has entered this market with Amazon Basics, which is its house brand that it manufactures itself, and sells at quite a bit cheaper price, ranging from 10-50% lower depending on what size and what quantity of batteries you buy. Interestingly the name brand batteries in this market are among the highest margin products sold in the U.S. Margins will in excess of 50% or 60%.

And Amazon Basics is putting real pressure on them to reduce their margins. But under the Warren proposal, Amazon would no longer be permitted to sell its own house brands on the Amazon platform. We’re not benefitting some small firm here, we’re benefitting Berkshire Hathaway who already has plenty high margins on its brand of batteries, and I don’t see any value in doing that sort of thing.

Knowledge@Wharton: Bill, your thoughts?

Kovacic: I think to take one other theme that comes out of this, there is not a lot of attention in the initial proposal to how this is actually going to be done. You notice that an additional element is that there is going to be a command to unwind a specific number of earlier transactions that Amazon, Google, and Facebook are engaged in. I think it is a total of six or seven earlier mergers.

And Senator Warren’s proposal says, I am going to appoint regulators who are going to take the jobs of the anti-trust agencies to unwind these transactions. And the strong suggestion is that anything short of going after all of them will be inadequate, will be a timid response. That is an extraordinarily difficult undertaking. Not impossible. But you are going to have to go into a federal court and explain a theory of competitive harm.

And as Herb has just been mentioning the other side is going to push back hard. And the U.S. jurisprudence allows you to provide evidence of consumer benefits, and to emphasize those benefits. There will be nothing automatic or easy about tearing out this program. There are no uncontested layups in this season of college basketball championships.

And the effort to go after these six or seven deals at one time will prove to be an extraordinarily difficult undertaking for the competition authorities. It would be hard to do one of them, but I think the mandate is going to be if you don’t do all of them at one time you have fallen short. And here again the implementation blindside I think comes into focus in a powerful way.

Knowledge@Wharton: So then how much of this, and I think is a relatively high number that I might be searching here for Bill, but how much of this issue being brought forth now by Senator Warren and Amy Klobuchar and others really does have the political element behind it?

Kovacic: I think it does have a strong political element, and the political environment creates the possibility for this debate to take place. For some of us who have been in the field for a while, it is to step back extraordinary to see how much attention and significance these types of proposals have gotten. And I don’t think one could have predicted that five years ago, that it would emerge in the way that it has. But the fact that it does have such political visibility and significance, and a constituency I would say in the background that finds it attractive, means that these are extremely serious proposals.

Knowledge@Wharton: Hemant?

Bhargava: A couple of things, Dan. First, I agree with one of Herb’s points about the effect of allowing platforms to also be in the markets. And in fact my colleague at UC Irvine wrote a paper on this 20 years ago demonstrating that when you do that, that bias can actually have positive effects for consumers. And in fact my friend Marshall Van Alstyne wrote a book, Platform Revolution, he is at Boston University just across from Senator Warren, and that book also supports the argument that Herb made.

The other point Herb said about these big firms acquiring small firms, and the effect of perhaps prohibiting that, I think that can actually have negative effects. Because many of these startups, small firms for innovation, the initial $100 million that they get are often with the intent that their products would be acquired by Google or Facebook, etcetera. And the reason for that is that in many of these technologies industry being able to build an end to end product is very, very beneficial. The synergies are huge if you can cover all of the pieces.

And therefore, for many of these firms their only way out is to get acquired by these big tech companies. And if you somehow tell them that is not going to happen, then they may never reach the point of developing that ambitious product. So I think that is also to be kept in mind. And then finally coming to the implementation point, I totally agree with William.

I think even with the Microsoft break up that was sort of mandated, it never happened. And Microsoft lost its lead in the Internet Explorer, not because of any help from the government or its other firms, but simply because other firms innovated better and faster. And in fact, Microsoft had a lot of hubris as the leader at the time. So I think we can rely on standard competitive forces, and in fact the evidence that the government actions lead to good results is really not there.

Knowledge@Wharton: So how much does what Hemant said, Herb, about that advantage, that competitive advantage that the Apples, the Googles, the Facebooks, or the Alphabets, have in society because of the resources that they have, and the ability to develop quicker than, say, a smaller company a startup would actually have?

Hovenkamp: It is actually very market specific. Search engines are a good example, where all of the traffic tends to gravitate towards Google or at least the dominant search engines, and secondary search engines have had a very hard time. If you look at other types, dating websites are a good contrary example, because they are able to product differentiate. They’ve got all kinds of dating sites for older people, for people of certain religious backgrounds, and so on.

And then there is more room for new entry at the differentiated level. But the important thing is you can’t come up with a blanket prescription for all of this. You have to look item by item. One thing that might work as an alternative in some of these startup acquisition markets is non-exclusive licensing as an alternative. That is, say, if a company like Google wants to buy a small tech company, maybe relegated to obtaining a non-exclusive license which will give Google everything it needs to practice in that market, but it won’t give it the right to exclude other people.

That will make the target firm less valuable, but it may not wipe out the advantages of entry altogether. I think we really are giving up something when we have too blanket an approach to categorically allowing all of these startup acquisitions, which number today in the hundreds.

Knowledge@Wharton: Bill?

Kovacic: And I think Herb and Hemant both underscore an important point here. The nature of the prohibition on large firms acquiring nascent threats in effect. Or deliberately using a strategy to buy them to suppress the development of new ideas. That is both of these proposals in many ways suggest a blanket prohibition of the kind Herb is recommending. It is not a case by case approach.

To make it effective, when you look at the words and the spirit of the proposals, it would mean that firms of a certain significance would not be allowed to purchase a new enterprise that, in some foreseeable way, or even speculative way, might come back to challenge them over time. So to put the proposals into effect as suggested really does involve a categorical ban on a significant number of acquisitions. And to do that I think really requires a change in the law.

So to make that truly effective, to have that complete barrier to acquisitions in place, requires going into the Congress and changing the law. As would the proposal that Herb was mentioning, that is a separation within the platform and products. That requires a new statute. So it is a long, long way to temporary to put all of those measure in place.

Knowledge@Wharton: Yeah, and right now Bill I don’t think there is an expectation of being able to try and get any kind of legislation that would include all of this through this particular Congress right now. We might be looking at several years before something would come up like this.

Kovacic: The only thing that would change is if you had some cataclysm, some external shock that is a result of a corporate scandal, a further set of revelations that calls into question the legitimacy of the sector. And that has already had a bit of a catalytic effect already. But if you look at the cataclysm it took to produce Dodd Frank, which Senator Warren had a major role in crafting that. She is very familiar with that process.

But if we think about the kind of political opposition, the backlash that will take place as you start accounting for a number of the factors that both Hemant and Herb have mentioned, the counterpoints, you will have a lot of resistance in the legislative process. So this really — if you start to look at the calendar and plot out how long it will take to do this, it’s a couple of years to get the legislation, it’s a couple of years to set up the new regulatory mechanism that will oversee the operation of platforms. A couple of more years to get implementing rules that put the separation into effect, the court challenges. This is a bit like building an aircraft carrier. It will take quite a while.

Knowledge@Wharton: Hemant?

Bhargava: Yeah, so the speed at which this industry changes is of course enormous, and that is why regulation gets left behind. And that doesn’t mean we should never start thinking of that regulation. We should have thought of it 20 years ago. But what it means is that it really should be focused on fundamental factors that will last for a long time and govern the evolution of competition in this industry.

And exclusive licensing or data sharing, that is something that Herb brought up as an example, that we normally think of data sharing as a really, really bad thing. But think about how it has really pro-competitive effects. So if one company spends millions of dollars to acquire all of the mapping data, and then dozens of applications can be built on top of that, if you prohibit sharing of the data, then you are actually preventing new firms from emerging and building those applications.

And the same thing could be, it could be data about consumers, it could be data about devices. So this is something that needs to be debated, that is data sharing good? Is it bad? And what should be the boundaries around it? How should that be regulated? And I think by debating on size or power we are actually barking up the wrong tree, because that is really, really hard to establish.

It is hard to control and achieve the right results through regulation, and then its implementation. And instead I think we need to have a rigorous debate on, what are some of the fundamental factors and how do they influence the outcomes that we are truly interested in. And as Herb said at the very start, the standard benchmark of consumer prices and welfare, if you simply evaluated the tech industry by that, there is really no case for any regulation or any action. And therefore it is clear that even the thinking has to change, and that other benchmarks need to be thought about.

Knowledge@Wharton: One of the things, Herb, that we have talked with you in the past when we talk about these issues is consumer harm. And that seemingly is kind of the standard that is used in many of these instances. And when you think about an Apple or a Facebook — now Facebook might be a little bit different because of the data privacy issues that they have been going through recently, but with an Apple where is the consumer harm by having the variety of products that they have in their realm right now?

Hovenkamp: I think the single biggest fundamental problem with the Warren proposals is that they do not sort out who is being harmed and who is being benefitted. And I found them somewhat distressing, because Elizabeth Warren has dedicated her entire professional career to protection of consumers. But these proposals seem calculated to me to raise consumer prices, reduce consumer variety, and it is kind of hard to figure out who is systematically benefitted by them.

I mean, there are plenty of small businesses who are injured by Amazon, but there are at least as many small businesses who are benefitted because Amazon acts as their broker and gets stuff out that they could never — they would not have the resources to get out on their own. So you’ve got to start with the question, who is getting hurt? And then decide whether that group is worthy of being your constituency.

Knowledge@Wharton: Bill?

Kovacic: I think Herb puts his finger right on the deeper, larger issue that is presented here, and that is what is the purpose of having a system of competition policy? Who are the intended beneficiaries? And maybe this is partly Senator Warren’s aim, along with like-minded advocates of basic reforms, to promote this debate.

I suppose if she were on the line she would say that what the three of us have been saying shows how timid and blind we are to the larger environment that we live in. That some larger mixture of concerns involving small enterprises, concerning the welfare workers, the impact of large corporation organizations on the integrity of the democratic process, those all should be part of the debate.

I suppose a healthy part of the debate could be, as Herb said, a much more specific discussion of how you intend to put this framework in place. And precisely what tradeoffs you are willing to make. The discussion so far takes place at such a high level of generality, that all of those tradeoffs and adjustments can be sidestepped. But as one goes further down the path of thinking about what you actually want to do, and what changes in the law you want to provide, I don’t think you can avoid a very specific discussion about who the intended beneficiaries of the law will be.

Knowledge@Wharton: Great having you all with us. Herb, great seeing you as well.

Hovenkamp: Thank you.

Knowledge@Wharton: Hemant, Bill, great to have you with us today. Thank you both, gentlemen.

Kovacic: Thank you.

Bhargava: Thank you Dan.