Nobel Laureate Paul Romer on How to Curb Big Tech’s Power

Sebastian Burca:
Good afternoon and thank you all for joining us. I am Sebastian Burca, the acting director of the Stigler Center. And today we are very happy to host Professors Paul Romer and Luigi Zingales for the keynote in our antitrust and competition conference, which this year has been focused on monopolies and politics. And we have been holding it in a series of webinars throughout the year. You can find more information on our website.

Before we begin please note we are on the record and we will be posting the video on our YouTube channel later. If you have any questions for the speakers, we will address them in the last 15 minutes or so and you can submit them via the Q and A at the bottom of your screens. As usual views expressed by guests are their own, not those of the Stigler Center or the University of Chicago. And we hope that you will join us for more upcoming webinars, so please check our website as well as our publication ProMarket.org and our Capitalisn't podcast.

Back to today. In the interest of allowing more time for the discussion, please allow me to very briefly introduce our speakers. Their full bios are available on our website. Paul Romer is a University of Chicago alum, a professor of economics at NYU, and also the core recipient of the 2018 Nobel Memorial Prize in Economics. And Luigi Zingales, our moderator is the Robert C. McCormack Distinguished Service Professor of Entrepreneurship and Finance at Booth and the Faculty Director of the Stigler Center. So now without further ado, I turn it over to our speakers. Thank you.

Luigi Zingales:
Thank you Sebastian. And as a short introduction, I want to say that the Stigler Center has become known for its trendsetting conferences on antitrust. In 2019, we presented a report on digital platforms that anticipated much of the work that Congress did in 2020. In May 2020, we're planning to have a conference on the threat that monopolies pose to democracy. Unfortunately, the pandemic forced us to move that conference online and we'll spread it out over time.

The first phase took place in the spring and fall of 2020 and brought together economists, legal scholars, historians and political scientists to explore the historical evidence on the interconnection between market power and political power. Our webinars discuss the role of industrial cartels in facilitating the rise of Nazi Germany and of Imperial Japan. The impact of Chaebol and other industrial conglomerates in the political dynamics in South Korea, Israel, Brazil and Mexico. And also the rise and fall of anti-monopoly movements in the United States.

All these webinars are freely available online on our Stigler Center YouTube page. In general, final conclusion to these discussions is that market and political power are intrinsically linked and a vibrant democracy only exists in an environment of vibrant economic competition and vice versa. We are however witnessing growing market concentration. In the last few days we also saw our concentration can facilitate the elimination of dissent. These trends raise the question, do we need to modify antitrust policy to take care of these concerns, and if so how?

This is the reason why the second part of our conference, which will start today is focused on policy. In the next 45 days, we will discuss whether antitrust should be designed to explicitly promote economic or political liberty, and what results are involved. Today's webinar marks the beginning of the second phase, we hope that you will follow us in this debate. Thank you. And Paul, the floor is yours.

Paul Romer:
Okay, great. So let me make a few remarks to just set up the discussion that will follow. I'll start by offering what I think should be the subtitle of my talk, which is that asymptotically big really is bad. When firms get big enough, it's always bad and it's bad in terms of efficiency if we understand efficiency correctly. This means we can't look just at narrow efficiency measured in terms of say purchasing power that people have, income, goods they receive, but we need to look broadly at the things that the market delivers at income, the traditional measures. But also things that people value like democracy and the rule of law for their own sake. And the instrumental value that the rule of law and democracy, which acts as the control, the check on rule of law, the guarantor of rule of law.

These also have an instrumental value in protecting the system that lets the market thrive. So I think we just look at a couple of examples to see the ways in which when firms get big enough, they evade the rule of law and threaten our democratic system. As an example of evasion of the rule of law, I think there's no better case to look at than Boeing, which was able to systematically gut the regulatory mechanisms that had been set up and which worked extremely well in the United States in terms of ensuring a very high level of safety, of flight.

Boeing has just now, in the wake of all of the chaos tried to kind of slip in their negotiated settlement where they admitted guilt in misleading the FAA. But the problems at Boeing go far beyond what they admitted and the fine that was imposed is really trivial compared to the benefits that Boeing was pursuing. So Boeing is effectively unregulated, has been and I think will probably continue to be because it's too big.

Now, let me take the other case. Facebook has a very large role that it can play in elections. When I learned about this, and this is not just in the US, this is in other countries. I was talking with the elected leaders in another country and asking them, aren't you a little bit nervous about having this foreign firm that so active in your elections. And then what happened was, is there was a discussion that followed where they were telling me their side is so much farther ahead in using Facebook that they're certain they're going to be able to win the next election, probably the next two before the other side catches up. And no kind of introspection at all about what this might be doing for the system as a whole, even as it might help them, because they're better at using Facebook.

I started asking people who know about Facebook, this was early in 2020. If Mark Zuckerberg decided to try and influence a close election, could he do so? Now, the way I asked that question kind of prompts people to give me positive responses. So I wouldn't pay too much attention to the level of the response I got. But what was striking was that the more people knew about the details of political advertising on Facebook and things like the newsfeed, the stronger was their assertion that yes indeed if he chose to Zuckerberg could tip a close election and could do so without violating any law.

So this, I think is the fundamental problem that we have to grapple with. One that is, I think, more important and more fundamental than this issue about moderation, whether or not moderation is censorship. My quick take on that is that if you think that moderation is censorship, moderation in the sense of a firm which moderates limits what can actually be said on its platform. If you think moderation is censorship, you got a competition problem and you need to fix the competition problem.

Now, turning to that competition problem, first I think what we're going through is a kind of a phase change where many markets which before were arguably competitive or effectively competitive are now turning into markets which will be characterized by a strong form of winner take all competition. This may have been true for a long time in say manufacturing aircraft, but it's now working its way through many other sectors of the economy as algorithms become more and more important.

As a side note, I don't think that it's data here which is causing the underlying increasing returns. It's an algorithm that you can use over and over again. And we could follow up on this in the discussion that follows. If you go through this phase change where every firm is trying to be the first down the learning curve, and one happens to win and then ends up dominating the market, antitrust it just fails to be an effective tool because you basically have to say that if the winning firm broke the law, every firm broke the law and every firm that wins broke the law. And this would be an unacceptable position to be in if we were to give prosecutors the discretion to go out and find in violation of law any firm they want to.

So I think that what judges have always done in the United States and will continue to do is set increasingly high bars that prosecutors have to meet because they're worried about abuse of prosecutorial discretion. So you can't fix this problem of pervasive winner take all competition through the judicial system and findings of the firms have violated the law. The thing we can do to restore efficiency is the thing we've always talked about doing, which is to use a Pigouvian tax. Let me describe this in the context of the Boeing example.

Suppose that we wanted to make sure that there were several different manufacturers of aircraft worldwide, if that was in the interests of the US and US consumers, we could put a tax on the sale of aircraft. We could put a tax on tickets. It doesn't matter where you levy it, but the nature of that tax is that it should be a progressive. The marginal tax rate should increase with the scale of the firm that made the airplane, that the flight uses. And that tax could start out at zero for some big exemption bracket. Basically for firms that capture say, up to a fifth of the total worldwide aviation market, and there could increase up to something like 10% for a firm that captures the entire global aviation market.

So this puts an upper bound on the efficiency losses you're going to bear if you think that having a single firm is dramatically better in terms of more traditional measures of competition, you'd have the effect of a 10% tax. But if you look at the details in the aircraft market, I think what you'd see is that even a tax that approaches 10% would be enough to give some advantage to smaller firms, make it less attractive for a firm like Boeing to go out and buy a firm like Embraer or Bombardier, might even make it attractive for a firm like Boeing or Airbus to spin out as a separate company, a new firm that produces one line of aircraft, even as it produces another.

And if 10% doesn't get you the diversity of firms you want, make it 20, make it 50. We can push these tax rates as high as we want, depending on the weight we put on protecting the rule of law, protecting democracy, versus getting an efficient production system within a firm. So with that I will say that I'm planning to write something about modeling this phase change. I think it's an easy thing to model. And so you can look to things I'll put on ProMarket and we'll distribute later.

And I've also written about using progressive Pigouvian tax on digital advertising. We could explore the details there, but I think the general principles are that big is bad enough that we've got to have an effective tool to keep firms from getting too big. And you can always design a Pigouvian tax, which makes sure that whatever the underlying technology is, the net of tax returns to scale are ultimately diminishing. So with that let's proceed into the questions.

Luigi Zingales:
So Paul, first of all, thank you very, very much for a very sharp, short and provocative speech in a very Chicago fashion. You have put a lot of meat on the fire. So let me try to pick a little bit of this meat in pieces and then try to analyze it. The first thing is you said something that is very, at least perceived as very anti-Chicago, which is big is bad. Now, for those of you who don't know the history, actually this has not always been the case for Chicago. In fact, George Stigler, the namesake of the center that I direct wrote in 1952 a piece on Fortune saying the case against big business in which he was advocating the breakup of big business because the break up is a better way to basically regulate them than to actually regulate them. He said, it's not only the right ways, it's also the conservative way.

Now, however after that Stigler opinion changed. And the main thrust of the Chicago opinion changed. And I think much of that was changed, if I'm not mistaken by Harold Demsetz, who really emphasized the efficiency of big business. And so the idea was, we don't want to penalize bigness for bigness because is penalizing success, is penalizing efficiency, is penalizing all the things that come with it. So to a traditional Chicago guy who say, but doing so you put a tax on success, how would you respond to that?

Paul Romer:
Well, I think part of what was right about the Chicago skepticism is a concern that legislation might be dictated by notions of fairness. And fairness is something that people care about, but I don't think economists have anything useful to say about it. So I think if we want to try and contribute to a debate about antitrust, we need to frame it in terms of efficiency. The place where I think I differ from what's known as the Chicago School is in my definition of efficiency. And I'm not even sure that people like Stigler, Demsetz, or Aaron Director would disagree with me if we make it in terms of an asymptotic argument.

As firms get so large that they can basically intimidate the government, prevent the enforcement of law and undermine the functioning of democracy, I think they would agree with me that that imposes an efficiency cost that we have to reckon alongside of the cost of the airplanes or the value of advertising or something like that. So I think they just didn't contemplate big, getting up to the kind of scale that we're dealing with and big in the kinds of ways that we're now seeing, which is having these big efficiency costs in terms of what this broad measure of efficiency that I'm describing. So I'm hopeful that the people who were in the spirit of Chicago as they think about these questions will agree that there is an issue that we need to discuss about the broad efficiency effects of big firms.

Luigi Zingales:
So that's very clever by redefining efficiency, you are bringing everything under the economy standards. It's a little bit like the Chinese. And I'm told that they say they've never been dominated by an old Chinese country, and part of the story is that they redefine as Chinese who dominate them. So that's kind of a very clever way to address the problem.

Paul Romer:
But if I can say, just in my own defense, we probably all think a little bit about the marketability of our arguments. But frankly, I really struggled with this question and because I have trouble understanding the role of economics beyond efficiency. So it clarified things for me, what I could frame it in terms of efficiency. And actually, let me just mention one thing I forgot to say. There's a long debate about is efficiency enough? Do we have to worry about the ex-post transfers and so forth? I think in practice; most economists are comfortable with the kind of expected Pareto improvement criteria for efficiency. It's what Rawls had in mind when he talked about the kind of the veil of ignorance.

So when I think about a policy change in the United States, and I contemplate, what will this mean for my unborn great grandchildren in the future? I don't know anything about where they'll live or who they'll be, but I can contemplate what would this country be like for them? And if on average people in their position are going to be better by some change we make, then I wouldn't get hung up over the narrow distributional effects of some policy that we adopt.

Luigi Zingales:
I understand that, and I didn't mean to say in a negative sense, I think it was clever in a positive sense because allows economists to discuss these topics. In a sense, my view is that there is a price I'm willing to pay for freedom. And I think that whether that price is exhaustively given by the political authority and the democratic system, or whether this is incorporated in a bigger notion of the efficiency, I think the bottom line is that we need to face a tradeoff and of course we don't want to be starving and free. But also we don't want to have an extra layer of wealth at the cost of not being able to speak up. At least I don't, it's a personal preference, but I think that your notion allows to incorporate that in a framework that at least we economists find it very useful. I don't know whether political scientist or other people find it useful, but I think at least we find it useful.

But let me push back a little bit because my understanding is that your notion is more about concentration rather than size itself. Because if I am a large multinational companies that have a teeny tiny share of market in every country, but I'm surely big on sheer size, I'm not as dangerous as for example a big company town. The town might be small, the company might be small, but as a company town, I own everything in town and I have a disproportionate amount of power.

Paul Romer:
Yep. So you're right. And part of why I framed it in terms of like asymptotically or in the limit, big is always bad, it's because really what I'm trying to do there is suggest that this becomes an empirical question. We need to look in different markets. And our notion of what is too big and threaten say the rule of law might be different in the market for aircraft, as opposed to the market for advertising services or search or ... So unfortunately this is going to be a somewhat messy, empirical exercise in a number of different markets. But I agree that it won't be that it's just dollars of revenue as a single measure of big that we can apply across all contexts.

On the other hand, I think we want to push back a little bit against this is kind of objection that's easy to anticipate, which is like, how do we really know what's exactly the optimal scale for firms in any given market? I mean, answers we'll never know, but we're talking about kind of adjustments on the top of some quadratic surface. And you get it a little bit wrong, too big, too small. Like it doesn't matter from the point of view of efficiency. But where you really can get bad losses is if you go kind of to the limit where you really do undermine the rule of law and democracy, which as I said, we're already seeing effects like this, or if you went in the other direction and every firm had to be like trivially small you could do some real damage with that too. But there's a broad middle ground that I think we could target with the right kind of progressive Pigouvian taxation.

Luigi Zingales:
So there is one aspect in which you are squarely in the Chicago School of antitrust tradition, which is you are seriously concerned about what is called administrability of the antitrust. And so your preference if I understand correctly from your speech is not to delegate this to any judge, we'll find it very difficult to establish this, but is to delegate this to the political authority. So in your view, you see Congress coming in and deciding you are too big to exist to some extent. And so I will tax you as a result of that. Now, problem number one and I think I have other concern, but let's start with problem one.

Problem number one with this implementation is that the political economy is really, really hot, in a sense that here you are Google or Amazon, the richest or close to richest person on the United States, the most powerful from every dimension. And you expect that some elected representative that depend on you as you said, to be elected, to be campaign finance, to get a job after you step down from Congress and et cetera, to turn against you and say, I want to target you and put a tax on you. How feasible that is?

Paul Romer:
Yeah. Well, let me go back to something you alluded to, which is actually important to keep in mind, but we have a federal system. There are local versions of taxation, local government. There's national government, there's even an international kind of consensus about some things about taxation. My analysis went like this. I don't think you can ask judges to make this decision about, like who broke the law and who didn't, when we're kind of trying to figure out where we are in the top of this quadratic surface. So this is very badly suited for the judicial system.

On the other hand, I think part of the lessons from the Chicago School about how easily regulation gets corrupted should tell us that just giving this as a kind of a discretionary power to the executive branch is also a very dangerous way to go. So I selected the legislators as the least bad of the three possibilities, but here where we have some advantages, frankly is with the laboratories of democracies, is the many states, even cities in the United States.

The practical reality is that a firm like Boeing or in a market like advertising has revenue and power and benefits, which are very strongly geographically concentrated in this country. So Maryland, for example, which is one of the first legislatures that's tried to implement one of these ad taxes, it makes perfect sense for them to try and claw back for the citizens of Maryland revenue and income, which has been sucked out of their state and is going to San Francisco and New York, Seattle, a few other places. So I think the political economy here is one where if we can allow local taxation, we can see some innovation and then some states will do it, some won't and then we'll kind of learn from their experience as we approach a kind of a national consensus.

Now, this does have a very sharp implication though, for the nature of the taxation. You cannot make this based on corporate income because income is the difference between revenue and cost and those are incurred in different locations. So there's no way in principle, even with infinite data to say where income is earned. So these taxes are going to have to be on something like revenue or units that are delivered where the geographic location is very well-defined and you can allow different levels of taxation in different places and not have to worry about the games that we see, where firms move income around to find the location with the lowest income tax rate.

Luigi Zingales:
I think you're raising an excellent point because what I learned looking the history is for example, the reason why we got tobacco litigation is because the attorney general of Mississippi decided to go after the tobacco company. And the only reason why attorney general of Mississippi went after the tobacco company is that you don't grow tobacco in Mississippi. And if you went North Carolina would be completely a different ballgame. So I think that in that sense, the federalist of the US system, there are a lot of other negative aspects, but this is definitely a positive aspect. Now, let me push a bit-

Paul Romer:
Luigi, before we leave this, let me just like stub out another conversation that I think is worth having. One of the conclusions I've taken away from the pandemic is that we've over-centralized government in the United States. We've created these single points of failure, most dramatically, the CDC and the FDA, and we should be trying to decentralize a lot more things in the United States. So I think there's many reasons to start to think about going back towards more decisions, more expertise at the level of the states. And I think one of the things that would come from that is that it would actually be harder for these big firms to have the kind of success that they've had through lobbying in Washington.

Luigi Zingales:
I am a little bit less optimistic than you are on that front because I look at, for example, Uber and Lyft. And they've been incredibly good at capturing every local console on the face of earth. So the fact of being spread does not prevent you from being captured. But I see your point and I think it's a valid point.

Paul Romer:
But at least that's a conversation worth having about how does the federal structure in decentralization fit into our general agenda of trying to protect our system from abuse by dominant firms.

Luigi Zingales:
Absolutely. But I want to bring back to the case against judges that you made. And I'm going to take a risk here because I want to bring up the Microsoft case. So I know that you are infinitely more of an expert than me on the case. In fact, you were consulted on the case. So I think that, that's where I fear that I shouldn't even go down that path.

Paul Romer:
No, no.

Luigi Zingales:
But maybe because today's coming up a podcast on that particular case that Bethany and I taped. But I learn in taping the podcast that the judges had to do a decision or to what extent in a exclusionary contract is in abuse of market power. So exclusionary contract, if it's done by a small company with barely the market power I'm not very problematic.

And judges tend to see them as efficient and they don't do anything about. But when they're done by a company like Microsoft with at the time 90% on the operating system, they regard them as problematic. So why can't we have a similar judicial system in which you say, if I am, for example Amazon, I have, or whatever company, I have the right to not do business with people I don't like. Okay. Because after all that's what freedom is about, and I'm not forced to do work with you. I'm not forced to provide you service, et cetera.

But when I become 40, 50, 80% of the market share in particular in sectors that have to do with freedom of speech, et cetera, then I really have to see this as violation of some form of antitrust, which is clearly not what has been interpreted in the recent years. But if you go back to the history of antitrust, history of antitrust is not only about consumer welfare, is about kind of keep power in check. And so the fact that if I am, take Apple and Google, two firms control 90% of all the apps in the United States and for what I know in the world. And so if they decide that they don't like you Paul, your app has no future. So why judges cannot interpret the original statue, the Sherman Act in that sense and apply it in a very consistent way?

Paul Romer:
Well, I think this is a very good question, and I'm going to recognize that reasonable people can differ on this, but here's where I came out. If you look at something like price discrimination, we had a series of attempts through kind of strengthening of the law where the legislature was trying to make it easier for prosecutors to win cases against the "Bad firms" the firms that were behaving badly. And what we saw was this pushback from the judiciary where they made it harder and harder to win a conviction just by virtue of the fact that the firms were charging different prices, because all firms do this, virtually all firms.

And the judges were very uncomfortable being in this situation where the law would let a prosecutor prosecute and win a case against almost any firm because they all charged slightly different prices. If I could use an analogy here, that's a little bit crude, but we have some sense that very young people aren't mentally competent to be considered like violent criminals, older people are. You could say to judges, well, if they're short, they're not guilty of murder, but if they're tall, then, yep they're guilty of murder. At the extremes, somebody who's less than three feet high, somebody more than five feet, six feet high, that seems to work fine.

But imagine being a judge and left with this kind of ... this very sharp decision you got to make. Okay, if somebody's four foot six, is that person guilty of murder given the facts here, or by virtue of being four six instead of four seven, is this person not? And I think what judges instinctively and collectively do is resist getting put into that kind of box. So they're going to, and the way they do it, I think historically has been to just raise the bar regardless of what the legislature says and make it much tougher for a prosecutor to win a case.

And then historically, I think what happened is, is they turned to economists and said, "Okay, economists give us some technocratic basis for making it tougher to win these cases." And what the economists said, "Oh, well just make it all about efficiency." The law never spoke about efficiency, but efficiency was a way I think for judges to get out of this box where it was too easy for prosecutors to win cases. So I don't think this was like a devious plot by firms to take over the judiciary, but I think it was really a sign that we were asking judges to make decisions that were too harsh and too hard to make when you're dealing with shades of gray.

Luigi Zingales:
I completely agree on your historical account. Again, I'm going way out of my expertise. My impression is in the criminal cases, for example, if you are below a certain age the burden to prove that you are really guilty is higher. But if you are of age you can be proven completely-

Paul Romer:
But notice what's interesting about that. That was one where the judges said, "Okay Mr. Legislature, you figure this out. You tell me the age and then I'll enforce it." So the judges didn't have to unilaterally decide on age. They have some discretion, but they got the legislatures to own this decision about, okay, what age are you going to pick?

Luigi Zingales:
Maybe we find a compromise here that the legislature start to set some parameters and then the judges then implement them. So we say, if number one, you finance 90%, I'm exaggerating, but 90% of all of our social privacy that's already a pretty bad sign that you are too big to be in fault or if you do. And the beauty of adding some parameters like this well, we'll have some of these firms shy away because, or if you employ every EPA chair after they step down that's a pretty good indicator they are too powerful. And so I think that could be a very interesting-

Paul Romer:
So I should be clear that I mean, I think it's useful to have this conversation to look for ways to use the existing antitrust law more efficiently. And I think the direction you're articulating is a good one, and I hadn't thought about it so clearly, which is really to take some parameters that you could specify in legislation, and then tell the judiciary to work within those parameters. Like look at the kinds of indications that you're talking about. I mean, another one I would put on the plate, it's just, do we actually have choice? Can a consumer actually make a choice? Like, can I go bank with a bank that isn't spying on me with all kinds of tracking software and sending information to these surveillance companies? I got zero choice in that matter right now.

And I think a few pieces of evidence like that could persuade a judge this is a case where we should find a violation. I still think it's going to be tough for them to know exactly what to do. And if we go back to the Microsoft case, at the time I was in favor of, argued for was the one who was going to testify in favor of these apps ops split, just split Microsoft into two companies. Let's set aside the question whether that was really the right intervention, but it's kind of at the time we thought, because we were worried about ongoing interference by the government and actions of firms.

What happened was, is that the trial judge was so convinced about the seriousness of the events and he was willing to accept that split, but then the appeals court was not comfortable making such of what was claimed from Microsoft side to be such a harsh, punitive measure. And so they sought something that was like less serious. So I think you just have to anticipate that on the level of things, like what could get a split up, a durable split up of some of these existing firms. I don't think the judiciary is going to be willing to deliver for us. But I think that the right progressive tax will motivate voluntary spin-outs because they ultimately will be a way to minimize the tax burden.

Luigi Zingales:
So I'm intrigued by your tax example, but I fear that the Boeing case is not the most compelling case to win. And the reason is that rather than taxing Boeing or Airbus, what we see is the opposite, we see subsidies. So my question is in a world of increasing returns to scale and international competition, so taxes are set at the national level, but then there is competition at the international level. And we don't have a taxation at the United Nation level. So the fear is that they are going to use the argument, oh, this is a national champion. We need to win against Airbus. And so what they're going to do is actually subsidy and not taxes.

Paul Romer:
Yeah. This is the first time I've tried the Boeing case. I think it's okay to put it out there, even if it's somewhat weakens my argument, because I had the same kind of thoughts as I was trying to game this. The international agreements between say the United States and the EU would get very complicated in this case. And it's hard to see. I mean, it's not impossible, but it's hard to see how the states could solve this problem as well. I got a noise in the background, I hope you can still hear me.

Luigi Zingales:
Yes. We can hear you, but you can also hear the noise.

Paul Romer:
I apologize, my neighbors.

Luigi Zingales:
That's all right.

Paul Romer:
It's a spillover, what can I do? But I think it's good to actually recognize that this tax idea is going to run into trouble internationally in a case like with airlines. I think with big tech right now, it's so US-based, a little bit China-based too, that many countries may unilaterally go for this idea of taxing the digital advertising, and I think many local jurisdictions might. And then in the aircraft market, we may have to look for some other mechanism. But you're absolutely right that we have historically subsidized Boeing.

And actually Boeing is much more experienced than this. Part of way, the tech companies are vulnerable is that their base of operations is so concentrated. Some of these big firms, especially the military firms have learned how to distribute production across all the different jurisdictions to protect them from local challenges. So there's going to be no perfect clean solutions here, but at least getting taxes on the table I think gives us one more direction that we can push in.

Luigi Zingales:
So I have a lot of super interesting question from the audience. So let me start with one. So Stephen Sklifas said, "Romer Pigouvian tax proposal seems to be premised on the expectation, the big highly concentrated firms will engage in significant harmful conduct in influencing governmental policy. Why not condition such attacks on firm conduct as well as in size and concentration? Wouldn't this modification create incentives for last successful firms to refrain from such harmful conduct?"

Paul Romer:
Yeah. What that means basically is pitching it back to the judiciary. And just think about the kind of conversation between a very large powerful firm. This is a very nice economy you've got here. You wouldn't want something bad to happen to your economy, would you? It's just all the rhetoric of like gangs and intimidation. And I think, proving bad behavior in these cases would be tough in a judicial system where the judges remember, are worried about prosecutors who can misbehave here.

We've seen prosecutions for antitrust that were based on a purely political interests of the executive branch. So the judges are going to be very strict about what constitutes proof of abuse of the political system. And Boeing could say, "Look, we've got a right to lobby. We've got Citizens United, we can give money. You can't accuse us of just cutting the rule of law because we got Congress to cut the funding so the FAA didn't have the staff capacity to actually see what we were doing." So I think kind of this idea that somehow we can pitch it back to judges kind of like King Solomon will solve it for us. That is just kind of an unrealistic.

And we just have to grasp the reality of as voters and as people trying to influence legislators that these are messy, hard questions, but we've got to just make some decisions. And at some scale, the risk of manipulation is too high and we want to prevent it without even waiting for sign that it's a problem. I mean, a good analogy here is conflict of interest. We don't say to a colleague, "Okay. Yeah. We're willing to listen to all your evidence about the safety of this drug, even though you're being paid millions of dollars by the drug company, we'll listen to you. And then if we can prove that you said something that was biased in favor of the company, well, then we'll start discounting it."

We just don't do that. We just have this hard rule that you don't trust somebody who is in a position where they have an incentive to abuse trust. So I think we need to take that kind of a stance vis-a-vis firms who always want to evade rules that will bind on them and who may also have a strong interest in influencing who wins an election or influencing the decisions that the elected officials make.

Luigi Zingales:
Paul, whether we trust or not people with the conflict of interest in economics, I can start a new debate on that, but I will postpone. I have very strong opinion on that, but I don't want to delay the conversation. I want to go back to-

Paul Romer:
A little bit, just give me one sentence. So you don't think just like disclosure and non-conflict is enough?

Luigi Zingales:
First of all, in economics, disclosure is not very well done-

Paul Romer:
I agree.

Luigi Zingales:
... number one, and then two, one of the problem is the future incentives. So if I know that by advocating certain position in the future I will have better jobs and better pay, there is nothing to disclose now, it's just the system. Anyway, as I said this will bring us-

Paul Romer:
But this is actually really important point in a bunch of context, so you'll have to have me back on. I refer to this as the good reads problem, because you see this with firms where they behave very well. They build their community. They're the good actors, open source, blah, blah, blah. And then at some point in the future, they cash out and a good reads gets sold to Amazon. Is really hard to commit in advance that you're not going to do that in the future. So I admit that's a real challenge for our system where we're trying to build trust.

Luigi Zingales:
Absolutely. And I have a couple of paper to sell you on the topic, but that's a different story. So let's go to some of the other questions. And then Anurag Pandey. "First of all, thank you for your great talk," and then say, "Can you please provide any example of an effective antitrust law or even a subset of it from around the globe, which is working and doing its job well, especially for large platform companies?"

Paul Romer:
Well, I think, what we need to think about here is a dynamic where some parameters have changed because of Moore's law. Suddenly, monopoly is the de facto. It was the expected outcome in these markets. And then what's happening is this regulators and legislators and judges are all trying to scramble to adjust. And frankly, I think all across the world, they're playing catch up. They've let this one go much too far in the direction of firms that have too much power.

So, frankly, I'm not that optimistic about even finding cases where it's working. The one part of the law that it seems like we could tighten up would be merger review. We could set much higher standards where the presumption is very strong then a merger should not be allowed. But I think merger review on its own is not going to be enough to prevent the kind of concentration that we see. And frankly, if the learning curve, winner take all dynamic is strong enough, even a structural remedy, you go in, you break a company up into three pieces, it's eventually going to go right back to one dominant firm. So we need to change the underlying dynamics. We can't just think about these usual mechanisms, like a breakup or limiting the mergers.

Luigi Zingales:
Speaking about merger review, there is a question here from Andrew Iman that say, "What can be done from a policy perspective about transaction that were approved by regulators, but in retrospect seem poorly consider, for example, a number of Facebook acquisitions, and how can we address those problems without undermining business confidence that transactions can be finalized?"

Paul Romer:
Yeah. Well, I think this is an interesting question, but let's put it in the context of say like regulation by the EPA of a new chemical. I think the EPA could say to affirm something like, this looks okay, but we're worried that it's going to be bad for health or the environment and so here are your choices. We'll just not approve it and we'll wait and see or we could approve it conditionally, but with an explicit option held by the regulators to come back and revisit. But if you understand, you start making investments based on this conditional approval, you're at risk of losing the value of those investments if it turns out that the evidence goes against you. So we can all wait, or you can go ahead, but you're the one who's putting your capital at risk.

I think it would be good to give regulators more flexibility here, because there are many cases where information emerges only over time. And it wouldn't be efficient to stick to a rule that once something's approved by the regulators, it's approved forever because ex-post is too likely that bad information will come out. And so if regulators were forced to make permanent decisions, they would inevitably be much more conservative about letting anything get approved.

So I think we should have these kinds of mechanisms for conditional approvals and look backs. But the problem is, is that we didn't have those. And so a firm that is faced with a demand, okay, we're going to revisit the merger, I think might have a strong, legal case to say the law doesn't allow that.

Luigi Zingales:
In addition, in some cases, at least in front of the European Union, they lie say they will not merge certain data sets and they did, so there is an easy way to get in those cases. But here I have a question from Yishay Yafeh who actually is following us from Israel. One of the few advantages of the pandemic is that we can really be global in our seminars. And he's asking, "What is the difference between Professor Romer's new taxation scheme and revise or expanded antitrust? If you know when a company is large enough and should be taxed, why can't you impose restriction on its activity through antitrust?"

Paul Romer:
What I like about kind of a progressive tax system is that it creates steadily increase in pressure in the direction of diminishing returns. So at some point it's going to make sense for this firm to just not keep getting bigger or to spin things out. But it doesn't force a bright red line that when some number crosses this specific threshold, boom, some big things happen. So the progressive tax gives you a kind of a continuous steadily increasing pressure, whereas antitrust inevitably gets stuck with these decisions like this firm is so big, that price discrimination is bad, and it's going to get whacked. This other firm, it's a little bit smaller. We don't know.

So I think tax systems and legislative decisions are better able to make these kinds of gradual adjustments, almost like automatic stabilizers in macroeconomics where you can kind of put these in place and they're not going to get it exactly right, but they'll within some broad range keep you from going to the extremes.

Luigi Zingales:
To your point that sometimes it's easier to enforce graduated punishment. In Italy, nobody was receiving tickets for excessive speed because they were too high. And the policeman felt like bad to give you like this huge liability. Then they introduced the points that every time you do an excess you, you have a point deducted from your driving license. And when you accumulate a certain number of points, you lose your license. And then the system worked perfectly because everybody got the little point and by the end, it's not my fault. If you lose the license, you have accumulated too many. So in that sense this is very much key.

Paul Romer:
But this is a really important a point about our judicial systems, which my former colleague, Mark Kleiman, who's now no longer with us, but was very insistent on, judicial systems will not enforce punishments that are too harsh. And I think for better, for worse many judges especially at the appeals court level, think that structural remedies are too harsh. Now, if you think about this from the point of view of shareholders, it's not true. It's not clear that that's always the case. But the firms, the CEO, the employees can always go in and say, oh, look how disruptive this is going to be. And we just spent the last year mangling everything together to make it really hard, so this is really is disruptive. So you can't go ahead and do it now. So I think, structural remedies through courts may tend always to be in that category where the judges think they're too harsh. And if so, if we want more firms, we got to find some other mechanism to get there.

Luigi Zingales:
So in the spirit of our international reach we have a Richard Tapper that thanks you for the talk and ask a question about smaller companies say, "What opinion do you add on the role of the competition regimes of smaller countries? Can competition authority do something meaningful there?"

Paul Romer:
Well, I think it's tough for a competition authority to do anything, but I think the legislature could tax like a business model that people think is bad for citizens. So especially this, I mean, it's not just that these tech firms are big, but they're using this business model based on advertising, which has these horrible consequences because the firms want to maximize engagement. So I think any small country in the world would be wise to take very seriously the possibility of putting a tax on all digital advertising revenues that are collected by these firms that are mostly outside their borders and undermining the functioning of markets.

And just by the way, when I grew up, when I was a boy, the way we thought about a market was that like, I give something up to get something from a firm and then if I don't like it, I can go to another firm. Like no part of that is working in our market for digital services. I don't know what I'm giving up. I have no ability to go operate outside of the advertising surveillance system. And you've got a whole generation of people who think the way the market works is that they should "Get things for free in exchange for giving away the right to be manipulated." This is not the way I thought markets were supposed to work. And I can't imagine any economist, if you'd ask the question 30 years ago, how do you feel about this kind of market? I don't think anybody would have been very optimistic about how well this was going to work.

Luigi Zingales:
But speaking of small countries, in your initial talk, you mentioned a point that I think is extremely important. And I would like to come back on, is the role that these multinationals add in the democracy of smaller countries. So when I grew up speaking of our youth memory, when I grew up in Italy in the '70s, there was always a fear of the American influence and how this will shape our democracy, which of course was relevant. Remember this was in 1973, the United States so they determine the fate of Chile.

But today it seems that the problem is even worse because you don't need even need the CIA, I think Facebook and Google do the trick. And so the fear is that in every country what you're going to have is you're going to have a government that is not appointed by the US government is appointed by Google and Facebook.

Paul Romer:
Yeah. Mark Zuckerberg gave a talk that I heard at the Aspen Institute in the summer of 2019. I think this is available online. In that talk, he was very clear about the role that he saw for Facebook as the company that was going to effectively administer elections all around the world. He didn't mean to administer in the sense of like collecting the votes and counting the votes, but everything else about the discourse leading up to the election, Facebook was just going to take over and run because it was better at it than any government could be.

And this was the point I was trying to ask the political leadership, it was actually in Chile. I was asking leaders in Chile. How do you feel about the fact that Facebook has this plan to control the whole system for your elections? And the party in power was actually fine with it in the near term because they thought they were better at using Facebook, it kind of gaming the system. But I think it poses a very severe threat around the world. And we can't tolerate this argument from Zuckerberg, nobody is as good as we are at identifying bad actors. So you basically have to give over this whole important part of police powers to a private firm that I, Mark Zuckerberg personally control.

Luigi Zingales:
Yeah. Yesterday I was listening a very interesting talk co-Organized with Stanford, with the CAHSI Center, Stanford, and was done by an Indian expert in digital platforms. And he mentioned the fact that in India, the world colonization started to surface when it comes to this issue. And as you can imagine, that's extremely sensitive world. And so I think that the risk that we have in anti-colonization around the world movement is something that I see very very likely.

Paul Romer:
So can I ask myself a question? I want to get it out there.

Luigi Zingales:
Please

Paul Romer:
Okay. Professor Romer, what else would you do besides your dumb tax idea which I don't like? Here's some other things, I think one transparency could be very helpful here. So I don't see why anybody can argue legitimately that we're doing advertising because of the information it conveys, but firms have to have the ability to keep secret what ads they're displaying. So especially in this realm of political and edging into political ads, I think countries should insist every single ad has to be in the public record, along with information about how many and what types of people it was shown to. There's no logical principal does justification for secrecy in this domain. The US should do this, but other countries should definitely do this as well.

I think another thing that's worth thinking about is certain kinds of just disclosure, forced disclosure of information. When you drill a well, and then you frack it, you have to make public information about what mixture of water and sand you used, and what kind of returns you got. This information actually means that firms learn from other firms in oil fields. So public disclosure can actually be beneficial for everybody. And there may be bodies of data like Uber's data and Google's data, as it grows about car trips where the government may say, part of the deal here is that you have to disclose that data, you can't keep it proprietary.

So transparency, disclosure, and then public options, I think some governments could run some backstop services, like a basic search service that any firm could register with that service. And then people could find it through that service without any ads and any payments that are gaming the system. The government won't even have to have any robots crawling the system, it would just be a place people know they could go. Like if I want to find the Marriott at the airport in Houston, I can't find it when I search with the existing search engines because I get a zillion ads and I can't even find the website for Marriott let alone the Marriott in the airport or near the airport in Houston. So I think public options are well worth considering. And if it takes revenue to support them well, collect that from attacks on the private ones.

Luigi Zingales:
Paul, thank you very much for asking also your own question. And in this way, it's clear that I didn't ask the question to actually promote some of the Stigler Center has done. Your idea of transparency is one of the idea that the Stigler report on digital platform was pushing forward because we do agree is an extremely important step in the right direction. And unfortunately with this, we're running out of time. So my first thing is to thank you immensely for a really insightful and provocative talk. I think that a lot of people will keep discussing this in the month to come. And to this point, I want to remind everybody that these talk, as all the previous talk will be available on the Stigler YouTube channel, so you can educate yourself throughout this process. Paul, thank you very much and thank all the participants.

Paul Romer:
That's good, but let's make a date. Let's come back and talk about this issue about federalism and like the different levels that which government can operate, because I think this is a really important one and it's frankly, pretty complicated. So let's come back and explore what kind of levels and what kind of actions we want to contemplate.

Luigi Zingales:
Paul anytime. Thank you very much. Bye-bye.