Capitalisn’t: Why we should tax digital advertising

Mar 11, 2021

The political power big tech firms wield, and the limited competition they face, have prompted warnings from politicians, regulators, economists, and others. Many have argued we should turn to antitrust laws as a way to address these concerns, but Nobel Laureate Paul Romer says they may not be enough.

On this episode of the Capitalisn’t podcast, hosts Luigi Zingales and Bethany McLean speak with Romer about how the implementation of a digital advertising tax could address the size and business models of these tech firms.

Paul RomerWe shouldn’t underestimate the ability of these firms to capture the apparatus of government and use it for its own benefit. 

Bethany: I’m Bethany McLean

Phil Donahue: Did you ever have a moment of doubt about capitalism and whether greed’s a good idea?

Luigi: And I’m Luigi Zingales

Bernie Sanders: We have socialism for the very rich, rugged individualism for the poor.

Bethany: And this is Capitalisn’t, a podcast about what is working in capitalism. 

Milton Friedman: First of all, tell me, is there some society you know that doesn’t run on greed?

Luigi: And, most importantly, what isn’t.

Warren Buffett: We ought to do better by the people that get left behind. I don’t think we should kill the capitalist system in the process. 

Luigi: The Stigler Center at the University of Chicago, which produces this podcast, last year had organized a conference with the ambitious title of “Monopolies and Politics.” At the end, we reserved a special place for people who have some proposals on how to move forward. Among those, a particular role is played by Paul Romer. So, we decided to invite to this podcast Paul Romer, who is a Nobel Prize-winner for his ideas about endogenous growth. Today, we’re going to focus on one particular aspect of his vast production, which is an idea he has about how to tax digital advertising and why this tax could be a way to deal with the problem generated by digital monopolies. So, it’s with great pleasure that we welcome to our podcast Paul Romer

Bethany: I read your essay in the New York Times on the tragedy of the commons. And I thought, for the noneconomists who are listening to this, it might be interesting to start with the concept of the tragedy of the commons and why it applies to big tech. 

Paul RomerWell, the commons is this evocative phrase and this analogy that helps us think about things we share, and where one of us could damage it. And that hurts all of our fellow citizens. And I thought that digital advertising was like a form of pollution. It was causing significant harm in the economy. The kind of harms included the incentives to collect all of this invasive information. This process, it’s been called surveillance capitalism, but “commons” here I think is partly trust that when people are afraid they’re being tracked, they lose the ability to trust firms. This became a real holdup in terms of our ability to do something like digital contact tracing during the pandemic. So, trust is something that we share but can be eroded. Also, some sense of, like, comity, goodwill towards those with whom we disagree.

One of the problems with this model of digital advertising is, is that the incentive for a firm like Facebook is to maximize engagement. They don’t want you to feel happy about using Facebook. They want you to spend a lot of time on it. And what they discovered is by getting you angry and in some acrimonious exchange back and forth with somebody else, this is the best way to keep you online for a longer period of time, thereby giving them the chance to expose you to more advertising. My proposed solution was this tax on advertising. We say, well, if you don’t want people to do something, tax it. And tax it, not because you’re looking for the revenue, because you want them to stop doing that bad thing. 

Bethany: Explain what the tax solution is.

Paul RomerThe idea is just to put a tax on the revenue earned from this digital advertising model, on the theory that it’s a dangerous business model. It creates bad incentives. And we want firms to shift to some other business model, like subscription fees in exchange for digital services. And the reason it makes sense to make it a tax on revenue is that you know the geographic location where revenue is earned. Income is the difference between revenue and cost. Cost is earned in other places. So, we don’t have, even in principle, any way to define where income geographically is earned. That means that firms can shift income to jurisdictions that have lower taxes. And so that means that a corporate income tax is always going to be subject to this kind of race to the bottom between different jurisdictions. But now, if a particular state like Maryland, which has decided to try this, passes a law that says we’re going to put a tax on revenue that a firm earns from showing ads to people in Maryland, Delaware can compete by saying we’re going to have zero tax, and that doesn’t cause any race to the bottom. 

Bethany: I was thinking about the ability of firms to either gut their regulators or evade them in other ways. And I think that’s true more often than not. When you think of Fannie and Freddie and the runup to the financial crisis with their regulator, or the big banks with the New York Fed, it does strike me that you need the government not to be the sleeping beast or asleep at the switch. You need the fear of the government. You want the government to be trustworthy and not to abuse its power as well, but you need firms to believe that the government can see and can act, or all of these things just evolve into all of these wrinkles and ways to evade and play games, right? 

Paul RomerI think on tax enforcement, we’ve seen some of this same process that the IRS’s ability to audit has been undercut by measures in Congress that limited funding, that may even have specifically restricted the use of funds for audits. So, the fraction of returns to get audited these days is way down. And there are these really striking numbers about a tenfold increase in revenue you’d get for each additional dollar spent on auditing. 

There is a risk that the audit function could be compromised, but in this case, I think it’s a little bit less likely that this could be used to, say, favor some firms or individuals relative to others. When Boeing was the only firm being audited in terms of regulation, gutting the auditing ability was basically all it took to free Boeing. But I think on something like taxation, well, the existing taxes apply very broadly. Undermining the audit function doesn’t have effects that are so concentrated. 

Now, having said that, I’m realizing that auditing advertising revenues would basically be something which is done really only to the top three or four firms in the US. So, we might be a little bit closer to something that looks like Boeing. And I guess the people in Maryland and other states that try to do this should think about a kind of a subsequent attack, which involves undermining the audit function, which is necessary for the system to work. 

Luigi: One of the reasons why I like your idea of the tax on advertising is that I have a more cynical view of advertising than most economists. Most economists think that advertising is just information and is great, et cetera, but there was actually an economist at Chicago in the ’30s, Henry Simons, who was a big supporter of antitrust enforcement and this idea of a tax on advertising, because he thought that advertising was polluting, that it was an arms race between companies and is a waste of money. And, as with everything that’s polluting, you tax it. So, what we don’t fully appreciate is while there are some clear benefits of targeted advertising—I discover things I would not discover otherwise—I think there is a big component of an arms race in advertising. We all pay indirectly that cost through our goods. And so, a tax on advertising—even before digital advertising it made some sense, but now, with digital advertising, as you said, it makes a lot more sense. 

Paul RomerYeah. For example, Facebook has resisted measures which were designed to force them to disclose more about what ads they’re displaying and to whom they’re displaying them. Under the usual theory that ads just provide information, nobody should resist disclosure. They should welcome it, because it gives more attention to the information they’re trying to disclose. So, something else is going on here instead of just pure provision of information, especially in political advertising. 

Luigi: But the tax will reduce this, going in the right direction, but it will not eliminate the business model. The business model will remain in this way.

Paul RomerYep. And one of the things that the tech industry did well was a lot of trial and error. And I think in policy, we need to be willing to do that as well. So, I think this tax is something we should explore. We can’t anticipate with perfect certainty what kind of responses that’s going to trigger by industry and by users, but I think we should be exploring this dimension, see what happens and keep pursuing it if it turns out well. 

One of the things, if I go back to like the ’90s, I had written this paper saying, new technology is going to be great, it’s going to save us all, it’s going to give us progress. There are really very few things that came out of the internet revolution that I think have had a net social benefit. The telling one is Wikipedia. But the critical decision that Jimmy Wales made was that he would never allow advertising as part of the business model for Wikipedia. What they’re relying on these days is donations, which surprisingly works OK. Another one that I’ve encountered more recently is Duolingo, but their business model is a subscription. So, we don’t have to have this targeted advertising model as the only feasible way to make viable, valuable digital services. 

Bethany: That’s a really big statement that not much that’s good has come out of this internet revolution. Are there things we could have done, with the benefit of 20/20 hindsight—which is always an enormous, unachievable thing in the moment—but could we have done things differently to have made this thing around us grow up in a way that would be healthier for society? 

Paul RomerYeah, I think, let’s consider two different paths. One would be much more aggressive antitrust. The other would have been to have banned digital advertising. It would have been hard to sell that ban to people, but if we had done that, then what we’d see is more services like Wikipedia, more services like Duolingo. I think that would have also preempted this big source of increasing returns, which is driving . . . at least one of the big sources of increasing returns that’s driving all of the growth and the number of dominant firms, because it’s that centralized pool of information about people that becomes the asset that the dominant firm has that it uses to target the ads at the individual. 

In my talk for the Stigler Center, I alluded to this idea of a phase change, that if some parameter changes because of Moore’s law, all kinds of new things become possible with code and with sensors that let you collect data. You can tip into this type of competition that we refer to as winner-take-all. And so, there’s this ferocious competition to be the one firm that succeeds at getting to scale and the only one that will survive. I think the intensity of that competition and the fear that I think is actually real, in an organization like Facebook, that somebody could come in and we could end up being the new Myspace, we just get destroyed as a company. That kind of fear makes them incredibly aggressive competitors and very aggressive about preempting competition.

Now, the question would be, what if regulators were more actively involved? If these companies still feel like they’re in a life-and-death battle for survival, it may be hard for the regulators to contain that, but it might be that the presence of the regulation reassures firm A that they don’t have to be as devious and vicious to protect themselves from firm B, which is being devious and vicious in its competition. If they know that there’s some limits on what firm B can do, then that might reduce the sense of threat that firm A experiences. 

If you listen to what I was just saying, it sounds a little bit like we’re getting to a kind of managed competition where the government is trying to aim for a certain number of competitors. It’s a much more active form of regulation than I think we’ve contemplated in antitrust in the past, but that might have some value if the alternative is to be stuck in this world of every firm feels like it’s in a life-and-death struggle to become the one firm that will survive. 

Luigi: There is, as you well know, a bigger part of the economic literature and a part associated with University of Chicago that says, look, we don’t care about standard concentration measures, because what matters is the threat of entry. And, in fact, I recently read an interesting book by a guy called Petit, who actually looked at the annual reports and documents that the threat of competition is very much present and alive in the large companies and suggests we are fine. We shouldn’t worry about the problem. And, when we distinguish between good competition and criminal competition, we should actually be worried, because if the competition is of the type of winner-take-all, then I might resort to criminal ways to block it in a way that I wouldn’t do for the other. And so, we actually have the worst of both worlds, because we have criminal actions and monopolies.

Paul RomerYep. I made that link between—“contestability” was the phrase that was used around the breakup of AT&T—to suggest that, well, if the market is contestable, then consumers will be fine. I made that connection that contestability is not working in the tech industry, but I think the missing piece is this one you flagged, and we need some comparable term for competitive measures that are efficiency-destroying or malevolent or predatory or something. I’m not quite sure what the right word would be there, but if you respond to contestability by undertaking this, not competition of a better product or a lower price, but competition in these kinds of dark measures, then, as you say, we really get the worst of all worlds. 

Bethany: I’d argue, unfortunately, it’s not as simple, though, as legal versus illegal or criminal versus noncriminal, because the worst kind of behavior comes about by the gaming of the existing rules in a way that actually does meet the technical requirements of the law. So, I think that brings us back to regulation or to the fear of some degree of government scrutiny as the answer, because deferring to the legal system never works in the end. 

Paul RomerLet me take this back to my thinking about the ad tax, because I’m worried about solutions that involve effective regulation, especially effective regulation that requires very highly paid technical people to implement it. In our system of civil service, we can’t pay salaries that are comparable to what people in industry pay. That’s a bad constraint, one we ought to get rid of, but it’s one we live with right now. I’m worried that other firms will figure out how to do what Boeing did with the FAA, which was basically to get the Congress to cut the FAA’s budget, so that the FAA didn’t have the technical expertise to do the kind of safety monitoring that was traditional in the aviation industry. So that more and more of the monitoring of the regulations actually was delegated to people inside Boeing. Boeing even got Congress to mandate some defaults, where things by default were supposed to be monitored by Boeing. So, Boeing had this situation where it had an incentive to cheat, and it was also people inside Boeing who were supposedly the ones who are going to catch the cheating. 

And I worry that even if we could set up a good internet regulatory agency with the right kind of technical expertise, I still worry that the industry will be able to either gut it or they might capture it. We should look at how Facebook was trying to respond to competition from TikTok. They realized they couldn’t buy TikTok. So, what they did was lobby to try to get people on the Hill and then in the executive branch worked up about how TikTok was a security threat and got the government to basically limit competition from TikTok that Facebook was facing. So, I think we shouldn’t underestimate the ability of these firms to capture the apparatus of government and use it for its own benefit.

Luigi: Why don’t we go back to antitrust, because we had David Boies on our program. I’m not a lawyer, but I understood from him that if you can prove that an action is done only to hurt a competitor, you are doing an antitrust violation. So, I think many of these actions that the companies do to basically entrench themselves have these characteristics. And if we had a sufficiently aggressive antitrust, they could prove it. Now, there is a problem that we probably need more technology experts and economists in the antitrust to make it work, but I think it’s a possibility. 

Paul RomerYeah, I wouldn’t be opposed to pursuing that path. I think this is one where we should be trying to pursue many paths in parallel and see how they play out. But I think it’s interesting to think about the challenge that the judicial system faces. If you go back to the founding of the United States, one of the most serious fears that people had was that the discretion that’s given to a prosecutor. Who do you try to prosecute, and who do you not prosecute? That discretion could be used to completely undermine political competition. In antitrust, we saw two cases where people alleged that the Trump administration was pursuing antitrust cases purely because of political reasons, not because of the purported reason about trying to protect competition. Whether or not that was true, this is now at least a live issue.

The way our system has always dealt with the risk of prosecutorial discretion is to make sure that it’s pretty hard to win a case. If the prosecutor decides to bring a case to court, judges and the process don’t make it easy to win one of those cases. I think if we use the judicial system, I’m afraid we’re almost always going to be stuck in this situation where either it’s very hard to win a conviction, or if it’s not hard to win a conviction, then we’re in a situation where the government could prosecute almost any firm. And then, if it just prosecutes firms that, for example, own a news outlet which is criticizing the administration, then the administration can basically intimidate firms into changing how they cover the administration. So, even though I’m willing to explore some of these slightly more aggressive methods of antitrust enforcement, the whole system based on conviction for violation of the law is just too crude a tool for these kinds of subtle issues that we’re facing. 

And we’ve always been worried that if we strengthen the law, what that really does is give unchecked power to the executive branch. And what I like about the tax solution is that it rather works through the legislature instead of regulators in the executive branch, or prosecutors in the executive branch, along with judges in the judicial system.

Bethany: To me, Paul Romer’s argument was pretty convincing, because I worry a lot about putting a regulator in charge of the kind of content we want to see. I wasn’t even 100 percent happy about Twitter’s decision to ban Donald Trump. I understand it, I can argue it both ways, but it is scary when a private company has that much power. But I think it’s even more frightening if a government regulator has that much power. So that, to me, isn’t a great solution, either. So, I like this idea of a tax, do you?

Luigi: Yes. We economists call this type of taxes Pigouvian taxes, after the famous British economist of the middle of the, maybe, early part of the 20th century, A. Cecil Pigou, who had this idea that whenever there is an externality, you tax that particular activity, because the tax can help you internalize that externality. The typical example is the example of a so-called carbon tax or CO2 tax. We don’t internalize as consumers the damage we do to the Earth by producing more CO2. The moment every activity that produces CO2 is taxed, we are going to start to internalize that cost. And so, we’re going to start eating beans instead of steaks, because beans produce less CO2 than steaks, especially because they’re going to cost much less. 

Now, what is interesting is that from an economic point of view, Pigouvian taxes are fantastic because they do the right trick. They intervene minimally to fix the problem without too much overhead cost. Politically, they are a nonstarter. Why? Because you are going to penalize a few people a lot, and you’re going to benefit a lot of people a little. And, as a result, political-economy calculus is exactly against those people getting taxed. Because if you start to tax steaks because they produce CO2, you will have the entire farming industry up in arms and screaming. Who is going to be on the other side defending this idea? In principle, all of us, but we know that dispersed interests are not exactly very vocal, while concentrated ones are.

Bethany: Yeah. That does get into a slightly different aspect of this conversation, though. There’s obviously a difference between what’s the best thing to do and what’s the feasible thing to do. And given what, you’re right, I still fear, despite all that we know about big tech, I just worry people don’t care because it appears that you’re getting something for free, and it doesn’t appear that there’s a cost to it. 

It’s funny, I was thinking about Paul’s arguments in light of, we’ve just seen more news about Boeing, and the whole idea that you can create a regulator that would rein in big tech, I think, is just undermined by the—that seems to be my favorite word of this podcast today, undermined— undermined by the history of regulation in the United States. And the Boeing story with the FAA shows how a regulator that is in charge of regulating just one or two very powerful big companies almost never succeeds. Is there an economic term for that? A regulator tasked with regulating only one or two very big and powerful entities? 

Luigi: Actually, it broadly is called regulatory capture. And actually, since we are at it, it was invented by Stigler, the namesake of the center, exactly 50 years ago almost to the day. Stigler did not make it specific to monopolies, but you’re absolutely right, that this is particularly so when you are in monopolies, because competitors also compete in capturing the regulator. And, by competing, they give some margin of maneuver to the regulator, at least in order to try to maintain their independence. But if you have one client, you don’t have a client, you have a boss. 

Bethany: The subject that I know the best, obviously, is Fannie and Freddie and the tragic history of their regulator, their former regulator, which was called OFHEO, the Office of Federal Housing Enterprise Oversight. And OFHEO’s sole task was to regulate Fannie and Freddie and make sure they were safe and sound. As we all found out in the 2008 financial crisis, that didn’t work out so well. And part of the reason is that Fannie and Freddie were just far more—Fannie Mae in particular—far more politically powerful than poor OFHEO. And so, OFHEO was outmanned, under-gunned, underfunded, everything. And you can see that same thing happening to any regulator put in charge of big tech, right? I mean, what a terrible job that would be. 

Luigi: Yes, absolutely. I think that a partial solution is to start considering paying the regulators a little bit better. There’s such a gigantic gap between the pay of the regulator and the pay of the people outside that it is very hard to retain talented people in a regulatory agency. And I love the cleanness of the Pigouvian approach, because it dispenses with all these things in between. But, to his credit, I think that we can go even farther and go to the Matt Stoller solution, which is to ban the business model. At issue here is that the real mechanism of financing everything through advertising is rotten from the very source. And maybe we should ban the business model. I don’t think that this has any chance of passing anytime soon, but I think it’s useful sometimes to recognize where the problem is. And so, taxation is probably the most feasible solution in that direction. 

What I like from a political economy point of view is that, as Paul says very clearly, you can do it individually. It’s very hard for the Congress of the United States to pass this law. But if you are a state where Google or Facebook are not very powerful, that’s a pretty appealing thing to do, because you get funding. Now, if you go down that route, I suspect that pretty soon Google and Facebook will sponsor every state election on the face of the earth.

Bethany: It is actually fascinating, though, because the national media tends to concentrate on national lobbying by corporations, but Facebook and Google, the thing they’ve undermined the most, it hasn’t been the national media, it’s been the local media. And so, Facebook and Google have actually created a perfect situation for themselves in the sense that they can become incredibly powerful, and they probably already are at the state level. And the organizations, namely state newspapers that once would have paid attention to that and kept tabs on them, have been destroyed by Facebook and Google. So, Facebook and Google now have carte blanche to do what they want at the state level without anybody paying attention. 

A different question for you that I was thinking about as a possibly different alternative, what if Google had to offer—and this wouldn’t work for Facebook necessarily—but what if Google had to offer a paid option where they absolutely didn’t track you or pay any attention? How much would you pay per month to be able to use all of Google’s services without them having any access to your information? And how many people do you think would opt in to a paid model? 

Luigi: That’s a very good question. I think there are some estimates. I suspect not a lot, but what would be interesting is that this would be highly selective. Probably the people who are willing to pay are the ones who are better off, who end up being, very often, the most valuable for targeted advertising. But I was reading the other day how much margin Facebook and Google make on every customer, and it is not that huge. It is on the order of $20 or $30 per customer. So, if you were to pay $30 a year, I would be very happy to not be tracked. 

Bethany: I’m wondering if the tax, as much as I’ve said I’m a fan of it, I’m wondering if the tax could actually backfire, in that, as it becomes harder for companies to make money, they have all the more incentive to behave badly. I mean, is it possible that this tax could incentivize the wrong kind of behavior instead of the right kind of behavior?

Luigi: That’s a very good point. And, two things, there’s that risk. There is also the risk that you are going to reduce their profits but not change the business model. In general, we think that by taxing something in economics, you’re going to reduce the consumption, and you’re going to see less of it and you’re going to see other behavior. However, I don’t think that unless the tax is really, really very large—which I find difficult to see that it could be approved—it’s not going to change the business model of a digital platform. So, you’re going to see more money in the coffers of Maryland, and this might be good or bad depending on your point of view, but not really a change in the business model, which is what we’re trying to achieve. 

Bethany: Yeah. It worries me that perhaps there’s no putting the proverbial genie back in the bottle. So, I’m getting more cynical, more skeptical, as we talk. 

Luigi: I actually like the Fukuyama proposal of middleware, this idea of trying to separate the element that is noncompetitive, because it has gigantic network externalities, from the part that is competitive, allowing competition at that level. And having a minimum level of regulated utility on the noncompetitive level, I think is very appealing, because much of the damage is created by the superficial level, the middleware, and we are not allowed to choose among the middleware because of the network externalities.

I was very optimistic when there was that revolt against WhatsApp, that even my friends in Italy started to open Signal as an alternative to WhatsApp. And I actually used that as an experiment to see whether there was enough force to move to a different equilibrium. You saw that there were some that operated on both platforms at the same time. And then the Signal community started to vanish. And now we’re back 100 percent into WhatsApp. It was a complete failure to move, which makes the point of how entrenched these network externalities are. So, given that, I have very little hope that there is real competition. 

So, a Glass–Steagall separation between the pipes, if you want—and these are, I understand, virtual pipes, they’re not physical pipes, but still the virtual pipes that give you the network—and the editing side that does the choosing, that will lead to competition on the editing side. So, yes, there will be nasty social media out there, but you and I are going to choose the kinder and gentler version of social media.

Bethany: I think you might be giving me too much credit, Luigi, but I’ll wave to you over in your nice part of the world from my nasty, evil, malignant part of the world. I think that does, though, on a more serious note, bring us back to the question of feasibility. I mean, Glass–Steagall only was passed in the wake of the Great Crash, and there was enormous political will toward getting something done. And I don’t know what the equivalent . . . If what we’ve seen with big tech so far, with the dangers of big tech so far, still aren’t enough to create that pressure for legislation. I don’t know what will. What more evil do they have to show they’re capable of committing to get Congress to be willing to do something radical? But I think we are a long way from a Congress that would do something radical to big tech. 

Luigi: I’m a little bit more optimistic if we find a Pecora for the situation. Behind Glass–Steagall, there was Ferdinand Pecora, an Italian-American, who actually went after the important businesspeople at the time, embarrassing them and exposing the tricks and the fraud, et cetera, and creating the political landscape for, basically, the New Deal. I always thought that what was missing during the Obama administration was a Pecora to do that job after the financial crisis. And I think we still suffer the consequences of that. And maybe the time has come for a Pecora in the digital sector, I think that maybe the name is Cicilline or somebody else, but I—

Bethany: Definitely Italian.

Luigi: We need an aggressive prosecutor to expose and an aggressive prosecutor who knows how to play the public-relations game, not just the court game. And I think that that might actually change the opinions of a lot of people, but it takes a Pecora to make a difference.

Bethany: Yep. OK. Well, I guess we can agree on that, for sure, which is that we need the Pecora of big tech. So, if you’re out there and you’re listening to this, please volunteer your services. 

Luigi: I want to bring up another point that Paul Romer raised, which I thought was very interesting, because we’re discussing competition, and he is one of the few people in economics who emphasizes not only the positive aspects of competition, but also the negative aspects of competition. In particular, the fact that when rivalry leads to people trying to boycott or sandbag each other in order to be ahead, and these incentives are the biggest, the more it’s a winner-take-all game. If I know that arriving second will still give me a good price, I might be less tempted to cap your knee in the race than if I know that if I arrive second, I die. 

I was actually looking at the etymology of the world competition. And here I’m bragging a bit about my Latin, not particularly well used, but competition comes from Latin and is the form of two words, cum and petereCum means together and petere means seek. So, it is “seek together,” which, if you think about it, the race for the vaccine, you have different players, they’re trying to get the same stuff, but they weren’t sandbagging each other. In some cases, they were also helping each other, but because they had a common element. This led to the current term competition that is also used for races. When you have a car race or track and field race, et cetera, this is called competition. But that’s more of an, if I win, you lose, kind of situation, where there’s more rivalry and where the temptation to sandbag you is stronger. 

And so, abusing the Latin term, I said that we should call that contrapetition, that you seek against the other person rather than with the other person. And I think that the words have meaning and allow you to think in a different way. Once you separate the idea of competition from contrapetition, you understand that some of the competition we see or that people claim exists in the industry is not really good competition but is contrapetition. So, in the digital platforms, we have a competitive threat, but that competitive threat leads people to bad behavior, not to try to seek a common good together. So, I think that that’s a very important insight that is not well known even among economists, let alone outside of them.

Bethany: So, contrapetition is a fancy way of saying race to the bottom, right? 

Luigi: It is actually worse, it is really kneecapping somebody else. Who was the famous ice skater who hired the boyfriend to kneecap the competitor? 

Bethany: I can’t believe I’m blanking. 

Luigi: Tonya Harding.

Bethany: Of course. It’s Tonya Harding. Yes.

Luigi: That’s the contrapetition—you literally kneecap your rival. That’s really bad for everybody, not just for the person receiving it, but for the entire system, if you go down that path. And I think we have more than our fair share of that. And, to avoid that, I think that, number one, you need to have regulation to prevent this, but also, it helps to have less of a winner-take-all market, because, exactly like ice skating, there is a gigantic difference between being first and being second. And so, people are more willing to do crazy things to get the extra price. 

Bethany: I still think we need a term that’s somewhere in the middle, though, or something different than contrapetition that isn’t kneecapping your opponent per se, but rather is what exists when the incentives are huge, and the way you capture those incentives is by behaving the worst. That was true in the subprime mortgage market, for instance, where the person who offered the worst product for consumers might be most likely to win the business, because they were making it possible for people who couldn’t qualify for a mortgage to get one. And that’s a little different, I think, than what you’re getting at with contrapetition. 

Luigi: Oh, absolutely. That is a case that is better understood in economics—I’ll use the technical term and then I’ll explain—where the competitive equilibrium is not efficient, it’s not what we say Pareto optimum. For example, in the case of the externalities we’re discussing, if I don’t factor in my CO2 costs. So, the more competitive the market is, the more I produce CO2, the worse the outcome. On the other hand, in a monopoly, I’m going to have fewer cars and produce less CO2. So, that’s a case in which a monopoly is actually good from an environmental point of view, because there’s less production of that CO2. 

Bethany: Right. That makes sense, back to Romer’s idea of the tragedy of the commons, right? 

Luigi: Yeah. The tragedy of the commons is a typical example of an externality, where the competition doesn’t lead to a good outcome.

Bethany: But wait, I think maybe this whole conversation might actually be beside the point, because you saw the announcement just a couple of days ago in the Wall Street Journal, which said that Google is going to stop selling ads based on our specific web browsing. 

Speaker 8: Google announced plans to phase out browser cookies. Those are small pieces of code that follow you around the internet and can be sold to advertisers for better targeting. Today, Google took that idea a step further and said that it would not use other invasive ways to track and target users as it phases out those third-party cookies. Instead, it will rely on anonymized and aggregated data using a program that it is developing internally called Privacy Sandbox. 

Bethany: So, we don’t even need to be concerned about this anymore, do we? Google is going to fix all our problems for us. This issue is going away. 

Luigi: Honestly, I’m still struggling to digest the news. The title seems to be very celebratory, the details much less so. If the only thing they do is stop third parties from putting in cookies to trace you, they’re basically increasing the monopoly they have on your information, not reducing it. You can argue that, at least in part, targeted ads based on geolocation are useful. If I want to find a pizzeria, I would like them to know where I am, because finding a pizzeria somewhere else on the planet is not particularly useful. So, there are definitely useful targeted ads. I would like to be able to choose when I receive them or not. And then, particularly, not be traced to know where I am in every moment of my life. But that’s a different story. 

Bethany: Yeah. I wasn’t sure what to make of the announcement, either. It struck me as one of those things where they might be getting more mileage from the headline than they deserve based on the actual substance of what they’re going to do. I didn’t quite understand the notion that while they’re not going to track you anymore, they are going to bundle your information into a cohort with those of other people like you, and then sell that to advertisers. So, that seemed to say to me, although I could be wrong, that Google is still going to have all your information. They may not be selling it individually, but they’re selling you as part of a group. I don’t know. Would you rather be sold individually or sold as part of a group? I’m not sure.