Capitalisn’t: Does Software Actually Slow Innovation?

James Bessen: The degree to which the dominant firms in an industry invest in intangibles, and in particular, proprietary information technology, the rate of growth of productive firms in that same industry is slower. Now, why does that happen?

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.

Bethany: One of the big questions in our economy and, therefore, our society is why productivity growth has slowed. In his new book, The New Goliaths: How Corporations Use Software to Dominate Industries, Kill Innovation, and Undermine Regulation, James Bessen, the executive director of the Technology & Policy Research Initiative at the Boston University School of Law, argues that the problem is really that big companies increasingly are able to dominate by mastering software, and software broadly defined. Software is the intersection between technology and data, and so, his argument is that productivity growth has slowed for that reason.

Not because fewer productive firms are being created, not because M&A is enabling big firms to buy small firms and take them off the market. The rate at which dominant firms have acquired other companies has actually declined since the late 1990s, but rather because of this mastery, essentially, of this intersection of software and data. And he argues—which makes his book, Luigi, right up my alley—that this mastery of software by major corporations helps explain rising economic concentration, increasing inequality, and slowing innovation. So, it has very big societal consequences.

Luigi: I want to add a factoid, because you explained the productivity slowdown very well. He has a number that I would like our listeners to keep in mind, because it’s quite stark. He said that the labor productivity growth—how much more each worker produces every year—went from 2.7 percent in the years from 2000 to 2007, to 1.4 percent per year ever since. And the difference is enormous, because if you grow your productivity at 2.7 percent a year, you double what you produce every 25 years. So, every generation sees a doubling in production. If you grow at 1.4 percent, it takes 50 years to double production. So, it takes two generations to achieve the same goal.

Bethany: There were a couple of other pieces of data that he had in his book that I also thought help clarify his argument. He points out that the total investment in proprietary software grew 74 percent to $239 billion over the decade that ended in 2019, and that the big companies are increasingly using this to manage complexity and gain competitive advantage. He had another number in here that, as we know and as we’ve had people on this podcast talking about, how investment in all types of intangibles has increased substantially.

But he pointed out that the stock of companies who have developed their own software increased eightfold for the top four firms, clearly outstripping the rest, and that the investments in this software by the top four firms even doubled relative to the mean investments made by second-tier firms ranked fifth through eighth. And so, this investment in what he calls own-developed software, this proprietary software, is dominated by large firms, and I thought that was another piece of data that helped point out what he’s getting at. So, we thought we would have James Bessen on our podcast to help explain his ideas.

Luigi: All organizations find it difficult to adapt. It is the history of evolution, to some extent. It is natural, and we regard it as healthy that young disruptors eat their lunch, because the fact that they eat their lunch means that higher productivity will trickle down throughout the economy. It will bring progress to all of us. But you notice that now, this relationship has changed in the last 20 or 30 years, and it has changed in a way that I find very different from the other stories I’ve heard.

I find it particularly intriguing that now, the superstar firms are defending themselves through a mode of complexity that makes it difficult for others to adopt, but maybe I’m butchering this idea. Why don’t you give your best explanation of why this is happening, because I think that this is a fundamental question, why productivity has slowed down so much, in spite of the fact that we all see a lot of technology coming about. What is causing that, in your view?

James Bessen: Right. My explanation for this is tied to the technology, and we’ve got some evidence that the degree to which the dominant firms in an industry invest in intangibles, and in particular, proprietary information technology, the rate of growth of productive firms in that same industry is slower.

Now, why does that happen? I think it goes along with an understanding of how these firms grew and the nature of the complexity of their technology. The problem we’re facing today is that we’re dealing with these very large systems that it appears aren’t diffusing. This very complex software, organizations that are built around that software, which are often very complex themselves, data—it seems to be very hard for firms to pose an effective alternative. You don’t see the kind of spinouts that you see in typical technology industries, where executives from Fairchild leave and start other semiconductor companies.

It’s just become very difficult for the technology to diffuse, and in that case, the property rights are not necessarily helpful, particularly as the pendulum has been going towards greater and greater restrictions on employee knowledge. We have noncompete laws, we have strengthening of trade-secret law. All of that limits the ability of employees to go and take what knowledge they’ve learned, often not patented knowledge, and use it elsewhere. It slows the diffusion process down.

When you get a situation where there’s such a gap between the leaders in the industry and their technology and the follow-on firms and their technology, that’s an indication that the balance that’s implicit in intellectual-property law is out of whack. It’s out of balance, and I’m suggesting we restore some of that balance.

Bethany: Is it possible to do that? What I took away from what you wrote is that this was always almost an implicit arrangement, that this diffusion took place, and the processes which enabled it were implicit, and now we have to take what was implicit and somehow make it explicit by putting in place new laws, new requirements. And given that the process by which it happened was diffuse and implicit, how do you make that explicit?

James Bessen: Part of the stuff I’m talking about is explicit, but you’re right. There’s a lot that’s very informal, and diffusion has never been something that’s been very clearly delineated: how it happens and what encourages it. One of the reasons it worked well in the past is that there are benefits that firms do spread their technology via license. When General Motors licensed its technology to the Ford Lincoln division, yes, GM would lose some sales to Lincoln, but the overall size of the market would grow so much that it was nevertheless beneficial to them, and they made money from their licenses.

We see something similar like that happening with Amazon’s AWS. They had a proprietary technology; they understood it was a source of competitive advantage. They had these huge websites that had to process an enormous number of transactions and had to do it very quickly. The slower it went, the less well it worked. So, they developed some tremendous capabilities to do that, and they realized this is a source of competitive advantage, and rather than keep it locked up, or entirely locked up, they opened it up, and that turned out to be a tremendous benefit.

Lots of small companies, as well as companies like Target or whatever, use their facilities, use their expertise. Amazon, of course, got paid. They’re still making a ton of money from it. They have some competition now, but it turned out to be a very profitable thing. So, I think, ultimately, one of the biggest factors is to what extent can firms be encouraged or, in some cases, compelled to unbundle, to open up.

And I think that may be the biggest policy challenge. There are opportunities out there where, if they open up like that, they can make tremendous profits, but I think it’s often very hard for managers to see that. It’s very risky. Government can play—and has, in the past, played—a role. I think one of the best examples is the IBM unbundling, where the threat of government antitrust activity nudged IBM to sell its software separately from its hardware and to allow other people’s software to run on their hardware. You’re basing it on a bet that a market will emerge that’s very large, and it’s often very hard to know. As it turned out, they did it. They were nudged by the government to just cross over that line and did it, and it turned out to be hugely profitable to IBM. They gave up part of their software business, but they gained a tremendous amount of business overall.

Bethany: Talk a little bit more about the role that complexity plays. I thought about your work as almost a Venn diagram between software development and complexity, and where we end up with a real problem is where software and complexity overlap, in that space.

James Bessen: Yeah. I think the key thing is that software enables firms to manage complexity. The economics of software is once you’ve written the code, you can reproduce copies almost costlessly. There are economies of scale. But there’s something else here, which is that modular software is extensible, meaning it can add features relatively costlessly. It can add features, it can add functions, it can add greater variety. So, when your automobile is based on software, it can have many more features and much greater control than in a traditional mechanical or mechanical-electrical system. Effectively, software lowers the cost of having a more-complex system. Complexity then has this added feature that it makes it difficult to copy or to imitate or to come up and challenge.

Luigi: Let me try a slightly different interpretation, and feel free to say that it’s completely wrong, because I just made it up, so it must be wrong. But the reason why we’ve seen the Walmarts of this world succeeding is also because the Sears of this world were inefficient, bureaucratic, not very adaptable. So, imagine that maybe with all that teaching in business school about improving the organization, et cetera, we have made the existing organizations more adaptable, more responsive, faster in seeing the new technology and adopting it.

After all, when we saw Snapchat coming in with a new feature, Facebook had copied it the next day and so on and so forth. Existing organizations are much faster in adopting. That would explain why we don’t see the young firms growing. What it would not explain is why don’t we actually see higher growth overall, because in Facebook, we saw Instagram copying TikTok and getting a lot of the features that were in TikTok immediately. All the customers get the same benefit instantaneously, so that would not explain the slowdown in productivity growth.

James Bessen: The largest firms, they’re not seeing an acceleration in productivity growth, but they’re seeing sustained productivity growth. The slowdown is coming with the smaller firms that can’t challenge. Now, is it a story that those smaller firms are having their ideas stolen? For one, I don’t think that happens as easily in all areas of technology as it did with Instagram. It’s very clear that you see with some companies, small and medium-sized companies, they have patent protection. They have other things that enable them to keep some sort of advantage, but they’re just not growing. I guess that’s really an empirical question at this point as to how much the large firms are able to take over technologies developed by others.

Bethany: One of the things I found most interesting about your work is the far-reaching nature of many of the implications. On one level, you can see this as a book about a change in the economy and the rise of technology that can shut down innovation, but on another level, it’s about a very profound change in our societal structure, and it does raise all these other interesting questions. Talk a little bit about that. What worries you the most about what you found and what you see?

James Bessen: This is very tricky, because in a sense, the problem is arising because these companies are benefiting society, definitely in the short run. It also poses a problem for regulation that a software-based economy is increasingly difficult for government regulators to manage, to handle. Regulation has turned increasingly dependent on information, and yet a software-based economy means there are ample opportunities for information to be distorted.

I give a couple examples, the Volkswagen . . . Well, it wasn’t just Volkswagen. There were a dozen companies faking the emissions on diesel cars, and this was done over many years, but if you took your Volkswagen in for emissions testing, where the EPA rates it and European regulators as well, they had various ways of detecting whether you were doing an emissions test or not. And if you were in an emissions test, they would make the emissions nice and low, but once the software figured out that you were on the open road, they boosted the emissions 20 times, so it was belching out serious pollutants. They were able to use software to obscure the nature of what’s being regulated.

A bigger problem is probably the 2008 financial collapse, where software wasn’t the only thing going on, but you had these financial instruments, increasingly complex, that could only be managed with software systems. Very opaque for someone to understand what’s really at risk. To alleviate that problem, they were rated by the ratings agencies, who used software models to rate them, but it turns out that financial institutions could pretty quickly learn how to game those ratings. And then you had the banks buying up this stuff who used financial-risk models—software models, again, that were also opaque—to evaluate their own risk, and it became a situation where nobody really knew what their level of risk was. The obscurity that came with software effectively made a significant part of the financial system unmanageable.

Luigi: You emphasize the software part a lot and not the data part a lot, but at the end of the day, the data are very important, and by the end of your book, I was actually expecting a call for sharing data ownership. Wouldn’t that be a solution?

James Bessen: Yeah. OK. This is my fault, I think, for the words I use. When I’m talking about proprietary software systems, that’s a combination of software and data and organization, all three, and so data is part of that, and I thought I did, in the book, talk about the importance of sharing of data.

Luigi: You do, but you don’t go so far as saying maybe the data should be automatically shared after six months. No? After all, Google and Facebook are telling us that after six months, the data are useless. If that’s true, why don’t they share it publicly?

James Bessen: Yeah. One of the problems about talking about regulation of data is it’s far more complex than it seems. I think it has to be industry-specific. Sharing data for social media for six months might make an awful lot of sense, but again, what you mean by data is fairly tricky. The Facebook case, for instance, if I like something, that’s connected to all these sorts of different data in terms of who my friends are, what they’re doing, what the information about them is. It’s not a simple thing of, oh yeah, you want our data, we’ll send you a disc. It’s far more complex.

I don’t pretend to have the answers on that, so I left it fairly open. I think on most of this, where I recommend policy, all I can feel confident in recommending is broad general directions. But certainly, I think opening data is part of it.

It’s also true that what makes these systems proprietary may be one piece or another. In some cases, the software may even be open source, which is true of some of the AI software, for instance, but it’s the data that makes it proprietary, or it’s the organization that makes it proprietary. In other cases, there’s proprietary software, and the data is publicly available data. So, absolutely, data is part of the equation here. Just how exactly we deal with that, how exactly we deal with sharing software and the information about software, these are difficult questions, and they’re going to take us years to figure out.

Luigi: One last question. I feel that it’s also important to understand the influence that large organizations have through grants and funding of research. I’ve seen—and kudos to you—that you disclosed that your research has been funded by Google and the Koch Foundation and so on and so forth. To what extent do you think that this has impacted or not impacted your research?

James Bessen: Well, I don’t think it’s impacted our research significantly. I think some of our funders may not be too pleased with where I’ve gone with some of the research, as a matter of fact. I have one funder who doesn’t talk to me anymore.

Luigi: That’s a good sign.

James Bessen: I think that’s always a risk. We took the gamble that we could take the funding and not get overly influenced. The funding doesn’t come with any strings directly attached to it, and I think it’s tended to be more a matter of we’re doing the kinds of research that organizations like, and therefore, they fund us more, but it’s a valid concern. It’s a valid question. I think different researchers have to answer it differently. I don’t think the answer is necessarily to forego funding from anybody who might be questionable, and the reality is even the well-known, seemingly independent foundations, they have agendas, and they have outlooks that may tend to mean that they fund certain sorts of work rather than others.

Luigi: Thank you very much. I really appreciate you answering this question.

This is where, in my view, he falls a bit short. I was trying to push him, but my view is that his reasoning leads to much more aggressive implications. The IBM example is actually pretty strong, because IBM was forced by an antitrust suit to unbundle the software and the hardware. And of course, there are a bunch of economists that whenever somebody forces, when the government forces a company to do something, and ex post, it turned out to be the right thing to do, they say, oh, but we were about to do it anyway.

It turns out that he cites enough sources to show that it was not the case, but in spite of that, he ends up saying we should nudge companies to do that. I don’t think you should nudge it. You should force the hell out of them.

There are two beautiful examples where antitrust spurs innovation. Number one is the famous case about the forced licenses of the transistor. The FTC, I think, forced AT&T to license transistors to everybody, and that really spurred the computer industry as we know it. That’s one case, and the second one is this one of IBM that forced the unbundling of hardware and software, and that is what led to the explosion of software, and basically, the ability to even have the personal-computer industry. I think that without that required unbundling or pushed unbundling, this would not have happened or would have happened many, many years later. So, his defense is that if IBM sees that it was to its advantage, they would have done it. So, I’m trying to push to make sure that IBM sees that it is to its advantage.

However, in his book, it’s very clear, this is to your advantage only if the dramatic reduction in prices will lead to such a high increase in demand that you can make more profits by taking a small slice of a much bigger pie than a big slice of a small pie. And this depends very heavily on the elasticity of demand of prices in that particular sector, which might be true, might not be true. I don’t think it is necessarily the case that this will happen. So, if that’s an incentive for companies, great. We should let companies know, and maybe in business school we should teach them to do so, but what about the cases in which it’s not in their interest to do so? In those cases, I think you should intervene, and that’s what antitrust is about. That’s problem number one.

Problem number two, I think he’s way too kind to platforms. When Amazon Web Services is opening up to everybody, or when Amazon Marketplace is opening up to everybody, there is a transition phase in which everybody’s going to benefit. There is an ultimate phase in which we all depend on Amazon for everything, and they’re going to extract a lot of surplus out of us, ignoring that eventually, the end is like jumping from the 50th floor and saying, the first 49 floors are great, and I ignore the last.

I don’t know, maybe I’m a little bit too harsh, but I have to say I’m harsh because I love his book. I think it’s very insightful, very novel. In my view, he doesn’t want to push it as hard as he could.

Bethany: I don’t disagree with that, but I guess I have a lot more understanding of why he wouldn’t want to push it as hard as perhaps you do, based on where you’re coming from, and I think one is that he didn’t begin as a critic of software. So, for him to take on the industry in this way was already a fair amount of movement and a very intellectually honest movement, and he wasn’t coming at this from the standpoint of somebody who thought something needed to be done differently. He was coming at this as someone who was looking to explain a problem and found a really novel and intellectually honest way of explaining a problem. I think his diagnosis is more compelling than his solutions, but I’m OK with that. I don’t think that takes away from the strength of the diagnosis.

The other part of the problem, I think, is that his diagnosis doesn’t lend itself to any kind of easy solution, in the sense that the issue can be different in any given industry, and I thought that was among the most compelling of his points, that in one industry, it might be the control of the data. In another, it might be the control of the software. This doesn’t lend itself to a prescriptive antitrust policy as much as it does a very nuanced, case by case, what’s the issue in this particular industry? What are we trying to solve? What is the core thing that is enabling this cementing of power?

It might be different from one industry to another. Therefore, the solution has to be very tailored, and you can bet, in our day and age, when corporations have so much lobbying power, that companies will be able to come up with a way of saying, we’ve enabled the solution. We’re sharing this piece of data, we’re sharing this piece of software, and they won’t be sharing it all, and the core problem will go on perpetuating itself if we start to lean into a one-size-fits-all solution, or even any kind of easy solution, because I think the core of the issue is going to be very difficult to diagnose, and it’s going to be very different in different industries.

Luigi: Actually, Bethany, I find your, if you want, cultural interpretation, very insightful and most likely right. The answer that there is not one size fits all, first of all, can be said for everything, but second, in fact, it is an argument in favor of antitrust and not regulation, because regulation is one size fits all. Antitrust can be more modulated if there is a strong principle behind it, and I think that the strong principle behind it that he’s making is, look, we have invested too much in protecting property rights, but intellectual-property rights have two sides.

There is the incentive to accumulate them, and there are the benefits of diffusion. Even in the US Constitution when it comes to copyrights, it says clearly that you need to trade off the two. We’ve gone way too far in one direction, and it’s pretty clear that you have to find the opposite direction. If these property rights are a form of monopoly—because after all, intellectual property is a monopoly—so, is a grant a monopoly? . . . But it’s a monopoly, and if you’re saying we should fight against the excessive use of that monopoly in order to promote more innovation, I think that you can derive a pretty compelling agenda for antitrust, if you want to.

Bethany: I agree with that. Doesn’t a core part of the problem here go back to this innate belief we have, which is turning out more and more to be wrong, which is that if a firm does something for its own economic benefit, somehow, the fact that it benefits them economically does, in and of itself, diffuse to the rest of society and that we need look no further than a firm maximizing its own . . . It’s a version of the Friedman-esque argument that we need look no further than a firm maximizing its own bottom-line profits because, therefore, that is going to automatically lead to the right outcome for society. And this is an interesting take on that argument or another example of how that is not necessarily so, and that what a firm does in its own economic interest may actually be to the detriment of society in a very broad and big way, and so, we need to think about things from a very conceptually different framework, and this is part of thinking about things from a very conceptually different framework.

Luigi: You said it very nicely. I think that the big distinction is we’re talking about innovation, which has some characteristics of a public good, versus production of widgets or any production of any other private good. When there is involvement of a public good or a semi-public good, because information is excludable . . . The two characteristics of a public good is that it generally is not excludable and is not rivalrous, in the sense that you and I can have it together, and my having it does not reduce yours. Now, information clearly fits the second criteria. It doesn’t fit the first one, because you can retain information and not share it with me. So, it is excludable. But the non-rivalrous nature of information and innovation is what determines the failure of many results of our standard economic analysis, including, of course, Friedman and maximization, et cetera, et cetera.

And this is why even the most conservative, pro-free-market individual can see that, for example, subsidizing innovation is a good idea, and government subsidization of R&D is a good idea. That argument could be pushed forward and, certainly, the part where he does insist, and where I agree 110 percent, is the fact that all these restrictions on limitation of movement of people and noncompete clauses and trade secrets and all this stuff is really, really detrimental to society at large, and the ability to sue should backfire, because some companies sue out of a matter of policy, before they even know whether they have an argument.

But if you are a young person, you leave to start your own firm, and I sue you, I destroy your startup, because the venture capitalist, the first thing that he or she does is to check whether you have a pending suit. And if you have a pending suit from Microsoft, Intel, et cetera, they say, bye-bye, see you later, whether there is merit or not. At the very minimum, you should say that if I sue and I don’t win, I owe you all the money I sued you for. If I sue you for a hundred million and then I can’t win that case, I have to pay you a hundred million. That would be a pretty good way to deter these kinds of suits.

Bethany: It does make me think that there’s a really good follow-on book from his book that would have to be done by some combination of an economist and an investigative journalist, but I believe that companies have probably long ago recognized this dynamic that he’s writing about. Whether they’ve recognized it implicitly or explicitly is an interesting question, but per your point, you think about how increasingly aggressive companies are around the movement of employees and the movement of this kind of information, that the policing of that is one of the ways in which you see that companies have recognized this dynamic.

I think another way in which you can see that they’ve recognized it is the way in which companies protect complexity. I know it the most in the case of the big banks that I understand, but the arguments over disclosure of derivatives positions have always been, that’s proprietary, that’s a trade secret. Other people could trade against us if we were to disclose that information in any more detail, all of which is some mixture of true and b——t, and you can see, again, that there’s something to protect in how protective they are about it. It would actually be really interesting to do a follow-on book looking at the various ways in which companies have recognized this dynamic and moved aggressively to change the environment in ways that enable them to protect the very things that he says are at issue.

Luigi: Yeah, you’re absolutely right, and maybe you should write that book. You seem to be highly qualified for that. But it’s funny, because you know that Intel was a spinoff from Fairchild Semiconductor. Basically, they de facto stole the wafer technology that was invented at Fairchild Semiconductor and sort of started at Intel. And because it takes one to know one, Intel sues the hell out of anybody who does anything coming out of Intel. The reason why it’s not possible to have these spinoffs is because of this aggressive litigation system.

And on trade systems, you probably know because you have studied the shale oil industry, what I read is that the combination of chemical agents that are used to extract shale oil, that are pumped into the Earth to extract shale oil, is a trade secret. And once, an employee was literally showered by those ingredients to a point that—I think it was she—she was taken to the hospital, and shivering on the verge of death, and the doctor said, I need to know what you’ve been taking a shower with. They called the company, and the company initially said, no, it’s a trade secret. I’m not going to reveal it. Eventually, it was forced to, but the person almost died as a result. This kind of protection is terrible.

Bethany: I have not heard that story, but that encapsulates everything we’re talking about. It’s probably the most pointed example you could possibly have about corporate power and the lengths to which they’ll go to keep these things secret, but for sure, that protection is also, in this case, on an environmental basis, because if they were to disclose the exact cocktail of ingredients, then that would give the environmentalists, most likely, even more fuel to add to their fire, no pun intended, or no oxymoron statement intended. Anyway, there is something in this that you can see the accuracy of his argument, in what we can tell about what corporations try so desperately to protect, and I think there’s an interesting nexus there.

Luigi: Now, it’s not because I’m subtly trying to have an episode dedicated only to Powell and the Powell memorandum, but I think that this subtle push in favor of more enforcement of property rights, more proprietary stuff, more this and more that, all started with a big push to our probusiness strategy at the beginning of the ’70s, and it took a long time to be fully in effect, but that’s exactly the consequence of that movement.

Bethany: I do like this idea of thinking about the world through the lens of what are public goods and what public goods need to be diffused in order to keep our economy and our society on a track that we all agree is for the best over the long term, and that leads to a very different set of outcomes or policies than thinking about the world from the standpoint of what will maximize an individual corporation’s bottom line, and I think those are, it’s—

Luigi: But this, in a sense, should be the framework with which patent policy is developed. The reason why patent policy is designed the way it is is because they say, we give you a monopoly, which is in general a bad thing, but we give you a monopoly, number one, as a temporary incentive to invest more, but number two, because you are sharing that information, so that by the end of the monopoly period, everybody will use the same information.

The example I heard made many, many times is the comparison between the Stradivarius violin and the saxophone. The Stradivarius violin is not imitable. Why? Because Stradivarius did not have a patent system, and so the secret of his company died with him, and to this day, people have not been able to replicate the Stradivarius.

The saxophone is an instrument named by a guy called Mr. Sax, who got a patent, and he made a bit of money while the patent was in place, but after the patent expired, everybody was able to produce saxophones according to a specific detail provided by Mr. Sax.

Now the question—and this is a bigger question that maybe we should analyze separately—is, are we right about that trade-off, and to what extent is the trade-off different across different areas? In software, it might be very different than in pharmaceuticals. Certainly, in the last 20 or 30 years, lobbying has pushed us more and more in the direction of extending everything. There is a vested interest in extending all these rights as much as they can and milking every benefit, even when there are no incentive reasons. When they extend the copyright of work by Walt Disney, it’s not creating incentives for Walt Disney to write more stories, because he’s dead. But they still extract a lot of surplus from all the little kids that have a balloon with the picture of Mickey Mouse.

Bethany: I did not know the Stradivarius/saxophone example, and I love that. That’s really interesting, but I couldn’t help thinking as you were talking, when you mentioned pharmaceuticals, that we’ve been trying for however many decades, wasn’t it 1970 that Hatch-Waxman was passed? We’ve been trying for 50 years to get the balance right in the pharmaceutical industry between the enforcement of patent rights and the public good of having lower-cost drugs more broadly available, and we’re still fighting that fight, and I don’t think we’ve gotten anywhere very good, in part because of a legal system that has enabled companies to game the patent system by coming up with all sorts of ways to extend the patent life of their products.

But it is a little bit of a caveat to what you’re saying and what we’re discussing about property rights, that the attempt to strike this balance in pharmaceuticals . . . I don’t know. Maybe I’m wrong. Maybe you’d look at it more broadly and you’d say we are striking a good balance, but I can’t help looking at the past 50 years as just an absolute morass that has enabled lawyers to make an incredible amount of money on either side of the argument while not necessarily doing the right thing for consumers. Maybe I’m being too negative. Maybe you could actually spin that around and say that the attempts to strike this balance in the pharmaceutical industry have been broadly successful, even if narrowly miserable.

Luigi: I don’t know whether we got it right. Certainly, the United States has a phenomenal capability in R&D in pharmaceuticals, and we have seen it with the vaccines against COVID, and we should be thankful for that in many, many dimensions. Now, whether we got that trade-off right, I think that remains to be seen, and that’s part of the bigger discussion.

I think he’s very thoughtful in this analysis, and I am still a bit puzzled by the fact of why firms don’t grow that fast, because his argument is that our runners-up don’t grow that fast, and one interpretation is that maybe they don’t have space to grow. One of the arguments he makes, even if not so explicitly, is that once you have a very strong incumbent, the only way you can enter is by differentiating yourself in some dimension, maybe quality or others.

If you have a very strong incumbent that is difficult to unseat, you’re going to only specialize in niches of the market, and so your growth is very limited. It is a very strong indictment of the fact that this time seems to be different. When you talk to people about antitrust, they say, oh, but IBM has come and gone, and Microsoft is not gone, but it’s certainly sort of gone down, and so, why are you worried? Technology will automatically produce a demise of the new firms. His argument seems to suggest that this is not that obvious, at least for now.

Bethany: Yeah. I love that, and I wonder if it would suggest an area for future research. I wonder if you were to look at the way in which we talk about new entrants entering a given industry, if we would see that those new entrants are increasingly just pulling off, per your point, one little piece of the business where they might be able to challenge, but they can’t challenge the whole thing because of this built-in dominance of the lead firm that we’ve been discussing, and so maybe you’d see that a new entrant isn’t really a new entrant the way Walmart was a new entrant in the business of retailing next to Sears. You’d see that these new entrants, instead, it’s being broadly defined as a new entrant, but in reality, what they’re doing is just being able to pull off a little piece, and maybe that’s part of the issue.

Luigi: Yeah, absolutely.