← LSE: Public lectures and events
LSE: Public lectures and events
How progress ends: technology, innovation, and the fate of nations
01:24:53
Loading summary
Transcript77 lines
- [00:02]
A
Welcome to the LSE Events podcast by.
- [00:04]
B
The London School of Economics and Political Science.
- [00:07]
C
Get ready to hear from some of the most influential international figures in the social sciences.
- [00:16]
D
Good evening and welcome to the London School of Economics. Those of you in the room and online for tonight's event. Tonight's event is hosted by the Department of Geography and Environment and co hosted with the program on Cohesive Capitalism. And it's also part of the ESRC Festival of Social Sciences which is a series of events which we'll be carrying on until November on bringing the social sciences to the world. So my name is Neil Lee. I'm a Professor of Economic Geography here at the LSE and I'm really pleased to be. Welcome. Well, to welcome an all star panel to talk about this event, talk about this book. So we've got Carl Benedict Frey from the University of Oxford, Jane Gingrich and Michael Stauper. And we're here to discuss Carl's book, How Progress Ends Technology, Innovation and the Fate of Nations. Now Carl is one of the. I'd say he's probably one of the leading scholars of technology and innovation in the world today. So we're really, really pleased that he could come here and talk about his book. He's Dieter Schwartz, Associate professor of AI and work at the Oxford Internet Institute and the Oxford Martin City Fellow at the Oxford Martin School, both at the University of Oxford. He's a Fellow at Mansfield College, the Institute for New Economic Thinking at Oxford and Lund University's Department of Economic History. His other book is the Technology, Capital, Labour and Power in the Age of Automation. I finished his other book, I haven't quite finished this one, but I can say his first one is excellent. So hopefully this one's going to be just as good. So we're joined by two brilliant respondents. So first of all we have Professor Jane Gingrich who's Professor of Social Policy at the Department of Social Policy and Intervention at the University of Oxford. She's also a Senior Research Fellow at Green Templeton College. Now, Jane's area of expertise, comparative political economy and social policy. She's got a particular interest in the contemporary restructuring of the welfare state. Jane is also a fellow of CIFAR's innovation, equity and Future Prosperity Program. As am I. And she's an Associate Fellow at the UK based Institute for Public Policy Research which is one of the UK's leading think tanks. So we have a sort of Oxford team. We also have a home team which is Michael. Michael Stauper is Professor of Economic Geography at the lse. He also holds affiliations at Sciences Po in Paris and UCLA's Department of Urban Planning. He's a corresponding fellow at the British Academy, winner of the Sir Peter Hall Prize, and he won the Royal Geographical Society's founders medal in 2016. So we're really pleased to welcome all three of you to speak today. Let me just talk a little bit about some of the housekeeping we have here. So first of all is if you could please, can you put your phones on silent? You are allowed to use your phone, but only to tweet. If you are sorry. If you're right wing, you tweet. If you are left wing, you use Blue Sky. The hashtag is LSE Events. I also need to remind everyone that we are. There is a podcast today of today's event which will be online shortly after the event, provided there are no sort of technical glitches or, or anything like that. It's also live streamed and there will be a sort of, you know, it will go on the LSE YouTube, so that's. You guys can all watch it again if you want another time. And for those of us on the panel, it's just a reminder to try not to go viral for doing anything too stupid because it's all online. There'll be a chance for questions at the end. Those of you online can sort of start now. Those of you in the room will have to wait until the end of the event when you can put up your hand in the usual way. We'll want your name and affiliation. I think that's it. I'm really excited about this book. I think it's a brilliant piece of work. One of the reasons we actually got Carl to speak here today is because we have a new Masters in Innovation Policy at the lse. So we wanted to do more innovation related events. And so I was really pleased that Carl, who's one of the world's leading experts on innovation, was able to come and talk to us about it. So, Carl, without anything else, without further ado, I can hand over to you. Thank you very much.
- [04:17]
B
Thank you very much, Neil, for that very generous introduction. Unfortunately, I have no way of living up to the expectations after that, but I'll do my best and try to give you at least a bit of a flavor of what the book is all about. Now, the book is not quite as gloomy as the title might suggest, so I'm not suggesting that necessarily technological progress is about to end, but what I am suggesting is that we shouldn't take it for granted if progress was Inevitable. It would not have taken humanity 200,000 years to have an industrial revolution. If progress was inevitable, most places around the world would be rich and prosperous today. If progress was inevitable, former technology leader like Britain would not be facing productivity stagnation. So there's nothing inevitable about progress. In fact, progress is always hard work in progress. Now, I'm not the first person to write a book explaining differences in patterns of innovation and income across space, nor am I likely to be the last. But since I'm at the university, let me at least try to situate the book a little bit in the literature though I'm not going to do justice to all the people who've contributed to this. They are cited throughout the book. But I think it's fair to say that there are three buckets of leading explanations for why some places are rich and poor. And they have to do with geography, they have to do with culture, or they have to do with institutions and different variations within those three buckets. I think if you look back to the Soviet Union, which grow very rapidly from for around four decades and then essentially collapsed in 91, could geography explain that? Well, only if it interacts with some other factors. Right. Geography is constant, so can't really explain spurts in growth followed by stagnation. Nor would you say that there were marked changes in culture in the Soviet Union during this period. If anything, the Soviet Union became more open and more exposed to new ideas because no amount of barbed wire could prevent the radio and television waves from Western Europe to travel right through the Iron Curtain. And obviously in the Gorbache of years through glass knots, open debate, scientific discourse. Technology transfer was more on the agenda. But it didn't halt decline or collapse. Nor did the Soviet Union during any period have anything like secure private property rights that economists have long emphasized as critical to growth. And if anything, again through perestroika, the Soviet Union became somewhat more decentralized. Firms were granted a little bit more of autonomy during the Gorbachev years. That didn't bring a halt to collapse either. And so the point I'm trying to make in the book is that we need to understand how these factors interact with technology because catch up growth is very different from innovation at the cutting frontier. The Soviet Union did very well when it came to military competition and launched some spectacular firsts. That's which led a lot of people to speculate that the United States would soon be overtaken. And they took ample advantage of the Ford factory's open door policy, sent across several delegation and transplanted a car industry to the Soviet Union. And they were very good at mobilizing resources at scale to catch up in key industries. When a technology is mature, it's more easy for authorities to hold people accountable. You could simply benchmark across factories and see how factories producing the same product were performing. And so the Soviet Union could hold factory managers accountable that way. But that is much harder at the technological frontier, because if something is new, what you benchmark against. People like Oscar Lange, writing in the 60s, thought that the computer revolution would, if anything, aid central planning. But what he actually saw is that it made it much harder, because when a technology is new, that kind of benchmarking is particularly hard. And needless to say, Soviet contributions to the computer revolutions were essentially none. And so what the Soviet Union needed when the mass production system essentially petered out in the 1970s was something new. And so to get to that something new, you need exploration. And exploration thrives in more decentralized environments. And so, for example, if you were an aircraft engineer in the Soviet Union, you could go to the Red army and ask for funding. If they declined, maybe you had two or three other options. If they declined, well, basically your idea would die with you. That's quite different from Google, where in which Bessemer Ventures famously declined to Invest back in 1999. It probably regrets that now, but it also illustrates that Google was no safe bet at the time. Yahoo and AltaVista were leading in search. But the fact that Bessemer Venture didn't invest didn't mean the end of Google as a company. And that's the virtue of decentralization, that you can take many different bets and you're not going to know exactly what will catch on before somebody has actually taken the risk to invest in it. Now, this is not just a story of private capital versus publicly funded innovation and R and D. Many state projects have been successful in achieving clear objectives. So the Apollo project put a man on the moon. The Manhattan program delivered the bomb before Nazi Germany. Operation Warp Speed succeeded in combating the pandemic. The RAD lab advanced radar technology. But that's very different from decentralized exploration. And here too, the government can play an important role. There was no plan to develop the modern Internet at arpa, yet ARPA seeded arpanet, which eventually became the modern Internet. But it did so by granting program managers considerable autonomy to build their own teams. And they in turn, had a decentralized network of universities to draw upon for talent. And so you wouldn't necessarily put ARPA and the Apollo project In the same sentence, they are a very different nature. Now the Soviet Union is obviously an extreme case, but it's also reflective of a broader pattern which is that throughout the post war period growth was very rapid in Europe, in Central Europe, in Eastern Europe and on the periphery. And in all those places Europe relied more on bureaucratic planning, whether it was through the state, like through the Monet Plan in France, which created a concerted economy whereby industries could coordinate and industrial partners could coordinate more easily, or whether it's through state holding companies like in Italy or in Spain. And what Europe did in this period was essentially taking advantage of ample opportunities for technology transfer through Marshall Aid, which essentially transplanted the manufacturing system, mass production system into Europe. And it already had institutions for catch up growth in place because during the 19th century Europe had used those institutions to catch up to Britain. And of course no place was better at coordination and catch up growth than Japan, which many people in the 1980s predicted would overtake the United States at the time. As well as the flag of the Soviet Union was lowered from the Kremlin and replaced with the Russian flag. PBS Frontline ran a documentary entitled Losing the War to Japan in which Chalmers Johnson famously remarked that the Cold War is over and Japan won. And there was certainly a sense that the way Toyota had out competed General Motors and Ford in automotive in the same way Japan was now out competing American firms in consumer electronics and even in semiconductors at the time. And to be sure, the Allied occupation authorities led by the United States, they tried to impose Western institutions after the second World War. But the anti monopoly law was relaxed very swiftly. The Zaibatsu conglomerates of the 19th century were replaced by Congrez conglomerates which did much to coordinate between different industries and to share technology which aided scaling. What Japan didn't do so well was dynamism and was entry of new types of firms like in Europe. They did very well when it came to catching up, but not so well at innovation at the frontier. And here I think it's important also for today to remember the impact antitrust policy has had on innovation and growth in the United States. Because had it not been for the antitrust's lawsuit against IBM which forced it to unbundle hardware and software, we may not have seen the entry of of Microsoft and a range of other firms competing in software in similar fashion, had it not been for the breakup of AT&T, we may not have seen the upsurge in patents driven by unrelated firms. And it's actually quite likely that when the national Science foundation released arpanet, the predecessor to the modern Internet, into the world. It would have been absorbed by AT&T and potentially bottlenecked in its boardroom if it saw no immediate commercial benefit. And the commercial benefits of the Internet at the time were certainly far from clear. And secondly, the United States turned out to be more flexible in its capacity to rewire its economy, create and integrate into global value chains, which was in large part driven by the computer revolution. There is no way that you can manage complex supply chains at distance without modern information and communications technology. And the integration with China that followed, I think many people today forget, was actually a response of American firms finding themselves out competed by Japan when it came to production and process improvements. And so integration with China was in large part a solution to this problem and obviously had a nice upside that lifted 800 million people out of poverty at the same time. Now, I think it's worthwhile to ask the question then, why China? Why not the Soviet Union? Didn't they, after all, both have the same communist past? Well, to some degree that's true, but to some degree they're also very different. So in the Soviet Union, essentially everything from steel to the railroads were managed by line vertical line ministries that essentially managed steel, railroads, electricity, etc. Centrally from Moscow. That's very different to China, where provincial governors and local authorities had a lot of autonomy in managing their own economies and their own industries. And what that meant is that China could experiment locally and scale up what worked without having to pull the rug under the entire system. And a key fact about Chinese industrial policy is that at the national level, it's actually a fairly recent phenomenon. It really takes off in 2006. And Philippe Aguion here at LSE, who won the Nobel Prize last week, was actually early to note a key feature of industrial policy in China, which is that it was driven locally with local authorities competing against each other. So it was pro competitive rather than anti competitive. And what you see in the early days of Chinese industrialization is that whereas Beijing continues to pursue a bit more top down industrial policy, all firms moved down to the south, which Deng's Southern tour in 92 showed was becoming increasingly prosperous and pulling ahead of the rest. Where China retained control was over key personnel and who got promoted inside the party. And growth was a key measure and a key variable underpinning promotion, which is to say that places that pursued liberal reforms which aided economic growth were afforded for it. And this is the system that Branko Milanovic has aptly called political capitalism. Now, the Problem is though, that when you come closer to the frontier, that system becomes harder, because how do you hold people accountable when it comes to frontier innovation? One way of measuring innovation is through patents. But what you see is that the local authorities begin to put out rewards for filing patents. That leads to patent inflation. And businesses are not even willing to pay maintenance fee to uphold their own patents, which suggest that they're not really covering valuable technology. And at the same time, we've seen forces of centralization in China in recent years, putting more political power in the hands of one man in particular and of Beijing in general. And we're also seeing growth taking a backseat in terms of priorities. Technological self reliance, common prosperity are becoming increasingly prominent goals as well. And it turns out that it's quite hard to incentivize private firms to pursue national objectives. And as a result of that, China is once more and more reliant on state owned enterprises, which have been a drag on productivity and innovation in the past. And so what we're seeing in China today is basically a decade streak of slow productivity growth. China has already begun to stagnate at income levels much lower than we saw in Japan or Korea. And in one critical aspect, or in many critical aspects, perhaps China is not too different from Europe or the United States. It's smaller foreign invested companies that tend to be most innovative. But as long as they serve national objectives, some deep seats can be permitted to succeed. The trouble is that because political priorities can change swiftly, these firms as they grow, need to invest more in political capital to have a seat at the table when the priorities begin to change. And that means more resources allocated away from innovation towards building political capital. A trend that we are unfortunately now seeing in the United States as well. And so it's not just China that's stagnating, although that may be more surprising given that they have greater scope still for catch up growth. If we look across the industrial economies, productivity has been trending downwards for a good two decades. And even more exception to that is is the United States, which sees a brief upsurge in productivity growth around the IT boom from 95 2004. And key puzzle at least in my view, is that not that Europe might be less dynamic, but it's actually even failing to catch up in digital. But maybe we'll leave that question for Q and A later. Now, as I point out in the book, a key virtue of the US system when it comes to innovation and technological change has been its diamonds. So if you take the five largest firms in the United States by market capitalization. They're all in technology, and they are less than 40 years old on average. If you look at Germany, they are 100 years old plus. And they are mostly firms that were created during the second industrial revolution. An exception to that is SAP, set up in the 70s by five former employees of IBM. And if you go to France, it's a similar picture, but most of the firms are in luxury goods instead. And so, if you want to understand the relatively brief productivity upsurge in the United States related to computer revolution, entry is a very big part of that story. But entry is fading across the board. So regardless of whether which metric you look at, whether it's productivity, whether it's reshuffling of market share, whether it's dynamism in terms of exit and entry of firms, even in the technology sector, the trend is pointing downwards. And it coincides by a doubling in lobbying expenditure by incumbents primarily. It's probably also driven by a revolving door between incumbent firms and parts of the state bureaucracy. One example of that is the Patent Office, where we see that patent examiners are more likely to grant low quality patents to incumbent firms and then move to work for them. And in addition to that, we've seen more lax antitrust policy, we've seen an upsurge in killer acquisitions whereby incumbents essentially buy up smaller firms just to shut them down. So this historical advantage is currently fading. And what that leads to is that we have more and more inventive talents clustered in a few large companies. And you might be forgiven for thinking that some firms in the tech sector are still inventive. And that's not very surprising given the amount of talent that they employ. Nor is it very surprising that the country with over a billion people like China has some innovation. It would be extraordinary if they hadn't. But the question is, are those inventors, is that talent more productive in these large incumbents than they were in the younger firms they previously worked in? And the answer to that is decisively no. Now, there's a lot of talk these days about whether AI will revive this productivity slowdown. And if it was the case that you get to better AI just by scaling up existing models, more data, more compute, more parameters, then some concentration might not be a bad thing. But I think basically all evidence now suggests that scaling is running out of steam and we probably need new ideas rather than just scale. And so many of you have probably heard of ChatGPT or whatever your favorite foundation model is, passing the bar exam, or this human benchmark or that other test or whatever. And that would be much more impressive if they weren't trained on the right answers, right? And one way of getting to at that problem and see if they actually just memorize or able to generalize outside the training data is to introduce novel problems and test their performance on them. And so ARC AGI is one such benchmark. And on that benchmark, none of the leading models do better than 15% get 15% of their answers right on tests where humans, any human, basically get 100% of the answers right. And to sort of illustrate this point a bit further, some of you probably heard, or probably all of you have heard about AlphaGo beating the world champion Echo back in 2016. I suspect fewer of you might have heard that human amateurs using standard laptops actually beat the best open Source Go Programs 2 years ago by exposing them to new concepts and new positions in training that they wouldn't have previously encountered. And so even in domains where AI achieved superhuman performance, we've seen an astounding comeback. And what this means, I think, is that AI is very unlikely, at least in the near future, to replace humans in frontier discovery. And so consider the following thought experiment. You're in 1900, you have a foundation model, has been trained on everything that exists up until that point, and you ask it, will humans ever be able to fly? What's in the training data? A long list of failed experiments of human flight. You might have looked to birds as evidence of it being possible to fly in principle, but even there you would conclude that there is no bird that weighs more than 30 pounds that is able to get off the ground. Right? Larger birds, like ostriches, are flightless. And so by any stretch of the imagination, you wouldn't have produced the technology introduced by the Wright brothers very soon after that. And so what humans do write is probably a book in itself, but a big part of it has to do with exploration and experimentation. Which is to say that I think that the old rules, patterns, whatever you want to call it, hold even in the age of AI. One of those key lessons is that decentralized organizations do better at periods of rapid technological change. If people at the lower levels of an organization, who are the ones that are closest to the technology, who uses it on a regular basis, experiments with it, knows what it's capable of. If those people don't have any decision making rights inside the organization, adjustment will be very slow. And a key reason why US firms did better during the computer revolution is that they were more decentralized compared to most of the European counterparts. Second entry is critical also because newer firms tend to pursue different types of activities. Large firms, they have scale. They are more likely to invest in process improvements and automation and bring what they already do to market at a lower cost. Smaller firms, more likely to develop new products, new lines of business, new kinds of industries. And I think a key reason why growth was so rapid in the post war period was that we saw the development of entirely new industries. The automobile industry was the largest industrial undertaking that the world had ever seen. And in addition to the automatic automobile industry itself, there was a host of industries producing components that go into making a car. A host of industries producing machine tools to make the components. A host of new industries related to electrical industries. Basically all the gadgets that you have in your kitchen today had an industry behind them. The rise of road commerce, mass tourism, et cetera created huge new activities, new kind of jobs.
- [34:31]
D
Right.
- [34:32]
B
And so if all we've done since 1800 was automation, we would have productive agriculture and cheap textiles, but that would be about it. Right. We wouldn't have vaccines, antibiotics, airplanes, rockets, etc. Right. So most prosperity comes from doing new and previously inconceivable things. And those technologies were mostly conceived in the US in the 19th century, where there was a very vibrant market for technology at the time. Most discoveries were made by independent inventors. The railroad companies didn't invent the railroads, the telegraphic companies didn't invent the telegraph. Dupont didn't have any internal R and D at all. They just screened the market for patents and tried to absorb external technology and build products around them. The same with most industrial firms at the time. And so the market structure and the institutions that support exploration can be very different from the ones that support scaling and the development of these new kind of industries that we see in the post war period. But that kind of institutional change, that kind of dynamism is not something and we can take for granted. And it critically depends on competition. I mentioned that the first industrial revolution happened in Britain. Britain was long the technology leader. In the 19th century, there was a lot of concern over Britain being overtaken by Germany. That never happened in the 19th century, but it did eventually happen in the post war period when Britain was overtaken by Germany, both Germany and France. Why is that? Well, part of the reason is that domestically competition policy was extremely weak. A third of manufacturing output was cartelized. And secondly, Britain remained sidelined from the European economic community, which meant that exposure to foreign competition was limited in this period as well. What are we seeing in the US Today, high levels of market concentration, external barriers to trade through tariffs. I haven't even mentioned restrictions on immigrations and cuts to scientific funding. Now, that is not to say that the United States is doomed to stagnate, but we should remember that nothing lasts forever. And in the grand sweep of history, even 100 years is a short period of time.
- [37:21]
A
Thank you.
- [37:33]
D
That was wonderful stuff. So we'll move on to our respondents. Michael.
- [37:48]
A
All right.
- [37:48]
C
Well, let's see.
- [37:50]
B
Can you hear me?
- [37:51]
C
All right. Well, it's quite an honor to be here for this debate because this is a major book, and I think I want to sort of make a general comment about why it's a major book and where it stands in social science. So what I would characterize it as is what we might call sort of macro comparative social science. And this can be done in different disciplines. It's still done in history, political economy, sociology, and now in innovation studies. And when people do these kinds of books, they're very ambitious because what they're doing is taking a huge variety of kinds of evidence, many different kinds of cases over long time spans with many different kinds of interacting factors. And what they're doing is the intellectual art of making sense out of them, which is something very hard to do. So there's a grand tradition in which this book fits. And I think why it is a grand book, we could call attention to, for example, the work of last year's Nobel, Acemoglou and Johnson, who tried to do it for state structures and economic development, you know, where they have this book on a thousand years of prosperity and poverty, and then they give us a model of why certain states enable prosperity to occur. We can go further back in time to great names in social science, the ones that come to mind, you know, for the young people in the room, maybe these are forgotten names, but we have, for example, a classic in why we get Dictatorship or Democracy, a man named Barrington Moore, or we get Trade, Fernand Braudel in the great Comparative Annale School in In France, Charles Tilly in comparative historical sociology, recently Peter Turchin about elite competition using a quantitative comparative method. The point is that these books have a sweep that's very unusual and very interesting in that they're really trying to take a huge amount of complexity and give us kind of like a framework for reading it in a in a relatively straightforward way where we can get some lessons out of it. So I think this book, Carl, I think it's going to become a central reading and for students of innovation, I See some of you, because you're in my classes. Hi, all you or some of you used to be in my classes. The bibliography is just astounding. It's really like, it really will become a reference book for those of us who are working, learning this field or trying to move forward in it. So yes, this is a big book, a very good book, a very important book. Now what I would have wished for is a model. And why am I saying that? If I think about the enormous sweep of the book, there's a point where when one reads it, one wants to see the model as it were, stripped down into kind of what we would call a simplified parsimonious form. And I'm going to try something and I'll be interesting to see if Carl thinks I got it right or not. But here's what I think the model is. So it seems to me that the model, or no, no, I'm using the word, let's call it the framework, okay? Which is something like the. It's like the, the system by which all this evidence is read and read along sort of like different coherent axes and put together, okay. To kind of give us these general takeaway lessons that Carl was talking about for me. So there are these different elements. First out of three would be innovation. And the basic distinction that runs through the book is one between exploitation and exploration. So in other words, going along a sort of existing technological pathway versus developing new ideas. And those of you who've done innovation studies know that this is, it's like a slippery slope of what we call a polysemic field. There are lots of synonyms. We say things like incremental versus oops, incremental versus radical, applied basic, less disruptive, more disruptive, imitation versus breakthrough, second mover versus first mover, legacy versus new. And these are all like, they're subtle distinctions. But I think it's like, in other words, it's a very important gradient in the field, which to me is one of the right ones by which to read the history and economics of innovation. So that's the first thing is we are told to read the case studies by whether a country and its firms are principally doing exploitation and catch up, imitation, refinement, or whether they are doing sort of blue sky, pushing us into new areas, disruption and all that stuff. Okay, then I think the sort of second big axis running through the book is the causal one. And I think that's the main purpose of the comparative method used in the book. The basic causal interest is what makes for innovation success at any given moment in an economic development process. And what we learn, I think correctly, is that this kind of depends on whether a nation's institutions are well matched to being the kind of innovate to the kind of innovation that is right for that country at that place and time. So that means innovating in a way that propels our economic development. So there are these two different ways to be innovative. Exploitation and exploration have different institutional needs. Exploration or first mover requires more decentralized institutions. Exploitation. Catch up imitation has a wider range of possible institutional complementarities or ways you can do it. So ranges of something like decentralized but with good selection mechanisms for strengths and forces for scaling up or centralized with bureaucracy that's responsive to what works and eliminating what isn't working. So the trade offs are kind of tricky when you're trying to be a catch up power. The strong claim in the book seems to be that to be a leading explorer or first move, decentralization is necessary. And centralization and bureaucracies are all inherently rather problematic for doing that. The third element. So we've got innovation, we've got institutions, and I think then we have time. What a given country can do at any moment depends on first of all its own past. Is a country in reach, in reasonable reach, reach of being a first mover or not? This is basically about where the country is in the development and income ladder of the world. There are entrants to this club over time and at any given moment this determines what the mix of exploitation or exploration for a country could reasonably be. This kind of depends on so again the country past, but it also depends on the world environment, meaning the nature of the world technology frontier over time, from the first to the second to the third and now to the fourth industrial revolution. There are different limiting factors in general, as the scale of scientific talent and capital that's needed for world leading breakthroughs generally goes up, up over time. Like science and R D are just much bigger enterprises than they've ever been in the past. This is kind of like a linear. It's kind of like a linear expansion process roughly since you know, the year 1500. Right. So there's a strong bias that gets built in over time toward big places like today China and the U.S. or if Europe existed as an innovation system, it would would be Europe. Right. That's the European fragmentation problem. But in the first and second industrial revolutions, right. The textile revolution and the mechanical revolution. One of Carl's great heroes, Joel Moker, Right. What he told us was that back Then you kind of had the big things, but you had this possibility of small scale, decentralized, what mocare called tweaking. Right. And you could do it in small towns and small firms and all kinds of things could kind of happen around the edges of the system. Not so clearly evident in today's hugely scaled up R and D systems. Anyway, the question is whether countries can become explorers and first movers. It sort of depends on how their institutions are set up or evolve from one revolution to another as well as from their own stages of development to another, with the two intersecting in time in different ways. So I'll say this, I kind of wish, Carl, you'd given us a causal diagram. Right. I love causal diagrams. I think they're awesome for kind of like simplifying these great sweeps of things and then a bit more of like a parsimonious subject summary of this stuff and where the arrows are and kind of like what the big ideas are. It's all there. But I would have liked the diagram. Okay. And that's where I want to kind of like, I want to get to the last theme here, which is this kind of the end of progress theme that Carl was talking about toward the end of his talk. We can all agree that a decisive challenge for China and for the future of the world world economy is whether China has the institutions to get from being the world's most incredible catch up imitator to being the first mover, innovation economy. The book, the book argues that China may have a problem with, as Carl said, with the tension between the way that it kind of seeds innovation and gets all these companies going and also how it does it through, through the competition between companies, but also between regions and cities. And then it's scared of them basically is what I read, right, is that as these companies start to get, start to get toward that frontier, the risk is that they will get powerful. Right? And we know that this is a lesson from America. Those startup companies that Carl just showed, they're the key power players in not only in the American economy today, but in the American political economy. Right? This is a key lesson. They've overturned the entire institutional arrangement of American political power and they are constantly reworking in a detailed way the institutions of the old New Deal and manufacturing economy. So there's a relationship there that Chinese political power may fear. Last thing, what about Europe? Here we are in Europe, so it's obvious that we are in big trouble in Europe. Right? We've missed out on the third industrial revolution. There are no household names in it here in Europe. And we're probably falling way behind in the fourth in the books framework. We in Europe locked ourselves in into an exploitation legacy framework. We had very good catch up institutions in the middle of the 20th century. We built great sort of second industrial revolution firms that are still some of the best in the world for the quality of their products and even for their hourly labor productivity. The problem is we missed out on the third Industrial revolution, right? But we, and we borrowed the third Industrial revolution technologies to integrate into our products by the friendly American big sister or big brother. Right. And that was happy to share with us, to keep us sort of in a protectorate relationship with them and also to make money concomitantly though, okay, so now the question is for Europe, what could we do? And I think here the quest, what the book asks us to kind of consider is do we need radical institutional reform here in Europe away from kind of like our nation state institutions, but also those of the European Union, which may be better oriented toward exploitation and catch up. They're kind of centralized, they like indicative points, planning and guidance. There's strong influence of these old companies in the political life of every country here in Europe. What could we do or what are we capable of doing to build possibly the kinds of decentralized institutions that may be necessary to be at the cutting edge? That I think is a huge question for us here in Europe. Report after reporting calls attention to where we are, but hasn't offered us frankly much in the way of reform solutions. The last element of this is an issue that isn't very much brought up in the book. It's sort of touched on around the edges, which is this. So America is the leading world innovation economy. China's probably on the way to being one of them. And what both of them have seen is astonishing rises in spatial and interpersonal inequality. There's a big debate, at least for the American evidence, as to why inequality has grown so much. One hypothesis is simply that the legacy industries in America have become very monopolistic because they're is no more enforcement of antitrust action in America since Ronald Reagan, basically. But another hypothesis is that the inequality comes from the innovation sector itself, because the innovation sector generates a lot of very highly skilled labor and entrepreneurial rents. And those people, you have to pay them what it takes them, you have to incentivize them to have all of this creativity they expect. And now we're in the third generation of it and they have a lot of wealth and they're investing in it. Michael, we're gonna so question here in Europe is could we build new institutions without American inequality? What would that kind of combination look like? So thank you Carl, for a wonderful book. It allows us at least to kind of to have a discussion around these kind of big questions.
- [54:02]
D
Thank you very much, Micha. Looking forward to discussing those.
- [54:05]
B
And now Jane.
- [54:08]
A
Thank you so much. I too am really honored to be here and also share Michael's praise for this book. It is truly a tremendous piece of work in its historical sweep and its theoretical and empirical ambitions. And I think as we stare down two increasingly kind of fever pitched visions of the future, one of kind of an AI abundance on the one hand and the other of ecological and social collapse on the other, what this book reminds us is that progress is both neither inevitable but also very much possible. It's about thinking about the match between institutions and the incentives that people have on the ground. So Michael asked for a model in this book and here's how I understood the argument and I'll explain this in a moment. I think the argument is sometimes higher. Hayek is right, sometimes Gershon Cron is right. But Mansur Olson is always right. There's rent seeking up and down and we need to think about institutions that prevent that. So what does that actually mean? So the Hayekian insight of the book is that you need an openness to the new that the sort of control, particularly over flows of information, dampens innovation at the frontier and that any institution that tries to sort of fundamentally centralize control over information will fail to allow innovators that are thinking about things differently to emerge. But the Gershon Cronian insight is that oftentimes decentralized institutions don't allow scale. They don't allow scale, particularly at the rapid, at sort of the rapid introduction of technology. And that places the that need to move quickly to develop often need the forms of centralized control over systems, including information systems that allow development. But the Olsonian insight is that both decentralization and centralization on their own can become a source of weakness. These weaknesses are partly economic, a mismatch between the tasks, forms of innovation and the outcomes. But but they're also partly political. The decentralized institutions that allowed American innovation can also allow rent seeking actors multiple venues to shop and stop progress. From local governments to court to Congress. The centralized institutions that allowed rapid development in Europe in the post war period or in places like South Korea can lead incumbent actors to prevent novel ideas monopolizing new information. So the question then is how to create institutions that are both nimble enough to adapt over time to new economic tasks, but resilient enough to resist this rancid. And this is clearly not a new question. It is a kind of classic Hamiltonian question, right? It's how do you create a state that's strong enough to act, but not strong enough to repress? But in a sense, what this book really does is it opens up a new set of questions around this logic and I think raises questions about the fit between particular forms of decentralization and centralization and their linked outcomes. And I want to push on kind of three things here. So the first is the book argues that both decentralized and centralized approaches can fail even on their own terms, not just in relation to particular tasks. But centralization clearly doesn't always deliver growth. And Carl says this in one of the concluding chapters, he says, you know, it's a kind of a puzzle that the whole world's not developed if you can just centralize and do exploitation. But clearly that doesn't happen. And equally, decentralization doesn't always deliver innovation. It can just deliver very sort of weak sets of outcomes. So in unpacking the cases, I think we need to push on the concepts and ask how much they reveal and where they hide things. So both decentralization and centralization sort of refer in some sense to control over information and initiatives, but the institutions that underpin them, and I think this is particularly the case for decentralization, are very diverse. Sometimes decentralization appears to be a property of firms. So flatter management structures allow more innovation, sometimes of systems of financial allocation. So more venture capital or more decentralized banking might be allowing innovation. Sometimes it's about what I would think of as liberalism. Openness to immigration, particularly labor migration from across borders is what allows innovation. And sometimes it's politics, right? Political institutions, the structure of the state itself. And sometimes it's just as in the Silicon Valley case, It's the absence of incumbent firms that allow new firms to emerge. And these features may co vary. It may be that politically decentralized systems allow decentralized finance or more migration, but they don't always. And as the mid century consolidation of industrial firms in the US shows, you can end up with centralized players within decentralized politics. And so the question is, what kind of decentralization produces innovation? Are there particular types of institutions in education, finance or corporate governance that are complementary with particular technological affordances? Are there particular forms of decentralized political institutions that produce these? Or is it a story of complementarity if we look at modern Britain, it's unclear to me if it's decentralized or not. On the one hand, there's tremendous consolidation of power in Westminster, in this country. On the other hand, there are more decentralized systems of firm management, finance. Is it just the mismatch between the politics in the economics that leads to sort of stagnation? Would more radical political devolutions spur growth in this country? Or is there something else? Or is it a story simply of novelty places, as Olson himself suggested, that were newer, that had fewer incumbents, that had just gone through a war? These are the places that can truly be decentralized. And anything else is just sort of the layering of interest groups on top of it. But I think we're fundamentally back to Madison and Hamilton. Both state capacity and social networks appear to be a phantom variable underpinning the success of both decentralized and centralized models. A state both able and willing to break up monopolies is critical to preventing rent seeking in decentralized systems, and so are social networks that share information. A state able and willing to pursue developmental rather than chronic aims is central to is key to centralized systems. But the question is how you get the more Hayekian outcome in terms of lack of control over information with political institutions that are able to regulate and equally to avoid centralized models. Sorry, centralized models, to avoid the problem of capture, have always relied on what Peter Evans called a form of embedded autonomy. So some circulation of ideas from below. So what allows states to adopt these roles? Is there something that we can theorize independent of the structure of interest groups? Or are we caught in a circular loop where interest group accumulation is both a cause and effect of state capacity? So second, if the state matters, which I think it really does in this argument, whether we're in decentralized or centralized models, the question is whether the Olsonian framework is enough to explain it. The strength of the book lies in the way it shows the fragility of particular growth paths. Neither bureaucrats nor entrepreneurs are the heroes or the villains of this story. Both play different roles at different points in time. They're agents of stagnation and growth. And this approach, I think, provides a powerful corrective to many contemporary debates where both either bureaucrats or entrepreneurs are cast into the role of heroes or villains, but as a provocation. I want to ask what is the underlying model of politics that allows the state to play this kind of role? As we watch decades of Further in the US building moats that prevent competition, behave monopsonistically and so on. The oldsonian critique makes sense. But when we observe de globalization, attacks on universities, tariffs, attacks on immigration, these aren't just led by firms, but also by electoral politics. And at least in democracies, this closure and stagnation appears to extend beyond just the private sector. Political scientists Jacob Hacker and Paul Pearson describe the current moment in the United States in terms of plutocratic populism, an electoral coalition between populist voters and big business that pushes back against the institutions of growth. And where did this coalition emerge? And I think one answer to this question might come from Carl's first book, which is that too little was done in terms of either trade or technology to allow more sort of embedded forms of liberalism to survive. And this I think speaks to the question of inequality that Michael just raised. But perhaps a stronger critique, and not necessarily a liberal one, is that state capacity alone is never enough. There needs to be more structural equality between labor and capital in order to prevent extractive institutions from emerging. And this is where I'll channel my inner Marxist, although I'm not one. But how do you discipline capital in the contemporary moment? Perhaps you need sort of more alternative forms of power outside of both the state and the market that prevent things like the rise of kind of stock buybacks or extractive forms of financialization and so on. And so finally, and I know we're almost out of time, I also wanted to end on the Europe question, but Michael's done a really good job at this. I guess I'm a little more skeptical than perhaps either the book or Michael is about the difference between exploration and exploitation. I think the book itself shows that there's much innovation that occurs in a gray area between the two. And so Even in the 19th century, Carl shows, gives an example of chemical invention in the UK that was scaled by large German industrial firms. But it's not clear if the invention really was that significant beyond the sort of innovation that the firms themselves did in sort of allowing this to scale. And indeed, I think comparative political economists would argue that innovation that is not necessarily just right at the frontier constitutes something beyond just catch up. It constitutes a form of innovation that can produce growth and well being and equality in some cases for people. And the question is what kind of institutions allow this to emerge, but as we've seen also what kind of how this ends, what paths sort of lie forward? So I too was thinking about Europe and this slide really points out the question of sort of, what is it that these countries and their institutions were not able to adopt in terms of computing technology is the solution radical decentralization, the breakup of the eu? Is it a sort of pushback against the limits on migration and a movement towards much more sort of openness globally and particularly in Europe, to new people? Or is it something else that fits with the grain of the way these institutions have developed over time that would allow this sort of middle level innovation to continue? So thank you Carl for this fantastic book.
- [65:27]
B
Wonderful.
- [65:27]
D
Thank you very much, Carl. I'm just going to ask you for a few quick responses if you can, and then we'll open for questions.
- [65:36]
B
With critics like this, you don't really need friends, do you? So thank you both for, for this tremendously thoughtful comments and remarks. And I wish I had taken more notes and had more time to digest and reflect on the diversity of points you raise. But let me sort of make one or two observations. So first of all, I think you both touch upon this in different ways. You can have different kinds of centralization and a unifying theme. I think Jane was right in saying, you know, sometimes Hayek is right, sometimes Gershon Kron is right, Michael Olsson is always right. And I think they're broadly speaking, two ways in which you can undercut vested interests. One of them which was acknowledged by Olson himself through his concept of encompassing coalitions. And so one way which you see in the United States in the late 19th century is through social mobilization, which pushes through the Pendleton act, which gives rise to meritocratic civil service in the US which then uses its power to implement policy and introduce new acts of legislation like the Sherman Antitrust act, which gives the state the tools to check anti competitive practices. And so in that case it's an internal social mobilization globalization story. On the other hand, on numerous occasions you see political competition being the key driver. And in the case of the Stein Hardenberg reforms in Germany, for example, which abolished guilds, which made private property more secure, which abolished serfdom and other things, they were really enabled by losses to Napoleon in 1806 in the same way that the abolishment of serfdom in Russia was enabled by the free defeat in the Crimean War. And so there are instances where the check on the composition, on centralized concentration comes from within, comes from a mobilized society. And there are cases where the state itself is still sufficiently strong and has the incentive because of global competition to check vested interest groups within. And so I think those are broadly speaking the two power hats out of that. And you might think that competition with China would usher such reforms. Unfortunately, we're seeing that it is so far leading to dismantling of the capacity of the state. State. And we're seeing that it's leading to more arbitrary interventions in the economy favoring mostly incumbents. And so there's nothing automatic about that process either. And so that I think is what makes progress in part so fragile. And having said that, once again, thank you for your very thoughtful comments and remarks.
- [69:38]
D
Thank you. We're going to open up the questions. So if you're in the room, if you're, you know, accompany your question with your name and your affiliation, if you have one, please make them questions, not life histories. That'd be the helpful. And then we'll do some questions online. So if we start off, put your hand up if you've got questions. Quite a few questions at the back. I will go with the gentleman in the pink shirt there and glasses. Oh, and then the woman at the back afterwards.
- [70:06]
B
Thank you. I enjoyed your talk. Is part of the.
- [70:08]
D
I'm sorry, name and affiliation, if you don't mind. Affiliation. I'm named Seb.
- [70:12]
C
Is part of the slowdown in productivity.
- [70:14]
D
Since the period of 1950. Simply that we sort of plucked the low hanging fruit of productivity gains since then. So we can't really electrify the grid anymore.
- [70:21]
C
Great.
- [70:22]
D
We'll take three questions at a time. So forget the back lady at the back with the. The glasses on her head.
- [70:30]
B
So.
- [70:31]
A
Hello, my name is Patrice and I.
- [70:33]
B
Am a student of the Environment and.
- [70:35]
A
Development Program here at lse. So my question is, you wrote in your preface to the book that it's ironic that the Soviet Union collapsed just as the computer age dawned, which is.
- [70:47]
C
A time when computers could have made central planning easier.
- [70:50]
A
Yet you explain that new technology often makes decision making more complex rather than.
- [70:56]
C
Simpler, especially within rigid institutions that struggle.
- [70:59]
A
To bounce its centralization and innovation. With AI now be making many decisions for us, do you think AI could.
- [71:07]
C
Blur the line between human judgment and machine control?
- [71:11]
A
And if we imagine a hypothetical like World War 3, could AI reshape how we think about global power struggles in ways that reflect your arguments?
- [71:21]
D
Wow. Okay, great question. Quite a big one. Justin, we'll go one more if we just stay where you are. Just the gentleman there.
- [71:28]
B
Hi there.
- [71:29]
D
Great talk.
- [71:30]
B
My name is Andre, former student, LSE. In 1990, Japan's global share of GDP was roughly 19%. So on a par with the US.
- [71:40]
C
Today it's less than 5.
- [71:42]
B
Where do you see China in 20 years in comparison with its focus on.
- [71:47]
D
Outputs more than productivity and its declining demographics? Carl. So we've got low hanging fruit, right?
- [71:55]
B
So I think there's something that, I mean, it's always true to a certain degree that we plucked the low hanging fruit and the role of science is building ladders that allow us to pluck the fruit higher up in the trees. And I think, you know, we've seen some of the productivity slowdown, at least stemming from the ideas producing sector. And it's completely beyond me and that, you know, if you think about it, that the personal computer and the Internet, which gave most people the world store knowledge in their pockets, aided the research process tremendously and connected some of the best talent around the world. All it led to was a decade long productivity upsurge. And we now sort of seen two decades of productivity stagnation following that, which leads me to think that there must be some institutional bottleneck and probably several of them. So even if you think, for example, that AI is going to have a real impact on medical discovery, you still have to go through clinical trials. And unless that process becomes leaner, you still have to rely on a larger pharmaceutical company to go through that process. Right. And so unless we adjust our institutions there, AI is likely to have a more muted impact than many people think. I think there's probably something significant similar happening in academia which may not have to do necessarily with the same set of bottlenecks, but with incentives. And so with productivity tools you can either do more projects or you can drill deeper into one or two things. And I think what we've seen since the computer revolution is that people tend to do more projects. And the more things you do well, the less impactful any of those projects are likely to be. I think that's part of the story as well on AI. I think the AI that we currently have is far too unreliable when it comes to situations where precedent is thin. So when precedent is plentiful, AI does very well. And so these foundation models are very good tutors, but they're not very good at cutting edge research. And so if the world was just a static distribution of events, then you could probably just brute force AI through more compute, more data at the problem, you get better performance. But the world is changing all the time. And so a key question is how well does AI perform in those new circumstances? How resilient is AI? I think the example of AlphaGo illustrates that even when you achieve super human performance, there are valid concerns of how it will perform when circumstances change. And during that lengthy also, I forgot the third question.
- [75:10]
D
Japan.
- [75:11]
B
Japan. And so I have no way of knowing where China will be in 20 years. Let me just be clear about that. But you know, they have huge demographic challenges. About a third of Chinese GDP now is real estate. So they overbuilt tremendously. And Taiwan in principle shows that, you know, it's possible for China to get to much higher levels of GDP per capita. In the case of China, we actually have a natural experiment there as a comparison. So there is nothing in principle that should prevent China from growing other than political and institutional constraints.
- [75:59]
D
So we've on to online questions now.
- [76:01]
A
Hi, I'm interrupting this event to tell you about another awesome LSE podcast that we think you'd enjoy. Lseiq asks social scientists and other experts to answer one intelligent question like why do people believe in conspiracy theories?
- [76:17]
B
Or can we afford the super rich? Come check us out. Just search for lseiq wherever you get your podcasts.
- [76:27]
A
Now back to the event. We have a question from Jeffrey Thomas.
- [76:32]
C
Who asks what will happen to the.
- [76:34]
B
Innovation frontier given we face an aging.
- [76:37]
A
Population and a possibly falling global population.
- [76:41]
C
In the next 100 to 150 years.
- [76:44]
D
Great. So global depopulation then. We'll go with another question. There was one over here in the.
- [76:49]
C
Yeah.
- [76:50]
D
With your hand half up. Yeah.
- [76:54]
B
Thank you. Thank you for your talk. I wanted to just get your thoughts.
- [76:57]
D
Ethan, you need to introduce yourself. Oh, yes, thank you.
- [77:00]
B
I'm Ethan, Ethan Dodd. I'm in the local economic development program here at lse. I wanted to get your thoughts a bit more on responding to China in terms of competition because I think we can all agree creative destruction is a force for innovation. But how to respond when there's dominance in various key industries and whatnot, that we're seeing that with a roll out of export controls on rare earths and things, you know, so you know what to do there. That's still competitive.
- [77:33]
D
One more question if we could have is that we try and aim for a bit of balance, but it's. I would go with the gentleman in the middle there.
- [77:45]
B
Thank you. My name is Jay Tyagi. I am an investment analyst at Aberdeen Investments. My question refers to particularly the last 40 years because for me a key reason for this centralization and the domination of incumbents has been global integration of capital market markets. I think the economics of funding winners and the momentum based kind of characteristics of indexes like The S&P 500 is a key factor for the reason why you get such dominant companies. And I believe that was touched upon in the critiques. So my question would be with regards to policy implementation to counter this again, do you think that the decentralization of the market, say through, for example, as an extreme breaking up of monetary unions or capital controls as an example, is going to be policy maneuvers that potentially will be taken. And secondly, all of the examples I believe that have been mentioned have been from a game theory perspective to main players. Can you think of any more multipolar examples and how that might influence the kind of conclusions if you were to say that for these kind of market environments? Thank you.
- [79:17]
A
Thank you.
- [79:17]
D
I'm going to take one more question from the woman with the check shirt. Thank you.
- [79:23]
A
Hi.
- [79:23]
B
Thank you.
- [79:24]
A
I'm a student of the Master of Economics, Environmental Economics and Climate Change. I wanted to ask, how do you think all these framework and all these concepts could apply to developing countries in the Global south, like how could they either catch up or get to innovate, how they could adjust their institutions and so on. Thank you.
- [79:46]
D
Great question. Thank you, Carl.
- [79:49]
B
All are good questions. So on the first, I mean, I struggle to see how an aging population is not going to have a negative impact on innovation. I think it's debatable how big the impact will be. And so there's one strand of reasoning which is that we have this burden of knowledge and you need to absorb a lot more stuff to push the frontiers of science and innovation today, as there is some truth to that. But I've also seen some pretty compelling evidence that people still do their most creative work in the early years. And so I do think that is going to be a significant headwind on how to respond to China. I wrote a piece in Foreign affairs recently arguing that the US should basically use the playbook that it had with Japan, which relied on integration through global value chains. It should diversify those outside of China. And clearly rare earths is a key bottleneck. There are others too. And so diversification of supply chains are going to be important. And it should double down on its strength, which is dynamism attracting talent, which is enforcing vigorous competition policy to make it easier for firms to enter and to counteract killer acquisitions in particular. And I think that goes to the other questions around policy as well. I think it's important to note that if you look at the early days of the computer revolution, every firm from Microsoft and Apple to Google and Amazon IPO'd and became firms in their own right. Since the 2000s, we've seen WhatsApp, Instagram, YouTube, numerous other firms being acquired instead. I think we need to think about how to tilt that balance. A part of the reason, I think has to do with lax merger review, we should probably look more to valuations than just revenue in merger reviews. Part of it probably has to do with compliance costs of being public firm and we should look into that as well. Final question, Lessons for For the Global South I think it goes to Michael's observation as well that you know, there are a variety of tools and mechanisms that can drive catch up growth. And so in East Asia it relied a lot on the fact that they had a long tradition of a meritocratic bureaucracy that was capable of providing public services, imperial implement policy, build infrastructure, etc. In other places like Poland it relied more on good institutions and attracting foreign direct investment for example. And so I think that that's a very context specific question and you have to do case by case, but there are some things that always help. And so having institutions that support decentralized experimentation such as strong private property rights, having a meritocratic civil service certainly always helps as well. And having decentralized system of well funded universities certainly helps too. But it's only very few countries that do have that.
- [83:39]
D
Michael Jain, anything you want to add in that case? Let me just finish up by saying a few things. First of all, if you liked what Carl was saying today, or if you didn't, you can buy his book outside. Okay. So he'll be signing books of his. Sorry, copies of his books outside. It's kind of a limited time offer because Karl has to go and do CNN in a bit.
- [84:01]
B
Right.
- [84:01]
D
So I get in the queue sort of early and sort of, you know, this is obviously high pressure selling from me, but I would strongly recommend you buy this and if you don't do it now, you won't get it signed. There is also drinks for everyone afterwards. Hopefully there should be enough. But before we finish we need to say thank you very much to three fantastic speakers.
- [84:33]
C
Thank you for listening.
- [84:34]
A
You can subscribe to the LSE Events.
- [84:36]
C
Podcast on your favourite podcast app and help other listeners discover us by leaving a review.
- [84:41]
D
Visit lseac.ukevents to find out what's on next.
- [84:47]
A
We hope you join us at another LSE event soon.