The American Innovation System’s Storied History and Uncertain Future

Nicolas: Hello, my name is Nicolas Wittstock, and I will be your guest host for today’s episode of Talking Policy. I’m joined today by William Bonvillian, one of the most eminent scholars of American science technology and innovation policy. Bill is currently a lecturer at MIT (The Massachusetts Institute of Technology) and also the author of the 2024 book, Pioneering Progress: American Science, Technology, and Innovation Policy.

Together, we’ll talk about the U.S. government’s role in driving and organizing science and technological innovation, and the challenges and prospects of American science and technology leadership in the 21st century. Bill, first of all, welcome.

Bill: Thanks, Nic.

Nicolas: So Bill, I’m sure you know science and technology policy appear at first glance like a somewhat unusual policy concern, and I’m sure for many of our listeners, it’s not even necessarily clear what this policy field entails exactly.

So let’s start by understanding why we should care about science, technology, and innovation in the first place. What is at stake here?

Bill: Well, the stakes are big, Nic. There’s an economist named Robert Solow who won the Nobel Prize back in 1987 for essentially proving that around two thirds of historical U.S. economic growth came from what he called technological and related innovation.

In other words, technology has really driven the American economy—and for that matter, the worldwide economy—for a number of decades now. In fact, really since the industrial revolution and, you know, you can trace it back to what we can call productivity gains. In other words, your ability to make more with less labor, make more with less. 

That’s what productivity is, and economic wellbeing really traces from this. So you introduce a new technology in the economy, and you’re able to make, from it, more with less—that creates a real gain, and that’s how societies grow. That’s how economies stretch and grow. So, you know, right at the heart of societal wellbeing is the ability to do technological and related innovation.

Now there’s obviously an additional dimension. Another aspect that the U.S. is increasingly concerned about, but so is everybody else, is national security. And technological innovation was absolutely critical to U.S. technology leadership in World War II, which was a major factor in the outcome of that war, and certainly also to Cold War successes.

You know, our society is founded and organized in significant part around technology advance, harkening back to the industrial revolution. Think about what those new technologies—of railroads, and the telegraph, and electricity—meant in the 19th century. And look, we’re essentially doing the same thing over in the 20th century, and we’re doing that again [in the 21st century].

Nicolas: Absolutely right. I mean, I think a lot of people think about science or, I mean, maybe the most obvious relation to people’s lives on an everyday basis comes from medicines or something like that, right? And, as you say, technological innovation is at the root of material prosperity, not just in the U.S. but across the world. Thank you very much for starting us off with that.

Just curious, how did you get personally interested in working in this area and, maybe tell us a little bit about what your professional background is.

Bill: Sure. I mean, I was trained as a lawyer, and got a job working in the U.S. Department of Transportation, really when I was a kid. And, you know, one of the things that I had to wrestle with, I worked on a lot of legislation. That was a time of big change in the late seventies, early eighties, of transportation policy. One of the things I had to work on was the Chrysler bankruptcy. And that was a disaster. [A] huge number of jobs were at stake. Chrysler had lots of direct jobs, but many more indirect jobs that it was key to. And I got to think about the transportation side, not the financing side, but the transportation side of that set of questions and learned a tremendous amount about U.S. technology, U.S. problems in manufacturing, U.S. ability to adopt innovation.

After a while, I went to the U.S. Senate, where I worked for about 15 years. Did a lot of work for the Senate Armed Services Committee, so I learned the R&D system. I learned the R&D agencies [of] the federal government. I got to work with DARPA (the Defense Advanced Research Projects Agency) pretty extensively with the defense R&D agencies, but also the major civilian side R&D agencies.

And I worked on, I would say, the great bulk of the science and technology legislation in that time period. So playing a role as a senior advisor in the Senate was a great introduction to me to technology policy. And that interest in science and tech policy led me to MIT. And there I ran MIT’s Washington office, which handles MIT’s dealings with the R&D agencies, but also brought me into a university and I started teaching and doing writing and, you know, six books later on technology policy and innovation policy issues, the most recent of which you just mentioned. So that’s kind of the saga for me that got me into this field.

Nicolas: Yeah. That’s fascinating. Thank you so much. 

So, you mentioned R&D agencies, right? I think, in a funny way, the way in which the U.S. federal government is involved in science is almost sort of a subconscious pop culture touchstone, right? Where you know, you’ve movies like Asteroid City, or, I don’t know, in limitless amounts of TV, you know, you have these like secretive government installations somewhere that are doing some sort of science or technological policy research of some kind, right? But, it’s not really clear, I think to a lot of people what exactly happens there. To what extent policy is important to technological innovation in the United States—or anywhere, right?

And what, really, the role is of government in enabling innovation in any way. Right? Like, when you describe federal R&D agencies, what exactly is happening there? Maybe give us a sense of how the government influences innovation by giving us some historical context about that in the United States.

Bill: Sure Nic. It’s an interesting saga. I hope I can make it somewhat intriguing, but the extent of the federal government’s role in science and technology really can be in large part traced to World War II and having to mobilize to cope with that war.

Going into that war, the U.S. was not the science leader. Countries like Germany and Britain were the science leaders. But going into the war, we had very strong industrial capacity, but we had to build that science capability in the course of the war.

Nicolas: And excuse me, when you talk about science capability or that Germany was more innovative at the time, what does that mean? 

Bill: In terms of basic science—in areas like physics and chemistry [the] U.S. was not the basic science leader. I mean, we had important chemists and we had some important physicists, but these fields really were dominated by other countries, and particularly in the U.K. and in Germany, particularly in both those sectors that I just mentioned.

So the U.S. had to build that capability in the course of the war. There was a big influx of European scientists in the United States, immigrants. That was a huge help. So we built technology projects, and those in turn created the models for the kind of innovation system we would build.

We built something called the Rad Lab, which innovated a lot of radar advances. The British came up with the original ones, but the U.S. added a lot of innovations on top of those through its Rad Lab based at MIT, and it brought in engineers and scientists from around the country. And that work on radar really led, frankly, to the whole creation of the American electronics industry in the immediate post-war period. Obviously we innovated nuclear power, but there were major advances in aviation, both in Europe and in the United States. There were the beginnings of space rocketry in the course of that war. So there were lots of technological advances that got pushed by the war, but as importantly, the organizational mechanisms for how to organize science really came out of that war and are frankly still with us. 

So, for example, that Rad Lab model brought in really top scientists—I mean, 11 Nobel prize winners actually worked at the Rad Lab in the course of the war. Some of this was modeled on what Britain was doing, because we copied some of their models, but they were flat, non-hierarchical institutions. This was not a military bureaucracy. This was a very flat organization of scientists where they could really exchange and collaborate with each other in a non-hierarchical kind of way. And that turned out to be optimal, right? We got a lot out of that model, and we essentially adopted a lot of that model, as the war ended. 

What happened, as the war came to a close, was that the U.S. dismantled its military. We brought 16 million service people, that were serving around the world, back to the United States. It was a huge issue as to whether anybody was going to have a job, right? But President Roosevelt’s science advisor during the war was named Vannevar Bush, and he wrote for Roosevelt a book called Science: The Endless Frontier. And his idea was, you know, science is going to be an underpinning for this technological advance that we can have in that post-war period. And he created a system that essentially said, let’s invest in basic research. We’ll leave it to industry to do the follow-on steps, right? To actually get a technology into use. But let’s have the federal government invest in basic research. And during the course of World War II, the U.S. had really created federally funded research universities. 

MIT received, in four years of the war, 80 times more funding than it did in all of its previous 80 years of existence. Vannevar Bush saw the power of that model, the ability to put universities to work on this innovation system, and decided, let’s hang on to that, right? He wasn’t worried about manufacturing, ’cause the U.S. by the end of the war, was absolutely the king of mass production worldwide. We didn’t have to worry about that one. So he created a civilian R&D system that evolved over time. The National Science Foundation (NSF), the National Institutes of Health (NIH), the Science [Office] at the Department of Energy, a series of basic research organizations that do fundamental science.

But by 1950, the U.S. finds itself in a Cold War and the Defense Department has to go back to the system it was using during the war. And during the war, the Defense Department would invest in the research, it would invest in the development, it would invest in the prototype, it would invest in the demonstration, in the testing. It would fund the initial production and often create the initial market for a new technology. That connected model actually turned out to be of huge benefit. So out of that connected model being run on the defense side of the innovation system, we got, you know, all kinds of advances in aviation, space, nuclear power, lots of electronics, computing, the internet, all those things came out of the defense innovation system in that post-war period. And so that’s kind of, you know, a rough shot at, at some of the history here. 

You know, you also asked what are some examples of what we’ve got and how it works. Well, let’s take an example of semiconductors, right? The integrated circuit was the critical advance in semiconductors. And when the two great innovators, Robert Noyce at what became Intel—a company called Fairchild, later moved to Intel; and Jack Kilby at Texas Instruments. They developed the first integrated circuits. There was no civilian market for integrated circuits for four years after they came up with this innovation. NASA and the federal government’s Department of Defense, they both cared about the ability, this was directly related to making missiles and rockets, right, having onboard efficiencies for those systems. And so they were the buyers, right? They created that market. They helped drive the technology. 

On the health side, that’s another big story. I haven’t talked about that side yet, but, you know, recent stuff like CRISPR. That came out of the National Institutes of Health, NIH funding to the University of California, Berkeley, and MIT. That was basic research that developed this whole new ability to work with genes in a new kind of way, which we’re still experimenting with. 

Ozempic, right? The diabetes and weight loss drug. Where did that come from? Well, NIH-funded basic research, interestingly, by a scientist at the Veterans Administration named Jonathan Eng. I mean, just to throw a statistic out, between about 2010 and 2020, roughly something like 350 drugs were approved in the United States, and virtually all of them can trace their origins to federally funded research. It’s a remarkable story. 

I mean, recently, mRNA vaccines kind of saved the day in significant part from the recent pandemic. DARPA research helped on figuring out how you could stabilize mRNA so you could actually work with it. NIH Research helped on understanding how to stabilize the spike protein on the coronavirus, and when you put the two together, you can have a vaccine. So again, lots and lots of health advances. The vast majority of health advances in the last half century have really traced their roots to federally funded research. 

Nicolas: Yeah, absolutely. These are great examples. It’s really crazy when you look around and you try to think like, it’s basically impossible to look at your living room, right, and not, you know, feel the incredible influence of big federal government research initiatives.

I mean, the internet, given that we’re recording this conversation over Zoom, being like a really important example for sure, right? Even Google… I forget which agency; was it a DARPA grant?

Bill: It was NSF. Yeah, the search engine research funding came out of NSF, initially.

Nicolas: Oh, there you go. Yeah. It’s crazy. It doesn’t end. It’s really been incredibly influential. And more recently, as you say, as well, it hasn’t really stopped. 

However, arguably one would say that maybe that was a big mobilization during the Cold War, but we’re no longer in the Cold War, right? And, once you poke a little bit more, government funding relative to private R&D funding has dramatically decreased pretty much since, I would say the 1960s. So, pretty precipitously. Is the government still all that important today?

Bill: Yeah, and the reason why it is, Nic, is that industry takes on efforts where it can realize a gain, and therefore industry spends money really on the development side. The predominant amount of money  for research is on the government side, and particularly early stage research is just too far away from the marketplace. It’s just too risky. You don’t know what you’re going to come up with. There’s [a] tremendous amount of trial and error. You can spend a substantial amount of funding on basic research and you are really not sure where the results are going to lead you to. So industry, historically, has got to work on a timeframe no more than a few years long at most. Basic research is a much longer term project. We’re talking 10 years, at least, if not considerably longer. And industry is just not organized financially to be able to structure that kind of fundamental research. 

So that’s known in economic circles as a market failure, right? [The] market is just not going to intervene to support basic research. And that in turn became a governmental function, a governmental role. And that’s true in, you know, virtually every country around the world at this point. But you don’t get development unless you get research. Right? 

What the U.S. has been doing is we created an enormous amount of research advance in the last, say, four decades, five decades. And that’s been driven through to the early stage development side through federal research support. So industry has got a lot of stuff that it’s been able to invest in, hence enabling that huge increase in industry investment.

So the government puts a lot of stuff on the table. Industry has picked a lot of it up and moves it to that demonstration side. And so industry’s investment has been rising, but its [role is] predominantly in demonstration. Unfortunately, the federal investments in research have been really pretty stable, and as a share of GDP, as a share of the economy, they’ve actually been in significant decline for several decades. So what you want in a powerful, innovative economy, you want two parallel lines. You want a rising line in research investment. And then on top of that, you want a rising line of increasing development investment by industry. And the two sides play off each other, they’re completely interrelated. If you start slacking on one, like pulling down on the research side, industry has got much less to work with. So it’s eventually going to pull down your development side. 

Nicolas: And that’s going to lead to, you know, at the end of the day, less economic growth. Because we have fewer innovative ideas that we can build new products around that are going to hopefully make our lives better in some way. 

I would say, like, this is this sort of, like, very stark distinction between basic research funding by the government, and then basically, as you were putting it, the government puts things on the table for business to pick up and commercialize. That’s sort of the way that this is typically framed, I would say, in neoclassical economics, that there is a market failure that the government comes in and solves. 

Now, I would argue that governments often do significantly more than just putting things on the table. And, in your writing, you have also repeatedly pointed out that there are certain weaknesses in the American innovation system, specifically on the civilian side, and that there is arguably a need to reform some of those weaknesses to be able to take advantage of new opportunities or rise to new challenges.

There’s always been a big difference between the civilian side of the U.S. innovation system and the defense side. So could you elaborate on that a little bit?

Bill: Sure. The successful technology advances of World War II came out because we were creating what I call a connected innovation system. In other words, there was very close collaboration between government and government support and industry and universities. They teamed up, right? And they were very connected with each other, and the government funding, the university funding wasn’t just in research. Instead the team went to work, and research was undertaken, but also the technology development and all those kind of follow-on stages of, you know, development, prototyping, testing, and demonstration through initial production, all that kind of occurred. And look, we kept that system. You know, the United States has a theory that the government ought to keep its hands off except on defense research. Right? And there, the Defense Department, as I said before, plays a role in this highly connected system still.

And again, what do we get out of that system? Frankly, we got most of the advances of the second half of the 20th century in technology. So it’s not just enough to invest in research. You’ve got to make sure that there are the connections between the follow-on actors, right, so that you’ve got a set of collaborative networks that are going to pull off the technologies that you’re going to need.

And if you’re just investing in basic research, let’s draw a picture: The universities, they’re like monasteries. They’ve got walls around them, you know, like medieval monasteries, and they’re researching away with some federal research support. Every once in a while they’ll write an article and they’ll throw it over the wall, and they’ve got to hope that some capitalist, coming along in a stretch limo, is going to stop the limo, get out of the car, pick up the article, and actually read the thing. Obviously exaggerating greatly, but that’s the civilian R&D system as opposed to the defense R&D system, which is much more connected. You know, and again, that’s an exaggeration. But a better connected system will lead to better results.

You know, I spent a lot of time, Nic, worrying about U.S. manufacturing. So the U.S. used to have, in that post-war period, we used to innovate here and produce here, and we got the gains across the entire spectrum. Then we figured out how to innovate here and produce there. And the problem with that is that there’s an enormous amount of innovation in the production side. In the manufacturing side, you have to take a scientific idea and essentially turn it into a product, and that requires a tremendous amount of very creative engineering. You’ve got to design the product for quality, for resilience. You’ve got to design it to meet a certain price point. It’s a very creative process, and then often you have to rethink a lot of the underlying science to make a technology into a new product. 

So that manufacturing stage is very creative and very important, but the U.S. hasn’t seen it that way, and so we divorced any attention to manufacturing from our innovation system. Federal government research doesn’t support research in manufacturing processes and technologies, and that’s a gap in the innovation system, a big gap.

Nicolas: Yeah, so you have argued, or would argue, that the decline in U.S. manufacturing capacity, and competitiveness in U.S. manufacturing across the board, does constitute in many ways one of the main technological challenges of the 21st century for the United States, right? Because it’s related to all kinds of other issues.

Competition with China is increasingly seen as a major technological problem. And, on top of that, I think both the pandemic as well as the war in Ukraine have shown that there is a tremendous national security benefit to being able to produce domestically, things that might seem a little bit more mundane, right? Like, tank shells, artillery shells, things of that nature. Or even just PPE, right, when we’re talking about the pandemic. I think just to really drive home this connection between technological competitiveness, technological innovation and manufacturing capacity, the point that you’re making here, right, that these are ultimately part, or in the U.S. case, that there’s a hole in the technological innovation system, because there’s much too little emphasis on ensuring that there is domestic manufacturing capacity is, I think, a point that resonates quite widely.

And to your point, that innovation doesn’t just flow one way, right? It’s not just the case that if you invest in R&D and put it out, firms will bring it to the market.

But if you completely outsource manufacturing, the concern is that that reduces your ability to essentially do what people call learning by doing, which has a workforce development component, for sure, right. If you’re not manufacturing in the country, then you’re not essentially training a workforce on the job. But beyond that, on the application side of technology, on the engineering, the manufacturing side, that’s where a lot of creativity and innovation comes out on its own as well.

So it’s not necessarily something you can just cut off or, I mean, that’s the argument, at least, right?

Bill: Yeah. Look, I mean, once you lose the knowhow and how to make things, it’s very hard to get that back, so you can’t just kind of magically reshore technology production.

You’ve got to get into the game and get into the technologies and processes that bring you, I would argue, to these new advanced manufacturing capabilities. You know, we’re doing a lot of tariffs, but tariffs are a short-term measure. In the end, you’ve got to be able to compete better. 

The U.S., as I said, has had an R&D-led innovation system. Other countries have had manufacturing-led innovation systems. So Germany and Japan, for example, coming out of World War II, they had to get people back into jobs. So they focused their innovation systems on manufacturing. They had to get production back, and they built extremely strong manufacturing production systems, which, you know, remained with us to this day. But other countries adopted the same model: Korea, Taiwan, and now China at enormous scale. And we’re watching India start on the way. 

There’s no reason why we can’t do both. And look, we’re still the second largest manufacturing power in the world. It’s not that it’s completely vanished, there’s still a huge amount left here to build on it. So we’ve got a lot of sectors where the U.S. still has a very strong role, where we can build on that role and fill future niches, and then use our innovation capability and new critical technologies to help move those out with an improved manufacturing system.

Nicolas: So we’re recording this conversation at the end of March, 2025. Help us maybe to contextualize some of the recent policy developments in the United States and how they pertain to some of the topics that we’ve discussed here, specifically. You know, the capacity of the United States to remain a science and technology innovation leader internationally. But also to meet some of these specific challenges related to technological development, innovation, productivity. Specifically maybe in manufacturing but also, in general, the U.S. capacity to be a leader in those areas, in the context of the second Trump administration openly describing, but also acting, you know, on the desire to dramatically curtail the federal administrative state in a variety of ways that could potentially impact the capacity that is related to science, technology, and innovation policy.

What might the impact be here in your assessment?

Bill: Yeah, look, I mean, it’s still early days. The U.S. has built what I think is still the leading early stage innovation system, the R&D side of the innovation system. It still remains the world leader in that, that’s incredibly valuable, and, you know, systematic or chaotic cutbacks in that system worry me, because it’s a very strong foundation and we’re going to have to build on that foundation, if we tackle what needs to be next, which is, I would argue, a very hard look at the manufacturing sector and an effort to put advanced manufacturing, new productivity, new efficiencies into manufacturing. If you don’t have a strong foundation to build that next stage on, you know, we’re going to be in trouble. 

And then, you know, I’m concerned about a “tariff-only” policy. I think tariffs have their uses, but as I said before, a “tariff-only” policy only gets you so far. It doesn’t get you back into the competitive game where eventually you’ve got to be again. For example, when the U.S. imposed tariffs for the steel industry, about six, seven years ago, the steel industry was able to kind of retain its market share, but it actually declined in productivity. It wasn’t being forced to stay competitive. And you’ve got to keep these industries competitive, rather, you can’t just wall them off in an economy that inevitably is going to be highly global. 

You know, industries, multinational companies are successful because they’re able to compete and build their product bases on every continent. And if they can’t do that, they’re inevitably going to be weaker firms and less capable firms. So, in the end, we are a highly connected world. There’s no getting around that. And despite the talk of the end of globalization, the U.S. is running a $1.2 trillion deficit in manufactured goods at this point. So we’re trading more than ever. And that’s not going to get eliminated overnight. It can get chaotic, but the world is going to remain an increasingly connected world, and we’re gonna have to learn how to compete in that world. And I would argue that investments in advanced manufacturing, and new technologies for manufacturing, and new processes for manufacturing, are a key, as you put it earlier, Nic, a big hole in our innovation system. We got to move to fill those. And we’re not going to do that from tariffs alone.

Nicolas: Yeah, absolutely, right? I think it’s somewhat tragic. I think there’s a very persuasive case to be made, and you’ve made it here, you’ve made it in your writing, that what is really needed—or one thing that could for sure invigorate a lot of U.S. manufacturing specifically, but also address a lot of other national security issues, or issues related to emerging technologies—would be more industry-oriented innovation policy of some kind, right?

To maybe oversimplify it slightly, you need more government. You need, essentially, more spending on R&D for sure. But you need a more muscular government that is involved in a lot of these efforts at a closer level. But at the same time, if anything, we’re moving in the opposite direction right now. So I think there’s a political, cultural challenge here. 

Bill: Let me just give you an example. How did China go, in the space of about 30-35 years, from say five percent of world manufacturing to 31 percent, by far the largest share? China focused on scaling up its production and it provided a tremendous number of tools and incentives from the government to enable that scale up of production. China spends, according to a CSIS (Center for Strategic and International Studies) evaluation, about 500 billion equivalent in U.S. dollars a year in scaling up its evolving, emerging new manufacturing companies.

Nicolas: So this means small companies, as you were describing, trying to take things off the table and bring them into the commercial realm.

Bill: You got it. In this scale-up process. In other words, moving from a prototype into an actual good, produced at large scale. That’s a huge, complicated, difficult, expensive process.

China spends about $500 billion a year on that, and then in addition, it’s got something called guidance funds. Somebody got the idea, “Hey, maybe the private sector might be better at doing these investments than the government, so let’s put private sector folks in charge of creating funds where there will be government seed money.” And these guidance funds will invest in a host of what the government has identified as critical technology areas, and they’re the same critical technology areas we’ve got. 

There’s just no equivalent for this financing system. Now look, we’re not going to do China state capitalism in the United States for sure, right? We’re not going to adopt that model. It’s not going to work here. But we also have an incredibly powerful and rich set of capital assets in the United States. We have huge capital potentially available for investment. 

How do we incentivize and motivate that capital to invest in this critical problem we’ve now got in our innovation system, which is the decline of U.S. manufacturing? There are ways we can start to incentivize those kinds of investments, and we’ve got the capital to be able to do it. And the government can create incentives to encourage that capital to tackle this scale-up-of-manufacturing problem.

Nicolas: Yeah, and just to be clear, you were mentioning that the U.S. spends significantly less on similar endeavors, but none of these kinds of initiatives are unheard of in U.S. history, right? There are a lot of success stories to point to of similar kind of efforts.

It seems to me that the main barrier here in this context remains, in the United States, what I would describe as, like, a political/cultural problem, right? That there’s a sort of a barrier, because in part I think the history that we’ve discussed here is not that widely appreciated,?

There’s this, I would describe it as a political/cultural aversion to these types of policies. Would you agree with that assessment?

Bill: Yeah, I mean, look, the U.S. is a very innovative, very entrepreneurial society. So the culture is a great one for creating new things, as we’ve demonstrated for, you know, three centuries I think so far. So there’s a huge amount of potential here. 

And there are good important qualities here. The country is, you know, somewhat adverse to governmental regulation, which is frankly a positive for a lot of business growth and expansion. We created an incredibly strong education system, over a significant period of time. Mass higher education through the public university system has been key here, so there’s been a strong talent base. So there’s a lot of great cultural features that the U.S. has got to play with. 

You know, there’s been resistance to a governmental role in the later stages of economic development. But the government has played a critical role in many technology advances since 1780. So the governmental role in innovation gets traced back to Alexander Hamilton and the beginning of the country. It’s been ongoing, largely through the Defense Department side, but we’re going to have to broaden that base. It’s real. It’s part of us, right? It’s part of the culture. We’re going to have to broaden it beyond just a narrow national security outlook. Because, as China has been teaching us, economic security is directly related to national security and you can’t sever those two.

Nicolas: William Bonvillian, thank you so much.

Bill: Nic, thanks so much for hosting me here. I really enjoyed the discussion.