Fuqua Insights Podcast: Why America Needs Both Startups and Big Firms to Innovate

Sarah Kern  00:03

Welcome to Duke Fuqua Insights, a podcast where we explore faculty research and the actionable takeaways for business leaders at every level. What if the biggest driver of innovation isn't just funding, but the promise of future customers. Research shows that when companies expect guaranteed demand, especially from government contracts, they invest more in early-stage science. But that boost isn't evenly distributed, and it doesn't always translate into commercial products. I'm Sarah Kern, a current MBA student at Fuqua, and I'm joined by Professor Sharon Belenzon, a strategy scholar at Duke and a faculty research fellow at the National Bureau of Economic Research. His work focuses on innovation, corporate strategy, and the role of science in business. Professor, I'm looking forward to diving into this research. Thanks for being here today.

Sharon Belenzon  01:00

Thank you for the invitation.

Sarah Kern  01:02

Of course. So just to get started, I'm wondering what motivated you to look at government procurement as a driver of innovation, and can you clarify what government procurement means in that sense?

Sharon Belenzon  01:15

Yeah. So I think we all believe for many years that companies, if we just let companies invest themselves in R&D, they will not invest enough, and this is for different reasons. But the key reason in the economic literature, is that when a company creates knowledge, this knowledge benefits many other companies or benefits society at large in ways that does not allow the company that created the knowledge in the first place to capture enough returns. So companies will naturally always underinvest if we just let market forces dictate incentives. So this is like this has been like in the kind of key policy in the policy arena for like decades. But most of the academic literature focused on the role of the government as funding innovation. Essentially, it's called like more like push strategies. The government is pushing companies and universities to do more research by giving them resources by giving them money. While universities need money from the government because they are cash constraints, big corporations, especially in the U.S., it's hard to believe that the way in which governments can affect their incentives to innovate is giving them more money. Because if they need money, they can go to the stock market and raise money from public shareholders. So we have been thinking a lot, looking at historical evidence on what happens, like you know, after the Second World War and even before, what drove American firms in what we call the golden age of R&D, which is roughly after the Second World War until like 1970s and 80s, what drove these companies to essentially solve so many big problems, technological problems for the government, which essentially created the technological foundations of almost everything we have today, including like you know everything we have today. So Bell Labs is the most famous example. And when we look back, we realize that when you look at case studies, it wasn't really the money that the government gave for like paying for the research, but it was the promise that if you provide a scientific and a technological solution to a problem that the government needs, the government is going to reward the firm by giving it access to a very very big and lucrative downstream market for the underlying products and services that come from those innovations, that come from those scientific kind of solutions and technological solutions, and I can give you lots of examples. But this kind of idea that like “pull policies,” essentially kind of pulling companies towards specific solutions instead of just giving them money and hoping for the best, has not received a lot of attention in the literature, mostly because it's hard to understand what's going on, and data was not really available a lot. So one of the key contributions that we've made is to clarify what is, as you mentioned, what is an R&D contract from the government, what is a procurement contract, etc. So just to kind of answer your question, a procurement contract is something the government buys from essentially the private sector. Now, interestingly, the government can also buy R&D services. R&D services or R&D contracts are very different from grants. When I get a grant from the government, I write a research proposal. They give me money. I do the research, and that's it. When you get an R&D contract from the government, you are essentially signing a contract to deliver something that the government needs, and you have to meet specific milestones. And it's a contract. If you don't deliver, then you know they can punish you. And then we kind of thought about really deeply about what is like, what is the economic role? How do R&D contracts incentivize firms, and we can learn that R&D contracts, in my opinion, are one of the key mechanisms by which governments cause or lead or incentivize companies to solve the big challenges that that society faces. And I can explain more if you like.

Sarah Kern  05:17

Thank you. Is your research centered in certain countries?

Sharon Belenzon  05:21

So I focus on the American innovation ecosystem. We have a book coming out that looks at the past 150 years of American innovation. So I think today I learned a lot about how the American innovation system works and how it has evolved over time. What I can tell you is that what I learned is very specific to the U.S. and I know nothing about other countries because it would require like it took me like maybe I think--well I started to work on it in 2005 honestly so a long time—two decades just to kind of really understand some facts and document important facts about the U.S. and we still have so much work to do. So understanding how innovation ecosystems work and what is the role of different actors from the government to startups to big funds to universities is extremely context dependent. Really hard to do and measure at the level of detail that you need. And hopefully somebody is going to do the same for other ecosystems like Europe, maybe Asia. Asia is really interesting, very important, clearly, and other countries. But my expertise is solely focused on the U.S.

Sarah Kern  06:31

Your research explores the idea of guaranteed demand. For listeners unfamiliar with the term, can you explain what you mean by guaranteed demand?

Sharon Belenzon  06:39

So honestly, not many people understand this concept, also in academia. And the basic idea is, which is really important because it's actually part of the law. That companies, or the government can essentially have two different ways in which it can solve problems. One is it can give money to amazing researchers, research organizations, and say develop a specific body of knowledge that is relevant to--so let's take an example: the human landing. We just kind of just came back, like NASA came back from the moon. One of the key technological problem we have is a human landing system, so NASA needs essentially companies in the U.S. to develop a human landing system that will help kind of humans land in the moon on the moon and later on Mars. So there are two ways to do that. One is let's give money to organizations that specialize in research and they are going to do the science that we need to understand in order to develop technologies that would allow humans to land on the moon safely. So we give lots of money to universities, let's say billions of dollars, and then this knowledge becomes publicly available. And then we have another kind of contract. We have a bid or a contract, a procurement contract, where the best manufacturing companies in the world are going to take this knowledge and embed it and implement it in technologies that will eventually result in the HLS, the human lending system. This is one, and this is the most efficient way to do this, honestly, because it allows different experts to participate in different parts of the innovation process: the best researchers do the research, the best manufacturers do the manufacturing. In practice, however, it doesn't happen a lot. What happens is something very different, and this has happened, for instance, like one of the classical examples are the essentially integrated circuits, or like the version of the memory chips or semiconductors. The early applications were the Apollo program, and the only consumer for the early kind of integrated circuits was the government, and the government wanted to buy those integrated circuits because they had to put it on missiles to essentially to kind of fuel the Apollo program. Now the interesting part is that the way in which it was done was not by giving money to separate kind of parts of the ecosystem, participating in the ecosystem, but giving lots of resources to companies that did R&D while promising those companies explicitly, not contractually, not in writing, but making them the promise that if they deliver the best scientific knowledge, let's say if they deliver the best pilots or the best kind of if they meet specific very high-level technical specifications, they are going to be the only the only parties, the only companies who would be allowed to bid on a very lucrative market for essentially integrated circuits, which are going to be placed on the Apollo systems. So the price of the innovation was the downstream of the market itself, and this was enough to incentivize companies to invest massive resources in upstream R&D. Which raised, really complicated questions, which is why would the government find it optimal to give essentially a single R&D contract to the best, to essentially to one firm that is also going to essentially solve two problems, one manufacturing and R&D and not break it up by essentially giving the R&D contract to the best researchers and the manufacturing contracts to the best manufacturers, and it's a difficult problem to understand. And it took us a long time to kind of think about that. And it's hard to explain, like you know, to be honest with you, I can explain the intuition we came up with, but it's not straightforward, but the important point is that something that seems obvious that we are going to use a prize to incentivize innovation, which has been done historically, became essentially potentially illogical in the sense that it deviated from what we believe to be efficient, and the important point that people started to realize is that by not opening up bids for competition on these massive procurement contracts, essentially giving the contracts without any competition, just to a firm that kind of won the R&D contract, is potentially inefficient. Maybe corruption--I won't say corruption--but it's inefficient and can lead to high costs. It's unfair. It's not transparent. And starting in the mid-90s, under the Clinton years, the policy changed, and we started to see a big shift away from awarding those procurement contracts non-competitively into awarding them competitively, which means this kind of explicit promise, which I'm going to give, which means I'm going to give those downstream contracts to the firms who are going to give me the best R&D solutions. This link broke, and we argue that this had implications to the incentives of companies to engage with the government and try to solve the big problems they had--the government has.

Sarah Kern  12:06

So how does it work now?

Sharon Belenzon  12:09

So the HLS (Human Landing) system, which is the example I gave you, is a fascinating example. And here you have like intense competition between private companies. Actually, one of the interesting parts is like SpaceX currently won the big contracts for procurement, but one of its competitors actually petitioned to Congress saying it's unfair that you give the human landing system just to one firm, and they actually said that they are willing to invest $2 billion of their own money, own money to do the R&D. So not only that the government doesn't need to pay for the R&D, the company said, "I'm going to pay for the R&D myself. Now why? Because the company said by investing in R&D, I may have a chance to win the downstream contracts. So we do see that like at least in this example with space, with this kind of aerospace technologies, we see lots of this going on today. But in general, we've seen a decline in the use of those essentially bundling of procurement contracts with R&D contracts. We've seen an erosion in this idea of guaranteed demand, and this erosion kind of took place mostly for reasons of efficiency, trying to save costs because it's always more efficient to break down this contract and give the R&D contract, give the R&D contract to the best R&D firm and give the manufacturing contract to the most efficient manufacturer. But there is an economic logic that under some conditions, what we want is we want to kind of put them together, and when the economic logic is that the only way to kind of incentivize R&D is when we put them together, by not doing so, we are going to essentially break the system. And we have identified the conditions under which it is actually optimal for the government to bundle R&D contracts with procurement contracts. But unfortunately, I don't believe that policy currently distinguishes between those conditions. 

Sarah Kern  14:07

So economically, it makes sense for them to offer guaranteed demand, but in terms of efficiency, if they were to unbundle production and R&D…

Sharon Belenzon  14:17

Actually no. So this is not what I'm saying. I'm saying that under some very specific conditions, it is optimal and efficient for the government to bundle. But these are very specific conditions, and those conditions are when the firm that does R&D has also manufacturing capabilities. Essentially, when the firm that does R&D is also big, only under those conditions it makes sense to bundle. Any other condition, essentially, it's much more efficient to have two separate contracts. So, size in this respect, being big, is the only time where it is optimal for the government to actually have this bundling system, where you have guaranteed demand and you don't have the more efficient breakdown of activities.

Sarah Kern  15:05

Got it. So in the current system, when a company wins one of these government contracts, where do they tend to allocate those additional resources?

Sharon Belenzon  15:15

This is a great question, and honestly, I do not know. We don't know yet, so I'm working with some colleagues here. So my PhD student did amazing work on something we actually don't know nothing about, which is thinking about startups working for the government. So she's kind of asked the innovation ecosystem, which kind of has this contractual relationship with the government. We know that big firms have contracts with the government, and they do important things from like integrated circuits in the past to like human landing systems today. But what about startups? We tend to think that startups go and solve problems in the private market. But she showed that many startups, the first customer is the government, and then she goes and asks, you know, what are the benefits of working for the government, and to what extent those benefits translate into civilian applications, into things that is happening outside of the government market. And what she finds, which is really interesting, actually, what I'm not going to talk about what she finds. I'm going to talk about what you specifically asked about the benefits. She finds that essentially those companies that choose to work with the government follow a different path than those companies that choose to work with venture capitals, and this difference actually is something that stays and is persistent. Like this kind of difference is persistent. In other words, what you do for the government doesn't seem to allow you to pivot and essentially work with venture capital afterwards. And it's really interesting. So I find it fascinating to ask, like, what are the different--why those different kind of pathways exist for, and how companies that try to solve different types of problems find it optimal to kind of choose one or the other. Specifically, what are the benefits of going with the government, like working for the government as your first customer versus raising money from VCs and trying to go after private markets. So we don't know much, to be honest with you. We don't know much about, and this is why I think it's so important. Because if you look at numbers, the numbers, like the amount of money the government spends on procurements, on old procurements, but even procurements that involve technology and science are way, way, way larger than any amount of money the government gives for R&D in any form. So understanding how this money actually is used and spills over to different sectors of the economy is, in my view, a first order question.

Sarah Kern  18:00

Earlier, we were talking about the difference between big and small firms. How this research shows that large firms tend to respond much more strongly than smaller ones when it comes to procurement deals. What's driving that difference?

Sharon Belenzon  18:16

So this is the economic logic, and let me kind of explain it in the simplest way possible. So the key choice is between having two contracts. One is R&D, and the other one is manufacturing, versus one contract, just R&D. Those who win the R&D gets both the R&D contract and also the downstream manufacturing contract. Now, why would it matter whether the government does one or another? Like, in principle, it should not matter at all. Let me give you an example. So suppose you are a great researcher, and you are going to solve a great problem. But you don't know how to manufacture, you don't know how to produce. But I need you to solve something, and then I need you to kind of take this knowledge and implement it into a product which I need. So I can do two things: I can give you Sarah the R&D contract and pay you for that, and then you give me the knowledge, and then I run another auction to find the best manufacturing firm who can use your knowledge and build me a plane, all right? Or I can say, you know, Sarah, you want the R&D, you get both contracts, a bundle contracts. What would you do? Now you do not know how to manufacture yourself. So what would you do? Exactly what I do as a government. You are going to run the same auction exactly, and you're going to get exactly the same best manufacturer, and you're going to give them exactly the same money that I would have given as government to the same company. So whether I run the second auction or whether you run it, we are going to get the same solution. With one exception: if you can manufacture as well, you will get a better outcome than me when you run the second auction, because under very general conditions, if I run the auction as a government, I will get the most efficient firm, and they're going to charge me a high price, you on the other hand, if you run the auction, even if you're not the most efficient manufacturer, the most efficient manufacturer cannot charge you a very high price because you will say at some point, you know what, forget about it, I'll do it myself, given that price. The other firm knows it, so they're more limited in their bargaining power when they're dealing with another firm that has manufacturing capabilities, so the interesting part is that if the firm that does R&D is also big, can also produce, it can actually negotiate and get a better deal downstream. When they essentially, even if it's not the one who produces the plane, they can get a better deal, better deal with the company that produces the plane. Now, the government knows it, so the government needs to pay less money for the R&D for the firm to do it because the government knows that the firm that does the R&D will also get some additional profits downstream that the government itself cannot get. So, thinking about who has the comparative advantage in running the second auction, the government or the firm is key.

If we are dealing with a small firm, there is no difference between this small firm and the government. They're going to get the same solution. But with big firms, what gives us the same level of quality but low cost is to give the big firm also the ownership of the downstream contract, which is to some extent maybe not what many people want to see. They want to see more fairness, like more let's say less concentration of contracts within the hands of few firms. But in this specific case, and historically we see many examples of that, big firms have a natural advantage in being more efficient in how they essentially manage those two contracts.

And let me just say one last thing: the instinct is that this efficiency benefit comes from some knowledge. It's easier to kind of transfer knowledge within the same firm, or some benefits in the production function. It's not. It's really about the fact that you can get a better deal downstream, and this is the logic of what we argue, the simplest explanation possible, without any complication, and the thing which I think is really important to understand is that when the government has been pushing for fairness and transparency, in my opinion, this logic, this simple logic which I explained, that sometimes big firms are more efficient in managing those contracts, I think, has been lost or maybe poorly understood from the beginning.

Sarah Kern  22:48

So, with those complexities in mind, what do you think policymakers should take away from this research when it comes to designing these kinds of innovation incentives?

Sharon Belenzon  23:01

So, I think that we don't like big firms in the U.S. Let me just first say this: we don't like big firms. I think it's a mistake. I think to solve big problems, we need big firms. I think we should distinguish between big firms that help us solve big problems and big firms that do not. So I think policymakers should not kind of brush off all the firms or essentially view all big firms the same. We should try to think very hard about the social benefit we get from the R&D programs of large firms, and examine each case for itself, and not just put all big firms in one bucket and just make the assumption that because there are costs, and I'm and I'm the first one to say that there are costs associated with big firms--I'm not saying that big firms are you know are costless--clearly, there are costs and there are negative things associated with having very big firms in an economy. But at the same time, not all big firms are the same, and the specific and in the specific context of an innovation ecosystem, it is clear that some problems cannot be solved by startups alone, and we need to have those big firms as anchors in the innovation ecosystem. But again, it's not an argument about all big firms. It's an argument about us as a society, as policymakers, having the responsibility to identify the scientific benefits that we get and the technological benefits that we get from different types of firms, even especially the big ones. So this is kind of my message. I think.

Sarah Kern  24:33

Before we wrap up, it's important to mention that government-backed innovation is just one piece of a much larger innovation ecosystem. You're exploring this broader system in your upcoming book. How does this research fit into a bigger story about translating science into impact?

Sharon Belenzon  24:51

Yeah, this is a great question, and thank you for raising this about our book. So we have been working on this book for the past, I would say, 20 years, and the key argument, I can say, I can summarize it in a single sentence, that in order to understand the weaknesses and the strengths of the American innovation ecosystem, we need to really understand how this ecosystem is organized and how this organization has changed over the past many decades--I would say 100 years, which we analyze in the book. The key point is that in the past four decades, the American innovation system has transitioned from being what is called vertically integrated, essentially few actors performing many different tasks within the ecosystem, into a system which is very fragmented, where different actors perform highly specialized tasks, and those tasks are connected to each other through markets, through the market for technology, and we argue that to understand the challenges that we face today, specifically the ability of the U.S. system to solve big challenges that we have as a society, to understand that, we need to understand organization. We need to understand why, in some sectors, what is called division of innovative labor--where each part of the innovation process is done by different types of organizations--why this can be extremely efficient in some sectors, like life sciences. Like if you look at the COVID vaccine, maybe the best example of the strength of a highly fragmented innovation system, where each expert does one thing, which is connected to other things via kind of market exchanges. This can work extremely well in in life sciences. However, the U.S. system is falling behind in many in other sectors. For example, in material sciences, we are behind. In clean energy, we are behind. In battery technology, we are behind. And in those fields--behind, I mean, relative to China, for instance--and in those fields, you tend to see that fragmentation may not be as advantageous. That essentially maybe in those fields, as we argue in the book, having some degree of vertical integration, some degree of an ecosystem which is anchored in big firms and in big size may actually be very important. And in the book, we explain why this has been the case. And just one thing I would say that if you look at the golden age of American innovation, as I mentioned before, the great innovations that this country has created after the Second World War until the 1980s, or the late 70s, they have been done by huge American firms that performed all activities of innovation, from basic scientific research, winning Nobel prizes, to applied research, translational research, development in manufacturing, everything was done within the boundaries of very large firms. Those structures largely disappeared, and we have transitioned into an economy which is dominated mostly by startups, and we show in the book that sometimes this kind of structure, creates tremendous benefits, but in other sectors, we are falling behind, and we need to rebalance the system. And in the book, we explain how we can do that.

Sarah Kern  28:15

Well, thank you so much for spending time with me today, professor.

Sharon Belenzon  28:19

Thank you, Sarah.

Sarah Kern  28:26

Duke Fuqua Insights is produced by the Fuqua School of Business at Duke University. You can learn more at fuqua.duke.edu/podcast.