While watching an old interview with Max Levchin and Peter Thiel from 2011, something Levchin said about innovation stagnation stood out to me. Levchin argued that one of the things that was noticeable in his time was the impetus and vision provided by heads of States and governments at large with regards to the economic activity that would take place over 15-20 year horizons. Since the Reagan administration in the US, the trend one has observed in Western governments has been one of the State moving away not just from production or innovation, but from the provision of a vision as well. For instance, federally funded R&D has fallen from 1.86% of GDP (1964) to ~0.63% (2022), outside defense, the state’s role has thinned. By 2022, business funded 75% of U.S. R&D, up from 31% in 1964. The State operates as if severed from the market, its duties primarily administrative while the load of innovation as well as deciding what one innovates falls on the private sector, with Venture Capital and angel investors being the primary means to do so. In this essay I will argue that this evolution has caused the stagnation in most fields that people see today, and on the contrary, the one sector in which progress has continued unfettered has been in the sector that the government has operated in and offered incentives to private producers, namely defence. Taking the United States, previously and perhaps even today considered the global hub of innovation as my case study, I will hope to demonstrate that the areas of growth in the economy have been largely tied to defence or some arm of the State, and the argument I will consequently advance is that Laissez faire systems slow down innovation if vision is left to their hands, and the solution is to once more bring back the previously happy marriage between State and industry.

First, I must defend the stagnation thesis in the US. With Thiel as its most outspoken proponent, it is simple. The argument goes that the rate at which we have been innovating has dramatically slowed down starting from the early 70s, and alongside that visions of the future have equally become uninspiring and unimaginative, as this trend of stagnant innovation continues (these two processes likely go hand in hand, but not as simply as a purely cultural explanation of the phenomena might suppose). This should be immediately obvious to any observer. The rate at which technology disrupted society at large in each decade of the early 20th Century as well as the rate at which new innovations were coming out was certainly remarkable. The cause of the decline of such innovation post the 1970s has been unclear, some argue it is cultural, others argue that the low hanging fruit have been plucked, and others still argue that it is due to overregulation by the State, a mix of cultural and institutional problems. The argument I am advancing here is an extension of Levchin’s statement, namely that the market might be good at coordinating resources and building, but it is reliant ultimately on the State to provide a vision and the safety net in order to organise around the long term interests that build innovative technology.

Why do free market agents not produce innovative outcomes (I am speaking of this as a norm, exceptions certainly exist)? The incentives of agents in the free market are geared around preservation of capital and its consequent accumulation. Therefore, in the free market, individuals are likely to invest in projects whose outcomes are clearly modelable, these are more often than not, projects that are short term, or at least projects where the investor’s exit horizon is in the short term. This is evidenced once again by looking at business R&D, which skews to development. In 2022, business outlays were 6% basic/15% applied/79% development. The reason for this is also a tendency towards risk aversion, as well as the fact that the positive externalities of investment in research are hard to capture privately, the social benefits far outweigh the private benefits. Therefore incentives are once again skewed on this front. Furthermore, in concentrated markets such as pharmaceuticals, disruptive innovations are often absorbed through acquisitions that shelve nascent products, here the incentive to preserve one’s own position in the market means disruptive products are eliminated before they can be a threat.

The exception to this comes once again in the form of technologies such as AI, where across the sector there is a common mission around which Silicon Valley has aligned, consequently allowing companies such as Google to invest a lot of money over years with little return until finally there was a breakthrough, or Musk’s funding of OpenAI to do the same. Even so, AI aligns with certain State objectives such as Project Maven, which utilizes AI and machine learning algorithms, specifically computer vision, to process and interpret video and imagery data, including identifying objects like vehicles, buildings, and people which Palantir, another private player has helped develop.

What precisely does the State signalling a mission do? It dramatically reduces downside risk. Often when the State signals intent to build something it is willing to offer infrastructural and institutional support. It allows for capital to congregate around a specific end which is to be attained over a long term horizon. Contrast this with free markets, where different agents are working towards different outcomes and capital is thus diffuse, along with the considerations in the earlier paragraphs about the skewed incentives present for private sector players with regards to funding innovation. A Venture Capital firm has limited rolls of the dice. The example of the US DoD remains the most striking, and the strongest instance I have in order to make my case. Take for instance, the fact that about 46% of all federal obligations were accounted for by the DoD in 2023, with federal funding overall accounting for about 40% of basic research.  The skew is clear, the US DoD remains overwhelmingly one of the stakeholders whose interests drive the direction of research, and consequently, private capital in this sector can safely be allocated.

Why is this so? The State providing a mission statement is not simply a vague statement of support. The State backs it actively, through various mechanisms such as advance purchase commitments as well as direct funding. One need only look at Lynas Rare Earths’ recent Texas facility for evidence of this, as well as the fact that the US government will be one of the chief customers of Lynas, given their need to diversify NdPr producers away from China. Furthermore, the ability of the US DoD to use OTAs (Other Transaction Authority) which allows them to sidestep Federal Acquisition Regulation, allowing for smoother offtakes of projects. When the state uses APCs + OTAs + multi-year procurement, downside risk collapses. For instance, Operation Warp Speed obligated ~$8.8B via OTAs out of ~$13B by December 2020, overlapping trials and funding at-risk manufacturing. The DoE’s $465M loan to Tesla in 2010 is another such de-risking move. If one looks at the major avenues of innovation in the last 20-30 years, they have happened in the internet economy, the internet itself an innovation by DARPA, another wing of US Defense.

The link between innovation in the real and virtual economy and defence is nigh undeniable. Take for instance Palantir, a firm set up with the CIA as its only client for the first few years, as well as one which was heavily funded by the CIA. Palantir’s technology has obvious applications in intelligence, and as DARPA develops technologies such as Mind’s eye or ARGUS-IS, Palantir’s ability to rapidly assimilate, organise, and analyse data proved increasingly invaluable. By the 2020s, Palantir possessed extremely useful commercial applications as well (Foundry) which allowed for it to be a company used as a consultant by many manufacturing sector companies, but this was an application of Palantir which was born years after its initial inception.  Another example is Starlink, which has direct military applications as demonstrated in 2023 when the US DoD enlisted it to provide support in Ukraine, as well as Starshield, an explicitly stated defence company. In an instance where the relation is not as obvious and a little more vague but equally pointed, one may look at DARPA’s lifelog program, which coincided with Facebook’s founding, Facebook happening to provide the exact same function lifelog was supposed to, as well as CIA contractor Thiel being one of the first investors in Facebook. The implication here isn’t that the DoD or DARPA or the CIA were directly involved in the funding or creation of Facebook, but certainly had an interest in its success.

One might counter with the example of software (AI and social media aside, which I’ve addressed above) and why it has remained high growth in the face of slowing growth in the real economy. The answer I propose to this is twofold. The first is that software remains an insular sector. What this means is that improvements in software do not typically require replication unlike in the material sciences, they require adoption. Software businesses also tend to not be very capital intensive, allowing Venture Capitals many rolls of the dice, therefore incentivising them to be more open to risk. This is also why one sees hype bubbles in Tech, be it the dot com bubble, the crypto bubble in 2021, or more recently (as I would argue) the AI bubble. Risk is not as costly in software, and value can be far more easily captured, incentivising firms to invest in innovation. A further natural objection might be that it is regulations that slow down innovation, and deregulation would fix this. While deregulation would not solve the coordinational problem highlighted above, of capital dispersed in various directions, it is also true that in sectors such as rare earths (mining, heavily regulated) recently, where there is State (particularly once again, defence) appetite for risk, capital follows. Lynas Rare Earths’ Texas facility remains the template in this context.

This brings us back to my initial thesis. Investment in R&D particularly in research slowed down from the 1980s owing largely due to the fact that the State stepped back from taking an active role in the economic activity of the country. In other sectors of the country, particularly in medicine, you do see this with singular projects, but nothing coordinated at the scale at which the US DoD accomplishes. Thus any upshots are unique and not systemic. The US post 1980 has not stopped innovating, but outsourced mission statements from the State to the private sector, which has cost long term alignment of interests as well as the ability to eat losses necessary for innovation to flourish. Most innovations are second order effects following an obscure breakthrough in a lab, doing research that may not seem immediately obvious. To bring this back is an institutional fix, not ideological. Markets are efficient at mobilising resources and building, but without a common goal to work towards, these tools are wasted and progress becomes incremental, rather than revolutionary.

While watching an old interview with Max Levchin and Peter Thiel from 2011, something Levchin said about innovation stagnation stood out to me. Levchin argued that one of the things that was noticeable in his time was the impetus and vision provided by heads of States and governments at large with regards to the economic activity that would take place over 15-20 year horizons. Since the Reagan administration in the US, the trend one has observed in Western governments has been one of the State moving away not just from production or innovation, but from the provision of a vision as well. For instance, federally funded R&D has fallen from 1.86% of GDP (1964) to ~0.63% (2022), outside defense, the state’s role has thinned. By 2022, business funded 75% of U.S. R&D, up from 31% in 1964. The State operates as if severed from the market, its duties primarily administrative while the load of innovation as well as deciding what one innovates falls on the private sector, with Venture Capital and angel investors being the primary means to do so. In this essay I will argue that this evolution has caused the stagnation in most fields that people see today, and on the contrary, the one sector in which progress has continued unfettered has been in the sector that the government has operated in and offered incentives to private producers, namely defence. Taking the United States, previously and perhaps even today considered the global hub of innovation as my case study, I will hope to demonstrate that the areas of growth in the economy have been largely tied to defence or some arm of the State, and the argument I will consequently advance is that Laissez faire systems slow down innovation if vision is left to their hands, and the solution is to once more bring back the previously happy marriage between State and industry.

First, I must defend the stagnation thesis in the US. With Thiel as its most outspoken proponent, it is simple. The argument goes that the rate at which we have been innovating has dramatically slowed down starting from the early 70s, and alongside that visions of the future have equally become uninspiring and unimaginative, as this trend of stagnant innovation continues (these two processes likely go hand in hand, but not as simply as a purely cultural explanation of the phenomena might suppose). This should be immediately obvious to any observer. The rate at which technology disrupted society at large in each decade of the early 20th Century as well as the rate at which new innovations were coming out was certainly remarkable. The cause of the decline of such innovation post the 1970s has been unclear, some argue it is cultural, others argue that the low hanging fruit have been plucked, and others still argue that it is due to overregulation by the State, a mix of cultural and institutional problems. The argument I am advancing here is an extension of Levchin’s statement, namely that the market might be good at coordinating resources and building, but it is reliant ultimately on the State to provide a vision and the safety net in order to organise around the long term interests that build innovative technology.

Why do free market agents not produce innovative outcomes (I am speaking of this as a norm, exceptions certainly exist)? The incentives of agents in the free market are geared around preservation of capital and its consequent accumulation. Therefore, in the free market, individuals are likely to invest in projects whose outcomes are clearly modelable, these are more often than not, projects that are short term, or at least projects where the investor’s exit horizon is in the short term. This is evidenced once again by looking at business R&D, which skews to development. In 2022, business outlays were 6% basic/15% applied/79% development. The reason for this is also a tendency towards risk aversion, as well as the fact that the positive externalities of investment in research are hard to capture privately, the social benefits far outweigh the private benefits. Therefore incentives are once again skewed on this front. Furthermore, in concentrated markets such as pharmaceuticals, disruptive innovations are often absorbed through acquisitions that shelve nascent products, here the incentive to preserve one’s own position in the market means disruptive products are eliminated before they can be a threat.

The exception to this comes once again in the form of technologies such as AI, where across the sector there is a common mission around which Silicon Valley has aligned, consequently allowing companies such as Google to invest a lot of money over years with little return until finally there was a breakthrough, or Musk’s funding of OpenAI to do the same. Even so, AI aligns with certain State objectives such as Project Maven, which utilizes AI and machine learning algorithms, specifically computer vision, to process and interpret video and imagery data, including identifying objects like vehicles, buildings, and people which Palantir, another private player has helped develop.

What precisely does the State signalling a mission do? It dramatically reduces downside risk. Often when the State signals intent to build something it is willing to offer infrastructural and institutional support. It allows for capital to congregate around a specific end which is to be attained over a long term horizon. Contrast this with free markets, where different agents are working towards different outcomes and capital is thus diffuse, along with the considerations in the earlier paragraphs about the skewed incentives present for private sector players with regards to funding innovation. A Venture Capital firm has limited rolls of the dice. The example of the US DoD remains the most striking, and the strongest instance I have in order to make my case. Take for instance, the fact that about 46% of all federal obligations were accounted for by the DoD in 2023, with federal funding overall accounting for about 40% of basic research.  The skew is clear, the US DoD remains overwhelmingly one of the stakeholders whose interests drive the direction of research, and consequently, private capital in this sector can safely be allocated.

Why is this so? The State providing a mission statement is not simply a vague statement of support. The State backs it actively, through various mechanisms such as advance purchase commitments as well as direct funding. One need only look at Lynas Rare Earths’ recent Texas facility for evidence of this, as well as the fact that the US government will be one of the chief customers of Lynas, given their need to diversify NdPr producers away from China. Furthermore, the ability of the US DoD to use OTAs (Other Transaction Authority) which allows them to sidestep Federal Acquisition Regulation, allowing for smoother offtakes of projects. When the state uses APCs + OTAs + multi-year procurement, downside risk collapses. For instance, Operation Warp Speed obligated ~$8.8B via OTAs out of ~$13B by December 2020, overlapping trials and funding at-risk manufacturing. The DoE’s $465M loan to Tesla in 2010 is another such de-risking move. If one looks at the major avenues of innovation in the last 20-30 years, they have happened in the internet economy, the internet itself an innovation by DARPA, another wing of US Defense.

The link between innovation in the real and virtual economy and defence is nigh undeniable. Take for instance Palantir, a firm set up with the CIA as its only client for the first few years, as well as one which was heavily funded by the CIA. Palantir’s technology has obvious applications in intelligence, and as DARPA develops technologies such as Mind’s eye or ARGUS-IS, Palantir’s ability to rapidly assimilate, organise, and analyse data proved increasingly invaluable. By the 2020s, Palantir possessed extremely useful commercial applications as well (Foundry) which allowed for it to be a company used as a consultant by many manufacturing sector companies, but this was an application of Palantir which was born years after its initial inception.  Another example is Starlink, which has direct military applications as demonstrated in 2023 when the US DoD enlisted it to provide support in Ukraine, as well as Starshield, an explicitly stated defence company. In an instance where the relation is not as obvious and a little more vague but equally pointed, one may look at DARPA’s lifelog program, which coincided with Facebook’s founding, Facebook happening to provide the exact same function lifelog was supposed to, as well as CIA contractor Thiel being one of the first investors in Facebook. The implication here isn’t that the DoD or DARPA or the CIA were directly involved in the funding or creation of Facebook, but certainly had an interest in its success.

One might counter with the example of software (AI and social media aside, which I’ve addressed above) and why it has remained high growth in the face of slowing growth in the real economy. The answer I propose to this is twofold. The first is that software remains an insular sector. What this means is that improvements in software do not typically require replication unlike in the material sciences, they require adoption. Software businesses also tend to not be very capital intensive, allowing Venture Capitals many rolls of the dice, therefore incentivising them to be more open to risk. This is also why one sees hype bubbles in Tech, be it the dot com bubble, the crypto bubble in 2021, or more recently (as I would argue) the AI bubble. Risk is not as costly in software, and value can be far more easily captured, incentivising firms to invest in innovation. A further natural objection might be that it is regulations that slow down innovation, and deregulation would fix this. While deregulation would not solve the coordinational problem highlighted above, of capital dispersed in various directions, it is also true that in sectors such as rare earths (mining, heavily regulated) recently, where there is State (particularly once again, defence) appetite for risk, capital follows. Lynas Rare Earths’ Texas facility remains the template in this context.

This brings us back to my initial thesis. Investment in R&D particularly in research slowed down from the 1980s owing largely due to the fact that the State stepped back from taking an active role in the economic activity of the country. In other sectors of the country, particularly in medicine, you do see this with singular projects, but nothing coordinated at the scale at which the US DoD accomplishes. Thus any upshots are unique and not systemic. The US post 1980 has not stopped innovating, but outsourced mission statements from the State to the private sector, which has cost long term alignment of interests as well as the ability to eat losses necessary for innovation to flourish. Most innovations are second order effects following an obscure breakthrough in a lab, doing research that may not seem immediately obvious. To bring this back is an institutional fix, not ideological. Markets are efficient at mobilising resources and building, but without a common goal to work towards, these tools are wasted and progress becomes incremental, rather than revolutionary.