Search the archive.

//DH

This application of innovation snowballing to climate change begs the question: when does innovation snowballing justify government intervention? Nearly any line of technology can, in theory, experience innovation snowballing, since the accumulation of knowledge could increase the productivity and profitability of most technologies.88 However, government intervention is not necessarily efficient whenever one technology’s knowledge stock is bigger than another’s. Instead, government should only direct innovation away from one technology and towards another if two specific requirements are satisfied: there is a policy

In addition, a green transition will not happen without the revolution this book is advocating in how governments operate and how the relationships between public and private organizations are structured. For too long, governments have over-invested in carbon and under-invested in renewable energy sources. Mainstream economic prescriptions of simply solving climate change with a carbon tax and some R&D subsidies (to let markets find the optimal pathway), combined with political economic impediments to carbon taxes, have left us with negligent carbon tax systems and a worryingly slow green transition. The political-economic constraints are not solely to blame here. Economists have failed to realize and set out the reasoning for transformative public investment and re-regulation that accomplish what CO2 tax cannot on its own. As behavioural science has shown (and the reader knows from everyday life), real persons do not normally optimally react to price incentives, and they tend to ‘satisfice’ (are satisfied) rather than maximize profits or measures of happiness at every step of the way.19 This leaves a sizeable scope for clever, well-informed regulations such as energy performance standards. Furthermore, industrial change through radical innovation and infrastructure investments cannot be expected to respond mechanically to CO2 taxes as a matter of supply and demand in a farmer’s market.20 A few governments have created markets in which renewable energy technology – such as wind power in Denmark and solar power in Germany – can flourish and develop through learning by doing, in order to give them opportunities to mature and compete with fossil-based technology.

Our lethargic transition pace, globally, is a lesson in what can happen if government leaves the market to sort out problems and abstains from assuming its entrepreneurial role in society. It also flies in the face of what happened in that most transformative technological and business upheaval, the IT revolution, when government initiatives and investments were pivotal. This time, however, we need more than just initiatives from government. Neither the US nor the EU proposals can operate in a vacuum within sectors like renewable energy, de-carbonizing vehicles, or any other sector-specific initiatives. Instead, innovative transformation is required across all sectors, constituting one of the largest shifts humans have ever attempted.21 Reducing the material content of heavy industries like steel, and introducing a circular economy for waste in sectors with a repurpose, reuse and recycle mentality will be critical. This requires changes to materials that can produce more durable clothing, changes to our nutrition, with more focus on locally sourced food, and switching to production techniques like 3D manufacturing.22

A clue to what is required to accelerate the green 'revolution' is found in the US, where government-funded initiatives are busy building on their understanding of what has worked in previous technological revolutions. But while the US has been good at connecting and leveraging academia, industry and entrepreneurship in its own push into clean technologies (historically with the Department of Energy and more recently with the Advanced Research Projects Agency-Energy, or ARPA-E), its performance has been uneven. As one of the 'first' countries to seriously push into wind and solar power in the 1980s (and the first crystalline silicon solar cells were invented in the US in the 1950s), the US failed to sustain support and watched as Europe (Denmark, Germany and Sweden, in particular), Japan and now China take the lead. Worse, the US failed to alter its energy mix significantly, setting up its position for decades as a world-leading C02 emitter. With world-class innovative capability, the world's largest economy and a massive energy grid, the US is ideally positioned to kick off a clean technology revolution, yet it has not. In the context of the 2012 election season, clean energy development was again facing extreme uncertainty, and the very real possibility of losing government support at a critical juncture.9 Jeffrey Immelt, CEO of GE, bluntly describes the current structure of the US energy industry and its lack of an energy policy as 'stupid', estimating that other nations already have a 10-year lead on the green economy (Glader 2010).

Pros and cons of the US model

To begin the green industrial revolution and to tackle climate change we are again in need of an active State that takes on the high uncertainty of its early stages, which the business sector fears. Yet, despite the buzz surrounding 'clean technology' as the 'new economic frontier', and the 'green revolution' as the imminent third 'industrial revolution', there is in reality little that is truly new about many clean technologies. For example, wind and solar power have histories reaching back well over 100 years (and further still if considering non- electrical exploitation of the power sources). While the industrial revolution is often told as a story of steam and fossil fuels (Barca 2011), we have relied in the past on what would today be considered biomass, wind and hydro power.3 Despite our past experience and current knowledge of 'clean' energy technologies, government support seeking to make clean energy a dominant part of the energy mix has historically either been non-existent, or tended to wax and wane. The lack of focus and commitment to a clean technology future is what is preventing a more rapid transformation of the fossil energy infrastructure into a clean energy infrastructure.

But there are some rays of hope. In this early part of the twenty- first century, governments around the world have once again taken the lead in pumping up the R&D of many clean technologies like wind and solar power, and efforts are being made to establish modernized energy grids. They also subsidize and support the growth of leading manufacturers that compete for domestic and global market leadership. Finally, governments deploy both policy and finance to encourage stable development of competitive markets for renewable energy. As has been the case in the development of other industries such as biotech and IT, private businesses have entered the game only after successful government initiatives absorb most of the uncertainty and not a little risk of developing new energy technologies in the first place.

The pharmaceutical industry is interesting because of the new division of innovative labour. Large pharma, small biotech, universities and government labs are all parts of the ecology of the industry. But it is especially government labs and government-backed universities that invest in the research responsible for producing the most radical new drugs—the new molecular entities with priority rating in Figure 5. The ex-editor of the New England Journal of Medicine, Marcia Angell (2004), has argued forcefully that while private pharmaceutical companies justify their exorbitantly high prices by saying they need to cover their high R&D costs, in fact most of the really 'innovative' new drugs, i.e. new molecular entities with priority rating, come from publicly funded laboratories. Private pharma has focused more on 'me too' drugs (slight variations of existing ones) and the development (including clinical trials) and marketing side of the business. It is of course highly ironic, given this sector's constant bemoaning of 'stifling' regulations.

Economists measure productivity by comparing the amount of input into production with the amount of output that emerges. In this sense the large pharmaceutical companies have been fairly unproductive over the last few years in the production of innovations. As Figure 6 shows, there has been an exponential rise in R&D spending by members of the Pharmaceutical Research and Manufacturers of America (PhRMA) with no corresponding increase in the number of new drugs, commonly known as new molecular entities (NMEs). This also holds for patenting: while the number of patents has skyrocketed since the Bayh-Dole Act (1980) allowed publicly funded research to be patented, most of these patents are of little value (Demirel and Mazzucato 2012). When patents are weighted by the amount of citations they receive (the common indicator of 'important' patents), the figure is relatively flat, meaning that there are few important patents.

The problem is also evidenced in industries, like pharmaceuticals, where there is a trend of increasing public sector investments in R&D, while private sector spending is decreasing. According to Lazonick and Tulum (2012), the National Institutes of Health (N1H) has spent almost $300 billion over the last decade ($29.3 billion in 2013), and become more involved in the D component of R&D, meaning they absorb greater costs of drug development (such as through clinical trials), while private pharmaceutical companies2 have been spending less on R&D overall, with many shutting down R&D labs altogether. Of course the total R&D spent may be increasing, because the development (D) part is getting increasingly expensive. But this hides the underlying issue. While some analysts have justified the decreasing expenditure on research in terms of low productivity of R&D (increased expenditures not matched by increased discoveries), others, like Angell (1984, 2004, ex-editor of the New England Journal of Medicine), have been more explicit in blaming Big Pharma for not doing its share. She argues that for decades the most radical new drugs have been coming out of public labs, with private pharma concerned more with 'me too' drugs (slight variations of existing drugs) and marketing (see Chapter 3 for more details). And in recent years, CEOs of large pharma companies have admitted that their decision to downsize—or in some cases eliminate— their R&D labs is due to their recognition that in the 'open' model of innovation most of their research is obtained by small biotech firms or public labs (Gambardella 1995; China Briefing 2012). Big Pharma's focus is thus turned to working with such alliances, and 'integrating' knowledge produced elsewhere, rather than funding R&D internally.

Health innovation is expensive, inefficient and unsustainable, while the pharmaceutical sector itself consistently puts profits before people. Largely, this is due to incentives which encourage pharmaceutical companies to set high prices and deliver short-term returns to shareholders instead of investing in riskier, long-term research that advances critical therapies. The high prices of medicines have prevented patients from using them worldwide, or have reduced access to them, with damaging consequences for human health and well-being. This is especially problematic given the very high public spending worldwide on drug innovation. In the USA alone, the National Institutes of Health spends $40 billion a year on it and yet drug prices do not reflect that, leading some to say – an idea we have encountered before – that while the costs are socialized, the profits are privatized.24

A case in point: the hepatitis-C drug Sofosbuvir was the product of over ten years of research funded by the US Department of Veterans Affairs and NIH-funded research at Emory University, as well as NIH small-business innovation grants.25 After acquiring the drug, Gilead Sciences originally marketed it in the USA at $84,000 for a twelve-week course at one pill a day. In the UK, the list price for a course of treatment was nearly £35,000 (excluding VAT).26 These prices presented a huge barrier to access, even in wealthy countries. But by the end of 2017, Sofosbuvir-based products had generated over $50 billion in sales for Gilead.27 And more recently, in July 2020, Gilead was charging $3,120 for its COVID-19 drug Remdesivir, which was co-developed with a $70.5 million grant from the government-funded NIH.28

The problem has been not only the lack of venture capital investment in the most critical early seed stage, but also its own objectives in the innovation process. This has been strongly evidenced in the biotech industry, where an increasing number of researchers have criticized the venture capital model of science, indicating that significant investor speculation has a detrimental effect on the underlying innovation (Coriat, Orsi and Weinstein 2003; Lazonick and Tulum 2011; Mirowski 201 1). The fact that so many venture capital-backed biotech companies end up producing nothing, yet make millions for the venture capital firms that sell them on the public market, is highly problematic. It creates a need to question the role of venture capital in supporting the development of science and also its effect on the growth process. The increased focus on patenting and venture capital (VC) is not the right way to understand how risky and long-term innovations occur. Pisano (2000) in fact claimed that the stock market was never designed to deal with the governance challenges presented by R&D- driven businesses. Mirowski (2011, 208) describes the venture capital- backed biotech model as:

. . . commercialized scientific research in the absence of any product lines, heavily dependent upon early-stage venture capital and a later IPO launch, deriving from or displacing academic research, with mergers and acquisitions as the most common terminal state, pitched to facilitate the outsourcing of R&D from large corporations bent upon shedding their previous in-house capacity.

THE NATIONAL INSTITUTES OF HEALTH: CREATING THE WAVE VS. SURFING IT

State support and involvement in biotech span a wide range of forms, the most significant being that the enormous knowledge base which biopharmaceutical companies are dependent on has developed more from government investment than from business. The knowledge base has been developed from the critical investment the government has given to funding basic science. At the forefront lie the NIH and other government programmes which have invested in many of the key scientific achievements that the industry's success has been built on. Drawing on NIH spending data compiled in Lazonick and Tulum (2011), it is easy to see how crucial this funding was for biotech innovation. From 1938 to 2013, NIH spending on life sciences research totalled $884 billion ([2013$], deflated by the Consumer Price Index). Every year from 1938 to 2013, with a few exceptions,3 NIH funding increased in nominal terms, in contrast to the widely fluctuating funds from venture capital and stock market investments.

In addition, governments should steer the innovation ecosystem to use collective intelligence and innovation for public-health impact. Scientific and medical innovation thrives and progress is made when knowledge is exchanged openly, building upon shared successes and failures. But proprietary science does not follow that logic: it promotes secretive competition, prioritizes crossing regulatory finishing lines in wealthy countries over wide availability and public-health impact globally, and erects barriers to technological diffusion. While voluntary technology pools can be helpful, they risk being ineffective as long as commercial companies are allowed to retain control over critical technologies and data – even when these were generated with public investment.

The public sector, as a critical investor and participant in developing crucial technologies from vaccines, therapeutics and diagnostics, should govern the drug innovation process more like a market shaper: steering innovation, getting fair prices, ensuring that patents and competition work as intended and safeguarding medicine supply. At a global level, this requires governments to join efforts to impose firm rules regarding intellectual property, pricing and manufacturing. In the case of COVID-19, the process must also be governed in ways that value collaboration and solidarity between countries – not competition – both for R&D and for the availability and distribution of products.

Block (2008, 188) identifies the four key characteristics of the DARPA model:2

• A series of relatively small offices, often staffed with leading scientists and engineers, are given considerable budget autonomy to support promising ideas. These offices are proactive rather than reactive and work to set an agenda for researchers in the field. The goal is to create a scientific community with a presence in universities, the public sector and corporations that focuses on specific technological challenges that have to be overcome.

Going way beyond simply funding research, DARPA funded the formation of computer science departments, provided start-up firms with early research support, contributed to semiconductor research and support to human-computer interface research and oversaw the early stages of the Internet. Many of these critical activities were carried out by its Information Processing Techniques Office, originally established in 1962. Such strategies contributed hugely to the development of the computer industry during the 1960s and 1970s, and many of the technologies later incorporated in the design of the personal computer were developed by DARPA-funded researchers (Abbate 1999).

Another key event during this period was the new innovation environment that emerged after a group of scientists and engineers in 1957 broke away from a firm started by William Shockley (Block 2011). The rebellious group of scientists and engineers, often referred to as the 'traitorous eight', went on to form Fairchild Semiconductor, a new firm that advanced semiconductor technology and continued 'a process of economic fission that was constantly spinning off new economic challengers' (Block and Keller 2011a, 12-13). Lazonick (2009) adds that the spinoff culture ultimately began with Fairchild Semiconductor— and the firm owed nearly all of its growth to military procurement. The spinoff business model became viable and popular for technological advancement following the 1957 revolt, yet would not have been possible without State involvement and its functioning as a major early customer. A new paradigm emerged that resulted in innovative ideas moving from lab to market in far greater quantity.

People often look to DARPA as a way to consider organizational innovation inside the public sector. DARPA – the key investor in the internet as well as innovations like Siri – has always been mission-driven and known for its ability to attract top talent even without paying the kinds of salaries offered, for example, by the Singapore government (up to $1 million for department heads), so attraction to the project means more than money. This mission mystique is determined by a strong sense of purpose, such as to win a war or implement a Green New Deal. It is also determined by the ability to invest, lead and create with confidence, words usually applied only to business leaders. Key to DARPA’s model has been the explicit recognition that internal ways of working must be flexible and adaptive.

DARPA is the perfect example of why it is wrong to think that government is simply there to de-risk private risk-takers. DARPA took enormous risks in funding the invention of the internet – and it did so with a problem in mind: so that satellites could communicate. Similarly, the US Navy funded the invention of GPS in an effort to target missiles better. The point is: to think in a mission-oriented way is revolutionary because it requires rethinking the role of government in the economy, putting purpose first and solving problems that are important to citizens. It means transforming government from being merely an ‘enabler’ or even a ‘stifler’ of innovation to becoming the engine of innovation. Usually this type of flexibility and outcomes orientation only occurs in wartime, when military organizations are given autonomy from more bureaucratic branches of government. And yet DARPA works in peacetime, too. Recently, for example, DARPA funded early R&D for two pharmaceutical companies, Moderna Inc. and Inovio Pharmaceutical Inc., to create RNA and DNA vaccines – technologies that many scientists and investors considered speculative and high-risk. However, DARPA believed that nucleic-acid-based vaccines could be developed much faster than conventional technologies, and their bet paid off. In 2020, after the COVID-19 pandemic hit, Moderna’s RNA vaccine became the first COVID-19 vaccine to be administered in a Phase I trial, and Inovio’s DNA-based vaccine was on track to start a Phase I trial shortly thereafter. Pursuing high-risk, high-reward technologies is routine for DARPA, which continually creates, manufactures and distributes prophylactics within sixty days of identifying a new viral pathogen.

To solve specific problems that catalyse investment and collaboration between many different people and organizations, we need policy instruments that focus on outcomes and foster experimentation. Procurement contracts, grants, loans and prize schemes should reward innovators for taking risks to solve public problems, whether the aim is to develop a vaccine or a new way to help hospitals share data. This means putting less emphasis on the need to support specific types of firms (such as SMEs and start-ups), less on specific technologies (such as ‘high-tech’, artificial intelligence, quantum computing), and less on specific sectors to promote (such as life sciences or the creative sector). At the same time, it means putting more emphasis on the big problems facing society and how all types of organizations, technologies and sectors can play a role in solving them.

Prize schemes, in particular, are useful for driving cross-sector innovation – and have been for hundreds of years. In 1675, the Royal Observatory was founded at Greenwich, on the River Thames to the east of London, to carry out astronomical observations. Part of the Observatory’s mission was to help with the thorny problem of how to measure longitude – the distance east or west of the prime meridian which runs north to south through the Observatory and girdles the globe. Determining longitude was becoming critical for commercial and naval reasons. Britain was locked in competition with rival powers, notably France and Holland. Ships were being lost, or their voyages were taking longer than necessary, because they could not accurately establish their position at sea. Being able to work out longitude would give a country a commercial and naval advantage.

Part III of this Article catalogued the manifold ways in which innovations can be bad for society. Even in the fields that have best exemplified technological progress, like pharmaceuticals, many new innovations are harmful. And the share of antisocial innovations will be even greater in other fields, like workplace surveillance, finance, and tobacco products. But none of this is apparent from two of our most important innovation governance tools—patent law and antitrust law. Both sets of doctrines treat all or almost all innovation as good innovation and worthy of encouragement.328

We refer to the uncritical embrace of innovation as the Inverse Precautionary Principle. The term derives from administrative law, where, according to the Precautionary Principle, the state should actively regulate behaviors that pose a risk of harm, even in the absence of empirical evidence of the risk.329 For example, the Precautionary Principle should be used as a guide for regulating environmental harms; according to the United Nations: “[w]here there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.”330 The guiding intuition is “better safe than sorry.” In the face of uncertainty about risks, the Precautionary Principle dictates, it’s best to regulate and avoid them.

There are various solutions to the Box 4 problem that we have already discussed. Intellectual property rights give innovators exclusivity for a time where they can charge supracompetitive prices for access to their innovations, thereby recouping their investments.184 Prizes, grants, and tax incentives are other mechanisms to decrease or compensate for innovators’ costs to make the innovation privately valuable and worth pursuing.185 In effect, these incentives turn Box 4 innovations into Box 2 innovations.

Importantly, though, innovation incentives create two major risks. First, if they are awarded to innovators who already find themselves in Box 2, where they have sufficient private motivation to pursue the innovation, the incentives are unnecessary and socially costly. Giving patents to firms that would be willing to pursue inventions in the absence of exclusive rights simply drives up costs for consumers and competitors without generating any innovation benefits.186

//DH

Nor is path dependency unique to energy innovation. Empirical work, for example, has demonstrated that innovation snowballing affects the direction of pharmaceutical inventions.75 And one carefully conducted study by economist Walker Hanlon provides strong empirical evidence that path dependency shaped the historical evolution of textile technologies.76 In particular, the American Civil War cut off the supply of American cotton, leading demand to temporarily shift to lower-quality Indian cotton. That temporary shock caused a jump in innovations for processing Indian cotton, which, in turn, made the quality of Indian cotton as good or better than American cotton. As a result, the impacts of the temporary innovation shock lasted past the end of the Civil War. Innovations in the technologies for processing Indian cotton continued to increase, alongside demand for Indian cotton, even after American cotton became available again.77

//DH

Innovation snowballing is more than just a theory. Recent quantitative work suggests it is an empirical reality. Economist David Popp, for instance, finds that the propensity to innovate in a given energy-efficient technology increases with the number of prior patents filed for each technology, indicating that energy innovators do, in fact, stand on the shoulders of past entrepreneurs, leading to innovation snowballing.70 Other research shows that oil shocks increase innovation in energy-saving technology like air conditioners and that the switch toward more innovation in energy-saving technology persists after the end of the oil shock.71 Path dependency also characterizes the trajectory of clean automobile innovations. In particular, a team of researchers led by Harvard economist Philippe Aghion shows that, when fuel prices increase, innovation rates in cleantech (e.g., hybrid and electric vehicles) increase and that these increased rates are self-perpetuating, while innovation rates in dirtytech (the internal combustion engine) decrease.72 Firms with a history in clean auto innovation are much more likely to channel their future research efforts towards clean technologies.73 The same is true for firms with histories in dirty car technologies. A firm’s future research agenda depends not only on its own innovation history, but also on those of its neighbors: a company’s propensity to innovate in clean (dirty) technologies increases if they are situated in countries where other firms focus more on clean (dirty) automobile innovation. Put differently, firms that are exposed to clean (dirty) innovation spillovers are more likely to engage in clean (dirty) innovation in the future.74

//DH

In addition to reducing the benefits of patents, innovation snowballing exacerbates the costs of patents relative to government-set tools that do not place intellectual property in private hands. Patents’ major disadvantage is that the monopoly rights granted to innovators under a patent system increase the cost of using the resulting intellectual property above its marginal cost, thereby reducing the technology’s use among two important groups: producers who use the innovation to make a product (e.g., electricity) and entrepreneurs who use the innovation to develop future innovations.147 Thus patents—especially those on foundational concepts148—can impede innovations from being added to the innovation stock (i.e., make the “shoulders of giants” smaller than they could be).149 For example, companies with patents can, in theory, entirely prevent others from using its technology for the duration of the patent; at a minimum, they will likely charge more than the zero marginal cost of the use of the idea. Likewise, patents can create an anti-commons problem: if a new potential innovation requires as inputs a variety of past innovations, each of which is patented by a different company, then the costs of acquiring all those licenses may be prohibitively high.150 Innovation snowballing makes the monopoly pricing and transaction costs under the patent system all the more perverse. Path dependency follows from innovators “standing on the shoulders of giants,” and patents make it even harder to stand on those shoulders, hindering innovation snowballing.

//DH

a. Patents vs. Government-Set Tools