How Tech Bubbles Accelerate Innovation
Is there a bubble in artificial intelligence? Are we in a virtual reality bubble? Are Uber and Airbnb overvalued? Or, are we even in another tech bubble?
Given the unprecedented valuations of the so-called “unicorns” — Uber is still valued at more than $60 billion, Airbnb at $25 billion, and Palantir at $20 billion — talk about bubbles has intensified in Silicon Valley. Bubbles are often considered to be negative: they are described as destructive, economically inefficient, and as generating financial waste. Indeed, almost $6 trillion evaporated after the dotcom collapse in the early 2000s, and the globalized financial system crashed when the financial bubble burst in 2008.
In economics and finance, bubbles form when unrealistic expectations about future cash flows decouple prices temporally from fundamental valuations. Fueled by self-reinforcing feedback, imitation and herding behavior, prices elevate until a crash drives them back to fundamental values. While considered negative — as unrealistic and excessively optimistic expectations about the future fail — they are a persistent and inevitable phenomenon: from the Dutch tulip mania in 1637 and the South Sea Bubble of 1720, through the great crash of October 1929, Black Monday of 1987, and the financial crisis of 2007, destructive bubbles punctuate our history.
Yet not all bubbles are necessarily negative. Rather, the formation of bubbles can be understood as an important process for innovation in various fields. Bubbles deploy the financial capital necessary to fund disruptive technologies at the frontier of innovation. By generating positive feedbacks — resulting in excessive enthusiasm and investments that are necessary for various social and technological enterprises — bubbles can be beneficial for society. Without these effects many technological innovations or large-scale societal projects would never have happened.
Bubbles deploy the financial capital necessary to fund disruptive technologies at the frontier of innovation.
The complex network of social interactions between enthusiastic supporters has catalyzed the formation of large-scale scientific or technological projects, including the Human Genome project and the Apollo program. Characterized as manifestations of collective over-enthusiasm, they constitute an essential element in the dynamics that give rise to great scientific discoveries and radical technological innovations. Bubbles have also been at the core of the technological revolutions that have fundamentally transformed our economic, social, and technological systems over the last centuries.
Bubbles and Technological Revolutions
What can we learn from the history of bubbles — as necessary elements in innovation — that could help us to anticipate and understand them better? Historically, it is singularly important to recognize their role in technological revolutions. In her now classic work on technological change, the socio-economist and historian Carlota Perez has identified five distinct technological revolutions, which she has called “great surges of development.” She defines a “great surge” as the process by which a technological revolution propagates across the economy and society. Emerging technologies — which might consolidate in what Perez calls a “techno-economic paradigm” — evolve around new infrastructures, energy sources, and materials. These major technological disruptions have lead to the emergence of new interrelated constellations of technologies. The industrial revolution resulted in the mechanization of factories and the development of water power in the late 18th century. The second surge, which erupted in 1829, was based on steam and railways. The third revolved around steel, electricity and chemical, industrial, and electrical engineering, which, then, intensified, due to trans-continental railways and intra-continental steamships, international trade and globalization. All these new technologies and infrastructural networks emerged from the wreckage of speculative bubbles.
As Perez has shown, each technological disruption is triggered by a financial bubble, which employs the necessary capital to fund the technological innovations. Perez has identified a regular pattern of technology diffusions. There is an “installation” phase in which a bubble drives the installation of the new technology. This is followed by the collapse of the bubble or a crash, to which Perez refers to as the “turning point.” After this transitional phase — which occurred, for example, after the first British railway mania in the 1840s, or, more recently, after the dotcom-bubble — a second phase is unleashed: the “deployment” phase, which diffuses the new technology across economies, industries and societies. The economic exhaustion of the technological revolution and excessive financial capital, which searches for new investment opportunities, can, then, give rise to the next technological revolution.
These bubbles, which have catalyzed major technological innovations and disruptions, share a central dynamic: the funding of these new technologies, energy sources, or networks decouples from expectations of economic return. Irrespective of quantifiable financial returns and economic values, bubbles mobilize the financial capital necessary to develop new transformative technologies. As venture capitalist and economist William Janeway has shown, bubbles are necessarily wasteful as technological innovations require “inefficient” trial-and-error experimentation.
Contrary to the assumptions of neo-classical economics — in which rational, that is, utility-maximizing agents allocate scarce resources on the basis of perfect information — the economic wastefulness of bubbles is essential in the deployment of radical technological innovations. The investments, which bubbles fuel, are often not justified on the basis of traditional cost-benefit analyses. As Janeway has highlighted, such productive bubbles are, initially, driven by state investments. From the construction of the railroad and electricity networks, which have been subsidized by the federal government, through the role of the Defense Advanced Research Projects Agency (DARPA) in the creation of the internet, to the public R&D funding of many technical components of the iPhone, such as the touch-screen display and the voice-activated assistant SIRI, many inventions and innovations — as innovation economist Marianna Mazucatto has extensively documented — have been initially government-sponsored. Historically, it was often the state that was able to provide financial support for large-scale technological projects, which were infused with radical uncertainty about future economic value. After the initial state investments, financial capital drives these productive bubbles, which provide usable commercial infrastructural networks and technologies, even after the bubble has burst. Although they destroy value and inflict pain on individual investors, such bubbles can nevertheless be net beneficial for society as they enable the exploration of new economic spaces. In other words, the overall long-term benefits outweigh the short-term costs of productive bubbles.
Bubbles extend beyond the domains of financial markets or technological innovations. As Didier Sornette and his research group have shown, bubbles are generic phenomena that occur in society at large. Similarly to financial bubbles, the essential ingredients are socio-behavioral mechanisms, such as herding or imitation, and exuberant over-optimism and unrealistic expectations. Such “social bubbles” have been the driving force for many large-scale political, scientific, and societal projects. Sornette and his collaborators have, for example, examined the Apollo program as well as the Human Genome project as they embody the essential characteristics of such bubbles. Identically to the productive ones that have fostered technological revolutions and major scientific breakthroughs, social bubbles can be characterized in terms of a reduction of collective risk-aversion. Standard risk-benefit analyses are disregarded, which results in extraordinary risk-taking in the face of radical economic and technological uncertainty.
Driven by political factors, such as the space race with the Soviet Union, as well as scientific curiosity, the Apollo program provides an interesting illustration of a massive social bubble. Still one of the costliest projects that have ever been undertaken by the United States at peacetime, it massively accelerated technological innovation and scientific advances. Fueled by self-reinforcing collective over-enthusiasm, massive financial, political and individual risks were taken to “put a man on the moon.” While the collective enthusiasm and public and political support peaked with the first moon landing in 1969 — which was then viewed as the first step in the colonization of space — the bubble subsequently burst as political and public support for the program decreased. Nevertheless, while the cost of the program might have been disproportional to the returns, many returns may still be exploited in the future.
Another vivid example of a social bubble documented by Sornette and his collaborators is the Human Genome project. Initiated and funded by government agencies to sequence the human genome, it became a hyped large-scale scientific project, which was fueled by scientific and financial interests. Intense competition between the public and private sector catalyzed the anticipation and results that generated extraordinary over-enthusiasm and media interest that further attracted even more financial and scientific investments, which, in turn, resulted in a proliferation of genomic start-ups. This network of self-reinforcing feedbacks between different actors — such as scientists, government agencies, entrepreneurs, and venture capitalists — generated a massive social bubble, which inflated expectations about medical and commercial applications of the Human Genome project. However, the scientific understanding of genes and the commercialization of some of the underlying technologies have not advanced as anticipated. Still, the Human Genome project accelerated the falling of sequencing costs and generated impressive advances in scientific understanding and technological innovation. We’re still in the early phases of this projects long term impact on society.
These bubbles share an essential dynamic that is driven by excessive public, political, and financial expectations about future outcomes, which are not justified by fundamental valuation or standard cost-benefit-analyses. But is this collective over-enthusiasm just another manifestation of irrationality? While expectations are often massively elevated, it seems misleading to label this collective exuberance enthusiasm purely irrational.
As the dotcom bubble illustrates, the fundamental expectations or beliefs behind the bubble can be rationally justified. The intuition that the Internet will radically transform our economic and societal structures was ultimately correct. While investing in startups such as Pets.com was a painful experience for some investors — $300 million evaporated when the company folded after the collapse of the dotcom bubble — companies such as Google and Amazon emerged from the debris. Although, it now seems to be the result of irrational exuberance — as Robert Shiller’s book, which was published at the peak of the internet bubble, suggests — new accounting methods, such as net present value or so-called “real-option” valuations of companies, have rationalized dotcom investments. By extending option theory to companies, investors were able to assess future investments in tangible assets and the value of future opportunities. It follows from real-option theory that the greater the uncertainty, which has infected the development of the Internet and related technologies and services, the more valuable the options may be. It was, thus, rational to invest in dotcom companies, such as Amazon, as they had real options in investing in emerging e-commerce markets.
Similarly, the current unprecedented valuations of tech unicorns might seem irrational. However, when considering another feature of tech companies — network effects and the economics of increasing returns that are characteristic of Silicon Valley — the valuations of some of these companies might be rationally justified. As with Google and search, Facebook and social media, or Uber and ridesharing, often one company will eventually control the entire market. While is it unclear whether network effects will continue to dominate, the potential scalability and increasing returns help rationalize some of the current unicorn valuations. In other words, while the expectations and enthusiasm that drive bubbles might be inflated, it does not necessarily render the extraordinary investments or commitments, which are a defining feature of such bubbles, irrational. Rather, these bubbles should better be characterized as phases of innovative exuberance, which enable the development and deployment of disruptive new technologies.
Productive vs. Destructive Bubbles
But how can we distinguish between productive bubbles — which are necessary elements in the process of innovation — and destructive bubbles, which destroy value and destabilize financial systems? As Janeway has shown, bubbles can be separated and categorized along two different dimensions. One dimension is defined by the object of speculation. Charles Kindleberger, in his important Manias, Panics and Crashes, mapped out the spectrum of assets that have historically been the objects of speculation, such as tulip bulbs and railroads, real estate, commodities, stocks, and bonds. More recently financial derivatives, such as credit-default swaps, options, exchange-traded-funds, and cryptocurrencies have all been objects of excessive speculation. The other dimension, which Janeway identifies, is the locus of speculation, that is, the economic space where the speculation occurs. Often, speculative activity in credit markets generates bubbles that are highly destructive. As they infect the real economy, these destructive bubbles can increase the systemic instability of the financial and economic system as it was the case during the credit bubble of 2004–2007. On the contrary, the dotcom bubble, for example, was confined to the capital markets and did not spill over into the real economy. While destructive for some investors, the dotcom bubble provided the technological foundation of the Internet and the emerging digital economy.
What distinguishes positive bubbles from negative bubbles is that they are transformative engines for economic growth and productivity. Historically, the five major technological revolutions, which disrupted the economic, institutional, and political structure of society, resulted in extraordinary increases in productivity. These technological revolutions propagated across the entire economic system. Not only have they all fundamentally transformed existing industries and infrastructures, but created radically new economic spaces to be explored and exploited. These massive technological disruptions — which can be aptly characterized by a process of Schumpeterian “creative destruction” — all generated novel interrelated sets of new technologies, which, in turn, have catalyzed further innovations, infrastructures, and industries.
In contrast to these innovative bubbles, which accelerated growth and productivity, the destructive bubbles, which have punctuated the global financial markets since the 1980s, have generated economic growth by the hyper-financialization of the real economy. As the financial crisis of 2007/2008 has shown, virtual finance diverged from the real economy. The bubbles that we have experienced in the past three decades were driven by consumption, rather than productivity. Virtual wealth has been created by debt and credit, which shifted the economic regime from innovation-driven wealth-creation to credit-based wealth-extraction. Fueled by an illusionary belief in eternal economic growth, financialization and explosive debt and leverage — which resulted in the creation of ever more complex financial products — increased the systemic instability of the global financial system and the real economy even further.
With the rise of virtual finance, financing of technological innovations has shifted from long-term high-risk investments — which accelerated radical innovations as they were not justified by traditional cost-benefit analyses — to short-term and exit-driven investments that tend to decrease radical innovation. Since the bursting of the dotcom bubble, there has not been a similarly productive bubble, fueling innovation and economic productivity and growth. Of course, there have been technological innovations and scientific advances, but they have not been driven by an extraordinary social or technological bubble. Due to a variety of technological, political, and regulatory factors, the clean-tech bubble, for example, which burst in 2007, was not as productive. Clean energy and artificial intelligence, however, have the potential to generate the bubbles of the next economy, which might trigger a new wave of innovation and productivity. They potentially even result in the next major technological revolution. An essential element of productive bubbles can be identified in these domains: a compelling core narrative about the future that advances the frontier of technological innovation. What is needed, then, is a bubble and long-term vision that unleashes the innovative exuberance that is necessary for the kind of radical technological innovations that accelerate technological change and economic productivity. We need an innovative exuberance that will increases massive risk-taking and fuel long-term investments in start-ups that combat ageing, disease, death, and climate change. This innovative exuberance will be necessary to advance the disruptive technologies, such as artificial intelligence or clean energy, from which a new golden age can arise.
William Janeway, Doing Capitalism in the Innovation Economy: Markets, Speculation and the State. Cambridge University Press, 2012.
Charles Kindleberger. Manias, Panics, and Crashes: A History of Financial Crises. Palgrave Macmillan, 2001.
Monika Gisler and Didier Sornette, “Exuberant Innovations: The Apollo Program,” Social Science and Public Policy (2009) 46:55–68
Monika Gisler, Didier Sornette, Ryan Woodard, “Innovation as a Social Bubble: The Example of the Human Genome Project,” Research Policy 40 (2011) 1412– 1425
Carlota Perez, Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages, Cheltenham: Edward Elgar (2002)
Didier Sornette. “Nurturing breakthroughs: lessons from complexity theory” Journal of Economic Interaction and Coordination 3.2 (2008): 165–181.