 The foundations of the digital revolution were all sponsored and funded by the United States Department of Defense. The context was the Cold War, which legitimized the design and implementation of perhaps the most significant industrial policy since medieval mercantilism. The history of ICT is a prime example of the economic power, not just the scientific and technological value of mission-driven R&T. So we start from where we were in lecture four. We must step outside of the framework of neoclassical economics to grasp the dynamics of the ICT revolution. DOD, the Defense Department, orchestrated the coordination that the market could not provide. We must step outside the framework of national income accounting to appreciate the value generated by mission-driven research and development. Now by definition of the national income accounts, the public sector output equals the input. No productivity gain is possible by construction. Plus federal financing of ICT as a public good had enormous implications for intellectual property and accelerated the diffusion of new knowledge. Contrast this open innovation environment with Watt and Bolton's holdup of the steam power general purpose technology. And note the importance of the procurement function and perhaps most important, the institutional innovation embodied in the belated establishment of American research universities. Now some specifics. First, digital computers. The military supported multiple projects, not a single national champion. The history of the first general purpose programmable computer, ENIAC at University of Pennsylvania, is exemplary. It was designed to calculate artillery firing tables for the Army and then repurposed to explore the feasibility of building a hydrogen bomb. John von Neumann was a universal genius based at the Institute for Advanced Study in Princeton. His notes were compiled from sessions at Penn exploring the possibility of a next generation computer with the ability to store and execute its own programs. The EDVAC. John Mawkley and Prosper Erkert, the designers of ENIAC, bitterly resented the transfer of their IP to the public domain. They were finally awarded a patent that in turn was voided because of the existence of prior art, art that was arguably their own. But letting out the model for a programmable digital computer broadly across the domain of interest radically accelerated the development of computing. Second, semiconductors. Now here note the disparity. The established firms, the Giants, AT&T and IBM, got most of the R&D contracts. But the procurement contracts went to the new firms like Fairchild Semiconductor. And then on to the birth of software, SAGE, the semi-automatic ground environment to manage the network of radar stations established to track a potential Soviet attack. It was obsolete by the time it was completed. And missiles succeeded bombers, which were all that those radars could track. But software as an independent craft was born. Finally, upstream investment in creating the enabling discipline, computer science. The same sort of federal support was also devoted to solid state physics, the basis for semiconductor processing. The digital revolution was also accelerated by a different kind of state intervention. Antitrust, again reducing the ability of the investors in R&D to control, appropriate, and limit access to intellectual property. The positive consequences were enormous. The antitrust interventions opened up the IT sector to competitive entry by new companies. From 1980 to 1983, the nascent information technology sector went through profound structural change. Recall the McNamara depression of the early 1960s that forced micro-electronic firms to seek commercial markets. By contrast, in the early 1980s, commercial markets, above all the new market for personal computers, pulled IT firms away from the military market. Millspec, designing products to meet the military specifications, became a niche, a narrow niche characterized by extreme performance requirements. Eventually, the military migrated to COPS, commercial off-the-shelf hardware and software, as the commercial market defined the priorities for all participants in the industry. The broadly defined computer industry has evolved through a succession of platforms. IBM System 360 was the dominant computing platform for what turned out to be a short generation. But the PC platform was delivered by the wind-tell duopoly of Microsoft and its software and Intel and its computer chips, not by a single computer manufacturer. Today, there is a standard platform, open-source software, the Apache Stack, riding on the cloud computing services from Amazon, secondarily Microsoft, and with Google struggling to catch up. A platform that is incredibly open and incredibly cheap. The result is the flood of software-defined, web-delivered services, indeed a Darwinian explosion of innovative applications. The first critical architectural transformation occurred in the 80s, the shift from a vertical, closed and proprietary industry to a horizontal industry, open network through published interfaces and protocols. Now this treats the Internet only as a communication medium, but of course it is much, much more as we shall see. Now the location of the Internet initially, within the public sector, contradicted the principles of market fundamentalists and constrained any commercial evolution of the network technology. The critical role of the National Science Foundation, however, a public lead user of the Internet, led the transition from the Defense Department to a privatized medium for commerce as well as for information. The open character of the public sector Internet carried over into the private sector, at least it has until now when net neutrality seems to be under threat. But a set of critical technical choices were made in the early days on which openness depended, especially what's known as TCPIP, Transmission Control Protocol, Internet Protocol, chosen in the face of proprietary offerings, especially from IBM, that would have closed the Internet and led to control of access from the beginning. Then the gift to the world from Sir Tim Berners-Lee of CERN, the World Wide Web. Shane Greenstein identifies the distinctive organizational characteristics of the Internet with respect both to governance and technological direction. It's the integration of communications with transactions, plus the application of real-time analytics to communicative and transactional behavior that distinguishes the Internet as a unique environment and enables the proliferation of the seemingly limitless range of applications. Early perception of these factors was catalyzed by Netscape's IPO with the first commercial web browser in 1995. The outlines of a new economy began to become visible and the basis for a bubble emerged. Today, the digital revolution has grown up. It's no longer dependent upon state sponsorship and support and in fact, the digital revolution has evolved to attack the authority of the state at the micro level, enabling assaults on the regulatory ecosystems of established markets and at the macro level, enabling a new wave of automation, globalization and financialization that challenge the state's ability to buffer its constituents from disruptive change and at the most fundamental level, the digital revolution has been undermining the integrity of the political process on which the authority of the state ultimately rests. Digitalization of physical activities transforms atoms into bits. Digitalization eliminates technological frictions, but other frictions remain. All of these real-world frictions can slow down the diffusion of innovation and the growth of the disruptors. They may even call into question the long-term sustainability of the digital disruptors. Recall discussion of the unicorn bubble in lecture three and note that a characteristic flaw in the world view of those who know they are inventing a new world is not to pay attention or to understand how this world came into existence and how it works until it bites them in the ass. During the 1990s, value in computing shifted from hardware to software. Hardware was commoditized and software became the vehicle for creating and capturing value. Now it's shifting from software to data as software in turn is commoditized to availability of open source system software and are bundled into cloud services. The feedback from more data to better algorithms in principle is complex. In some domains, such as identifying a unique individual, there appears to be diminishing marginal productivity from incremental data. Relatively few data points give you what you need to know whether or not you should know it. In other domains, such as natural language processing, there may be no such constraint or we certainly haven't come close to reaching it yet. But the first effect of more data in general is a lower signal to noise ratio. That is, the number of potentially false correlations rises exponentially with the amount of data and the challenge of identifying genuinely causal relationships becomes both harder and more urgent. This is why the promise of machine learning is at best somewhat problematic and at worst offers a veneer of objectivity to data that embeds prejudice. But even as the digital revolution pervades economic and social and political life, productivity growth has slowed markedly across the developed world, generating an intense academic and policy debate. There are multiple candidate sources for mis-measurement of productivity, including missing output due to economic activities crossing the production boundary. That is, work that used to be paid for and captured in the national income accounts become self-service. Second, final goods become intermediate services that drop out of GDP. For example, rental of cloud computing services substitutes for the purchase as final goods of computers themselves. Overstated inflation as a third factor, cost-reducing innovations come too fast for correction. But mis-measurement on its own appears insufficient. And in fact, average productivity growth itself can be very misleading. The OECD has tracked the best versus the rest across both manufacturing and services sectors. Productivity growth for the top 5% has continued on trend as with the diffusion of all technologies. It takes time to master the new. The same pattern shows an analysis of some 30,000 UK firms. Look at the enormous gap between the 99th and the 90th percentile of productivity growth. Multiple studies show that the productivity puzzle is bound up with several related observables. Increased industrial concentration, rise in market power, decline in labor's share. This paper by a brilliant young graduate student at Cambridge University examines the impact of the increase in intangible capital that is not stuff you can kick but intellectual property services, IT algorithms due to the adoption and exploitation of digital technologies. Intangible capital reduces marginal costs and increases fixed costs. It confers competitive advantage on the first movers whose lower costs in turn chill competition from potential new entrants and cut aggregate productivity growth in the longer term. So this is a story about how digitalization has increased market power and reduced productivity growth. Thomas Philippon's important new book, The Great Reversal, examines in detail the sources and consequences of slower growth of productivity and increased concentration in US industries and markets accompanied by increased share of revenues going to profits and lower share going to wages. Philippon in his book sorts through a range of alternative explanations to find that two hypotheses, the rise of superstar firms and decreasing domestic competition are consistent with the pattern of increased concentration and increased profit margins. But the pattern of investment and productivity growth reject the superstar hypothesis as dominant. And a recent paper by Axeget and Ateves takes the argument one step further, reduced dynamism expressing itself in slow productivity growth and in other observable ways once again reflects the perverse consequences of an excessively tight intellectual property regime with patents concentrated amongst a relatively small set of incumbent market leaders. Even while the productivity puzzle has unfolded in the statistical economy, numerous studies from different sources protect increased impact on the labor market from increasingly intelligent automation. Like previous technological revolutions, the digital revolution has transformed both the content of work and the management of work. So did the introduction of the factory system in the first industrial revolution. So did the introduction of the production line in the second industrial revolution. Indeed, the digital revolution may have more disruptive impact through the algorithmic management of work versus changes in the content of work. The rise of the gig economy is generating endogenous labor market and political responses from guilds of Uber drivers and the political response has actually reached a segment of Republican-aligned conservatives. This is a statement, very recent statement from a Republican think tank arguing that workers need to have a seat at the table. When we turn to the macro consequences of the digital revolution, they transcend national borders. They have transformed the global economy and spilled over to disrupt the political process across the developed world. The first globalization peaked at the start of World War I. It too was enabled by new technologies, steamships, railways, the telegraph and telephone networks. It involved free movement of people and capital, trading goods with limited by protectionism in the top follower nations, the United States and Germany and gross trade peaked at about 30% of world GDP. The second globalization, our globalization was enabled by IT. It involved free movement of goods and services and capital. Movement of people was clearly not so free but IT enabled the movement of work internationally also known as outsourcing. Gross trade peaked at no less than 60% of world GDP. The pursuit of efficiency in the construction and management of global supply chains directly and necessarily reduced their robustness as we have learned under the impact of the coronavirus pandemic. IT driven globalization in the real economy eliminated buffer stocks and redundancy and left the production system of the world dangerously fragile. Now, financialization is the third dimension of the digital revolution's transformation of economic life. This chart shows the sheer scale of the financialization of the US economy. From 1950 to 1980, the value of financial assets relative to GDP rose only from 1.3 to 1.8 times. As a compound annual growth rate, you could barely see it. But from 1980 to 2007, the ratio of financial assets to GDP rose from 1.8 times to 4.8 times. As discussed in lecture five, the great credit super bubble that burst in 2008 could not possibly have emerged without IT, expressing cataclysmically in the huge and unsustainable increase in leverage. For each individual bank, more leverage increases its capital efficiency. Each dollar of capital supports more loans and other assets and generates more income in absolute terms and in higher returns on capital. Oh, yes, it also generates higher bonuses. For the system, increasing leverage reduces the robustness of the system just as with physical supply chains. Immediately prior to the global financial crisis, the banking system was so leveraged that a mere 3% decline in asset values would have rendered it insolvent and did so. Again, efficiency and robustness are at war with each other. And of course, globalization and financialization interacted. The key measure in this chart is the volume of cross-border claims of banks on other banks. The thicker the link between regions, the greater the volume of loans. Between 2002 and 2007, just five years, US claims on Europe rose from half a trillion dollars to more than one and a half trillion dollars, a factor of three. And Europe's claims on the US rose from one trillion to two and a half trillion. As with the global supply chains in the real economy today, a decade ago, the global financial networks enabled by IT proved shockingly fragile. Globalization of the world's financial economy has generated the political dilemma defined by Danny Rodrick of the Harvard Kennedy School. You can have deep economic integration. You can function as an autonomous nation state. And you can have representative, responsive government, two out of three. The Eurozone has been a prime stage for the acting out of that trilemma. Responses to the political stresses generated by the trilemma express themselves both domestically in the rise of polarized populism and internationally in the rise of barriers to trade and to migration. At the most fundamental level, the digital revolution is undermining the integrity of the political process on which the authority of the state rests. And so in one generation, the relationship between the state and the digital revolution has been reversed. And so the IT revolution sponsored by the state and funded by speculation feeds back not only to transform the market economy, it also conditions the political dynamics that shape the capacity of the state to offset and balance the coordination failures and self-destructive outcomes of markets disrupted by those same digital technologies as their participants operate under conditions of radical uncertainty. The retreat from hyper-globalization was already well underway before the impact of the coronavirus pandemic. Managing that process through and beyond the pandemic will continue to test to the utmost the competence and authority of states around the world even as they are urgently called upon to address the next needed technological revolution, the green revolution in response to climate change, the subject of the final lecture in this series.