Friday, September 4, 2020

"The Knowledge Economy: A Critique of the Dominant View"

From American Affairs Journal:
The knowledge economy is the science- and technology-intensive practice of production, devoted to perpetual innovation, that has begun to assume a commanding role in all the major economies of the world. It is present in every sector of these economies—in services and even agriculture, as well as in advanced manufacturing. In each sector, however, it remains a fringe, excluding the vast majority of workers and firms.

The insular character of this new vanguard of production has become a powerful driver both of economic stagnation, seen in slowing productivity growth, and of economic inequality, aggravated by the increasing distance between advanced and backward parts of the production system. It matters to the future of the United States, and of every country in the world today, that we understand both what the knowledge economy is now and what it can become. We have a stake in its deepening and dissemination: in the development of an inclusive productive vanguardism.

Paul Romer authored two papers, published in 1986 and 1990, that laid out what became—and what has remained—the most influential account of the knowledge economy within economics.1 Charles Jones has recently presented an elucidation and defense of Romer’s approach.2 An analysis of Jones’s spirited case helps reveal the deficiencies of Romer’s view as a basis for understanding the knowledge economy and its possible futures. These deficiencies speak to the inadequacy of the ways in which contemporary economics has addressed economic growth. They go more broadly to the limitations of neo-marginalist economic theory, which arose from late nineteenth‑century marginalism and remains at the heart of economics today. We need an alternative account of the knowledge economy, one that will have significant implications for economics and its methods.3
Economics and legal thought are the two major disciplines of power. When they misrepresent both how the market economy works under established institutional arrangements, and how it might work under alternative arrangements, they become enemies of prosperity and democracy, as they are now. 

The Dominant View of the Knowledge Economy
The dominant approach to the knowledge economy, inspired by Romer’s early work, consists of three elements. The first element begins with the seemingly uncontroversial claim that, unlike all previous forms of production, whether advanced or not, the knowledge economy is organized around ideas. Ideas, unlike the inputs and products of past forms of production, are “nonrivalrous”: their consumption fails to deplete them. When a new computer code or a new treatment for malaria is used by some people, it does not become any less usable by others (unless, that is, intellectual property law intervenes to limit access to its use).

The nonrivalrous character of ideas, which are supposedly the main object of the knowledge economy, in turn explains the possibility of increasing returns to scale. Without increasing returns to scale, the possibility of continued breakthroughs in productivity and growth remains doomed, as it is under the theories of economic growth that were most influential when Romer made his proposals.
The second element in Romer’s treatment of the knowledge economy is the role of profit-maximizing entrepreneurs and imperfect competition. Although ideas and their resulting material benefits may be nonrivalrous, the law of intellectual property may intervene to limit or discipline access to their use and make them, under certain conditions, “excludable.” Because ideas—or the goods and services in which they are embodied—are partly excludable and thus are not pure public goods, there is a place for private economic incentives and self-interested entrepreneurs in their development. There is also room for policies and arrangements that can bridge the gap between what private gains support and what social gains recommend.

The third element of Romer’s account is its statement in a form that allows it to be adapted to the familiar formulas and equations of the production function: the economic concept describing the relationship between the quantity of output and the different quantities of inputs used in the production process. To fit the new protagonist—idea-based production—into the linear production function of growth theory, it suffices to substitute knowledge for physical capital in the respective equations. In the revised formula, the productivity parameter, as reshaped by the knowledge-based potential for increasing returns to scale, the amount of human capital in the economy, and the equilibrium fraction of human capital allocated to research (the expansion of knowledge) determine the path and rate of economic growth.

In this way, what seemed to be a fresh approach, focused on the importance of ideas to the new economy and on their nonrivalrous character, could be taken up without disturbing the settled habits of mind and the accepted methods of the discipline. Substantive novelty could be achieved without methodological disruption.

Of these three elements, the first plays the decisive role: it makes a claim about what is distinctive to the knowledge economy and why this difference matters. The second element builds a bridge to the incentive-based workings of a real-world, imperfectly competitive market order. The third element reconciles this story of the knowledge economy with what the economics departments of the leading American research universities regard as the hallmark of the field: its diversification of mathematical models, defined by the specification of new variables and of their values, against the background of a consistent theoretical conception of economic phenomena—indeed, a background that has barely changed, in its essentials, since the late nineteenth century.

It follows from this understanding of the structure of Romer’s view that any criticism must focus on the first and most consequential of its three elements. Romer’s account of the knowledge economy stands or falls on the merits of his description of what is distinctive about the knowledge economy. 

Ideas and the Knowledge Economy
Is there—or can there be—a practice of production that, alone among all practices of production, gives a commanding role to ideas and exhibits in its development the consequences of their nonrivalry? And does that practice lie at the heart of today’s knowledge economy? The answer to both of these questions must be no.

On the one hand, there has never been a practice of production that fails to rely on ideas, and to translate them into a way of combining technology and labor. Consider the immediate precursor to today’s knowledge economy as the most advanced productive practice of its time: mechanized manufacturing and its development into industrial mass-production. The mechanical inventions and organizational innovations that sustained nineteenth- and twentieth-century industry relied on the science of the time. Like every technology before or since, industrial production connected a way of mobilizing nature (in the form of energy) with a way of organizing work (in the technical division of labor). And like every technology before or since, it developed formulaic devices and work practices aiming at increasing efficiency by having machines perform many repetitive tasks.

If mechanized manufacturing and industrial mass-production failed to achieve their full potential to improve productivity, part of the reason may have been that they evolved in ways that consigned workers to a narrow stock of repetitious movements, leading the workers to work as if they themselves were machines. The shortfall from potential productivity had more to do with the distribution of property and power than with the intrinsic limitations of the technology of the time, or of the science on which it depended.

There are no idealess practices of production. But it is equally true that there is no practice of production that traffics in disembodied ideas and their nonrivalrous, and thus potentially inexhaustible, use. Neither is today’s knowledge economy such a practice; it is not an economic equivalent to nature without friction.

Those who claim that the knowledge economy represents a jump into such a world argue that, because ideas (such as a computer code or a treatment for disease) can be used without depletion, they yield increasing returns to scale and consequently lay the basis for exponential growth. Jones makes such an argument when he distinguishes production under the knowledge economy from production as it previously evolved and was studied in the literature of economics: “Solow had constant returns to scale and therefore diminishing returns to capital, and this is what dooms growth in the neoclassical model. Now that we have increasing returns to objects and ideas together, it is not clear that diminishing returns . . . dooms us in the same way.”4 There should be no diminishing returns to the use of ideas in the way that, under the Solow growth model, there were diminishing returns to capital against the background of constant returns to scale.

Of course, the need to reconcile the economic consequences of nonrivalrous ideas with the logic of economic incentives remains: the law of intellectual property will have to allow certain economic agents to control, in their own interest, access to otherwise inexhaustible ideas. By being made excludable, the ideas—or the products in which they result—become objects of an economic interest and even the sources of a rent. There is always the danger that the legal logic of excludability will interfere with the advantage of nonrivalry and with its promise of increasing returns to scale. In this narrative, however, the risk of such interference is a price that we must pay to bring the revolutionary advantage enjoyed by nonrivalrous ideas down to the earth of interests and incentives.
Before exploring the confusions and fallacies that lie at the heart of this narrative, it is important to pause on a prior point: the conception of the knowledge economy as a practice of production that is not only organized around ideas but that also shares in the attributes of ideas. Because it shares in their attributes, it differs from all previous forms of production, which had no such internal and intimate relation to ideas.

Today’s knowledge economy does differ from earlier practices of production, but it is no easy matter to settle on an accurate view of what distinguishes it. In each sector in which the knowledge economy has taken hold, it remains an insular vanguard, underdeveloped and even elusive. A practice of production reveals its deepest features and higher potential only as its spreads across a wide range of forms of economic activity, but the knowledge economy has yet to do that. We cannot do justice to its nature—in the arrested version that it now takes or in the deeper and widespread form that it might take in the future—by representing it as an economy of ideas that happens to be given material expression and to intersect with economic arrangements and incentives. Just as no practice of production has ever been idealess, no practice of production can ever work, or be understood, as bodiless—if by the body we mean the detailed ways in which we bring together our discoveries of the workings of nature, our mechanical devices, and our regimes of cooperation.

Consider the example of the application of artificial intelligence to the task of organizing the way in which high-skill workers use 3-D printers and other numerically controlled and adjustable machine tools in advanced manufacturing. Among the tasks to be addressed is the allocation of machine and labor time among three main tasks: (1) the manufacture of variable batches of relatively standardized products; (2) the occasional development, by analogical extension, of new variants of these standardized products; and (3) exploratory tinkering with the goal of creating products that no one ever envisioned before but that people working with advanced machines discover that they can make. This third category evokes the difference between a national defense strategy that is organized around foreseeable combat situations (for example, a conventional war on a particular frontier or an asymmetrical proxy war against a distant power manipulating a troublesome neighbor) and a strategy that is built around military capabilities and their alternative uses. The allocation of resources and time among these three tasks is a problem that must rank high in the concerns of a contemporary knowledge economy. As in the role of algorithm-based investment decisions in finance, the division of labor between machines replacing human minds and minds monitoring and using machines remains open to experience and experiment.

How can the development of artificial intelligence in computer science and mechanical or electrical engineering be driven by attempts to accomplish these three tasks? How can production be set up so that it becomes part of the process of perpetual innovation rather than a passive, occasional beneficiary of scientific discoveries and technological breakthroughs external to itself? What kind of worker, educated in what way, can rise to this challenge? What way of organizing work can make the most of technical and economic opportunities? And under what legally defined institutional arrangements and cultural impulses is such a way of organizing work most likely to prevail and flourish?

None of these questions speaks to the conventional account of the knowledge economy as a practice of production organized around disembodied, nonrivalrous ideas that the law of intellectual property allows profit-seeking entrepreneurs to turn into sources of rents. All these questions concern the nature of innovation—its episodic or perpetual character, the extent to which it is or is not integrated into productive activity, and the degree to which experimentation is favored or inhibited by established institutions and culture. It is there, rather than in the specious contrast between production-as-ideas and idealess production, that we must look for what makes—or can make—the knowledge economy distinctive. And the distinctiveness is likely to show and to count only as the knowledge economy begins to break out of the insular vanguards to which it has thus far remained confined. 

Increasing Returns to Scale and Diminishing Marginal Returns
If the first basis of Romer’s view is an untenable contrast between a practice of production centered on ideas (supposedly today’s knowledge economy) and all earlier productive practices, the second basis is a thesis about the potential of the knowledge economy to yield increasing returns to scale. To understand the content of this thesis and the mistake that it makes, we need to address an unrecognized source of confusion in standard economic theory: the relation between constant returns to scale and diminishing marginal returns. The confusion is all the more significant and influential—yet largely disregarded—because it forms part of the basic education of economists (as well as that of the much larger number of young people who are exposed to introductory courses in the discipline but go on to specialize in other fields). One can often find this confusion, misrepresented as simple logic, in an early chapter of many textbooks.....
....MUCH MORE