Harnessing Science and Technology for America's Economic Future: National and Regional Priorities (1999)

Richard R. Nelson

Professor, Columbia University

In this brief talk, I will touch on three themes. First, I shall make some observations about the history of economic thought on technical advance and economic growth. Then, I will identify a few salient features of technological change that research has highlighted, which seem highly germane to the present policy discussion. Finally, I will lay out several matters bearing on the connections among technical change, economic growth, and public policy, that I personally would stress.

First, some history of thought. Reflecting on what some early economists had to say about technological advance and economic growth seems valuable for at least two reasons. First, it brings out clearly that much of modern growth theory in fact has been the understanding of economists for a long time. Second, several very important elements of the earlier articulations have been repressed, or lost, in the more contemporary ones. They need to be brought back in. Adam Smith's The Wealth of Nations, written in 1776, is largely a book about economic growth. At the time he wrote it, the first industrial revolution was moving into full swing, and Smith, the keen observer, makes a number of insightful remarks about technical advance and economic growth. In reading Smith, one can see that the current emphasis on increases in capital per worker and on technological advance as sources of growth is no new conception:

But note that in Smith's account, growth of capital per worker and invention are not separate sources of growth, but are tightly connected. And both are related to the organization of industry. Modern growth theory has only recently been scrambling back to these important insights.

Thomas Malthus is another great classical economist whose central interest was economic growth. The ''limits to growth'' discussion often is regarded as a modern conception, but nearly two centuries ago Malthus raised the issue in his prediction that the fixity of land would, ultimately, limit population growth and economic progress. A key part of Malthus' analysis involved his proposition that technical change in manufacturing would be significantly greater than technological advance in agriculture. Historically, that has turned out to be wrong. But here, nearly two centuries ago, one can see economists recognizing that technological advance is unlikely to proceed uniformly across all areas of economic activity, and considering the implications. Unfortunately, much of this perception is missing in modern growth theory.

And, of course, technological advance was central in Karl Marx's analysis. In Marx technological advance is an essential element of the competition among firms. Under the force of competition, firms are inexorably driven to adopt new technologies that substitute capital for labor. The result for Marx was as much rising unemployment as it was rising productivity. One can see here the origins of the modern dispute about the effects of automation. By and large, technological advance seems not to have caused widespread unemployment. But the issue is repressed in most modern growth theory, which simply assumes full employment.

For a variety of reasons, toward the end of the nineteenth century, the interest of economists in economic growth diminished. While there are important exceptions, it is fair to say that strong interest in technical change and economic growth only returned to economics after World War II. Here one can recognize two quite different strands, which were then, and still are, somewhat at odds.

One was due to Schumpeter. Schumpeter was strongly influenced by Marx, at least in his insistence that technological innovation and industrial competition are closely intertwined. While Schumpeter's argument that innovation and change are central to economic activity goes back to his Theory Of Economic Development, written in 1911, his writing that has had the greatest influence on contemporary analysis is his 1943 volume, Capitalism, Socialism, and Democracy, where he made two important arguments.

The first is that technological change and economic growth involve disequilibrium in a fundamental way. Technological advance, and competition in industries where technological advance is important, proceeds through a process of "creative destruction." Schumpeter's second important contribution was to call attention to the relationship between industrial structure and technological advance. In particular, he called attention to the fact that, while innovation was a central form of competition in many industries, the structure of those industries never was "perfectly competitive" in the sense articulated in standard macroeco-

nomic theory textbooks. The second of these themes spawned a large stream of empirical research exploring the relationships between economic structure and technological advance, which I shall refer to shortly. The former theme, that technological advance must be understood as a disequilibrium process, has been ignored by most of my colleagues in economics, although it now is an understanding shared by virtually all empirical scholars of technological advance.

The strand of economic research that undoubtedly has had the greatest influence on thinking about the relationships between technological advance and economic growth, however, did not stem from Schumpeter, but rather from the work of a group of economists working in the late 1940s and early 1950s for the National Bureau of Economic Research, using the new national product data. Solomon Fabricant, Moses Abramovitz, and John Kendrick were among the most prominent of this group. These early, postwar studies led on the one hand to modern growth accounting, an art form pushed much harder by Edward Denison, and on the other hand to the empirical research based on neoclassical growth theory, stimulated by the original work of Robert Solow. While different in the details of the work, it is these kinds of studies that led, early in the game, to such propositions as "technological advance accounts for 80 percent of the productivity growth the U.S. economy has experienced," and this kind of analysis continues to provide the standard measures of technological advance.

In any case, the early work of Schumpeter, of Abramovitz and colleagues, and of Solow, set in train a significant body of research by economists focused directly on trying to understand technological advance. Prominent among the group of economists who got into the field in the 1960s were Jacob Schmookler, Edward Mansfield, Zvi Griliches, Nathan Rosenberg, and Christopher Freeman. My colleagues at this symposium—Richard Rosenbloom and David Mowery—also have done important work on technological advance. Here I want to call attention to four different features of technological advance that this body of research has highlighted.

First, the process of technological advance involves uncertainly in a fundamental way. The processes are full of surprises. There are winners and losers. And generally, it is not possible to guess in advance who and what will win. Schumpeter was right in arguing that technological advance is a disequilibrium process of creative destruction. I long have been completely persuaded of this. The fundamental uncertainty involved in technological advance seems to be the basic reason why detailed, long-range planning is doomed to frustration and often disaster, and why, to get rapid advance of technology, society generally needs a variety of different parties trying out different bets. Regrettably, this issue is repressed in much of contemporary analysis.

Second, while many economists read Schumpeter in his Capitalism, Socialism, and Democracy as proposing that technological advance always is associated with large firms with a considerable amount of market power, we now know that no generalizations of this sort hold up. In some cases the Schumpeterian

proposition seems to hold. Consider the chemical products industry, pharmaceuticals prior to biotechnology, and mainframe computers. But we also have experienced rapid technological advance in industries marked by the absence of large dominant firms, and considerable entry of new firms. This characterizes semi-conductors in the era after the invention of integrated circuits up until a few years ago, software, and biotechnology, at least in the United States. The kind of firm, and the industrial structure, that is conducive to technological advance tends to vary from industry to industry, and even within an industry to vary over time. Thus reflect on the history of computers.

Third, more generally, there are very considerable differences across the fields of technologies and industries in the way technological advance proceeds. Broad generalizations here tend to get one into trouble. Thus consider the importance of patents. Patents indeed are important in pharmaceuticals, and in certain other areas of fine chemical products, but in other high-tech industries it would appear that firms can profit handsomely from their innovations through mechanisms that do not involve patents in an essential way, like exploiting a headstart. Thus much of the history of semiconductors, and computers, has proceeded in a regime of weak intellectual property rights. And industries and technologies also differ significantly in their links with science.

Which leads me to the fourth proposition I want to highlight. It is that virtually all technologies that have experienced rapid advance are connected to various fields of science, or engineering research, that undergird them. There are a variety of different ways to measure the connection of a technology to fields of science. But using any of these measures, the correlation between technical advance and the strength of the science ties is high.

Let me turn finally to a few matters relating to technical advance, economic growth, and public policy that I personally would stress. As I reflect on these propositions, they all have an element of warning against oversimplification.

First, I am very uncomfortable with the attempts of my colleagues in economics to "divide up" credit for economic growth between capital formation, education, and other input increases, and technological advance. I think it important to understand the economic growth process as involving strong interaction among various elements. To go back to Adam Smith, it makes little sense to ask how much growth we would have experienced if capital per worker had increased as much as it has and we had had no technical advance, because it is technological advance that enabled the growth in capital intensity that we have experienced. Similarly, the rise in educational attainments that has been experienced by modern growing economies has not been a source of growth independent of technical advance, but rather has been an essential input to and complement of the technological advance we have experienced. Technological advance must be understood as part of a package of ingredients that generate economic growth.

Second, regarding the roles of public policy toward technological advance, I think it generally a mistake to think of a choice between government and markets,

or of government policies being justified because of "market failure." On the one hand, as Schumpeter argued long ago, technological advance itself is associated with and generates all kinds of deviations from the economists' benchmark condition of "pure and perfect competition." But that is no reason for having government try to control the process. On the other hand, virtually every economic sector that has experienced rapid technical change has been supported in one way or another by range of government programs that, at the least, have nurtured the underlying sciences. Sometimes, as in agriculture, public programs have supported quite applied research. In other cases, such as in chemical products and pharmaceuticals, the role of government research support has primarily been focused on more fundamental work. But there is a lot of variety from sector to sector regarding the division of labor.

In my view, technological change has been the central driving force behind the economic growth we have experienced. However, the relationships and mechanisms involved are complex not simple. I hope my remarks today have brought into view both some of the key relationships and their complexity.

Smith, A. 1994. The Wealth of Nations: An Inquiry into the Nature and Causes. New York: Random House.