Published by EH.NET (August 2003)

Nathan Rosenberg, Schumpeter and the Endogeneity of Technology: Some American Perspectives. London and New York: Routledge, 2000. vii + 125 pp. $85 (hardcover), ISBN: 0-415-22652-X.

Reviewed for EH.NET by Avi J. Cohen, Department of Economics, York University, Toronto, Canada.

This slim but important and compelling book is required reading for anyone who (wrongly) believes that endogenous growth theory has successfully incorporated a theory of technological change, or who is interested in (correctly) understanding the historical process of technological change.

The book consists of five chapters, four given as the Schumpeter Graz Lectures, plus a shortened version of a paper from Helpman (1998): 1) Joseph Schumpeter and the economic interpretation of history; 2) Endogeneity in twentieth-century science and technology; 3) American universities as endogenous institutions; 4) Innovators and “mere imitators”; and 5) Chemical engineering as a general purpose technology.

As a happy accident, I happened to read Nathan Rosenberg’s book at the same time as reading the second edition of Robert Solow’s Growth Theory. Solow, in his Intermezzo (pp. 97-105) on endogenous growth theory, presents the standard representation of labor-augmenting technical progress:

y = A(t) f(k/A(t)),

where y and k are per-unit of labor output and capital.

To endogenize technical progress is to provide a theory of the evolution of A(t), but Solow stress the “arbitrariness” in “every one of the major contributions to the theory of endogenous growth, [where] one can spot the moment when the key assumption is planted that makes A(t) or its moral equivalent grow exponentially, so that it can be said that the model determines the growth rate.” And one “usually gets no justification” for this assumption, whether in the Lucas model, Romer model, Grossman and Helpman model or Aghion and Howitt’s Schumpeterian model (p. 100).

This is where Rosenberg steps in — with his more comprehensive description of the process of technological change from Schumpeter’s (and Rosenberg’s) perspectives. That Schumpeterian process of technological change, outlined in Chapter 1, is a quite radical shift in perspective — involving historical analysis, a focus on disequilibrium and with tastes and social institutions as endogenous — rather than a single endogenous mechanism to be added to an otherwise exogenously determined growth model.

Rosenberg (p. 3) quotes Schumpeter on why history is more important for economic analysis than either theory or statistics: “First, the subject matter of economics is essentially a unique process in historic time. … Second, the historical report cannot be purely economic; therefore it affords the best method for understanding how economic and non-economic facts are related to one another and how the various social sciences should be related to one another. Third, … most of the fundamental errors currently committed in economic analysis are due to a lack of historical experience more often than to any other shortcoming of the economist’s equipment.” For this view, Schumpeter expresses his “admiration for and intellectual indebtedness to Marx”: “He was the first economist of top rank to see and to teach systematically how economic theory may be turned into historical analysis and how the historical narrative may be turned into histoire raison?e.”

This historical perspective also involves a disequilibrium focus. For Schumpeter, “capitalism has to be understood as an evolutionary system rather than as a system that continually reverts to some equilibrium after small departures from it. … this evolution is a reflection of certain dynamic forces … inherent in the incentive structure, the pursuit of profits, and the competitive institutions that lie at the basis of capitalism.”

In Chapter 2, Rosenberg sketches three inter-related phenomena that endogenize the process of technological change — R&D labs, the engineering professions and focusing mechanisms for profitable innovation. Economic forces shape the activities of industrial R&D labs as follows: “The role of industrial scientists and engineers is to improve the performance and reliability of those technologies and reduce costs, as well as to invent entirely new technologies. Thus, the industrial research lab has had the effect of subjecting science to commercial criteria. In so doing it has rendered science more and more an endogenous activity, whose directions are increasingly shaped by economic forces and concentrated on the achievement of economic goals” (p. 25). Thus, the engineering professions are key to endogenizing science because of the possibility of converting scientific research into marketable products and processes. Finally, a major innovation, whether by scientists or engineers, focuses the quest for further profitable innovations. “A major technological breakthrough typically provides a powerful signal that a new set of profitable opportunities has been opened up in some precisely identified location. Consequently, it is understood that scientific research that can lead to further improvements in that new technology may turn out to be highly profitable” (p. 30). This is a theme Rosenberg has long explored in Rosenberg (1972, 1976,) and Mowrey and Rosenberg (1989).

For example, in the iron and steel industry, the search for methods to neutralize the deleterious effects of phosphorous iron ore not only produced technological change, it also made the range of useable ore inputs endogenous. In electronics, the invention of the transistor in 1948 led to improvements in solid-state physics, instead of the imagined reverse causation from science to industry. The transistor dramatically upgraded the potential payoff to research in the solids state, catapulting this minor subdiscipline that was not even taught at most universities into the largest subdiscipline of physics. And the invention of laser technologies led to a major push in fiber optics research (and innovations), because the scientific and profit potentials for telephone and data transmission were clear early on.

Chapter 3 discusses universities as producers and transmitters of economically useful knowledge, primarily technological knowledge. In the nineteenth and twentieth centuries, American universities were more competitive and decentralized than their European counterparts, benefiting from the complementarities of teaching and research, particularly in advanced graduate education. Rosenberg documents the crucial role of land grant universities in disseminating agricultural knowledge and responding to the problems and concerns of farms. In aeronautical engineering, the emigration of German aerodynamic professors to Cal Tech in Pasadena in 1930 ultimately led to the commercial Douglas DC planes. There are other examples from electrical engineering and computer science that result in prototypes for industry, leading Rosenberg (p. 56) to find that “American universities constitute, among other things, a huge economic enterprise, one that has been powerfully shaped by, and highly responsive to, economic forces. But the universities are also, by their very success, reshaping the structure and performance of the American economy.”

In Chapter 4, Rosenberg takes issue with Schumpeter’s distinction between entrepreneurial innovators and the “mere imitators” who carry out the diffusion of the new technology. Rosenberg’s contrary hypothesis, which accounts for the empirical difficulty of connecting technological innovations and subsequent productivity growth, is that numerous incremental improvements with cumulative significance are more important that breakthrough innovations. “Schumpeter was fond of speaking of ‘mere imitators,’ because in his view all they need to do is to follow in the footsteps of the entrepreneurs who have led the way, and whose earlier activities have resolved all the big uncertainties. My own view is that, on the contrary, these so-called ‘mere imitators’ may be far more than imitators. In fact, they have commonly been the essential carriers of an improvement process that decisively shapes the eventual contribution of new technologies to productivity improvement” (p. 62). There are huge uncertainties at the start of new technologies, which can persist — the impact of the camera, which is over 150 years old, is again being transformed by digital imaging. To make connections between innovations and productivity, one must understand the trajectory of later improvements of a technology (in the hands of Schumpeterian “imitators”). Technological complementarities often shape the eventual consequences and productivity of new technologies.

Rosenberg identifies key source technologies that have been built upon by complementary developments — steam engines, machine tools, electricity, transistors, computers, lasers. He classifies these as general purpose technologies (GPT), each of which “makes possible an increase in the productivity of R&D in a number of ‘downstream’ sectors of the economy. … as the GPT advances, it enlarges the range of opportunities for downstream applications, and the awareness of such possibilities, in turn, feeds back upon the incentive to perform R&D in the GPT sector as well as downstream” (p. 66).

This leads Rosenberg (p. 78) to exhort us to “”pay more attention to the extremely disorderly process that follows upon the first introduction … of a technological innovation. This should include not only the critically important subsequent improvement process, but also the innumerable and subtle ways in which technology is sorted, matched, and modified to suit the huge diversity of ultimate user needs.”

In Chapter 5 Rosenberg argues that GPTs are not limited to hardware. He takes a detailed look at chemical engineering, a set of ideas, as a GPT: “the engineering disciplines [are] the repositories of technological knowledge, and their practitioners [are] the primary agents of technological change in their respective industries” (p. 82).

In pointing out the limitations of endogenous growth theory’s analysis of technological change, Rosenberg, like Solow, calls for a deeper understand of process. If technological change is endogenous, “then surely that Endogeneity must include the growing body of technological knowledge that provided the intellectual basis for the design and construction of new technologies” (p. 79). This is part of the radical nature of Rosenberg’s approach that goes far beyond a simple endogenization of one variable in a model.

Solow (2000, p. 101-2) describes “growth theory ‘proper’ [as] the study of the long-run behavior of an economy conditional on A(t). But then there is a separate, though closely related, field of study that is concerned with A(t) itself, or more generally with the understanding of the process of technological change. … the only reasonable basis for a more nearly endogenous theory of growth would appear to be a serious analysis of the determinants of innovation and technological process.” This description is exactly in line with Rosenberg’s (p. 35) view of the contribution that economic historians can make, being “interested in Endogeneity, not only as a theoretical modeling exercise, but also as an empirical phenomenon of growing significance throughout … the twentieth century.” The end result Rosenberg (p. 81) envisions, of “more fully endogenizing the growth of knowledge upon which technological change depends, calls for a joint enterprise involving theorists, historians, and engineers.” To that I say, “Amen.”


Helpman, Elhanan (ed.), General Purpose Technologies and Economic Growth. Cambridge, MA: MIT Press, 1998.

Mowrey, David and Nathan Rosenberg, Technology and the Pursuit of Economic Growth. New York: Cambridge University Press, 1989.

Rosenberg, Nathan. Technology and American Economic Growth. New York: Harper and Row, 1972.

Rosenberg, Nathan. Perspectives on Technology. New York: Cambridge University Press, 1976.

Rosenberg, Nathan. Inside the Black Box: Technology and Economics. New York: Cambridge University Press, 1982.

Solow, Robert M., Growth Theory: An Exposition, second edition. New York and Oxford: Oxford University Press, 2000

Avi J. Cohen was trained as an economic historian and authored “Technological Change as Historical Process: The Case of the U.S. Pulp and Paper Industry, 1915-1940,” Journal of Economic History, September 1984. Recent work in the history of economics includes “The Hayek/Knight Capital Controversy: The Irrelevance of Roundaboutness, or Purging Processes in Time?” History of Political Economy, forthcoming, Fall 2003 and “Whatever Happened to the Cambridge Capital Controversies?” (with G.C. Harcourt), Journal of Economic Perspectives, Winter 2003.