Learning and Technological Change

Technological change has become a major public concern. Some countries face problems of declining competitiveness, stagnant wages and regular trade deficits. Many more face pressures from slowing growth of productivity and national income. Third World countries see technological change as a means to develop, while former Eastern bloc countries explore the relation of the market to productivity growth. Governments everywhere try to formulate productivity-enhancing policies. Clearly, in order to understand economic trends and formulate constructive policy we must understand technological change.

This chapter develops the following themes. First, technical advance needs to be understood as a cultural evolutionary process. Models by economists that treat it as a maximizing process miss the point, as do the models that treat it as just like a model in biology. Second, what provides private incentive to invest in research and development (R&D) is that the fruits are at least partly a private good. However, new technology has a public aspect as well, and this is what makes the evolutionary process cultural. As a result, technology advances through the work of a community of technologists who are both rivals and colleagues. Third, to a good first approximation the private part of technology is the specific application, and the public part generic. This is so for two reasons: the means that firms have to protect their inventions largely shield the specifics, but the generic aspects of their inventions are almost impossible to protect; and in addition, universities are scientific and technical societies and are in the business of making generic knowledge public. Fourth, technological communities (and the technological change they engender) have become more science based, and more transnational, in recent times. The consequences are profound.

At the core of business history is the firm. There the practices and activities of business are developed, modified and reshaped. There the processes of industrial production, distribution, communication and finance are carried on in modern capitalistic economies. And there much of the learning so central to technological change takes place. To paraphrase Joseph Schumpeter, capitalism has been an engine for technological change and for-profit firms in rivalrous competition are the featured actors in this evolutionary drama (1942, Chapters 6–8). The primary role of the business firm in technological change has been that of the development, not the discovery, of new products and processes.

History of technology in the post-medieval Western world has to be a history that deals with dynamic change, one that attempts to explain how positive feedback took place among increasingly integrated material and social systems, so that technological change became not an occasional, but an incessant, event. Since the concept of intellectual property protected by patents for invention emerged in history at the same time as this increasing dynamism of technology, it is tempting to assume there is some connection between the two.¹ On that assumption, students of technological change have frequently viewed patents as incentives to invent or as a measure of invention. This chapter does not debate the validity of those views, but attempts to outline quite a different feature of a patent system in action: as an institution it elicits behaviour aimed at patent management, which in turn acts as an invisible college of technology, in which learning stimulates further technological change.

For over three decades now historians of technology have been exploring the dynamics of technological change. Unravelling the paths by which new technologies are discovered and implemented has proved to be an intellectually engaging enterprise that has largely sustained the discipline. Much of the fascination with the process of technological change results from its complexity. On the other hand, this complexity has discouraged the development of more general approaches to the subject. Lacking explicit alternatives, most historians of technology have fallen back upon a linear model of technological change which links invention, development and commercialization in a sequential progression. Because the last two steps depend on the first one (invention), it is invention that has received the most scrutiny by historians. In fact, the concept of invention is tightly bound up with the linear model. Below it will be argued that the use of the linear model has distorted our understanding of technological change and that an alternative model, a Darwinian evolutionary one, more accurately expresses the complexity that emerges from the narrative histories.

Institutions generating new techniques form a principal engine of capitalist growth, but what determines the output of this engine? One promising answer, with a heritage extending back to Adam Smith, interprets technological change as an outcome of the very growth processes it engenders. In a positive feedback system, economic growth leads to expanded markets and learning which fuel further technological change.

A notable feature of the post-war experience of productivity growth is a tendency to convergence in productivity levels among a selected sample of advanced capitalist economies. This tendency has been identified and discussed by a number of observers.² However, the strength and generality of this tendency are a matter of dispute.³ Even if it is accepted for a specific subsample of countries, it remains evident that there is great diversity in the actual pattern of experience of a wider class of countries, including the less developed countries, observed over the same period. Among this wider class, one finds the coexistence of both convergence and divergence, with no clear and unambiguous case for either tendency to prevail across the whole set of countries.⁴ This result indicates that the picture is more complex than appears at first sight and, correspondingly, calls for a deeper investigation.

In his recent assessment of growth theories, Moses Abramovitz refers back to the Wealth of Nations as the illustrious ancestor of a long stream of investigations on the determinants of economic growth (Abramovitz, 1989). In fact, in the famous proposition on the dynamics linking division of labour, productivity and market growth, Adam Smith identifies one of many positive feedback loops between innovative learning and economic development. Since then, the evidence on the microeconomics of learning and innovation has got much richer, especially in recent years (a survey in Dosi, 1988). However, standard growth theory has proved to be hardly suitable to incorporate the microevidence on, for example, dynamic increasing returns, path dependent learning, ‘disequilibrium’ search processes, interfirm and international differences in technological capabilities and so on, notwithstanding recent increasing returns equilibrium models (see Romer, 1986, 1990; Lucas, 1988; Aghion and Howitt, 1989; Grossman and Helpman, 1991). Neither has standard theory focused on the institutions and behavioural norms underlying economic coordination and development: that is, what are the institutional mechanisms that allow the ‘Invisible Hand’ to operate in a world that continuosly innovates? Again, that question can be traced back to Adam Smith, where he asks — in the Wealth of Nations and, especially, in the Theory of Moral Sentiments — what are the ‘moral inclinations’, beliefs and behaviours that make non-destructive interactions possible in market societies. However, since Adam Smith, attempts at an answer have mainly been left to disciplines other than economics. In that, the typical economic assumptions of ‘perfect rationality’ and equilibrium have often hindered any proper account of the sociology and ‘political economy’ of development.

Conventional economic theory usually assumes that information about new technologies and products is a kind of public good that can be costlessly and timelessly acquired. Accordingly, best-practice techniques are considered to be instantaneously adopted and implemented by existing firms. Much recent literature has departed from this view and regards the process of innovation and diffusion as more complex. Technological change, it is argued, takes place under conditions of competing technologies and imperfect (or only locally available) information, with innovation or diffusion costs for the new technology, returns to scale in adoption and learning, and high risk and uncertainty for the innovating firms. These conditions admit a wide range of outcomes in which the best and most efficient techniques do not always corner the market. These new approaches give a fresh and more realistic outlook to the study of innovation and diffusion processes. This chapter will identify three strands of this literature and will then utilize them to build a dynamic model of innovation and diffusion.

To produce goods and services requires knowledge. To acquire knowledge requires learning. Given a society’s endowment of physical resources, the more it learns, the greater its productive potential. Learning can enable producers to use existing technologies more effectively. Learning can also enable producers to discover ways to substitute for those resources with which they are meagrely endowed. Learning, therefore, can both improve a society’s productivity on the basis of available technologies and its ability to generate new technologies.

Great Britain, and then the USA and Germany, became world industrial leaders by generating pioneering technology. They either invented new products and processes or were the first to commercialize them on a large scale. By contrast, late-industrializing countries in the twentieth century have evolved as ‘learners’, by borrowing and improving technology that had already been commercialized by experienced firms from more advanced economies.

However disparate in content and method, the contributions in this book have all interpreted technological change as a process of institutionally structured learning. Supportive institutions — firms, occupations, patenting systems, governments and universities — foster the acquisition of knowledge in ways that lead to changing techniques. Learning is thus the medium of ongoing productivity increase. Not all institutions support such learning; indeed, some obstruct it. The presence, type and outcome of technological change therefore vary with the structure and activities of institutions. It is thus fitting to conclude by considering the ways institutions, learning and technological change are connected.