Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
1. Relevance of innovation studies to developing countries
1.1
Introduction
1.2 Innovation and technological change
1.3 Implications for developing countries
1.4 Concluding remarks
Acknowledgements
Notes
References
Charles Cooper
Studies of industrial innovations in industrialized countries, and similar studies in developing countries, tend to be done in isolation from one another. The isolation is not total, of course: there is a small number of authors for whom the relevance of studies of innovation in industrialized countries to the situation in developing countries is virtually taken for granted - see for example Katz (1987) or Pack and Westphal (1986). Nevertheless, the isolation is sufficiently noticeable in academic writing, to raise the question whether researchers concerned with technology policies in developing countries might not benefit from a more systematic exploration of what have been called 'innovation studies' (Dosi, 1988). The purpose of this chapter is to map out the territory of 'innovation studies' that could be useful in research on developing countries. This chapter attempts to relate studies on innovation and technological change in developed countries and those in developing countries. It also attempts to connect technology studies to the broader stream of development economics.
By their nature, maps are imperfect reflections of the scale and detail of underlying reality. That, as Joan Robinson once pointed out, is their whole point.¹ It is, nevertheless, quite possible to criticise some maps for being on a scale too small to be as useful as they might be or to be arbitrary in the detail they select. Perhaps the map drawn in this article is vulnerable to those criticisms, for it is the outcome of a preliminary reconnaissance, and not a comprehensive survey. If that is so, we must hope that it at least reveals those parts of the terrain worthy of further exploration.
There have of course been good surveys of the issues of technology policy as they relate to developing countries. A particularly valuable, and quite recent, one by Martin Fransman (1986) covers part of the area of this chapter, but not all.
This chapter is set out as follows. The next section is a summary and discussion of recent literature on innovation in industrialized countries, which focuses on attempts to synthesize 'innovation studies' into a coherent theory. The theoretical structures that emerge from this effort go well beyond the explanation of industrial innovation phenomena per se: they raise rather fundamental questions about traditional theories of the firm and its behaviour. The emphasis is on innovation as a mode of competition, and as a source of sustained disequilibrium in the Schumpeterian style, an approach which is in contrast to the equilibrium stories which are at the foundation of received theory.
The subsequent section then deals with the possible relevance of this body of theory for developing countries. It addresses two questions. The first is how relevant are innovation studies to processes of technological learning in developing countries.2 The second is how important is technological change in the trading relations of developing countries. The final section draws conclusions.
1.2 Innovation and technological change
The microeconomic processes involved in the adoption of innovations, which we are inclined to describe today as Schumpeterian, were, it seems, clearly recognized by Classical economists. In his chapter 'On Machinery' in the Principles of Political Economy, Ricardo (1830; edition 1971) remarks:
He ... who made the discovery of the machine, or who first usefully applied it, would enjoy an additional advantage, by making great profits for a time.
(Chapter XXXI, pp. 378-379)
Marx expanded considerably on this notion in Book One of Capital (Marx, 1858; edition 1961, p. 312) in the theoretical discussion of the origins of 'relative surplus value'. If an individual capitalist ('some one capitalist', ibid., p. 316), doubles the productivity of labour, whilst the value of the means of production remains the same, then
The individual value (of the articles produced) ... is ... below their social value: in other words, they have cost less labour time than the great bulk of the same article produced under the average social conditions.
(ibid., p. 317)
However,
The real value of a commodity is ... not its individual value, but its social value; that is to say the real value is not measured by the labour time that the article in each individual case cost the producer, but by the labour time socially required for its production.... If therefore the capitalist who applies the new method sells his commodity at its social value ... he sells it ... above its individual value (i.e., cost) ... and thus realises an extra surplus value.
In both of these accounts, adoption of the innovation which leads to the generation of extra profit or surplus value is implicitly assumed to happen against the background of some initial equilibrium position. The subsequent story (that is, after a period of extra surplus value) essentially concerns a return to this equilibrium situation. Thus in his discussion of the adoption of a new type of machine, Ricardo concludes:
But, in proportion as the machine ... comes ... into general use, the price of the commodity produced, would, from the effects of competition, sink to its costs of production, when the capitalist would get the same money profits as before, and he would only participate in the general advantage as a consumer.
(Ricardo, op. cit. p., 379)
Similarly, Marx observes:
On the other hand, ... this extra surplus value vanishes, so soon as the new method of production has become general, and has consequently caused the difference between the individual value of the cheapened commodity and its social value to vanish.... The law of the determination of value by labour-time, a law which brings under its sway the individual capitalist who applies the new method of production ... this same law acting as a coercive law of competition, forces his competitors to adopt the new method. The general rate of surplus value, is therefore ultimately affected by the whole process, only when the increase in the productiveness of labour has seized upon those branches of production, that are connected with ... the necessary means of subsistence.
In short, technological change results in a general increase in surplus value, which for the present we may identify with profits, only when it increases productivity in the production of wage goods (or in the production of the means of production). The basis for this conclusion is that innovations in general ultimately leave the 'rate of surplus value' unaffected due to the reassertion of equilibrium; but innovations in the wage goods sector reduce the costs of labour time in all other sectors.
In its early form, Schumpeter's own analysis of innovation as a microeconomic process, especially in its early form, owes much to Marx (Schumpeter, 1912; edition 1961, Chapter IV on 'Entrepreneurial Profit'). In particular, in his early writing, which was much concerned to explain (even to justify) the additional profit generated for the innovating firm as a return to entrepreneurship, he placed considerable emphasis on the tendency of the industry to return to equilibrium. 'The second act of the drama' of innovation (op. cit., p. 131) comes when imitators enter production, thus driving prices down and leading to a 'complete reorganisation of the industry'. Consequently, after an innovation, 'that process of reorganisation occurs which must result in the annihilation of the surplus over costs' (p. 133). The idea that reorganization takes place in the reestablishment of an equilibrium might suggest that Schumpeter had in mind a more considerable process of adjustment than that described by Ricardo or Marx; but the return to equilibrium is still the keynote.
Subsequently, however, Schumpeter's thinking moved towards the notion of continual change as a result of a succession of innovations, leading to 'continual reorganisation of the economic system' (op. cit., p. 156), in which the reestablishment of equilibrium is preempted by further rounds of innovation. In his 1934 Preface to the English Edition of Theorie der wirtschaftlichen Entwicklung, he remarks:
The conclusion suggested itself that this body of theory might usefully be contrasted with the theory of equilibrium which explicitly or implicitly always has been and still is the centre of traditional theory.
(op. cit. p. xi)
Later in Capitalism, Socialism and Democracy, he wrote
the capitalist economy ... is incessantly being revolutionised from within ... existing structures and all the conditions of doing business are always in the process of change.... Every situation is being upset before it has time to work itself out.
(Schumpeter, 1966; our italics)
It seems fair, then, to distinguish two distinct but closely related respects in which Schumpeter takes issue with the conventional Marshallian microeconomic theory of the firm. First there is a fundamentally different view of the nature of competition. In Marshall, the technology of production is given and available to all firms; the technology defines the parameters within which firms' minimum cost levels of production are to be determined by adjustment to competition. For Schumpeter, competitive behaviour, led by innovators, is primarily concerned with changing the parameters themselves: that is, with a search for new technologies, which, temporarily, are available to the innovative firm alone and confer the advantages of monopolistic rents. It is also concerned with the imitative process which then ensues. The search for the optimum level of output and minimum cost of production is largely overshadowed by the process of competition between technologies.
The second major difference between Schumpeterian and traditional views of the firm resides in Schumpeter's view that competition based on the search for new technologies generates a stream of innovations which preempt the attainment of microeconomic equilibrium altogether.
Christopher Freeman puts these points in the following terms:
In Schumpeter's framework it is disequilibrium, dynamic competition ... between entrepreneurs, primarily in terms of industrial innovation, which forms the basis of economic development.
(Freeman, 1989, pp. 209-210)
Intuitively, the Schumpeterian model of competition gives a plausible interpretation of competition in a range of important sectors-especially, of course, science-based sectors such as electronic capital goods, chemicals, pharmaceuticals, biotechnology, and the like. This explanation may seem less convincing in what development economists often call the 'traditional' sectors. Nevertheless, as I will argue later, even here innovative competition occurs, mainly in the form of adoption of innovative plant and equipment originating in the capital goods sector.3 Later we shall discuss intersectoral differences in the sources of innovation and relate these to the apparently differing incidence of innovative competition between industrial sectors.
As well as intersectoral differences, there have been historical differences (i.e., changes within sectors over time) in the incidence of innovative competition. It is interesting to speculate whether Schumpeterian modes of competition are perhaps more characteristic of the modern (i.e., twentieth-century) industrial economy than they were, say, during and after the Industrial Revolution. The emergence of Schumpeterian thinking may reflect-with a considerable time-lag, of course-historic changes in the nature of competition itself.4 Perhaps Marshall has been overtaken by events rather than shown to have got it wrong.
Recent theoretical approaches to innovation have been based importantly on empirical observation of firms' behaviour5 and have been informed by the Schumpeterian concept of how competition takes place in the industrial sector. In particular, they draw on Schumpeter's notion that, at the level of the firm, competition is about creating a stream of disequilibrium situations, in which there are quasimonopolistic rents.6 The idea that firms continually search out innovations in this way has been shown in a seminal study by Nelson and Winter (1982) to generate a plausible explanation of economic growth processes.
There has, of course, been considerable development of these basic notions. At the risk of doing an injustice to the conceptual richness of the discussion of innovation, we will select three especially important and related developments for more in-depth discussion:7 (1) the idea that technological change is localized; (2) the notion that innovation at the level of the firm is the outcome of a cumulative process; and (3) the different incidence of factors determining the appropriability of new technologies. After discussing these three elements, which primarily concern conditions within innovative firms, we will go on to look at two characteristics of the environment within which the firms operate. These are the technological and institutional contexts.
The idea that technological change may be localized was put forward in a theoretical article by Atkinson and Stiglitz (1969), who contended that a localized 'bulge' in the neoclassical industrial production function may represent technological change better than simply a uniform shift of the whole frontier. The location of the bulge depends essentially on the point at which firms were producing initially-in short, upon their prior technological choices.8 At the same time as Atkinson and Stiglitz, Nathan Rosenberg (1969) put forward an economic historian's empirically founded notion of localization. These ideas were subsequently used by David (1975), who proposed an explanation of localization based on 'learning' processes in production:
Because technological 'learning' depends on the accumulation of actual production experience, short-sighted choices about what to produce and especially about how to produce it using presently known methods, also in effect govern what subsequently comes to be learned.9
(David, 1975, p. 4)
It is helpful, at this stage, to keep in mind that learning cum localization phenomena take place at the level of the firm. It is quite possible, therefore, that individual firms within an industrial sector have different 'vectors' of technological change; that is to say, firms have different patterns of localization within the particular technological fields relevant to the industry's production activities. There are many cases in which different patterns of localization coexist. An example is the simultaneous emergence of the Apple Macintosh computer system and the IBM PC system; another is the coexistence of several different methods of 'catalytic cracking' in various major chemical firms during the 1970s. Sometimes one or more of the competing variants on the basic technology will prove to be a dead-end, but this does not always happen. Implications of these differences between firms in their vectors of technological change, especially implications for market structures, are discussed later.
Second, there is the question of cumulativeness. David's account of the origins of localized technological change leads naturally to the concept of cumulation in the innovation process. Innovation processes are cumulative in the sense that David sketches out: technologies of production used today influence learning processes and the nature of accumulated experience. These, in turn, influence the uses to which innovative inputs [like research and development (R&D) at the level of the firm] are put, and so also the nature of tomorrow's production technologies. And so on.
The simplest examples of cumulative innovative processes at the level of the firm are the processes of 'learning by doing' (Arrow, 1962), and the other more empirically founded and realistic variants of those processes described in the literature.10 These processes have the characteristic that the productivity changes they generate depend upon accumulated experience of actual production. They are aptly described as irreversible, dynamic economies of scale.
The cumulative processes in question in theories of innovation may include learning by doing in its traditional form, but they also refer to other learning processes which may not be so simply related to the experience of production per se: technological learning,11 for example, or learning about effective resource allocation for innovation. By straightforward extension of the discussion of 'localization', the process of cumulation of 'problem solving capabilities' (Dosi, 1988), is likely to take firm-specific forms. The theory of innovation as it has developed in the recent past thus endows individual firms with histories, and these histories are of more than antiquarian interest: a firm's history determines what it is good at technologically, and that in turn has a direct influence on the rate and direction of innovation it pursues, and on the differences in performance between it and other firms in the same industry. Learning process of various kinds and the pattern of intrafirm accumulation of technological capability connect the firm's past with its present. The contrast between this and the ahistorical 'firm' of neoclassical microeconomics whose whole existence is supposedly described adequately by a set of cost curves identical to those of all other firms in the industry, is rather stark.
The third characteristic of innovation is that the knowledge incorporated in new technologies can, to varying degrees, be appropriated by the innovating enterprise. Appropriation of technological knowledge is essential to the innovative process, since it is appropriation which allows a temporary preemption of imitation and hence monopolistic rents (also temporary). It is the anticipation of these rents which induces enterprises to innovate in the first place.12 Appropriation is achieved in a number of ways. In some sectors (drugs, for example, or fine chemicals) patents are especially important; in others, secrecy is enough. The lead times which imitators face can be important, as may learning curve effects in the innovator firm. Appropriability may also depend upon the extent of tacit knowledge associated with the innovation (for a discussion of tacitness, see Dosi, 1988). As implied above, appropriability differs between sectors.
The rate and direction of technological innovation are explicable up to a point by localization and the cumulative nature of the innovative process or by opportunities for appropriation. But these intrafirm factors are, as we indicated earlier, only part of the story. A more complete explanation depends also on 'technological trajectories' (the constraints imposed by the logic of technological development itself) and on institutional factors (which are usually nation specific). The institutional side is receiving much attention at present. Its parameters are set out with particular clarity by Nelson (1988 a,b).
Intersectoral differences in innovation, and the flows of knowledge related to innovations between sectors, are important, especially in relation to developing countries (see below). That sectors differ both in the frequency and extent of innovation and in sources of innovation has long been recognized.13 Initially, the literature of the 1960s and 1970s distinguished 'traditional' and supposedly non-innovative sectors from more 'science-based', innovative ones. More recently, ways of grouping industrial sectors in relation to the nature and sources of innovative activity have consciously or otherwise returned to a pattern very similar to that implicit in Marx's historical account of relations between 'science' and production (Marx, 1858, Book 1, Chapters 13 to 15). Marx essentially differentiates the capital goods (machine-making) sectors as primary points of contact between science and production.14 In the same spirit, Robson et al. (1988) distinguished 'user sectors' and 'producer sectors', whereby the latter are chiefly science-based sectors and the more traditional capital goods sectors. Pavitt (1984) and later Dosi et al. (1990; pp. 92-98) distinguished between three types of sectors: (1) supplier-dominated sectors, which in the main receive innovations embodied in producer goods and are essentially innovation-user sectors (they include traditional sectors of manufacturing); (2) production and scale-intensive sectors (largely machinery-making sectors, consumer durables, automobiles, steel, etc.), which are predominantly innovation producers; and (3) the science-based sectors, which are producers and a source of innovation for many other sectors (these include electrical and electronic sectors and chemicals). These sectoral distinctions will be useful in later discussion on developing countries.
Before discussing the relevance of these theoretical approaches to developing countries, we shall discuss two important implications: first, for market structure; and second, for international trade and competitiveness.
Innovation is seen as based on the cumulation of firm-specific technological skills, leading to localized technological changes. It is therefore preeminently a differentiating process, in which firms attempt to establish control over markets by developing new products and new processes. This, of course, is in contradiction to the conditions of 'perfect competition' and to firm behaviours, which are conventionally presumed to follow from those conditions. It suggests market structures closer at first glance to those of the Robinsonian model of imperfect competition.15 There are, however, important differences. First, the Robinsonian model is commonly associated mainly with the explanation of trivial product differentiation, whereas the product differentiation associated with innovative activities is generally nontrivial and essentially depends for success on real technical advances. Second, imperfect competition models, Robinsonian or otherwise, are usually concerned with underlying trends to equilibrium-albeit equilibria different from those of perfect competition. Innovative competition, in Schumpeter's vision, is a process in which there is a continuing search for and attainment of new disequilibria in markets, from which rents flow.
Given this distinction, it would seem that innovation theory as it has evolved recently can be more convincingly related to concepts of market structure and competition which do not rely on tendencies to equilibrium. One obvious example is Kalecki's concept of the 'degree of monopoly' (Kalecki, 1971, Chapter 5, 'Costs and Prices', based on an earlier essay published in 1943). Alternatively, Sylos-Labini's analysis of oligopolistic market structures as a meta-stable configuration of competing firms might also provide a more convincing framework for describing and predicting innovative competition (Sylos-Labini, 1967).16 A strength of the innovation theories discussed here is that they provide an explanation of market imperfection which derives from the competitive process itself- market imperfection is, as it were, explained by an endogenous process (Nelson and Winter, 1982).
The implications of these Schumpeterian approaches to innovation for international trade have been explored by Dosi et al. (1990). The theme, which is in direct confrontation with conventional trade theory, is sketched in their introduction (Dosi et al., 1990, p. 11):
In so far as technology gaps and their changes are a fundamental force in shaping international competitiveness, their impact on domestic income, by inducing and/or allowing relatively high rates of growth via the foreign trade multiplier, will be significant.... It is the relationship between technology, trade and growth which is at the centre of the analysis, rather than the question about the short-term gains from trade stemming from the open economy allocation of resources, so crucial in the conventional view.
Later (ibid., Chapter 6), Dosi et al. present a detailed econometric study of the member countries of the Organization for Economic Cooperation and Development (OECD). The study covers forty industrial subsectors and addresses the relationship between competitiveness (measured by each country's exports in each of the forty product groups as a proportion of total OECD exports in that industry) and an index of technological innovation (each country's U.S.-registered patents in each of the industry groups as a share of total OECD U.S.-registered patents in that industry group). The regression equations suggest that the technology variable is strongly associated with competitiveness in most of the science-based sectors, but the association is not significant in the so-called traditional sectors. These results support the idea that absolute trade advantages may be built upon superiority in innovation.
From the point of view of the discussion which follows, these results need to be handled with some care. In particular they must not be taken to show that innovation and the increases in factor productivities to which it gives rise are unimportant in determining competitiveness in the traditional sectors. In these sectors innovation does not show up in patenting, precisely because they are 'supplier dominated' sectors as far as innovation goes: they receive their innovation from suppliers of producer goods who do the 11&D and hold the patents on the equipment used. Nevertheless, competitive position in these sectors depends importantly on the application of innovations developed in the supplier sectors. The traditional sectors may well be innovative despite the lack of patenting, but innovative in the sense of using new technology rather than producing it-a process which also requires a mastery of technology. This has special importance at the present time in view of the increasing application of innovations in microelectronics in such sectors.
1.3 Implications for developing countries
1.3.1 Innovation studies and the accumulation
of technological capabilities
1.3.2
Trade and technology
In principle the theoretical framework which has been developed from empirical studies of innovation and of the behaviour of innovative firms could provide useful guidelines for policy studies in developing countries from at least two points of view.
First, innovation theory contains insights into how and why technical capabilities are developed in the industrial sectors of advanced countries. In effect, they give some new dimensions of meaning to the concept of 'accumulation of local technological capabilities', which has come to play an important role in technology policy in developing countries. Second, innovation studies have much to tell us about the structure of international industrial markets. This kind of information is important in defining strategies for industrial export development.
This section in therefore presented in two parts: the first examines innovation studies and the accumulation of technological capabilities; the second focuses on trade and technological change.