Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Even in the absence of the supply-side constraints described in Section 3.2.1-that is, even if the new technologies were freely accessible to the developing countries-demand-side factors alone are capable of preventing any generally rapid and widespread diffusion of these technologies in the Third World. The experience of the 'early follower' countries after the Industrial Revolution is well worth citing in this regard: at that time, from many points of view, the forces of demand were much more conducive to rapid adoption of the then-new technologies (which were apparently also easily accessible). These relatively favourable circumstances on the side of demand in the early follower countries (of continental Europe) have been described by Landes.
Their supply of capital and standard of living were substantially higher than in the 'backward' lands of today. And with this went a level of technical skill that, if not immediately adequate to the task of sustaining an industrial revolution was right at the margin. Culturally, of course, the outlook was even brighter. The continental countries were part of the same larger civilisation as Britain; and they were certainly her equals, in some respects her superiors, in science and education for the elite. In short if they were in their day 'underdeveloped', the word must be understood quite differently from the way it is today.... Nevertheless, their Industrial Revolution was substantially slower than the British. Although they were able to study the new machines and engines from the start and indeed acquire them in spite of prohibitions on their export, they were generations in absorbing them.... In view of the enormous economic superiority of these innovations, one would expect the rest to have followed automatically.
(Lances, 1969, pp. 125-126, emphasis added)
As we turn in this section to review the major determinants of demand for the current microelectronics technologies by the contemporary developing countries, it is worth noting that, in at least one major respect, the composition of this demand will differ from that prevailing during the historical experience just cited. This distinctive feature of today's situation arises from the very considerably enlarged role of the state in contemporary latecomers vis-à-vis the 'early followers' of the Industrial Revolution (ibid.).18 In addition to these (now more numerous) agencies of the state, however, we shall also need to consider the demands for new technologies that arise from the other major groups of agents in the developing economies, namely, privately owned firms and individual households. The necessity for a classification of this kind arises not merely from the fact that each group generally pursues a different set of objectives within a diverse set of constraints,19 but also because the demands of each may often relate to a guise different set of microelectronics-based innovations. Whereas, for example, the public sector has to take decisions about the 'public goods' aspects of the new technologies (such as telecommunications), individual household demands relate mainly to consumer electronics products.20
3.2.2.1 Agents on the demand side: the public sector
The state plays an indirect role in determining the nature and extent of demand for the new technologies through the overall development strategy that it chooses to pursue. For example, an outward-looking strategy generally imposes greater pressures to acquire the latest innovations than do inward-looking strategies, in which older vintages tend to be more prevalent. Governments also vary in the strategic role that they assign to the electronics complex in the modernization process as a whole (see Jacobsson and Sigurdson, 1983). Nochteff (1985, p. 42), for example, has noted that
The strategic character which the Brazilian State assigned to national development in the field of data processing ... contrasts with a much less precise and aggressive attitude in the other two countries [Argentina and Mexico] where, historically speaking, the position in this sector had not been regarded as one of the nerve centres in the pattern of capital accumulation in modern societies.
More specifically, government policy towards importation of microelectronics technologies influences the range of and price at which these technologies are available domestically. For example, 'The nature of the protective policies for the machine tool industries in the NICs restricts access to, and therefore the scope for the introduction of, NCMTs' (Edquist and Jacobsson, 1988, p. 149). Government policy with respect to infrastructure, education, and skill formation also impinges in an important way on demand, since these variables often seem (as noted below) to be relevant to decisions to adopt the new techniques.
Governments influence demand in a direct way, however, through the agency of the state-owned enterprise, an institution that is currently of considerable importance in many parts of the Third World (see James, 1989). A series of studies has thrown considerable light on the factors that govern the technological choices of these firms. What emerges from this research, essentially, is a view of the decision-making process that is substantially at odds with the traditional theory of the firm (see James, 1989; Deolalikar and Sundaram, 1989; Fleury, 1989). For instance, managers appeared to use highly simplified decision rules, were often poorly informed, and were subjected to a variety of subtle motivations that led neither to profit maximization nor to fulfilment of the goals of the state.
While it is evident that state-owned enterprises also play a prominent role in the demand for microelectronics technologies,21 it is not yet clear what underlies these demands. The reason for the uncertainty is that much of the literature focuses on the applications of microelectronics technologies by various government agencies, rather than on the nature of the institutional decisions that underlie these applications.22 Several observations, however, suggest the need to examine the role of foreign (i.e., developed country) influences on these decisions. 'At present', for example 'much of the importation of microcomputer systems is done with donor agency funding' (Munasinghe et al., 1985, p. 37). Lahera and Nochteff (1983, p. 172) contend that 'computers and related systems were introduced in Latin America as a result of impulses which may be considered to be exogenous. The suppliers of computers ... pressured customers into buying products for which there was no local demand'.23 Other evidence points to the need to examine deficiencies in the process by which decisions for the purchase of microelectronics technologies are made by state-owned firms in developing countries. For example, during the 1960s and early 1970s, 'a principal cause of India's inability to obtain more efficient computer systems was the information-gathering and -analysis problems of the government institutions responsible for the acquisition of computers by major Indian users' (Grieco, 1984, p. 108).
3.2.2.2 Private-producing units
Profitability is obviously a very important goal of private units and considerable literature attests to the fact that the speed of adoption of innovations is positively related to their impact on profits (see, e.g., Davies, 1979; Rogers, 1983; Stoneman, 1983). To examine patterns of adoption of microelectronics innovations by private units in developing countries, it is accordingly essential to identify the factors that determine the profitability of these innovations and it is to this task that we now turn.
Factor costs Critical to the question of profitability is whether new technologies entirely dominate the old (i.e., when they are more profitable at all relevant factor price ratios) or whether they are profitable only at the factor prices prevailing in developed countries (Bhalla and J. James, 1986). What evidence that exists on this important question24 suggests (in the case of electronic capital goods) that dominance is not a pervasive aspect of new technologies, although there are cases in which this phenomenon does seem to occur. One important source of evidence was derived as part of a complicated simulation exercise conducted by Mody and Wheeler (1990) to determine changes in international competitive advantage in three different industries (semiconductors, automobiles, and textiles/ garments). For one of these industries, textiles and garments, the study reveals that the efficiency of new technologies depends very directly on relative wage levels across countries. Thus, 'The new microelectronics technology does offer savings in labor and time related costs, but these currently seem sufficient for dominance only in garment pre-assembly activities. In garment assembly and post-assembly, by contrast, sites in countries like China seem to retain a substantial cost advantage' (ibid., p. 64).
This finding is interesting because it mirrors in a broad way the outcome of previous generations of technological progress in this industry, which did not lead to the widespread disappearance of older techniques in developing countries.25 The findings are also illuminating from the standpoint of the contrast that they present with the semiconductor industry, in which the highly automated technology did appear to be dominant at all relevant factor price ratios.
Other detailed research on the effects of new technologies on factor costs has been conducted in relation to computer numerically controlled (CNC) machine tools (CNCMTs) (Chudnovsky, 1984; Edquist et al., 1985; Jacobsson, 1986) and in relation to four flexible automation techniques (Edquist and Jacobsson, 1988). Of this work, Jacobsson's (1986) and Edquist and Jacobsson's (1988) contains the most detailed quantification of factor-saving bias and it is especially important in drawing attention to the extent of skilled labour that appears to be saved through the use of both CNC machine tools and computer-aided design (CAD).
Product characteristics Product characteristics may have an important bearing on the returns that can be expected from the adoption of a new process. The reason is that process innovations usually cause changes in the product. And as Davies (1979) points out, this influence can cut both ways, by either raising or reducing profitability.
In the development context, the literature has focused predominantly on increased profits. For a variety of countries and innovations [e.g., CAD, computer-aided manufacturing (CAM), and CNC machine tools], enhancement of product quality is a well-documented phenomenon (see Chudnovsky, 1984; Tauile, 1984; Edquist et al., 1985; Kaplinsky, 1982a, 1985; Hoffman and Rush, 1988; O'Connor, 1989). What also seems clear is that this phenomenon is most important to firms competing in markets in developed countries (particularly those in which the basis of competition is quality rather than price) and in developing countries where competition is on much the same basis (for example, in market segments dominated by the affluent minorities in these countries) (see James, 1987a).
Boon (1986a,b) has drawn detailed attention to the manner in which product characteristics of new technologies may constrain the adoption in developing countries of the process innovations with which they are associated. In a study of the adoption of CNC machine tools by firms in Mexico, he shows that these machines are suited to a particular degree of piece complexity and heterogeneity (and one that is characteristic of production in the developed countries). Plants that produce to a different pattern (for example, with more simple and homogeneous pieces) were found to be less attracted to the new machines. Boon's argument, however, has been questioned by Edquist and Jacobsson, who contend that 'there is really no empirical evidence showing that the NICs have a radically different output mix than the developed countries' (Edquist and Jacobsson, 1988, p. 147). They conclude, correctly, that the issue can be settled only on the basis of far more empirical evidence than is currently available.
The changing economics of time Closely related to the discussion of products is the question of the changing economics of time: that is, the more rapid production response (or shortened production cycle) that is possible with new technologies. This consideration allows, on the one hand, a reduction in working capital costs, and, on the other, it affords the competitive advantage of a more rapid response to consumer demands.26 The latter is, of course, particularly relevant to industries (such as textiles and automobiles) in which demand tends to change relatively rapidly. Tauile's study of the Brazilian automobile industry, for example, shows that 'The need to assure uniformity of quality and respond quickly to modifications to orders (both regarding quantity and specifications) is definitely a powerful factor inducing auto parts makers to adopt more flexible and efficient production methods, i.e. ME [micro-electronic] equipment' (Tauile, 1984, p. 20). Moreover, insofar as rapid taste change is more typically a phenomenon of high-income societies (in these and other industries), this advantage of new technologies may again be most relevant to firms in developing countries that export to markets in the developed world.
The connection between product characteristics and flexibility on the one hand, and (developed country) export markets on the other, emerges strikingly from a set of recent case studies conducted for the International Labour Organization (ILO) (Pyo, 1986; Schmitz and Carvalho, 1987; Dominguez-Villalobos, 1988; Onn, 1989). These studies cover a range of industries, such as electronics, electrical, automobiles, general machinery, and ship-building, and they sought to determine, among other things, the reasons for adoption of flexible automation techniques. The results indicate that product quality and flexibility of production were uniformly among the most important factors in adoption decisions. These considerations, in turn, appear from the ILO studies to be closely related to exports, not only through quality requirements for competition in final goods, but also through the more stringent standards for inputs that such goods often induce.
Findings such as these are not unfamiliar in the literature on choice of technology. On the contrary, the need to produce to international standards is one of the reasons most commonly cited for adopting capital-intensive techniques in developing countries, especially in the case of relatively large firms (Stewart, 1987). Some of the issues that have been raised in this context may therefore also be relevant to new microelectronics technologies. From a policy point of view, for example, 'the extent and benefits from exporting must be clearly assessed and weighed against the cost of adopting inappropriate technology' (ibid., p. 291). Policy-makers also ought to consider whether exports could be directed towards markets that are less demanding in terms of product quality and which may thus lessen the degree of automation that is required. Trade between developing countries, perhaps on a regional basis, is one way in which this could occur.
Complementary infrastructure and skills The profitability of many innovations in microelectronics depends heavily on the availability of skilled labour and complementary infrastructure. Hoffman (1984) has described the design, maintenance, and managerial skill requirements that are often both essential, and lacking, in the Third World.27 Much the same can usually be said of infrastructure requirements;28 in the case of microcomputers, for example, maintenance support, existing telephone systems, and erratic electrical power supply all act as severe constraints on expanded adoption (Munasinghe et al., 1985; Munasinghe, 1989).
Other factors The factors considered above are those that appear to be generally relevant to the profitability of new techniques. A wide range of other variables also arise in specific contexts. Edquist and Jacobsson's (1988) study of flexible automation techniques, for example, contains an extensive discussion of a large number of factors that influence the adoption of these techniques in NICs. Only some of these factors fall into the four previous categories.
Taken together, the various factors that influence the speed of adoption through their effects on profitability tend to reinforce, on the demand side (and for some of the same reasons), the conclusion that was reached on the supply side: namely, that the bigger, more developed countries of the Third World are mostly better placed to take advantage of microelectronics-based innovations. For, except where they are dominant, these innovations tend to be more efficient at relatively high wages, when the product mix with which they are associated approximates patterns observed in developed countries (in the case of CNC machine tools), and when they are combined with a well-developed infrastructure (including technical and other skills).29 It is no doubt at least partly for these reasons that the diffusion of many of the flexible automation techniques is confined largely to the more advanced developing countries (Edquist and Jacobsson, 1988; Onn, 1989). The diffusion of computers appears to follow a similar pattern.30
Information imperfections Profitability is not the only factor that influences adoption, as Edquist and Jacobsson emphasized in their study of flexible automation techniques in the engineering industries in the OECD and NICs (Edquist and Jacobsson, 1988). Indeed, the finding that these techniques are generally being diffused less rapidly in the latter than in the former group can, according to this study, be explained only to a limited extent by differences in potential profitability (especially with regard to CNC machine tools and CAD, which, because of their skill-saving features, are often attractive from this point of view). Edquist and Jacobsson point instead to information imperfections-about and how to use the new techniques-as being 'especially important' for the developing countries and 'above all' for firms that are not subsidiaries of MNEs (ibid., p. 187). It follows from this conclusion that there is considerable scope for policy to provide information, especially to national firms, in these (and also, by extension, in other) developing countries.
3.2.2.3 Household demand
Corresponding to the literature on new processes are numerous studies that deal with the diffusion of product innovations (mostly consumer durable goods). These studies (see Davies, 1979; and Stoneman, 1983, for reviews) suggest that although the economic agents involved are different in the two cases (households as opposed to firms), a number of important parallels can nevertheless be drawn. First, it is clear that diffusion of new products, as with new processes, often requires a considerable period of time. Ironmonger's (1972) study of the diffusion of more than sixty new commodities in the United Kingdom, for example, found that although diffusion sometimes takes only about twenty years, in other cases it may take much longer. As a determinant of these disparate durations, moreover, factors akin to the role of profitability in the diffusion of new processes seem to be important. In particular, the cheaper is the new product and the greater the extent to which it constitutes an improvement over existing versions, the more rapid is its diffusion likely to be (Davies, 1979; see also Rogers, 1983).
In the specific context of developing countries, a small but quite separate literature has sought to explain the patterns of demand for new (and imported) products in terms of factors such as advertising and international demonstration effects (see James and Stewart, 1981; James, 1987b). This literature emphasizes the role that MNEs and other foreign influences play in shaping patterns of demand in the Third World. The relevance of this literature (as well as of that described in the previous paragraph) to the comparative speed with which consumer electronics products have been diffused within (and between) developing countries, has not, however, yet been explored. Indeed, the whole question of household demand has suffered from considerable neglect compared to the demands of producing units. Such neglect is especially serious in the light of the fact that household demand tends to be fairly substantial, even in the very poorest developing countries which do not undertake any productive activities in the electronics field.31
3.3 Impacts of microelectronics
3.3.1 Sectoral versus economy-wide impacts on
output and employment
3.3.2 Impact of adoption on non-adopters of
new technologies
3.3.3 Acquisition of technological
capabilities
3.3.4 Applications of technological
capabilities
The impacts of new technologies in general, and thus of microelectronics in particular, can be thought of in terms of two main categories: effects on output and income, and effects on income distribution. In analysing these effects, we shall pursue the integrative theme of this chapter by relating them, on the one hand, to the patterns of diffusion discussed above, and on the other hand to a broad range of relevant literature.
3.3.1 Sectoral versus economy-wide impacts on output and employment
Sectoral (or partial) analysis attempts to explain the behaviour of producers and consumers in determining price and output in a given market. For this purpose, it is necessary to assume that there are no major changes in any of the other sectors of the economy. In this type of framework, the forces that determine the output and employment response to technical change may be briefly described as follows (James 1985, 1988). Technical change, by definition, increases output per unit of input. If output remains constant, then the derived demand for at least one factor input (e.g., labour) must necessarily fall. But since increased productivity lowers costs, more output will be produced at a given price by the profit-maximizing firm. This incremental supply will mean that output is higher and price lower than prior to the technical change. The extent of the increase in output depends on the degree of cost reduction and on the elasticity of demand. (The greater these are, the stronger will be the tendency towards expanded output.)
To determine the new factor demands, whether these have increased or decreased as a result of technical change, the increase in output has to be compared to the increase in productivity. In some cases-if, for example, the increase in output is large and the technological change is not strongly biased towards saving a particular factor-demand for that factor will increase. (In other circumstances, such as when output growth is constrained and technological change is strongly biased, demand for that factor will tend to fall.) Assuming that demand for the factor does increase, this will be reflected (to an extent that depends on the elasticity of the supply curve of the factor) in either increased employment (in the case of labour) or a rise in wages (or some combination of these).
This discussion is based on the assumption that technical change does not alter the market structure and, in particular, that competitive conditions continue to prevail after the change. But insofar as temporary market power is created for the innovating firm, this power will tend to cause a reduction in output compared to the competitive situation. This, in turn, will lead to a reduction in the demand for inputs in general, and labour in particular. Alternatively, if several firms adopt the innovation at a similar time, some form of oligopoly will be created and in this case the effect on the volume of output (and hence the demand for inputs) is less clear.
It is apparent that applying partial analysis to understanding the employment effect of microelectronics innovations in a particular sector requires knowledge of not only the factor-saving bias of the particular innovation and its effect on market structure, but also the elasticity of final demand and the elasticity of (the various types of) labour supply. Although, as we noted earlier, progress has been made in estimating the first of these requirements (mostly in relation to CNC machine tools), no sectoral studies have combined these estimates with those for the other requirements. Some ILO case studies, however, do attempt to relate the labour-saving bias of flexible automation techniques (such as CNCMTs and CAD) to an increase in export demand that adoption of these techniques often seemed to induce (in large part because of changes in the characteristics of the products associated with the new techniques). In fact, several studies (Dominguez-Villalobos, 1988; Onn, 1989) suggest that for some firms increased export demand may have more than offset the labour-saving effect, thus leading to a net increase in labour demand. It is clear, however, that much more empirical research is needed to determine the effects on output and employment at the partial equilibrium level.32
A similar need arises in relation to impacts on income distribution at this level. A conceptual essay on these impacts by James (1985) concludes that they arc likely to be inegalitarian, but the data required to support this view are almost entirely lacking. Nevertheless, what needs to be emphasized is that this type of research would be relevant only to certain instances in which microelectronics technologies have been diffused in the Third World. In particular, partial methods make methodological sense in those developing countries where diffusion of technologies is confined to a very limited part (sector or region) of the economy (such as, for example, an electronics assembly enclave), since under those circumstances, the various effects of the given sector or region on others in the economy (and the effects that these, in turn, have on the original location) can quite reasonably be ignored. That is, under those circumstances, one is entitled to neglect the economy-wide effects of the diffusion of microelectronics-based innovations. But in countries (such as India, China, the NICs, etc.) where this is not the case-where, instead, these innovations are widely diffused across sectors-it becomes imperative to incorporate these effects into the analysis of impacts.
Because this task has not yet been undertaken in a modelling exercise, we are able only to offer insights from some existing economy-wide models of developing countries as to how these economy-wide effects might operate and how they may yield results that differ from those obtained using partial analysis. Binswanger's (1980) model, for example, suggests and quantifies several important mechanisms through which the economy-wide effects of technical change are transmitted. In particular, that model demonstrates how technical change that occurs in one sector affects the demand for output (and thus inputs) of other sectors, via so-called 'price' and 'income' effects. Technical change in a given sector, that is to say, increases per capita income (the 'income' effect) and tends to decrease the price of the good (the 'price' effect). If the former outweighs the latter, output in sectors in which technical change does not occur will increase, despite the relative rise in prices of the goods produced in that sector. In Binswanger's two-sector model, which simulates the effects of labour-saving technical change in the nonagricultural sector (representing, broadly, some of the likely technological and sectoral features of microelectronics innovations), the income effect is shown to outweigh the price effect so that agricultural output actually falls.33
Another important set of economy-wide mechanisms is incorporated in the well-known dual-economy model of Kelley et al. (1972). These authors criticize the framework within which the effects of technical change in developing countries were considered in the early dual-economy literature. In particular, they criticize the attempt by Fei and Ranis (who formulated one of the first such models) to examine the conflict between the output-enhancing effects of technical change and the employment effect 'while holding both the rate of capital formation and the rate of population growth constant'34 (Kelley et al., 1972, p. 205, emphasis in the original). Kelley et al. (ibid., pp. 205206) contend instead that 'any meaningful analysis of the conflict must take account of the impact on capital accumulation of population growth and the manner in which these variables interact with the improvement of technology'. They seek to demonstrate quantitatively, in the context of their simulation model, that the dynamic effects of technical change on these variables may be considerable, leading to conclusions very different from those based on the early dual-economy literature. Specifically, 'the advantage of rapid rates of technical change in developing economies now becomes much clearer. Rapid rates of technical change tend to raise achievable rates of capital accumulation and to lower rates of population growth by stimulating urban-industrial development' (ibid., p. 210).
The correct inference to be drawn from this conclusion is not that these economy-wide effects will necessarily accompany microelectronics-based innovations that are widely diffused in a particular developing country. It is possible, for example, to think of applications of these innovations (e.g., in the health sector) that will increase population growth rates,35 and it is equally reasonable to debate the favourable effects of technical change on investment in various Third World contexts.36 The important point is that these critical issues ought to be the subject of a great deal more empirical attention than is currently the case.
The final economy-wide mechanism we shall consider is backward linkage, or input-output effects, in production.37 Evidence from the United States based on a simulation exercise carried out by Leontief and Duchin (1985) points to powerful backward linkage effects associated with microelectronics, and highlights the additional employment created indirectly in sectors that produce new capital goods, especially computers. According to several observers, however (see Hoffman, 1985b; Schmitz, 1985), comparable effects cannot be expected in the Third World since most developing countries import rather than produce innovations in microelectronics. Hoffman (1985b, p.269), for example, argues that 'most Third World nations will be net importers of the new equipment and lack the strong software and capital goods sector that are the source of many new jobs in the developed countries'. While this distinction is important, it must be qualified in several ways.
First, as Schmitz (1985) and Edquist and Jacobsson (1988) have pointed out (and as was also apparent from the earlier discussion of the diffusion of the new technologies), there are important segments of the Third World in which production is important (though the extent of backward linkages from this will vary a good deal according to the specific form that it takes: assembly production, for example, is almost entirely devoid of linkage effects; see Lall, 1983). Second, the employment impact associated with an expansion in output depends not only on how integrated is the productive structure in the developing country - which determines the strength of linkages between sectors-but also on the labour: output ratios (or labour requirements per unit of output) in each of these sectors. The significance of this distinction has been stated lucidly by Stern and Lewis (1980, p. 39):
In all industrial sectors the direct labour coefficients, or labour-output ratios, decline as an economy develops. This decline ... will tend to reduce the employment impact of an expansion in output. At the same time, however, the increasing interindustry linkages that accompany economic growth serve to increase the employment impact of output expansion.
Determining the relative size of these conflicting effects in the specific context of innovations in microelectronics is yet another economy-wide issue that warrants empirical investigation.38