Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Critics and extensions of conventional trade theory
The approach adopted by the critics is basically one of analysing the outcomes and trade implications of the behaviour of firms operating in conditions which fall short of the ideals of perfect competition (monopolistic competition, imperfect competition, increasing returns to scale). Much of the literature in this category represents sympathetic attempts to relax the basic assumptions of the H-O model and test its robustness (Kierzkowski, 1987).8 In this context, monopolistic competition and other forms of imperfect competition have come to be central to the literature on trade theory' largely reflecting the persistence of intra-industry trade in reality. At one extreme there are those who equate countries to single firms, analyse their oligopolistic interactions, and attempt to link the instruments and concepts of industrial organization with the general equilibrium model (Caves, 1980; Brander, 1981; Brander and Krugman, 1983).9 Other analyses have tried to formalize equilibrium trade patterns, with endogenous technological change and monopolistic competition as the innovative intermediate inputs (Ethier, 1979; Krugman, 1987; Grossman and Helpman, 1989, 1990a).10
The link between trade theory and industrial organization was first proposed simultaneously by Dixit and Norman (1980), Krugman (1979) and Lancaster (1980).11 It has even been suggested that it is the contribution by Dixit and Stiglitz (1977) and Lancaster (1979)12 that provided the foundations for a theoretical framework for analysing economies of scale and product differentiation in a general equilibrium setting (Greenway, 1991).13 Since then, developments in this area have taken two directions: modelling the role of economies of scale and analysing market structures. In the latter case various forms of imperfect competition are taken as a starting point and the possibilities of strategic behaviour and interactions between firms are treated in the analysis.
Grappling with the presence of economies of scale
Economies of scale have recently come to be seen as more important (e.g. Scherer, 1980).14 Various explanations for this trend have been given in the literature (e.g. Helpman and Krugman, 1985; Alcorta, 1994).15 First, where industries produce multiple products, many products may be produced at less than optimal scale. Second, there may be important economies of multi-product operation which are not captured by plant-based estimates of scale economies. Third, there may be important dynamic scale economies internal to firms.
The existence of economies of scale provides an incentive for international specialization and trade. This incentive may complement the explanatory power of differences in factor proportions, and may even give rise to trade in the absence of such differences (Helpman and Krugman, 1985). Except under special circumstances, a world of increasing returns to scale will not be a world of perfectly competitive markets. However, in the absence of a generally accepted theory of imperfect competition, the admission of economies of scale would make it difficult to generalize on trade. More specifically, the presence of economies of scale implies that the H-O model can guarantee neither gains from trade nor the existence and uniqueness of a free trade equilibrium. The problem is that the persistent presence of economies of scale is inherently inconsistent with competitive equilibrium, as marginal cost pricing would in that case imply losses. Thus the admission of economies of scale calls for an analysis based on a market structure that allows prices above marginal cost. It is in this context that more explicit consideration has been given to alternative market structures in the analysis of international trade.
Three different approaches to the analysis of increasing returns to scale under alternative market structures (other than perfect competition) can be identified in the literature: the Marshallian approach, the Chamberlinian approach and the Cournot approach (Krugman, 1987).
The Marshallian approach
In the Marshallian approach, increasing returns to scale are wholly external to the firm. In this special case the competitive model is still operative at the level of the firm.16 Economies of scale are then introduced in the general equilibrium models in ways which allow for the existence of a competitive equilibrium. However, the laissez faire competitive equilibrium is no longer Pareto optimal, to the extent that the private marginal rate of transformation deviates from the social marginal rate of transformation of any two commodities. It is in response to this rather undesirable outcome (for the advocates of free trade) that considerable literature on trade has grappled with the problem of optimal tariff policies.
It has been shown that working from the allocation of resources to production and trade rather than the other way round clarifies the role of economies of scale in determining the pattern of specialization and trade (Ethier, 1979, 1982).17 If external economies arise from economies of scale in the production of intermediate goods which are cheaply tradable, it is argued, economies of scale should apply at the international rather than national level. Economies of scale arising from increased specialization (rather than from plant size) depend (at the aggregate level) on the size of the world market rather than on geographical concentration of industry (at national level). Such international increasing returns to scale were shown to be free of the recurrent indeterminacy and multiple equilibria characteristic of national increasing returns to scale. This implies the possibility of a theory of intra-industry trade in intermediate goods in accordance with the basic H-O model. Intra-industry trade in manufactures is viewed as complementary to international factor movements as predicted by the H-O trade theory. However, the basic assumptions, that economies of scale arise solely from fixed costs and that intermediate components are symmetric, can be questioned on empirical grounds.18
There is a possibility that external economies may arise from the inability of firms to appropriate knowledge completely. In such cases information may be viewed as an externality. However, innovative industries will ordinarily not be perfectly competitive. An emphasis on the generation of knowledge calls for a dynamic rather than a static model. The question of the applicable unit of analysis arises. If external economies are assumed to result from the incomplete appropriability of knowledge, the applicable unit for the analysis of externalities will depend on the details of how innovations diffuse: are they likely to be confined to a local area, to a nation, or are they international? Recent advances in information and communications technologies are likely to tilt the relevant unit of analysis towards the international arena.
The Chamberlinian approach
In the Chamberlinian approach, the possibility of product differentiation and product variety is introduced into the analysis. The resulting interaction between demand for product variety and economies of scale leads to intra-industry trade (Helpman, 1981; Lancaster, 1980).19 Similar results have been demonstrated for differentiated intermediate goods to satisfy the demands of producers who use these diverse intermediate inputs (Ethier, 1982). Developments to the Chamberlinian approach have taken two directions. The first has assumed that each consumer has a taste for largely different varieties of product (e.g. Dixit and Stiglitz, 1977; Dixit and Norman, 1980). The second has approached product differentiation by positing a primary demand for the attributes of varieties (e.g. Lancaster, 1980). Both approaches introduce the possibility that the response to market expansion may be greater product variety. Consequently, gains from trade may occur in the form of greater choice of product varieties and in the form of lower prices. It has been shown that these basic results and their implications are retained even when demand for variety is allowed for at the level of the firm (Dixit and Norman, 1980).
The Cournot approach
The Cournot approach invokes economies of scale to explain the existence of oligopolies and treats imperfect competition as the main actor. An extension of the H-O model along these lines has introduced increasing returns to scale into the analysis and related it to the role of protection. This extension has opened up the possibility that protection of the domestic market can help the local producer to generate a higher level of output resulting in enhanced competitiveness in terms of lower average costs (Krugman, 1984).20 This approach shows the effect of trade on increasing competition and demonstrates the possibility of interpenetration of markets, because oligopolists perceive a higher elasticity of demand for exports than for domestic sales.
The real world, characterized by economies of scale, the accumulation of knowledge and the dynamics of innovations, is not incompatible with the ideals of free trade. But if the analysis of such dynamic phenomena is formalized in static models there is a risk of major potential pitfalls where static and dynamic analyses are mixed (Helpman and Krugman, 1985). If a static model has to be used as a proxy for a dynamic world, it should be viewed as a representation of the whole time path of that world and not a snapshot at a point in time. In particular, the comparison of equilibria involved in comparative statics exercises should be understood as a comparison between alternative histories and not a change that takes place over time. Using static models to think dynamically is even more risky in imperfectly competitive markets, in which games over time can have many possibilities not seen in one-period games.
Trade theory and various market structures
The analysis of trade issues in the context of a variety of market structures has explored several issues which could not have been addressed adequately in the framework of the perfectly competitive model. The most notable approaches to the analysis of various market structures include: trade policy and the power of domestic firms, the role of price discrimination and dumping, and the role of governments in giving domestic firms a competitive advantage. Another group of analysts raise questions of the implications of the link between market structures and trade theories for new arguments for protectionism. Further extensions along these lines have attempted to capture more complex insights such as the role of intermediate goods (Ethier, 1982), non-traded goods (Helpman and Krugman, 1985), market size effects (Krugman, 1980; Helpman and Krugman, 1985)21 and attempts to demonstrate that economies of scope and/or vertical integration lead to the emergence of multi-activity firms such as multinational corporations (Helpman and Krugman, 1985).
The literature within this strand has basically examined alternative theories of market structures which deviate from perfect competition (Helpman and Krugman, 1985). Such analyses assumed various kinds of imperfectly competitive market structures such as contestable markets22 (Baumol et al., 1982),23 Cournot oligopoly and monopolistic competition. 24 Two main strands can be identified here: those assuming various forms of Chamberlinian monopolistic competition and those analysing various forms of oligopoly.
Various models based on Chamberlinian monopolistic competition have been developed, mainly analysing the interaction between economies of scale, product differentiation and different forms of monopolistic competition (Krugman, 1987). These models have demonstrated welfare gains from increased product variety and from lower prices. Essentially, however, the insights of the H-O model are shown to hold quite well under conditions of product differentiation and such economies of scale.
The question which the analysts of various forms of oligopoly have asked is whether firms with market power act in a cooperative or non-cooperative manner. Since formal cartel and price-fixing arrangements are generally not legal, such cooperation arrangements are to a large extent tacit. Partly for this reason, the theory of cooperative behaviour in oligopolistic industries is not well developed. Most of the contributors on this subject have therefore restricted their analysis of markets to non-cooperative behaviour.
The outcome of non-cooperative behaviour by firms has largely depended on the strategic variables with which the game is played and the conditions of entry into and exit from the industry. Most theoretical work on oligopoly has tended to take as the strategic variable either outputs (the Cournot assumption) or prices (the Betrand assumption). In more general terms, various forms of market imperfection permit firms to earn returns exceeding those that are tenable in purely competitive industries, suggesting that trade policy can be used to influence the share of international profits accruing to domestic firms (and in that way to the economy). For instance, subsidies can be used to shift profits in favour of domestic firms, implying enhancement of their strategic position versus foreign rivals in competition for world markets.
Cournot's equilibrium is tenable when each firm is doing its best to maximize profit by choosing its output level, given the output levels of its rivals. In equilibrium, no aggressive threat by any firm is likely to be believed by its rivals. However, if one firm manages to reduce its costs (or get a subsidy), a new equilibrium would be set at a higher level of output and market share for that firm. Reference has also been made to the strategic use of R&D expenditure (or subsidies) to lower costs and shift the reaction curve outward. These results open up the possibility that government action can alter the outcome of the strategic game played by rival firms. It is in this context that possibilities for strategic trade policy have been proposed. The policy of protection to promote exports is one outcome of the presence of economies of scale. Movement down the firm's learning curve, leading to higher output (facilitated by protection) and falling marginal costs, is expected to enhance the firm's competitive position in world markets (Krugman, 1984).
Trade theory and accumulation effects: introducing new growth theories
Accumulation effects in the medium term can be derived from the neoclassical models of growth. The neoclassical model without exogenous shocks (e.g. productivity increases or population changes) presents a steady state capital-labour ratio determined by equating the real marginal product of capital and the discount rate. Any policy which raises the marginal product of capital will also raise the steady state capital-labour ratio, inducing output to grow faster in the medium term as capital accumulation takes place at a higher level (Solow, 1956; Baldwin, 1993).25 In the neoclassical model, medium-term growth results from the connection between trade and the marginal productivity of capital. Recent developments have augmented the Solow model by including human capital as a separate factor along with physical capital and unskilled labour (e.g. Mankiw et al., 1992).26 In this case, any policies (e.g. streamlining licensing procedures or easing foreign exchange restrictions) which would raise the marginal product of physical capital or human capital accruing to investors (e.g. by reducing the difficulty and expense of making investments) will also raise capital accumulation and growth in the medium term.
Although in the neoclassical growth models, continual accumulation takes the form of productivity-boosting knowledge, the rate of productivity growth is taken as given (determined exogenously). One contribution of the new growth theory is to endogenize the rate of productivity growth itself. A distinguishing feature of endogenous growth models (endogenizing investment) is that continuous accumulation requires that the return to accumulation does not fall as capital stock rises. Endogenous growth models have differed as to the type of factor which is thought to play a dominant role in the accumulation process: physical capital, human capital or knowledge capital.
The Marshallian concept, of increasing returns which are external to a firm but internal to an industry, was most widely used in static models, presumably because of the technical difficulties presented by dynamic models (Romer, 1986).27 Following Smith and Marshall, most authors explained the existence of increasing returns on the basis of increasing specialization and the division of labour, but these cannot rigorously be treated as technological externalities.
Evidence from observations made over almost three centuries, from 1700 to 1979, shows that productivity growth rates have been increasing for the Netherlands, UK and US (Romer, 1986). Similar evidence has been found for individual countries over shorter periods. Other evidence coming from growth accounting exercises and the estimation of aggregate production functions shows that the growth of inputs alone does not fully explain the rate of growth of outputs. Interest in dynamic models of growth driven by increasing returns was rekindled by Arrow's work on learning by doing (1962),28 in which increasing returns are supposed to arise from the new knowledge generated in the course of investment and production.
Some studies, closely related to the analysis of long-term growth, have focused on patterns of trade and their linkages with the patterns of innovation across countries, across sectors and over time. These studies have found some robust evidence regarding the impact of innovation on international competitiveness and on growth. This trend also relates to those neo-technology models which have attempted to endogenize technical progress within equilibrium open-economy development models (Krugman, 1979; Spencer, 1981) 29 Krugman's modelling of the technology gap between the North and the South and Spencer's analysis of the learning curve have contributed to bringing out some dynamic considerations in the discussions of international trade theory. Such approaches can be reduced to analyses of either learning curves or the generation of new intermediate inputs under monopolistic competition.
New growth theories which take technical progress as the driving force have extended Solow's insights by endogenizing technical progress. In the earlier models in this category, the rate of return to investment was prevented from falling due to technological spill-overs in production (Romer, 1986). The evidence on productivity growth over time coupled with the inadequacies of previous growth models motivated Romer (1986) to present a competitive equilibrium model of long-run growth with endogenous technological change, and with knowledge as an input which has increasing marginal productivity. In his model, long-run growth is primarily driven by the accumulation of knowledge by forward-looking, profit-maximizing agents. Knowledge is accumulated by devoting resources to research. New knowledge is assumed to be a product of research technology and this research exhibits diminishing returns. Romer combined three elements (externalities, increasing returns in the production of output and decreasing returns in the production of new knowledge) to constitute his competitive equilibrium model of growth. His model deviates from the Ramsey-Cass-Koopmans model and the Arrow model by assuming that knowledge is a capital good with an increasing marginal product.
A major problem with analyses undertaken in the equilibrium framework is their assumption of the existence of price- and/or quantity-based adjustment mechanisms which ensure clearing of all markets and the attainment of equilibrium in that sense. The assumptions based on the presence of maximizing agents become an inadequate representation of the general behaviour of agents when fundamental features of technological change (uncertainty and various irreversibilities) are invoked (Nelson and Winter, 1982; Dosi et al., 1990; Cooper, 1991).30 As North (1994) has suggested, the problem here reflects on the neoclassical body of theory. A theory of economic dynamics comparable to general equilibrium theory would be ideal for understanding economic change, but we do not have such a theory. The neoclassical theory is inappropriate for analysing and prescribing policy measures that will induce development. It is concerned with how markets operate rather than how they develop. Even when it attempts to take account of technological development and human capital investment, it still ignores the incentive structure, embodied in institutions, that influences societal investment in those factors.
Attempts to correct this deficiency have been based on more evolutionary micro foundations, whereby firms with different technologies and organizational traits interact under conditions of persistent disequilibrium. The essential aspects of Schumpeterian competition are highlighted, especially the diversity of firm characteristics and experience and the cumulative interaction of that diversity. Contributions in this category have focused more explicitly on the micro foundations of innovation by addressing firm-level decisions to invest in product or process innovations. The ceaseless search for better product quality and for cost-lowering process innovations continuously leads to productivity improvements. What Schumpeter referred to as 'creative destruction' is a process whereby the impulse coming from new products, new processes and new markets revolutionizes the economic structure from within, destroying the old one and creating a new one.
Contributors in this strand are more heretical and heterogeneous in nature and scope and their models are not always thoroughly formalized. Following Dosi et al. (1990), they may be classified into four broad groups: post-Keynesians (e.g. Posner, Vernon, Kaldor), structuralists in development economics (e.g. the dependency school), institutional economists such as Douglass North, and economic historians (e.g. Kuznets, Gerschenkron, Balough). Much of the management literature focusing on firm-level capabilities may be included in this approach (e.g. Porter, 1990). Studies which follow this approach agree on several points: that international differences in technology levels and innovative capabilities are crucial in explaining the trade flows and incomes of countries, that the general equilibrium mechanisms of international and inter-sectoral adjustment are relatively weak, that technology is not a free good and that allocation patterns induced by international trade have dynamic implications in the long term. The questions have implications regarding the causes of industrial development and growth, the linkages between these processes and their micro foundations, and the understanding of the on-going transformations and restructuring of world industry.
The evolutionary theory of economic change attempts to provide a formal theory of economic activity, driven by industrial innovation (consistent with the Schumpeterian view). It seeks to understand technical change, its sources and its impacts at micro and macro levels (Nelson and Winter, 1982). Evolutionary theory consists of heterogeneous modelling efforts which emphasize various aspects of economic change, such as the responses to market conditions of firms and industries, economic growth and competition through innovation. Many of its underlying ideas can be traced back to classical political economy (e.g. Smith, Marx, Schumpeter). In addition, contributors in this field have adopted tools of analysis from other fields. For instance, from the managerialists they have taken over a more realistic description of the motives that directly determine business decisions. From the behaviouralists they have taken an understanding of the limits of human rationality which make it unlikely that firms can maximize over the whole set of conceivable alternatives. The linkage of a firm's growth and profitability to its organizational structure, capabilities and behaviour is adopted from industrial organization (e.g. Coase, Williamson). The view that the histories of firms matter, because their previous experience influences their future capabilities, and that firms adapt to changing conditions is largely adopted from evolutionary theorists (e.g. Darwin, Lamarcker, Alchian) and economic historians berg. Rosenberg, David).
The version of evolutionary theory presented by Nelson and Winter (1982) questions two pillars of neoclassical theory. First, the maximization model of firm behaviour is questioned with respect to the way it specifies the objective function and the set of things that firms are supposed to know how to do. In addition, evolutionary theory objects to the way firms' actions are viewed as resulting from choices which maximize the degree to which the objective is achieved, given the set of known alternatives and constraints. Second, objections are raised to the concept of equilibrium which is used to generate conclusions about economic behaviour within the logic of the model.31 The general term that Nelson and Winter use for all the regular and predictable behavioural patterns of firms is 'routine'. Routine consists of well-defined technical routines for producing things, procedures (e.g. for hiring and firing, ordering new stocks), policies (e.g. for investment, R&D, advertising) and business strategies (e.g. on diversification, overseas investment). These routines are categorized into the operating characteristics governing short-run behaviour, those determining investment behaviour (period-to-period changes in the firm's capital stock) and those which operate to modify over time certain aspects of the operating characteristics (e.g. market analysis, operations research, R&D). There are also aspects of the behavioural patterns of firms which are essentially irregular and unpredictable. These are regarded as stochastic elements in the determination of decisions and decision outcomes. Evolutionary theory attempts to model the firm as having certain capabilities and decision rules and choice sets (through which the main objective is pursued). These choice sets are not well defined and exogenously given. The core concern of evolutionary theory is with the dynamic process by which firm behaviour patterns and market outcomes are jointly determined over time.
The emerging theory of dynamic firm capabilities is presented by focusing on three related features of a firm (Nelson, 1991):32 its strategy (a set of broad commitments made by a firm that define and rationalize its objectives and how it intends to pursue them), its structure (how a firm is organized and governed and how decisions are actually made and carried out) and its core capabilities (core organizational capabilities, particularly those which define how lower-order organizational skills are coordinated, the higher-order decision procedures for choosing what is to be done at lower levels, and R&D capabilities, particularly regarding innovation and how to take on-going economic advantage of innovation). Since the real world is too complicated for the firm to understand in the neoclassical way, firms will choose somewhat different strategies which will lead to their having different structures and different core capabilities.
When a new and potentially superior technology comes into existence in a relatively mature industry, the evidence suggests that what happens depends on whether the new technology is able to conform to the core capabilities of specific firms (competence enhancing) or requires very different kinds of capabilities (competence destroying) (Nelson, 1991). A change in management, and presumably a major change in strategy, is often necessary if an old firm is to survive in the new environment. Organizational change must be seen as the handmaid of technological advance and not as a separate force behind economic progress (Tidd, 1991; Nelson, 1991).33 In the long run, what has mattered most has been the organizational changes needed to enhance dynamic innovative capabilities. However, there is little in the way of tested and proven theory for predicting the best way of organizing a particular activity, and there is considerable disagreement about what features of a firm's organization are responsible for certain successes and/or failures. These can only be unveiled in concrete situations through empirical studies which seek to understand firm-level strategies, structures and capabilities and the environment in which they are operating.
Attempts to model the way allocation patterns of international trade influence the long-term dynamics of an economy have been made, incorporating the main ideas from the two-gap models and hypothesizing that world growth is determined by asymmetrical patterns of change in technological and demand structures (e.g. Kaldor, 1970, 1975; Pasinetti, 1981; Dosi et al., 1990).34 The locus of these approaches has largely been on the relationship between trade, levels of activity and growth.
It has been pointed out that the neoclassical premise of efficient markets is undermined by the existence of transaction costs, which create a need for institutions to define and enforce contracts. If trade is to prosper, certain informational and institutional requirements become necessary. Market institutions are expected to induce the actors to acquire information that will lead to the correction of their subjectively derived models. But if institutions shape the incentive structure of society, the learning process which occurs in the interaction between institutions and organizations shapes the institutional evolution of an economy (North, 1994). Institutions may be the rules of the game but organizations and entrepreneurs are the players. While the bulk of the choices made by players are routine, some involve altering existing contracts and some may even lead to alterations in the rules of the game. Thus the most fundamental long-run source of change is learning by the players.
Some implications of new trade theories for Africa
The main contributions of new trade theories are basically in the recognition that economies of scale (and the associated market structures) and differences in technological capabilities are important. As regards technological capabilities, some strands of these theories have contributed to setting (or resetting?) the stage towards endogenizing technological development and innovations in the analysis of trade issues.
One important message which comes out of the new trade theories is that technology differences are a fundamental force in shaping comparative advantages. This implies that, in the design of trade policy and the formulation of industrialization policies, it will be necessary to give explicit consideration to the changing role and conceptualization of technological change.
In the process of industrialization in the less industrialized economies, elevating the importance given to technological change could contribute to improving the efficiency of the import substitution industries and improving the international competitiveness of the export industries. In the field of trade, in particular, lessons from the new trade theories seem to be relevant in highlighting the role of technological development and innovation and the importance of being forward-looking in assessing trade potentials, in contrast to the implications of the static comparative advantages model. In addition, the new theories demonstrate the centrality of technology in trade and, in particular, the need to gain detailed knowledge of the structures and capabilities of exporting industries and firms as a basis for formulating export promotion policies and policies intended to promote industries with dynamic comparative advantages.
Although the focus of the new trade theories is primarily on North-North trade, some elements of the role of economies of scale, product diversity and explanations for intra-industry trade may be applicable to the place of South-South trade in the world economy. As regards intra-industry trade, the available evidence suggests that the average levels of such trade have been low in developing countries and even lower if we consider only the non-NICs (newly industrialized countries) (Greenway, 1991; Havrylyshyn and Civan, 1985).35 One problem with such evidence is that it is derived from static analysis and does not take into account the directions in which such intra-South trade could evolve. Such a dynamic conceptualization would require that consideration of South-South trade should address innovations such as the development of more appropriate products and processes for the South as a basis for South-South trade (Stewart, 1984, 1991)36 and should pose the question of the conditions under which South-South trade is feasible and viable, drawing lessons from the emerging patterns of trade as unveiled by the new trade theories. That policy has important implications for the evolution of intra-industry trade can be inferred from the evidence that intra-industry trade tends to be higher among countries (whether developed or developing) with some kind of integration arrangement (Balassa and Bauwens, 1988).37 This could be due to the lowering of trade barriers and/or the ability to exploit economies of scale, factors which are often associated with integration and cooperation arrangements.
The actual trade relations of SSA can be classified as a hub and spoke' arrangement, with Europe and North America as the hub and African countries as the spokes. The key feature of this arrangement is that trade between the hub and any spoke is easier than trade among the spokes (Baldwin, 1993). The hub-and-spoke trade arrangements exert a marginalizing effect on the African economies. Although the cost of production in SSA countries may be lower due to lower labour costs, higher trade costs (due to small markets as compared to those in the hub countries) are likely to more than offset the lower production costs. Trade liberalization between the hub and the spokes favours location of investments in the hub, especially if intra-SSA trade is not also liberalized. Hub-and-spoke trade liberalization artificially deters investment in the spoke countries. Reduction of intra-SSA trade costs would increase the chances that the lower SSA production costs will outweigh these trade costs, and so would facilitate investments in SSA.
Within the broader context of forging inter-firm linkages and cooperation arrangements, special attention will need to be paid to the possibilities of promoting investments, not only by TNCs from developed countries but also among countries in the region and with other developing countries. There is evidence that TNCs from developing countries can also have capabilities to share with other developing countries. As this study has shown, many exporting firms in Africa have undertaken various technological modifications and adaptations in response to country-specific conditions. Other studies have shown that adaptations have been made by developing country firms with respect to the characteristics of raw materials such as type, quality and input mix, scaling down, product quality and product mix, simplicity and capacity and factor intensity (Lecraw, 1981).38 As is noted in Chapter 4, these firms have tended to produce simpler, easily marketed products.
South-South inter-firm linkages and cooperation arrangements should be viewed as complementary to the kinds of benefits which can be obtained from inter-firm networks and cooperation arrangements with TNCs from the North and not necessarily as substitutes. The Abuja declaration on the establishment of the African Economic Community is an encouraging step. Its implementation, however, should involve taking steps towards establishing the institutional framework to spearhead the development of these kinds of inter-firm linkages and cooperation arrangements, not only within Africa but between Africa and other regions.
Differences in income levels, historical backgrounds and other environmental considerations suggest that there are bound to be inter-regional, inter-country and intra-country differences in demand structures. This diversity of demand structures means that African firms could find and exploit windows of opportunities in export markets inside and outside the region.
Intra-regional trade in Africa has remained low, at perhaps 5 per cent of total African trade, and has not shown any clear signs of increasing. However, there are at least two sources of optimism regarding the prospects of growth in regional cooperation in trade and investments. First, there are indications that a substantial volume of intra-regional trade is unrecorded. Second, the degree of complementarity is often understated. That the potential for intra-Africa trade is higher than trade figures suggest is indicated by the results of demand and supply studies. For instance, the project 'Promotion of Intra-regional Trade Through Supply and Demand Surveys' was part of a more comprehensive programme aimed at promoting trade between the member states of the Preferential Trade Areas for Eastern and Southern African States (PTA) (ITC, 1985).39 The data shows that many of the items which are imported into the PTA are also exported from the PTA, indicating that there are potentials for intra-regional trade in a wide range of sectors.40 In spite of having the necessary raw materials, the region continues to be a net importer of various light manufactures.
Even within agriculture, which is often cited as an example of competitive structures, the potentials seem to be greater than official figures acknowledge. For instance, differences in climatic and agronomic conditions in the Eastern and Southern Africa region have been shown to be a source of complementarily of production (Koester and Thomas, 1992).41 Similar results have been obtained in the case of West Africa, where it has been found that the differences between countries in their access to international trade, and their historical conditions, technological accumulation and consumer preferences, suggest that there is a good potential for specialization (Badiane, 1992).42 The difference between potential and actual intra-Africa and South-South trade is due to the many obstacles to such trade: weak industrial structures, the lack of intra-industry linkages, inadequate infrastructure, tariff and non-tariff barriers, perceived unequal distribution of benefits and the instability of the real exchange rates between African countries.
The role of dynamic learning processes and the competitive pressures in export markets are relevant to the extent that such dynamic economies are industry-specific. Contrary to the neoclassical premise that all activities are equally important, new trade theories highlight the existence of strategic activities which could be developed through policy. One case is that of protection as a form of export promotion, as argued by Krugman (1984). The new trade theories' exposition of the possibility of 'import protection as export promotion' introduces a way of using trade policy strategically. The case for selectively supporting specific high-potential industries through government policy has been demonstrated to varying degrees in the experiences of the developed countries and the NICs.43