Technology and National Competitiveness 9780773562844

Citation preview

Technology and National Competitiveness: Oligopoly, Technological Innovation, and International Competition

JORGE NIOSI, Editor

McGill-Queen's University Press

Technology and National Competitiveness

This page intentionally left blank

Technology and National Competitiveness Oligopoly, Technological Innovation, and International Competition EDITED BY JORGE NIOSI

McGill-Queen's University Press Montreal & Kingston • London • Buffalo

McGill-Queen's University Press 1991 ISBN 0-7735-0827-9 (cloth) ISBN 0-7735-0859-7 (paper) Legal deposit first quarter 1991 Bibliotheque nationale du Quebec

Printed in Canada on acid-free paper Publication of this book has been assisted by a grant from the Publications Committee of the University of Quebec in Montreal. This book was typeset by Typo Litho composition inc. in 10/12 Baskerville.

Canadian Cataloguing in Publication Data Main entry under title: Technology and national competitiveness Papers presented at the International Seminar on Oligopoly, Technological Innovation and International Competition, held at Universite du Quebec a Montreal in 1987. ISBN 0-7735-0827-9 (bound). — ISBN 0-7735-0859-7 (pbk.). 1. Technology - Economic aspects. 2. International trade. 3. Leontief, Wassily W., 1906. 4. Comparative advantage (International trade). 5. Technology and state. I. Niosi, Jorge, 1945. II. International Seminar on Oligopoly, Technological Innovation and International Competition (1987: Universite du Quebec a Montreal). HD45-T42 1991

382'.1042

091090007-8

Contents

Tables vii Figures xi Contributors

xiii

Introduction/JORGE NIOSI

xv

PART ONE: TECHNOLOGY, ENERGY, AND INTERNATIONAL TRADE 1 Technological Change and International Trade FAYE DUCHIN

3

2 Technology Intensity of us, Canadian, and Japanese Manufactures Output and Exports LESTER A. DAVIS

11

3 Canadian Industrial Energy Consumption and External Trade K.E.

HAMILTON

44

4 Technological Clusters and Competitive Poles: The Case of Canadian Energy CHRISTIAN DEBRESSON

60

vi Contents PART TWO! STATE, TECHNOLOGY, AND COMPETITIVENESS

5 Technological Innovation and International Competitiveness GIOVANNI DOSI and LUC SOETE

Ql

6 The State and International Trade: Technology and Competitiveness JORGE NIOSI and PHILIPPE FAUCHER 119 7 Technological Competitiveness Considered as a Form of Structural Competitiveness FRANCOIS CHESNAIS

142

8 Indicators of Industrial Competitiveness: Results and Limitations THOMAS HATZICHRONOGLOU 177 PART THREE: TECHNOLOGICAL DEVELOPMENT AND GOVERNMENT STRATEGY 9 New Modes of Competition in the Textile and Clothing Industry: Some Consequences for Third World Exporters LYNN KRIEGER MYTELKA

225

10 Engineering, Design Services, and Technology Transfers: The Case of the Republic of South Korea JACQUES PERRIN

247

PART FOUR: INDUSTRIAL STRUCTURE AND INNOVATION

11 Oligopoly, Innovation, and Firm Competitiveness BERNARD BONIN

267

Tables

1 Capital per Unit of Labour in the Production of the Total Final Bill of Goods 5 2 Capital per Unit of Labour in the Production of us Exports 6 3 Capital per Unit of Labour in the Production of us Competitive Imports 7 4 Capital-to-Labour Ratio for us Competitive Imports Divided by Capital to Labour Ratio for Exports 8 5 Total R&D Expenditures on Manufactures Production, and Technology Embodied in Manufactures Output to Final Demand 21 6 Source of Total Technology Embodied in Manufactures Output to Final Demand, 1972 and 1984 23 7 Total Technology Intensity of Manufactures Output to Total Final Demands 24 8 Shares of Manufactures Total Final Demands and Embodied Technology Accounted for by Exports and Domestic Demand 25 9 Canadian Total Technology Intensity Components of Manufactures Output to

viii Tables Final Demand, with and without Contribution from Imported Inputs 27 10 Contribution to Canadian Total Technology Intensity of Domestic and Imported Intermediate Inputs to Manufactures Final Demand 28 11 Canadian Product Sectors Ranked by Import Share of Total Technology Embodied in Canadian Exports 29 12 Top One-third of Product Sectors, Ranked by Technology Intensity in 1984 32 13 Percentile Distribution by Export Value of Total Technology Embodied in Manufactures Export Classes in 1972 and 1984, Ranked by Technology Intensities in Each Year 33 14 Distribution by Number of Export Classes of Export Value of, and Total Technology Embodied in, Manufactures Exports in 1972 and 1984, Ranked in Top Third of Classes by 1984 Technology Intensity 34 15 Percentile Distribution of Total Technology Embodied in Direct and Indirect Inputs to Manufactures Exports, Ranked by Technology Intensity in 1972 and 1984 35 16 Primary Energy Production and Trade 46 17 Aggregate Economic and Energy Indicators 47 18 Canada's Total Primary Energy Requirements with and without External Trade 48 19 Energy Intensiveness of Exports and Imports 49 20 Selected Energy Intensities, 1976 50 21 Import Energy Intensity Change Decomposition 51 22 Decomposition of Commercial Energy Use per Dollar of GDP 52 23 Domestic and International Oil Prices 53 24 Correlations with Sectoral Energy Intensity 54

ix Tables

25 Secondary Sectors with High Import Competition 55 26 Innovation Matrix: Canada, 1979 68 27 Quantum Index of Exports by Large Groups of Products: Market Economies, 1948-1983 121 28 Value of Exports in Current us Billion Dollars by Large Groups of Products: Market Economies, 1948-1983 122 29 Unit Value Index of Exports: Market Economies, 1948-1983 123 30 Recent Trends in High-Technology World Trade: Market Economies, 1970-1984 124 31 Estimated Breakdown of Industrial R&D in the OECD Area 161 32 Reaction of Export Prices to Variations in Competitors' Prices 183 33 Relative Weights with Respect to Totals of the Seven Countries 187 34 Export Market Shares of Services 188 35 Export Market Shares of Manufactured Products 189 36 Export Market Shares in Manufacturing Industries as a Function of Their Technological Content 190 37 Relative Weights in the Total Flow of Outward Foreign Direct Investment of the Seven Countries 190 38 Relative Weights of Technological Receipts with Respect to Total Receipts of the Seven Countries 191 39 Apparent Rate of Import Penetration for Total Imports of Manufactured Goods 191 40 Apparent Penetration of Manufactured Imports as a Function of Their Technological Content 192 41 Inward Direct Investment Flows 192 42 Technological Payments 193 43 Real Growth Rates 193 44 United States: Foreign Penetration, 1977-1982 195

x Tables

45 Canada: Foreign Penetration, 1972-1980 196 46 United States: Trade Balance 202 47 The Evolution of Factors of Competitiveness Linked to Price in Manufacturing Industries with Respect to the Twenty-four Countries of the OECD 204 48 Trade Balance with the World in Manufactured Products 205 49 Commercial Trade Balance with the World in Services 206 50 Trade Balance with the World in All Goods and Services 206 51 Growth of Industrial R&D Expenditures and GDP 208 52 R&D Expenditures of Manufacturing Industries as a Percentage of Value Added 209 53 R&D Expenditures of Firms with Respect to the Grosss Fixed-Capital Formation of the Manufacturing Sector 209 54 Specialization by Niche 211 55 Revealed Comparative Advantage 212 56 Contribution to the Trade Balance 212 57 Import Elasticities with Respect to Global Domestic Demand, 1974—1985 214 58 Contribution of Technical Progress and the Substitution of Capital / Labour to the Apparent Labour Productivity, 1973— 1983 215 59 Variation Coefficient of Industrial Expenditures on R&D 216 60 Rate of Modernization in Third World Spinning Industries 232 61 Rate of Modernization in Third World Weaving Industries 234 62 A Comparison of Manufacturing Costs in Spinning and Weaving, 1979, 1983, and 1987 239 63 KOPEC Participation in Nuclear Engineering 259

Figures

1 Sources of Technology in Output 17 2 Basic Directed Graph and Types of Clusters 66 3 The Energy Innovation Cluster 70 4 The Energy Patent Cluster 72 5 Tree Structure Showing the Technological Function of the Company 157 6 Simplified Diagram of the Technological System of the First Half of the Nineteenth Century 164 7 The State/Research/Industry System: Transfer and Diffusion of Advanced Technologies 167 8 Rate of Apparent Penetration of Imports of Manufactured Products for the us, Japan, and the EEC 194 9 R&D Expenditures of Manufacturing Industries in the us, EEC, and Japan as a Percentage of All Such Expenditures of OECD Countries 210 10 The Functions of Design Organizations 250

This page intentionally left blank

Contributors

BONIN, BERNARD Economist, presently deputy governor of the Bank of Canada, Ottawa CHESNAIS, FRANgois Administrator, Directorate for Science, Technology and Industry, Organization for Economic Co-operation and Development, Paris DAVIS, LESTER A. Analyst, Office of Trade and Investment Analysis, United States Department of Commerce, Washington DEBRESSON, CHRISTIAN Professor, Department of Administrative Science, Universite du Quebec a Montreal, and researcher, Centre for Research on the Development of Industry and Technology (CREDIT) DOSI, GIOVANNI Professor, Institute of Economics and Statistics, Universita di Roma DUCHIN, FAYE Director, Institute for Economic Analysis, and Research Professor, Graduate School of Public Administration, New York University, New York FAUCHER, PHILIPPE Professor, Department of Political Science, Universite de Montreal, and researcher, CREDIT HAMILTON, K.E. Researcher, Division of Analytic Studies, Statistics Canada, Ottawa HATZICHRONOGLOU, THOMAS Researcher, Directorate for Science, Technology and Industry, Organization for Economic Co-operation and Development, Paris

xiv Contributors MYTELKA, LYNN KRiEGER Professor, Department of Political Science, Carleton University, Ottawa NIOSI, JORGE Professor, Department of Administrative Sciences, Universite du Quebec a Montreal, and director, CREDIT PERRIN, JACQUES Researcher, Conseil national de la recherche scientifique and at Universite de Lyon 2, France SOETE, LUC Director, Maastricht Economic Research Institute on Innovation and Technology (MERIT), Rijksuniversiteit Limburg, Maastrich, Holland

Introduction

This volume is a selection of papers presented to the international symposium "Oligopolies, Technological Innovation and International Competitiveness" organized by the Centre for Research on the Development of Industry and Technology (CREDIT) of the University of Quebec at Montreal in October 1987. It is a collection of original works concerning the role of technology in international trade and the importance of industrial structure in technical innovation. The first part of this collection is comprised of four chapters inspired by the works of Wassily Leontief. The American economist and 1973 Nobel Prize winner is well known for the application of his input / output model to factor-content analysis of American external trade. In 1954 he published a study in which he noted that the United States, despite its capital endowments, exported labourintensive goods and had a trade deficit in capital-intensive products. His famous contribution, the "Leontief paradox," changed the focus from the highly theoretical analysis of international trade to more empirical considerations. Leontief s publication was soon followed by significant adjustments to the list of "production factors" (land, labour and capital); technology was added as an independent factor of production, the intensity of which needed further analysis. The contributions in Part One, inspired by this empirical tradition founded by Leontief, analyse the export-factor intensity of several industrial economies. The chapter by Faye Duchin, Leontief s principal collaborator, applies the input/output model to an analysis of us international trade from 1963 to 1977 with projections to the year 2000. These

xvi Introduction

data and previsions indicate that technical change and transformations in international trade have significantly altered the configurations brought to light in the 19505. The United States exports less and less labour and makes progressively higher use of capital in its industrial production, a change that is reflected in its external trade. Lester Davis compares the United States, Canada, and Japan in terms of the technological intensity of their production and exports in manufactured goods between 1972 and 1984. He finds that the technological intensity of industrial production significantly increased for all three countries over this period. As for the technological intensity of industrial exports, those from Canada and the United States remained constant, while Japan's showed a rapid increase. Canadian production in the manufacturing industries showed an increasing use of mostly imported technology. These conclusions show the growing importance of technological factors in industrial production and world trade and illustrate the narrowing technology gap among industrialized countries. K.E. Hamilton's research analyses the energy intensity of Canadian exports and compares it with imports to Canada for the 1971-76 period. Hamilton finds that the energy content of the country's exports is considerably higher than that of its imports because of the predominance of semi-finished and primary products (such as primary metals, pulp, and paper) in exports and a higher import level of finished products with a lower energy content. This diachronic study demonstrates that the energy intensity of exports remained constant between 1971 and 1976 but that of imports declined in the same period. Hamilton's contribution therefore tends to support the Heckscher-Ohlin neoclassical thesis of international trade: Canada exports factor-intensive products with which it is most abundantly endowed. Christian DeBresson examines not the factor content of Canadian exports but rather the effect of external trade on the organization of technological innovation "clusters." Basing his analysis on the theory of technology gaps, DeBresson takes the case of Canada's energy sector, which appears to be the primary focus of such innovation clusters. This is also the sector in which Canada is most competitive in international trade, since products developed by innovating firms are often exported. DeBresson's contribution complements that of Hamilton and supplies it with an explanatory base. The volume's second section examines current international trade theories from a technological perspective. In their neoclassical form, these theories have neglected the role of technology (and of the state as one of its primary producers). Luc Soete and Giovanni Dosi ex-

xvii

Introduction

plain that an increasingly abundant literature on technology and international trade now seriously questions the validity of neoclassical comparative advantages. Neoclassical theory claims that technological differences between nations can be represented by production functions that are understood to be identical throughout the world. Perfect competition is another assumption of this model that has significant consequences in terms of the notion of perfectly divisible, accessible, and free technology. Soete and Dosi concede that critics of neoclassical theory are a many-hued mix of unorthodox theorists, none of whom can lay claim to the depth or rigorous analysis of David Ricardo or Paul Samuelson, but all of whom vigorously defend the relevance of their findings. Dosi and Soete classify the nonconformists into three major groups: (a) those who accept the basic principles of neoclassical theory but wish to modify one or another of its prepositions; (b) those, further from the model, who observe differences in the wealth and trade structures of nations and explain them by international technological disparities; and (c) those who focus on the international diffusion of technology. J. Niosi and P. Faucher's contribution complements the chapter by Dosi and Soete. The authors present their own critical review of the literature on international trade, noting the scant attention paid the role of the state beyond its erection of protectionist barriers and setting of exchange rates. Niosi and Faucher maintain that the state plays a crucial role in national factor supply, particularly in the case of technology. The state intervenes, for example, in manpower training, in public research-and-development financing, in setting up an institutional environment for research and innovation, and through the creation of technologically competitive public enterprises (or by promoting mergers of private firms). Faucher and Niosi conclude that the state must be an integral part of any analysis of the competitiveness of national economies. In the following chapter, Francois Chesnais defends the notion of structural competitiveness. He rejects the neoclassical analysis that deals with competitiveness in terms of wage and price rates; the neoclassical approach is based on the conception of the state as an abstract reservoir of production factors. For Chesnais, as for the previous authors, the state is a central factor in the workings of the national productive system and its absence in the literature is a major failing in neoclassical international trade theory. Drawing inspiration from the French theory of regulation, he emphasizes the central role of the state in the constitution and maintenance of the national productive structure. This structure is based on the technical system (a concept developed by Bertrand Gille), a key factor in the deter-

xviii

Introduction

mination of inter-industry relations. In this network of interdependencies, the manufacturing of industrial machinery determines the structural base of competitive national economies. Thomas Hatzichronoglou uses three indicators to measure the competitiveness of industrial economies: country exports as a percentage of world trade, the importance of imports to domestic demand, and multinational market penetration via subsidiaries. The author makes use of these results to draw conclusions on the competitiveness of the United States, the European Economic Community, Japan, and Canada within the Organization for Economic Cooperation and Development (OECD). Part Three deals with the analysis of government technological strategies in developing countries within the context of transformations in the world economy and accelerating technological change. Lynn Mytelka's analysis of the textile and garment industries underlines the growing role of production (design and organization) and marketing. These changes, combined with increasing protectionism in the industrialized world, have made access to markets in the North more difficult for newly industrializing countries (NICS). In order to avoid these difficulties, South-East Asian NICS have accelerated technology adoption and have advanced to the production of more sophisticated and higher-value-added goods. But several other NICS have been unable to keep pace on the learning curve. Some, like Brazil, have fallen off the pace of new technology adoption due to a lack of capital investment brought on by the country's debt crisis; others, like Morocco, lack domestic design capability. The renewed competitiveness of several developed countries in the textile and garment industries will present a major obstacle for many NICS in their industrialization process. Jacques Perrin analyses the importance of engineering design services in technological development and their transfer to South Korea. Perrin notes that the design for machinery and integrated production units is generally provided to developing countries by engineering firms from the industrialized world. However, these installations are regularly modified over the course of the production process through the introduction of innovations that, although minor, are crucial for the firm's competitiveness. This underlines the necessity that developing countries master engineering technology. A good illustration is South Korea, which since 1962 has employed a successful industrial policy of "de-packaging" imported technology. As a result, South Korea has managed to rapidly improve its local engineering capability, which it now exports to other developing countries. These findings provide empirical support for the hy-

xix

Introduction

pothesis of Faucher and Niosi and of Chesnais that the state plays a significant role in modifying technological production factors. The final section of this collection deals with industrial structure and technological innovation. Bernard Bonin focuses his study on the market conditions and the characteristics of enterprises best suited to engage in innovative activity. His point of departure is the Schumpeterian theory, which proposes that large firms operating in oligopolistic markets are the ones most likely to innovate. In a comparative empirical analysis, Bonin concludes that the Schumpeterian theory has neither been proven nor totally repudiated. Large firms are more active in R&D but do not necessarily innovate. This conclusion also holds on an international level; large multinational firms are not necessarily more innovative than small and medium-sized enterprises. These theoretical and empirical contributions underline the growing importance of technology as a factor of international trade and the central role played by the state in its development. This conception goes beyond the traditional and simplistic view of the limited importance of the state and recognizes its direct implication in the supply of production factors that have significant impact on the structure of international trade. The authors have not taken a position on the still unanswered question, What has become of the neoclassical theory of international trade? The Leontief tradition does not directly refute Ricardo, but the input/output-model empirical studies reveal some of the doctrine's incoherencies. Authors of political economy are increasingly hesitant to defend neoclassical theory, in part because of its neglect of the role of state. While no new comprehensive theory has yet emerged, it is nevertheless crucial to take into consideration technological change in the analysis of international movement of trade, capital, and technical knowledge. We hope that the texts in this collection will serve as a positive contribution to future debate. Jorge Niosi Director, CREDIT

This page intentionally left blank

PART ONE

Technology, Energy, and International Trade

This page intentionally left blank

CHAPTER ONE

Technological Change and International Trade FAYE DUCHIN

This paper first replicates the findings established by the "Leontief paradox," namely, that over a long historical period the United States has regularly absorbed labour and economized on capital through its international trade. Next, the paper shows that technological change has exercised a systematic influence on the factor content of us trade in recent decades. An economic interpretation of these findings is offered, taking a dynamic rather than a comparative static point of view. Finally, recommendations are made for extending the conceptual framework used in this paper in the direction of an operational theory of international trade. When I was invited to give the opening address at a colloquium whose major themes were provided by the work both of my colleague Wassily Leontief and of the late Francois Perroux, I was not yet familiar with the thought of the latter. Having now read his last book (A New Concept of Development], I am struck by his ideas about the appropriate role for theory in economic analysis. While his concerns were for the problems of the real world, he held theory in high esteem — theory leading to action, not to abstraction — that is, what I could call operational theory. He also recognized the power of mathematics and the possibility of using formal models not merely for illustrative purposes but to provide guidance for addressing complex practical problems that would involve changes in economic structures. The results reported here are based on work supported by the us National Science Foundation under grant PRA 8311407.

4 Technology, Energy, International Trade

Economists do not often write papers examining the role of theory, but in the area of international trade - especially since the computations that established what has been called the Leontief paradox - this subject has necessarily arisen. The dominant view is that constructing theory is essentially an exercise of formal logic that should not be constrained by factual analysis. In this view, it is, by contrast, the role of empirical analysis to collect the necessary factual information and to use such information in testing hypotheses that pit alternative theories against each other. Perhaps the fundamental contribution of Wassily Leontief has been to offer a different agenda for a science of economics, one based on operational theory, where constructing the conceptual framework and using it in the analysis of the actual economy are practically inseparable. This close interaction is made possible by the very particular role played by data in the input/output approach created by Leontief. Many of the phenomena studied in the natural sciences involve a handful of strongly interrelated variables, and the parameters describing the strengths of these relationships tend to be constant over relevant stretches of time and space. For these reasons, models in the natural sciences, with some exceptions, may involve extensive computation but use relatively little data. In economics, by contrast, the phenomena that interest us involve large numbers of often weakly interrelated variables, and the parameters measuring the strengths of these relationships tend to be different in different economies and at different times. It is precisely these parameters - describing the technical structure of an economy over successive stretches of time — that are the contents of input/output databases. Input/output theory is formulated directly in terms of the parameters comprising the technical database, and implications of different experiments (or scenarios) are quantified through actual computations. I would like to turn now to the Leontief paradox, which is one of the principal themes of these meetings. It is generally acknowledged that at least in the late 19408, the period examined by Leontief, the United States was well endowed with physical capital in comparison to its actual and potential trading partners. In virtually all sectors, it tended to use techniques of production that put a relatively large quantity of capital at the disposal of the average worker. Within the theoretical framework that then dominated virtually unquestioned, such a country would be expected to use more capital relative to labour in producing its exports than would be required, in the absence of international trade, to produce its competitive imports. But

5 Technological Change and International Trade Table 1 Capital per Unit of Labour in the Production of the Total Final Bill of Goods (dollars in 1979 prices/person-year) Bill of Goods Technical Matrices

1963

1967

7972

1977

1990

1963 1967 1972 1977

$29,796 32,534 39,088 41,229

$30,019 32,783 39,498 41,688

$30,052 32,822 39,658 41,924

$28,453 31,373 38,381 40,683

$29,354 32,146 39,179 41,572

1990 2000

49,558 59,257

50,123 59,943

50,073 59,642

48,836 58,195

49,704 58,979

Leontief found that the capital-to-labour ratio in the United States in 1947 was about 30 percent higher for the hypothetical production of imports than for the actual production of exports. The paradox ushered in unprecedented activity in the empirical analysis of international trade flows and has led to new conceptual work in several directions, notably efforts to distinguish the characteristics of different kinds of capital and labour and to represent explicitly the roles in production of factors other than capital and labour. In my view, the most significant new direction was taken by Professor Leontief himself in his construction of the World Input/ Output Model (Leontief, Carter, and Petri 1977). It is perhaps ironic to identify this model as a follow-up to Leontief s earlier work on trade, since the model does not incorporate a representation of comparative advantage. I will return to this point later, but first I would like to describe some recent computations carried out at the Institute for Economic Analysis, New York University. The purpose of these computations was to investigate the factor content of us exports and imports over the period 1963—2000 and to examine the possible implications of technological change over that period for the factor content of us trade. This analysis makes use of the database assembled for implementing our dynamic input/output model; the data are documented in other places. The fix ideas, the diagonal elements of Table i, show the capital (in constant 1979 prices) per person-year of labour required to produce all final deliveries in the United States in 1963, 1967, 1972, 1977, and — given projected technical parameters and the projected bill of goods - in 1990 and 2000. This ratio is seen to grow over the twenty-seven—year period through 1990 from just under $ 30,000

6 Technology, Energy, International Trade Table 2 Capital per Unit of Labour in the Production of us Exports (dollars in 1979 prices/person-year) Bill of Goods Technical Matrices

1963

1967

1972

1977

1990

1963 1967 1972 1977

$37,476 42,056 56,813 61,055

$36,198 40,784 55,376 59,748

$33,215 38,211 52,963 57,943

$28,947 34,618 49,875 55,954

$18,325 24,457 41,178 54,331

1990 2000

71,400 83,401

70,392 82,830

68,664 81,124

66,821 79,393

67,105 82,487

per worker to almost $50,000 per worker. The other numbers in the table are the results of experiments intended to distinguish the effects of technological change from those related to changes in the composition of final deliveries. Looking down any column, say the first, we see a monotonic increase in capital per worker to produce a given bill of goods with more recent technologies. However, the numbers across each row are virtually constant: while the composition (at the Sg-sector level of detail in terms of which the computations were based) of final deliveries changes with time, these changes appear to be unrelated to the intensity of factor use. This will not turn out to be the case for either exports or imports. Capital requirements per unit of labour for the production of us exports are shown in Table 2. Here again there is a systematic increase in the use of capital per unit of labour from one benchmark year to the next (along the diagonal) that is explained by the systematic increase attributable to more recent technologies (down each column). However, over the entire historical period, there is a systematic, offsetting decline across each row. The effect of the shifts in the composition of us exports over this period was to slow down the technological displacement of labour required in their production. The capital per unit of labour that would be required for the domestic production of us competitive imports is shown in Table 3. The reasons behind the monotonic increases across the diagonal and down the column are now familiar. The figures across each row show that there were significant but non-monotonic changes in factor proportions attributable to the changing composition of us imports. A more detailed analysis reveals that it is mainl

7 Technological Change and International Trade Table 3 Capital per Unit of Labour in the Production of us Competitive Imports (dollars in 1979 prices/person-year) Technical Matrices

Bill of Goods 1977

1990

1963

1967

1972

1963 1967 1972 1977

$49,208 54,957 67,095 70,383

$43,997 48,930 61,728 65,618

$42,049 46,629 59,613 64,475

$47,127 53,704 69,432 74,587

$18,843 25,208 43,453 59,069

1990 2000

79,767 90,315

75,806 87,358

75,271 87,671

86,119 100,211

72,941 90,222

the changing physical amounts and unit price of petroleum imports that explain this erratic behaviour. Such non-monotonic changes across the rows may characterize both the exports and the imports of many, especially smaller, economies. Finally, the capital-to-labour ratios for us competitive imports (Table 3) divided by the same ratios for exports (Table 2) are shown in Table 4; this is the statistic originally computed by Leontief for 1947. All the numbers in the table are greater than i .o, which means that over the entire period the United States was still exchanging labour for capital with each million dollars' worth of exchange of goods and services. The patterns down the columns and across the rows are, however, of considerable additional interest. The consecutive numbers in the first four columns of Table 4 are systematically falling. This means that despite the significant changes in the composition of both imports and exports over the period from 1963 to 1977 technological changes over the period reduced the difference in factor proportions of imports and exports for all of these bills of goods. Furthermore, these reductions in the difference of factor proportions can be expected to continue under the technological changes that have been projected for 1990 and 2000. The lowest figures in Table 4 are arrayed not only along the last row but also down the last column, corresponding to the composition of exports and of competitive imports that have been projected for 1990. Thus, according to our projections, both technological change and the changing composition of us international trade can be expected to diminish the exchange of American labour for foreign capital that the United States has now experienced for at least a number of decades.

8 Technology, Energy, International Trade Table 4 Capital-to-Labour Ratio for us Competitive Imports Divided by Capital to Labour Ratio for Exports Bill of Goods Technical Matrices

1963

1967

1 972

1977

1990

1963 1967 1972 1977

1.31 1.31 1.18 1.15

1.22 1.20 1.11 1.10

1.27 1..22 1..13 1.11

1.63 1.55 1.39 1.33

1.03 1.03 1.06 1.09

1990 2000

1.12 1.08

1.08 1.05

1.10 1.08

1.29 1.26

1.09 1.09

These computations yielded a number of specific conclusions about us international trade in addition to those indicated above. In more general terms, from a dynamic perspective the exchange of labour for capital on the part of the United States appears not paradoxical but rational in view of the fact that technological change has involved the displacement of labour by capital, and at a faster rate in the United States than in other countries. Through trade the Americans have managed to absorb the displaced factor. The comparative static input/output framework for a single country, while it has proven fruitful both in hypothesis testing and in empirical exploration, is not adequate for investigating fundamental questions about the international division of labour - notably, how it is likely to change and with what implications for different economies. An operational assessment of a country's international comparative advantage has never been made because the massive informational requirements to cost out feasible production alternatives in all countries simultaneously cannot yet be met. The basic requirements include full, reasonably compatible input/output matrices and related data for all potential trading partners as well as indices relating the unit price of a sector's output in one country to that of the corresponding sector (if any) in other countries.' These are entirely analogous to the informational needs of an operational, multi-sectoral, dynamic input/output model in order to distinguish - in the latter case on an intertemporal rather than spatial basis physical differences of goods and services from differences in relative prices of essentially the same goods and services.

9 Technological Change and International Trade

The World Input/Output Model, which divides the world economy into fifteen regions, each described in terms of about forty-five goods and services, represents the first important step in this direction. Since only a very preliminary database could be assembled when this model was constructed in the mid-19708, no attempt was made to measure, let alone to compare, costs of production in different regions. Instead, a region's exports were represented in terms of sector-specific international market shares, and competitive imports were assumed to be proportional to total domestic use. Thus, the model does not explicitly represent the links between, on the one hand, the changing availability of different types of labour and capital and the adoption of new technologies and, on the other, the determination of each region's pattern of trade. (No other operational model does this either, of course.) The form of of these equations remains to be worked out. This interregional input/output model of the world economy nonetheless represents a tremendous achievement for two reasons. First, it has been constructed and used for exploring economic alternatives — for instance, how different levels and regional patterns of arms trade, or how changes in the patterns of use and trade of raw materials, could be expected to affect various economies. The model requires exogenous trade parameters but allows for the computation of endogenous trade flows. Second, from the point of view of a longer-term research strategy, the great strength of this model is that it provides a framework for accommodating the large, highly structured database that will be needed for carrying out a more fundamental investigation of the changing international division of labour based on changing cost structures. In addition, the experience of building and using the model has gone some distance in solving a number of conceptual problems that need resolution in order to facilitate the assembly of this information. The existing model and database and the various empirical investigations that have already been carried out within this framework demonstrate both the feasibility and the promise of proceeding in this direction. NOTE

i If all inputs could be measured in physical units, like tons of (standard quality) steel or number of (standard) insurance policies, only factor costs would need to be specified, and sectoral cost-based prices could

io Technology, Energy, International Trade be calculated. In practice, differences in quality and product mix make such a simplification impratical.

REFERENCE

Leontief, W., A.P. Carter, and P. Petri. The Future of the World Economy. New York: Oxford University Press, 1977.

CHAPTER TWO

Technology Intensity of US, Canadian, and Japanese Manufactures Output and Exports LESTER A. DAVIS

Innovations and technological advances are important factors in the global race for export markets. These contributions to a nation's competitiveness are made possible by research-and-development expenditures. Indeed, R&D expenditures can be thought of as investments in the technology that is embodied in a nation's industrial output and as determinants of the international competitiveness of its industries. The value and intensity of technology embodied in output can be estimated at national and sectoral levels for economies that report detailed R&D and shipments data. These estimates allow international, intertemporal, and intersectoral assessments of the contribution of R&D expenditures to output for domestic use and export. A full accounting of the R&D contribution to output requires that estimates be based on the total technology that is embodied in output from all input sources. These sources include (a) technology embodied directly in the final product from R&D expenditures by the final producer and (b) technology embodied indirectly in the final product from upstream R&D expenditures by domestic producers of intermediate and capital goods inputs and by foreign producers of imported inputs. These estimates require application of national input/output methods to national R&D data. The methods used in this research paper extend those used in the widely cited OTIA (Office of Trade and Investment Analysis, us Department of Commerce) staff research report "Technology Intensity of us Manufactures Output and Exports," published by the department in 1982 and reprinted by the Organization for Economic Co-operation and Development (OECD) in 1983. The earlier methods are extended here primarily to account for technology embodied in domestic capital goods inputs and, in the case of

12 Technology, Energy, International Trade Canada, technology embodied in imported inputs. In this paper, us manufactures technology performance is compared with that for Canada and Japan in 1972 and 1984.' Principal results of the analysis are as follows: • The value of total technology embodied in manufactures output of all three countries rose greatly between 1972 and 1984. Nevertheless, the total technology intensity of us and Canadian output for exports remained flat and the technology intensity of output for domestic use actually declined. In sharp contrast, the total technology intensity of Japan's output - both for export and domestic use - rose sharply. The Japanese increas cut by more than half the technology intensity lead of us exports. • Imported inputs contributed a very large share of the total technology embodied in Canadian output. Furthermore, the import share for Canadian output rose sharply between 1972 and 1984. • Between 1972 and 1984, the concentration of technology embodied in most us and Canadian technology-intensive manufactures export classes increased only slightly. However, the technology concentration in Japan's top technology-intensive export classes increased sharply.

INTRODUCTION

Major recent international trade trends have renewed the demand for a greater understanding of the part played by technology. Rapid international technological developments have become a driving force in the growth of national standards of living and in the intensification of international competition. Nations have become increasingly dependent on each other for technological information and for the technology embodied in intermediate and final products. Adding to the complexity of these developments, there has been an apparent rapid de-linking of the national locations where R&D funds are expended from the locations where the results of that R&D are embodied in production. Since 1980, these developments have drawn increasing attention as a result of the deterioration in overal us trade performance, particularly in regard to its manufactures trade. This concern was bought home even more recently by the deterioration of the us trade performance in so-called high-tech products, which dipped in 1986 into a first-time deficit of $2.6 billion.2 The magnitude of total R&D embodied in output and the corresponding measure of the technology intensity of output provide important indicators of nations' efforts to enhance living standards

13 Technology Intensity of Output and Exports

and competitiveness. These indicators allow significant insights into the extent that R&D expenditures account for international differences in competitive performance, particularly differences in technological performance across sectors, across nations, and over time. To fully account for the total R&D embodied in output, it is necessary to account not only for the R&D expended directly on that output, but also for the R&D expended upstream in the production process that produces inputs used to produce that final output. These upstream R&D expenditures often account for a large share of the total technology intensity of final goods, contributing significantly to their competitive performance. The purposes of this paper are (a) to expand methods developed earlier by this author that use input/output models to estimate total technology embodied in total output and exports of manufactures and their total technology intensity,3 (b) to provide some new estimates for the United States and initial estimates on a comparable basis for Canada and Japan, and (c) to identify the structural shifts in technology embodied in those outputs and exports and the resulting changes in their technology intensity between 1972 and 1984. With the increasing international integration of production, imported inputs have become a growing source of technology inputs to domestic production. Imports are a very large source of technology for some countries and particularly for some sectors' outputs. Thus, an additional purpose of this paper is to introduce a method for estimating the technology embodied in domestic output and exports via the technology embodied in imports and, as an example, to show the very considerable Canadian dependence on imports for technology inputs into manufactures output. Technology embodied in imports was not included in this paper's estimates for the United States and Japan, since imports are a relatively small technology source in those countries compared with Canada and the author's resources were limited. However, the technique used here for Canada is just as applicable for Japan, the United States, and other countries. Three key areas relating to recent trade-performance trends are also addressed in this paper. The estimates that follow demonstrate the extent to which us efforts to enhance the technological competitiveness of its output and exports, in the face of growing international competition, have been sufficient to raise us technology intensity. These estimates also demonstrate, on a comparable basis, the extent to which Japanese efforts — whose purpose, as announced for the past two decades, has been to raise the knowledge content of Japan's output and export - have been successful in terms of

14 Technology, Energy, International Trade

increasing technology intensity and the extent to which Japan has caught up with the efforts of the United States. The paper also shows the extent to which Canada has been able to maintain the technology intensity of its output and exports and the contribution to that effort by imported inputs. The total technology embodied in the output of a product sector is the sum of the company intramural R&D expended directly on production and upstream (indirect) R&D expended by other companies on inputs to production. The upstream technology includes R&D expenditures on domestically produced intermediate and capital goods and, for Canada, R&D expenditures embodied in imported manufactures inputs. The total of upstream technology inputs include not only those embodied in intermediate goods and services inputs used by the final producer, but also those embodied in inputs several iterations upstream. These technology inputs are econometrically estimated using total-requirement input/output matrices. The direct and indirect R&D expenditures include expenditures on products and processes. The technology embodied in output is also analysed in terms of its "total technology intensity" (hereafter frequently referred to only as TTI) for the purpose of making comparisons across product sectors, years of output, use of output to final demand (domestic use and exports), and country of production. The total technology intensity of an output is defined as the ratio of its total embodied technology to the shipments value of that output. The desirability of producing estimates of the TTI of exports on the input/output method, compared with those based only on direct technology inputs, was recognized by the OECD but was not undertaken by it because of, among other problems, the complexity of harmonizing industry classifications with the need to expand the estimate basis to account for imported inputs.4 The following estimates are restricted to comparisons for the United States and its two largest trading partners - Canada and Japan - because of the limitations of the author's research time and data availability. These estimates are based on separate input/output models of the three countries and are presented in terms of each country's national nominal currency values. The results are limited to manufactures output, which also appears for all three countries to embody the major share of their national R&D expenditures embodied in output that is readily traceable with current R&D data. The estimates are presented in terms of technology embodied in final demand rather than in gross domestic output of product sectors in order to avoid double-counting technology embodied in output

15 Technology Intensity of Output and Exports

when aggregating across sectors. Furthermore, final demand is a more suitable basis, as it is in terms of the total output, including imported inputs, that is required to meet such demand. This is important because the estimates of total technology embodied in domestic output must be augmented by the technology embodied in imported inputs. For simplicity, the technology embodied in output to final demand has been separated into the two contrasting groups: (a) that in exports and (b) that in all other output to final demand (hereafter referred to as "for domestic use"). Both are important in assessing the competitiveness of a nation's output. Other researchers have produced estimates of the technology intensity of output and exports using other indicators in addition to direct R&D expenditures, such as number of R&D employees, level of employee education, and number of patents issued. In this paper, total embodied technology is used as a proxy for the total of all major sources of technology input, including R&D employees. METHODOLOGY

This chapter discusses the relation of R&D expenditures to output in terms of several basic measures: (a) the value of technology embodied in total manufactures and individual product groups delivered to final demand, (b) their total technology intensity, (c) the share of the technology embodied in the total accounted for by individual product groups delivered to final demand, and (d) the share of the total accounted for by individual product groups as inputs to that output. The concept of using R&D expenditures to estimate technology embodied in output, and the resulting TTI, depends on the use of those expenditures on individual product groups as a component of value added and thus as an "output multiplier."5 In turn, the use of these multipliers in conjunction with the input/output methods requires the assumption of the homogeneity of the TTI of a product group's output that is used as inputs to all other product groups, as well as the homogeneity of the TTI of a product group's output to all final demands, such as exports versus domestic use. The total value of the technology embodied in an individualproduct or product-group output is the sum of the directly and indirectly embodied R&D expenditures. To estimate the total value of technology embodied in a nation's output, it is sufficient merely to sum all direct R&D expenditures, since that total includes all indirect technology inputs to all output with no double-counting. However, to estimate the technology embodied in an aggregation of products,

16 Technology, Energy, International Trade

such as manufactures, without double-counting that technology, it is necessary to estimate the technology embodied in outputs delivered to final demand. The total technology intensity of an individual product or product group is the ratio of the R&D expenditures directly or indirectly embodied in its output to the value of its output measured in terms of shipment value f.o.b. point of production. The TTI of a particular product or product-group output is the same regardless of whether the output is delivered to exports or to any other final demand. However, the average TTI for an aggregate of those products or product groups will differ depending on the final demand, since the individual TTIS are weighted by the value of shipments to that final demand. A number of important data considerations are involved in estimating the technology embodied in a nation's manufacturing output and exports and their TTI, including the considerations outlined below. Need to Account for Total Technology

As will be demonstrated later, a large share of total technology embodied in final output is incorporated upstream as the indirect inputs to production of other products. These upstream inputs include, in turn, direct and indirect inputs. All of these direct inputs include intermediate and capital inputs and also imports. Inclusion of imports is demonstrated here only for Canada. See the following figure for a simplified view of these technology flows. The importance to downstream producers of upstream technology developments in both process and products can be readily seen in the example of aircraft producers' dependence on computers for the aircraft's navigation systems, the computer producers' dependence on semiconductor products, and the latters' dependence on rare earths and silicon and on the capital goods used for slicing the silicon and making the semiconductors. Accounting for the total upstream technology inputs requires estimating the total technology requirements generated by an input/ output model. A direct input requirements table fails to capture the further upstream input requirements. Product versus Industry Data

Because firms in a given industry often differ widely as to the products they produce, R&D data on an industry basis are generally a hodge-

17

Technology Intensity of Output and Exports

Figure i Sources of Technology in Output

podge of products. Thus, R&D data on a product basis are much more useful for making consistent technology estimates across sectors. All of the data in this paper are on a product basis. The us and Japanese R&D data are directly available on a product basis. However, R&D data for Canada are reported by Statistics Canada only in the form of expenditures by industry. In this paper, these R&D data were converted to estimated R&D expenditures by product group by distributing the industry expenditures across products in the same proportion as the output is reported by industry in the official Canadian input/output industry "make" tables. This conversion assumes that the two distributions are roughly proportional, a credible assumption given the substantial concentration of most Canadian industries' output in only a few major product groups. Product Disaggregation

The degree of product disaggregation desirable for computing estimated product-sector TTI is a major issue in this type of analysis. The decision is critically dependent on the particular question the resultant data is intended to answer. Unfortunately, the level of disaggregation is primarily settled by the aggregation level of the available data. The current levels largely available in the us, Canadian,

i8 Technology, Energy, International Trade

and Japanese R&D data seem adequate for addressing broad tradeperformance issues, such as whether a nation's trade performance is being adequately supported by its R&D expenditures, and for comparing the TTI of exports with output for domestic use. In contrast, these current levels of disaggregation are entirely inadequate for determining the technology intensity of individual products at the cutting edge of technology and comparing them with other products. The identification of the technology intensity of individual products has been virtually prevented by the lack of R&D data disaggregation; their identification has thus been thrown into the laps of more nebulous criteria. Some positive statistical efforts may help overcome this problem by producing data on measures such as the life span of the principal cutting-edge innovations embodied in products. However, the greater the disaggregation, the closer the trend in TTI reflects the product life cycle of the innovation-embodying product. That trend begins at an infinitely high TTI point, with the first unit of output, and rapidly falls, declining for the rest of the product's life as R&D assets are shifted to other products - a trend approximating a rectangular hyperbola. Such trends are not suitable for broad comparisons across products or countries. To the extent that TTI is a major factor yielding comparative advantage, the greater the TTI, the greater will be the comparative advantage. Therefore, to the extent that product groups are increasingly disaggregated down to individual products, the greater will be the diversity in their technology intensity and the more that TTI can be expected to explain differences in their competitive performance. For example, better competitive performance can be expected by recently developed electronics products than by electronics as a group. Conversely, the greater the aggregation, the more that competitive performance reflects the influence of factors other than comparative technological advantage. Comparable levels of disaggregation are desirable for crosscountry comparisons of particular product-group TTIS. However, a perfect fit was not feasible for this paper. Disaggregation of available R&D data is greater for the United States than for Canada and Japan. For this reason, the cross-country comparisons here are confined to total manufactures and to the percentile distribution of that total and are not disaggregated down to individual product sectors. Technology Not Constant through Time

For any one product group, the technology of production (in terms of total-requirements coefficients) is constantly changing- and prob-

ig Technology Intensity of Output and Exports

ably not very smoothly - over short periods. It is well recognized that the technology coefficients used in input/output analysis need to reflect the year-to-year changes in the relative amounts of inputs required to produce the output of a particular product or product sector. The longer the period covered, the greater the degree of change that can be expected. Realistic analysis of technology embodied in output depends to a large degree on taking account of changes in production technology, particularly where large changes in product or process are critical competitive factors, such as in high-tech sectors. Accounting for these technological changes is even more important for exports than for output for domestic use, as the production technology for the main export sectors changes even more rapidly than output for domestic use. Where significant changes in technology are expected, it is important to base the analysis on input/output tables that account for changes in total-requirements input/output coefficients. This chapter is based on separate input/output tables for 1972 and 1984 for the United States, Canada, and Japan that reflect estimated changes in those coefficients. Domestic Total Input Requirements

Estimates of domestic technology embodied in domestic output and the corresponding technology-intensity coefficients should be based only on total-requirements coefficients tables for domestically produced goods and services inputs — the so called domestic tables. The use of only domestic total goods and services requirements is necessary in order to assign domestic R&D company intramural expenditures only to the domestic output on which they are expended, and to avoid incorrectly assigning a share of those expenditures to imported inputs. This chapter's estimates are based on domesticrequirements tables. The inputs of technology embodied in imported-goods inputs must be estimated independently and then added to those in domestic inputs. Imported Technology Inputs

For any country with significant reliance on manufactures imports as inputs, account should be taken of the technology embodied in imported-goods inputs, particularly where imported R&D is relativel large, since such imported technology is not accounted for in domestic R&D expenditures. This paper uses the case of Canada to demonstrate a method for accounting for this additional technology

20 Technology, Energy, International Trade

source. The technology embodied in total Canadian imported inputs is estimated using the TTI of exchange-rate-adjusted technology embodied in inputs imported from the United States as a proxy for the TTI of all imported-technology inputs. Estimates of technology embodied in imports into the United States and Japan were not feasible for this paper. Company Intramural R&D Expenditures For cross-product group comparisons, it is reasonable to limit coverage of R&D expenditures to those reported as expended directly on a particular product group. Most national R&D data available on this basis are variations of company intramural R&D expenditures - those for work performed within the reporting company, which includes work financed by others, including by the government. In the case of the United States, the data are slightly more restrictive, covering only applied R&D. However, company basic research is relatively small and not very directly linked to domestic output in a particular period or for specific product groups. Expenditures on research by government and other non-product—producing entities are probably even less directly linked to commercial output in a particular year. Data in Nominal Terms All data in this chapter, including input/output coefficients, are in nominal, not real, terms. Official R&D expenditures data reported by product for the United States and Japan are only reported in nominal terms. A nominal basis is sufficient for this chapter, as only total embodied technology shares and TTIS, not technology values, are compared between years. More information on the methodology used here is in the Appendix. TOTAL MANUFACTURES TECHNOLOGY

National Total Technology Embodied in Manufactures Output Analysis of the data shows that the great bulk of total R&D expenditures on manufactures output is embodied directly or indirectly in manufactures delivered to final demand by most major developed countries. For example, in 1984, 95 percent of the us company

21

Technology Intensity of Output and Exports

Table 5 Total R&D Expenditures on Manufactures Production, and Technology Embodied in Manufactures Output to Final Demand"

R&D expenditures* Embodied technologyc

United States (million US$) 1972 1984

Canada (million CAN$) 1972 1984

Japan (billion yen) 1972 1984

17,730 17,706

370 523

785 669

60,712 57,907

2,122 2,060

4,126 3,499

"Includes all types of company intramural R&D expenditures - obtained from all fund sources on manufactured products, including expenditures on capital, with the exception that for the United States excludes basic company R&D expenditures. * Expenditures in year prior to year shown. r For Canada, embodied technology includes that in imported inputs. In 1972 these technology additions far exceeded the technology embodied in manufactures used as inputs to non-manufactures final demand. Imports were not included in the us and Japanese data.

intramural applied R&D expenditures was embodied in manufactures delivered to final demand. The balance was embodied in manufactures used as intermediate and capital goods to produce non-manufactures delivered to final demand, such as services and agricultural products. (See Table 5.) Even when the share of R&D embodied in non-manufactures is very large, this dilution of domestic expenditures on R&D embodied in manufactures can be largely offset by R&D inputs embodied in imported intermediate and capital goods. Indeed, for some countries the sum of the domestic and imported technology embodied in manufactures can even be substantially greater than their total direct domestic R&D expenditures on manufactures. Such was the case for Canada in 1972, when total R&D embodied in manufactures output to final demand was 41 percent greater than domestic R&D expenditures on manufactures output due to the large R&D input addition from imports.6 Importance of Indirect Inputs of Embodied Technology Inclusion of indirect technology requirements is essential for estimating the total technology embodied in individual product groups and their aggregate in total output to exports and domestic use and their estimated technology intensity. The importance of accounting for upstream intermediate and capital inputs can be seen from a

22

Technology, Energy, International Trade

comparison of the input sources of technology embodied in output across countries and across products. Furthermore, the relative importance of indirect in comparison to direct inputs differs widely between countries due to differences in composition of output and differences between countries in R&D expenditures on that output. These cross-country comparisons are also sensitive to levels of disaggregation. In 1984 - at the levels of disaggregation used in this report - for the United States, direct inputs on average accounted for two-thirds of the technology embodied in manufactures output to final demand. Even so, indirect technology inputs averaged one-half the amount of direct inputs. (See Table 6.) For Japan in 1984, the indirect technology inputs averaged nearly three-fourths the direct inputs. This was a much larger share than for the United States, even though the Japanese data are less disaggregated. For Canada in 1984, the average indirect share actually exceeded direct by 17 percent in 1984 when imported inputs were included. The value of Canadian indirect inputs, excluding imports, was 42 percent as large as direct inputs (about the same as for the United States). In this paper, the Canadian total embodied technology estimates, because they include technology embodied in imported inputs, have an upward bias compared with those for the United States and Japan, which include no import input estimates. Conversely, the importance of indirect inputs in the us and Japanese estimates is understated. The degree of understatement depends on the relative importance of imported inputs to that of output, the composition of the inputs, and the TTI of the country of the imports' origin. This diversity in importance of imported inputs is even greater for individual products, as will be shown later in the section on Canadian technology imports. Cross-product comparisons are also sensitive to levels of disaggregation. In general, the greater the disaggregation, the greater will be the indirect share of the total. Nevertheless, even at the high level of aggregation at which the most disaggregated official R&D data are available, not only is the average indirect share large, it is much larger for some individual product sectors. For each country, the indirect shares for individual products range widely. In 1984 the indirect share of the total technology embodied in individual us manufactures classes averaged 32 percent and ranged widely, from 6 to 80 percent. For Canada, the indirect share averaged 54 percent, ranging from 18 to 87 percent. For Japan, the indirect share averaged 44 percent, ranging from 12 to 87 percent.

23 Technology Intensity of Output and Exports Table 6 Source of Total Technology Embodied in Manufactures Output to Final Demand, 1972 and 1984a (in percentages) United States Technology Source

Japan

Canada

7972

1984

1972

1984

1972

1984

100.0 65.6 34.4 29.4 5.0

100.0 67.7 32.3 28.0 4.3

100.0 57.7 42.3 18.7 4.4 19.3

100.0 46.0 54.0 17.1 2.4 34.6

100.0 51.1 48.9 38.6 10.3

100.0 55.9 44.1 37.4 6.7

100.0 68.1 31.9 28.1 3.8

100.0

100.0 54.6 46.3 18.9 4.6 21.8

100.0 45.2 54.8 15.6 2.0 37.2

100.0 50.8 49.2 40.1 9.6

100.0 56.3 43.7 37.8 5.9

100.0

100.0 67.5 32.5 28.0 4.5

100.0 60.4 39.6 18.6 4.2 17.2

100.0 47.3 52.7 19.6 3.0 30.2

100.0 48.8 51.2 38.0 10.6

100.0 55.4 44.6 37.1 7.4

TOTAL FINAL DEMAND

Total Direct Indirect total Intermediate Capital Imports EXPORTS

Total Direct Indirect total Intermediate Capital Imports

68.8 31.2 27.7

3.4

DOMESTIC DEMAND

Total Direct Indirect total Intermediate Capital Imports

67.3 34.7 29.6

5.2

" May not add due to rounding.

Total Technology Intensity of Final Demand

The estimates of embodied technology yield estimates of total technology intensity that vary widely between the three countries for any one year. However, in terms of the impact on shifts in international competitiveness and trade balances, the extent of change over time in each country's total technology intensity is probably of equal if not of greater importance than differences in the absolute TTI level. This appears particularly so for us-Japanese competition. The total 1984 manufactures technology intensity for the United States was 56 percent greater than that for Japan and nearly two and one half times (239 percent) greater than that for Canada. By themselves these large differences in TTI of manufactures output to final demand suggest very large national differences in competitiveness, standards of living, and other measures. However, this is far from the whole story. The dramatic differences between the three countries' trends in manufactures TTI suggest that the differences

24

Technology, Energy, International Trade

Table 7 Total Technology Intensity of Manufactures Output to Total Final Demands" (in percentages) Japan

United States

Canada

1972

1984

1972

1984

1972

1984

4.797 6.283 4.615

4.838 6.101 4.643

1.581 1.583 1.579

1.427 1.609 1.201

1.858 2.141 1.773

3.107 3.791 2.757

—

+1 -3 +1

—

-10

-

+2 -24

-

+ 67 + 77 + 55

FINAL DEMAND

Total Exports Domestic CHANGE, 1972-84

Total Exports Domestic

" Includes technology embodied in intermediate and capital inputs and, for Canada, imported inputs. Intensity ratios (in percent) can be expressed in cents of R&D expenditure per dollar of shipments or yen of R&D expenditures per 100 yen of shipments.

in degree of change may be much more important than differences in absolute TTI level in explaining technology's contribution to changes in relative competitiveness. Indeed, the change in Japan's TTI of total manufactures output to final demand between 1972 and 1984 was strikingly different from that in the United States and Canada. The rapid growth in Japan's TTI inescapably made a major contribution to improved Japanese competitiveness. Japan's TTI for total manufactures output to final demand was 67 percent greater in 1984 than in 1972, while the United States' TTI was not quite i percent greater in 1984 than in 1972 and Canada's was actually lower in 1984. (See Table 7.) Similar major national differences occurred in the separate technology-intensity time trends of the three countries' output to final demand for exports and for domestic use. Export Shares of Embodied Technology

The shares of total national technology embodied in total output to final demand accounted for by exports rather than domestic use rose significantly for all three countries between 1972 and 1984. Indeed, for all three countries the rise in the export share of total embodied technology outpaced the rise in the export share of total goods output. Both technology and goods shifts were particularly large for

25 Technology Intensity of Output and Exports Table 8 Shares of Manufactures Total Final Demands and Embodied Technology Accounted for by Exports and Domestic Demand (in percentages) Japan

United States

Canada

7972

1984

1972a

1984

1972

1984

100.0

100.0

100.0

100.0

FINAL DEMAND

Total Exports Domestic

13.4 86.6

45.5 54.5

55.3 44.7

100.0 23.0 77.0

100.0

10.9 89.1 100.0

100.0

100.0

100.0

100.0

100.0

14.3 85.7

16.9 83.1

45.5 54.5

26.5 73.5

41.3 58.7

33.8 66.2

EMBODIED TECHNOLOGY

Total Exports Domestic

62.4 37.6

" Similarity of the 1972 output and embodied technology shares is not an error.

Canada and Japan. However, decomposition of the data shows that for both Canada and Japan roughly two-thirds of the growth in the export share of total embodied technology was accounted for by more rapid growth of their goods exports rather than by technological growth. In contrast, for the United States the export technology share grew more slowly than the goods export share, resulting in a decline in the technology intensity of us exports. As a result of the dramatic acceleration in Japan's export TTI, the export share of technology embodied in Japan's total final demand jumped from 26 to 41 percent between 1972 and 1984. (See Table 8.) During the same period, Canada's export share rose from less than half to nearly two-thirds of its total technology embodied in its total output to final demand, mainly reflecting a rise in the export share of goods output to final demand and a fall in the TTI of output for domestic use. Technology Intensity of Exports

Assuming homogeneity of inputs within individual product sectors' output to all final demands, total technology intensity is also assumed constant across those outputs. Even so, average TTIS differ widely between those of exports and those of output to domestic use due to wide differences in product composition of manufactures in those demands. Indeed, it is safe to expect that most countries' average export TTI would be higher than the average TTI of their output to domestic uses. This is clearly reflected in the wide differences between the final demands of all three countries concerned here. In

26 Technology, Energy, International Trade

1984, the TTI of us exports was 31 percent greater than that of domestic use, 33 percent higher for Canada, and 38 percent higher for Japan. (See Table 7). The level of export TTI also differs widely across countries. For example, in 1984 the Japanese export TTI was two-thirds that of the United States, and the Canadian export TTI was one-fourth that of the United States. Again, to the extent that technology is an important national trade determinant, large changes in relative national export TTI levels appear more important to trade performance than do differences in those levels. Between 1972 and 1984, the us export TTI fell slightly and that of Canadian exports rose slightly. In sharp contrast, Japan's export TTI rose by 77 percent. The Japan export TTI rise was so large over that period that it dramatically reduced the us export TTI lead — rising from one-third of the 1972 us level to two-thirds of the 1984 us level. Japan, between 1972 and 1984, nearly doubled the extent its export TTI exceeded that of output for domestic use due to the combined effects of (a) major growth in its overall absolute TTI level and (b) growth in Japan's TTI of exports relative to that of output for domestic use. This major relative growth in export TTI lends support both to the long-announced Japanese intention to target high-tech sectors for development and export and to the rewarding results of those efforts. This dramatic acceleration in Japan's export TTI was so rapid it greatly outdistanced Japan's own rapidly growing TTI of output for domestic demand. Between 1972 and 1984, even though Japan's domestic-use TTI rose by 55 percent, its export TTI rose from 24 to 38 percent more than its TTI of output for domestic use. This superior technology performance of both Japan's export and domestic-use output contrasts sharply with us export technology performance, which remained about flat for both types of final demand. The us R&D expenditure effort barely kept pace with output for both exports and domestic use, while Japan's R&D effort relative to output nearly doubled for exports. The overall Canadian technology performance was markedly weaker than even that of the United States. For Canada in 1972, TTIS for exports and domestic use differed little, reflecting little special export technology effort. Furthermore, between 1972 and 1984 the Canadian export TTI rose only slightly, while the TTI for domestic use dropped by one-fourth between 1972 and 1984. The Canadian export TTI rose slightly only because the rise in import TTI was more than sufficient to offset an 18 percent drop in Canada's

27 Technology Intensity of Output and Exports Table 9 Canadian Total Technology Intensity Components of Manufactures Output to Final Demand, with and without Contribution from Imported Inputs (in percentages) With Imports

Without Imports

Final Demand

1972

1984

1972

1984

Total Exports Domestic

1.581 1.583 1.579

1.427 1.609 1.201

1.276 1.239 1.307

0.934 1.011 0.838

domestic technology contribution to exports. The contribution of imports to the Canadian TTI is covered in more detail below. Canadian Import Technology Contribution

Between 1972 and 1984, the Canadian economy greatly increased its reliance on imported-goods inputs to produce output for export and domestic use. This study provides an interesting opportunity to examine how this reliance on imported-goods inputs translates into the contribution of technology embodied in imports to total technology in Canadian output. Between 1972 and 1984, Canadian dependence on technology embodied in imports rose sharply. The import share of total technology embodied in manufactures final demand output nearly doubled - rising from 19 to 35 percent. (See Table 6.) This import-share growth was primarily due to three factors: (a) greater dependence on imported-goods inputs, (b) decreased technology intensity of domestic inputs, and (c) increased technology intensity of imported inputs. The relative decline in the contribution of domestic technology to Canadian output was most striking. The share of domestic technology in the TTI of total manufactures output to total final demand dropped by 27 percent between 1972 and 1984. That in exports dropped by 19 percent, and that in domestic-use output by 36 percent. (See Table 9.) The major contribution of Canadian imported technology is even more evident when the import contribution is compared with its equivalent in terms of use from domestic sources - the contribution from technology embodied in domestically produced intermediate inputs. In 1972 the contribution of technology embodied in imported intermediate inputs to manufactures output slightly exceeded

28 Technology, Energy, International Trade Table 10 Contribution to Canadian Total Technology Intensity of Domestic and Imported Intermediate Inputs to Manufactures Final Demand (in percentage) Intermediate Inputs Final Demand

Domestic

Imported

0.295 0.244

0.304 0.493

0.300 0.250

0.344 0.599

0.291

0.272 0.363

TOTAL

1972 1984 EXPORTS

1972 1984 DOMESTIC

1972 1984

0.235

that from domestic sources. By 1984 the Canadian import contribution outpaced that from domestic intermediate inputs. While the domestic intermediates' technology contribution decreased between 1972 and 1984 to both export and domestic-use output, the importshare contribution to intermediate inputs more than doubled for exports and even rose by 50 percent to domestic-use output. (See Table 10.) As a result of these shifts, between 1972 and 1984 the import share of technology embodied in Canadian output to final demand rose from 22 to 37 percent in exports and from 17 to 30 percent in output for domestic use. By definition, the import shares of total technology intensity rose by the same proportions. These results reflect a dramatic rise in Canadian dependence on imported inputs for output both to export and to domestic-use final demand. Taken along with the growing export share of total goods output, they also clearly reflect a growing international integration of the Canadian economy. Although lesser in degree, a similar development appears to have occurred in the United States and Japan. For all three countries, there was commensurate growth in the export share of total final demand for goods and the export share of total technology embodied in total final demand. The difference between countries is largely in the share of total goods output going to exports in both 1972 and 1984, with Canada exporting a much higher share than Japan, and Japan a higher share than the United States. The growth in the technology intensity of Canadian imported inputs to manufactures final demand was due both to a rise in the

2g Technology Intensity of Output and Exports Table 11 Canadian Product Sectors Ranked by Import Share of Total Technology Embodied in Canadian Exports (in percentages) Product Sector

Other transport equipment (excl. motor vehicles; incl. aircraft engines & parts Motor vehicles Instruments Radio, TV, & other home appliances Rubber products Misc. manufactures Printing products Textiles Non-electrical manufactures

Share 63

59 49 41 37 36 33 32 27

Product Sector

Metal products Other electrical machinery Stone, clay glass products Medicines Paper products Chemicals Food & beverages Iron & Steel Non-ferrous metal products Petroleum refining products

Share 26 22 21 17 14 12 12 10 7 2

technology intensity of the individual products and a shift in the composition of Canadian output. Between 1972 and 1984, the TTI of Canadian imports alone rose by 62 percent. The dependence of Canadian exports on imported technology varies very widely across products. In 1984 the estimated import contribution to total technology embodied in manufactures exports ranged from 2 percent for refined petroleum products to 63 percent for "other transport" (excluding motor vehicles and including aircraft engines and parts), with the weighted average for manufactures exports at 37 percent (Table 11). The high degree of Canadian dependence on technology embodied in imports parallels that in Canadian imports of proprietary technology information. According to Daly, about 95 percent of new Canadian patents are granted to foreigners. This parallel between Canadian information-technology imports and merchandisetechnology imports suggests a close functional relationship between them. STRUCTURAL CHANGE IN MANUFACTURES EXPORTS' TECHNOLOGY INTENSITIES

Not only should changes be expected in technology intensities of individual products, changes should also be expected in their rank-

30 Technology, Energy, International Trade

ing in terms of those intensities. These changes are largely produced by yearly relative changes in levels of R&D expenditures and shipments. The changes in R&D expenditures mainly reflect changing technologies, changes in R&D unit costs, and new innovations. Among other factors, changes in shipments reflect changes in output and demand for the products embodying those innovations. Business cycles and changes in relative national competitiveness are likely to have greater impact on such ranking in the short run, but technological change is likely to have a far greater impact over the longer run, such as in 1972-84. Changes in the product-group structure of the technology embodied in a nation's manufactures output and its technology intensity can be examined from two perspectives: (a) the total technology embodied in the individual products delivered to final demand, including that embodied indirectly in inputs from all other sectors and that in imports and (b) the total technology contributed by individual sectors directly to their own sector or indirectly as inputs to other sectors. The former emphasizes the vertical interdependence of technology and industry structure; the latter, horizontal cross-industry interdependence. For brevity, the following only covers the changes in technology embodied in the structure of exports.7 Change in Export Technology Structure

Shifts in the total technology intensity of manufactures exports are accounted for by changes in the product-sector composition and the technology intensity of the exports. The relative importance of these two factors in explaining the change between 1972 and 1984 in technology intensity of total manufactures exports of the three nations was identified by alternately holding constant the 1972 exportproduct composition and export-product technology intensities. While most of the changes in the total manufactures exports' TTIS for the United States and Japan were due to changes in the individual product sectors' TTIS, for Canada the main cause was a shift in product-share composition. United States. As pointed out above, little change occurred between 1972 and 1984 in the us technology intensity of total manufactures exports. The average TTI of manufactures exports decreased by 5.5 percent. Decomposition of the results shows that most of the decrease in TTI was due to a major decrease in the individual TTIS, while only about 16 percent of the overall decrease was due to a shift toward exports with lower individual TTIS.

31 Technology Intensity of Output and Exports

Canada. The average TTI of Canadian manufactures exports was slightly higher — 2 percent — in 1984 than in 1972. This change was produced by a shift in export composition toward higher individual TTIS that more than offset a decline in the individual TTIS. While motor vehicles account for a large share of the total technology embodied in Canadian total output (over 25 percent in 1984) and exports (nearly 40 percent in 1984), the presence of this class does not exert a strong upward bias on those intensities. Excluding motor vehicles, Canada's technology intensity of total output to final demand and exports is only 5 to 7 percent lower for either 1972 or 1984. Japan. The 77 percent rise in the average TTI of Japanese manufactures exports was a product of both large increases in the individual TTIS and a shift in the composition toward those exports with higher TTIS. About 63 percent of the rise in average TTIS was due to increased individual TTIS; about 37 percent was due to a shift in export composition. Change in Export Technology Concentration

Change in the structure of a nation's exports implied by changes in the national average technology intensity can be identified in terms of (a) changes in rank order of the technology intensity of those product sectors or (b) changes in shares of total embodied technology value accounted for by groups of product sectors comprising a given share of all export classes. Although the data for the United Sates are much more disaggregated than that for Canada and Japan and therefore product classes do not exactly match, the export structure of the three countries can be readily compared by dividing those exports ranked by technology intensity into equal thirds. Between 1972 and 1984, the rank order of the technology intensity of manufactures exports — particularly in the top roughly one-third - was relatively stable for all three countries. The same product sectors remained in the top third for all three countries in both years, except for the shift of one sector each in Canada and Japan. (See Table 12 for the list of classes in this category in 1984.) Technology embodied in exports is very concentrated in the top third of each country's product sectors. These top-third sectors could be referred to as high-tech product sectors.8 Note, however, that the estimated rankings of these product sectors differ for the United States, Canada, and Japan, as they are based on the technology intensities of each country's output. This listing is not the result of

32 Technology, Energy, International Trade Table 12

Top One-third of Product Sectors, Ranked by Technology Intensity in 1984" (in percentages) United States

1 Guided missiles and spacecraft 2 Radio, TV, communication equipment 3 Aircraft engines and parts 4 Ordinance

5 Office and computing 6 Drugs and medicines 7 Engines, turbines, and parts 8 Agricultural chemicals 9 Instruments 10 Plastics and resins

Canada

1 Radio, TV, and other home appliances 2 Other electrical machinery

Japan 1 Medicines 2 Paint, soaps, and explosives

3 Motors and generators 3 Non-electrical machinery 4 Radio, TV and other 4 Other transport home appliances equipment (excl. motor vehicles; ind. aircraft engines and parts 5 Computers and other 5 Medicines electrical machinery 6 Motor vehicles 6 Chemicals 7 Instruments 8 Chemical fibres

" Out of a total number of manufactures product classes of 30 for the United States, 19 for Canada, and 23 for Japan.

applying one country's list to the others, nor of applying a common international list to them. As can be expected for all three countries, by far the largest share of the technology embodied in exports is concentrated in the top one-third of their exports ranked by technology intensity. (See Table 13.) As also might be expected, the concentration is far higher for the United States than for Canada or Japan. In terms of export value. Over 70 percent of the technology embodied in us exports is in the top third of the exports ranked by technology intensity; the parallel figure is 56 percent for Canada and 49 percent for Japan. Perhaps not so expected for all three countries was the lack of increased concentration in the top thirds' shares of technology embodied in their total exports between 1972 and 1984. Indeed, for all three countries the top thirds' shares of embodied technology declined slightly. For the United States and Canada, this lack of

33

Technology Intensity of Output and Exports

Table 13 Percentile Distribution by Export Value of Total Technology Embodied in Manufactures Export Classes in 1972 and 1984, Ranked by Technology Intensities in Each Year (in percentages) United States Percentile

Canada

Japan

1972

1984

1972

1984

1972

1984

74 17 9

70 20 10

60 29 11

56 32 12

56 28 16

49 35 16

13.913 3.242 1.697

12.858 3.585 1.862

2.845 1.398 0.505

2.706 1.535 0.587

3.686 1.836 1.071

5.607 3.993 1.773

SHARE OF TOTAL EMBODIED TECHNOLOGY

Top third Middle third Bottom third AVERAGE TOTAL TECHNOLOGY INTENSITY

Top third Middle third Bottom third

increased concentration appears mainly due to the top-third percentiles' decreased technology intensity coupled with their greater export growth rate than that for the less technology intensive exports. Japan, in contrast to the United States and Canada, experienced the benefits of a dramatic broadening in the technology base of its exports in terms of value. While the technology concentration in the top third of Japan's exports (by value) dropped, the technology intensity of all three of Japan's export percentiles rose sharply. The rise was particularly high for the middle third — more than doubling. In terms of product classes. Little change occurred in the technology concentration of us and Canadian exports. In contrast, for Japan, although large technology-intensity increases occurred across most manufactures exports, concentration of embodied technology increased sharply in the top roughly one-third of the export-product classes. The top eight of Japan's twenty-three export classes, ranked by technology intensity in 1984, increased their share of total technology embodied in Japan's exports by over half, mainly because their export growth rate was more rapid than that for the remaining twothirds of the classes — those that were less technology intensive. The technology-intensity growth rates of both groups were nearly equal. (See Table 14.)

34 Technology, Energy, International Trade Table 14 Distribution by Number of Export Classes of Export Value of, and Total Technology Embodied in, Manufactures Exports in 1972 and 1984, Ranked in Top Third of Classes by 1984 Technology Intensity (in percentages) Canada

Japan

10 of 30

6 of 19

8 of 23

29.0 35.9

18.0 20.5

33.6 52.6

70.4 72.5

41.2 42.0

54.3 72.9

United States NUMBER OF CLASSES SHARE OF TOTAL EXPORTS 1972

1984 SHARE OF EMBODIED TECHNOLOGY 1972

1984

Structural Change in Technology Inputs to Exports The technology intensity of exports in the aggregate as well as for individual product sectors depends on both indirect as well as direct contributions of technology. The concentration of technology embodied in exports can be examined separately in terms of the total technology embodied in export classes (described above) or in terms of total direct and indirect inputs to exports (without regard to export class). As might be expected, the latter reveals just as high a technology concentration as the former in those classes with the highest technology intensities. This similarity also tends to increase with greater class aggregation, since such aggregation reduces the indirect share. Furthermore, for all three countries the ranking appeared relatively stable between 1972 and 1984 for the top third of the classes. In both years the top third of each country's total direct and indirect technology inputs to exports accounted for the major share of technology inputs to exports, with little change between years in their shares. (See Table 15.) Although the embodied-technology composition of the top third of the input classes was significantly stable, the relative importance in several of them as inputs changed substantially. As might be expected, this was particularly notable for product sectors that included electronics components. Such product sectors, as inputs to other sectors, accounted for a major rise in the concentration of technology embodied in exports for both the United States and Japan. For the

35

Technology Intensity of Output and Exports

Table 15 Percentile Distribution of Total Technology Embodied in Direct and Indirect Inputs to Manufactures Exports, Ranked by Technology Intensity in 1972 and 1984 (in percentages) United States Percentile

Canada

Japan

1972

1984

1972

1984

1972

1984

73 18 9

70 20 10

62 27 11

59 29 12

59 29 13

53 35 13

14.763 3.681 1.886

13.254 3.772 1.941

3.300 1.411 0.563

3.011 1.478 0.618

4.021 1.979 0.867

5.661 3.741 1.338

SHARE OF TOTAL EMBODIED TECHNOLOGY

Top third Middle third Bottom third AVERAGE TOTAL TECHNOLOGY INTENSITY

Top third Middle third Bottom third

United States, the communications and electronic components sector increased its share of total us technology inputs embodied in all manufactures exports - electronics and otherwise - from 15 to 24 percent between 1972 and 1984. Similarly, the broader electrical machinery sector of Japan increased its share as important technology suppliers to export output from 18 to 26 per cent. In contrast, the share of "other electrical machinery" in total Canadian technology inputs decreased slightly and that sector's technology intensity dropped by one-third, even when the rising contribution from imports was included in those technology inputs. CONCLUSION

This chapter demonstrates some advances in techniques for analysing technology embodied in, as well as technology intensity of output to exports and domestic-use final demand. Also provided are some important statistical results concerning those measures and the structural change in technology embodied in those outputs. In Terms of Methodology This chapter provides a method for comparing across countries on a comparable basis the total of the direct and indirect contribution of R&D expenditures to the estimated total technology embodied in

36 Technology, Energy, International Trade

particular final demands. A way also is provided for estimating the contribution to the technology embodied in one nation's output by technology embodied in imports from another nation. A means is also shown for making comparisons of structural change in technology embodied in output across countries, even though product classifications are not strictly comparable. This chapter also demonstrates how the use of input/output combined with R&D multipliers helps to explain changes in the technology interdependence of sectors by providing a basis for estimating and incorporating with the direct R&D expenditures the often very large upstream (indirect) R&D expenditures on capital and intermediate inputs, including those embodied in imports. This methodology serves to clarify technology's contribution to longer-term developments in international competitiveness. In Terms of Results An analysis of sources of technology embodied in manufactures output demonstrates the vital importance of including upstream (indirect) sources of technology inputs in the form of domestic intermediate, capital, and imported inputs. Upstream sources accounted for over one-third of the inputs to manufactures output by the United States and for even more by Canada and Japan. The results produced by this method offer some interesting insights into differences between the average technology intensity of manufactures output for export and that for domestic use and, inferentially, into the competitiveness of us, Canadian, and Japanese output. Some of these insights are given below. The technology estimates indicate that while the value of technology embodied in us exports rose between 1972 and 1984, the degree of increase was not sufficient to keep pace with the growth in the value of those exports. As a result, the average technology intensity of us exports decreased slightly, instead of increasing in the face of intensified foreign competition. The results also support expectations about the growth in Japan's R&D targeting of high-tech sectors to improve export performance; they also reveal the not unexpected poorer us, as compared with Japanese, technology performance in exports. The total technology intensity of Japanese exports rose by 77 percent in just twelve years. The results show that the technology intensity of Canadian exports remained flat between 1972 and 1984. In addition, there was a major rise in Canadian dependence on technology inputs for both Cana-

37 Technology Intensity of Output and Exports

dian manufactures output to exports and that to domestic use. The import share of technology embodied in output nearly doubled between 1972 and 1984, mainly reflecting an absolute contribution decrease from domestic sources and an increase from imports. Finally, a number of interesting results emerged about structural change between 1972 and 1984 in the composition of technology embodied in manufactures export classes ranked by technology intensity. In terms of the top third of exports by value, for all three countries the concentration of technology in that third decreased. In terms of the top third of the classes, for the United States and Canada the concentration increased only slightly. However, for Japan, this concentration rose sharply, mainly because exports in classes with higher technology intensity rose far more rapidly than others. In Terms of the Data A number of advances in available source data are needed to improve the type of estimates made in this discussion. Much of the improvement depends, however, on overcoming the limitations imposed by the lack of adequate data disaggregation. Greater disaggregation, along with improved consistency of R&D product-class definitions across countries, would make possible a more direct comparison of individual product-sector technology intensities across countries and, therefore, much improved analysis of structural change in the composition of technology embodied in output. Greater disaggregation would also help minimize the defect inherent in assuming homogeneous technology intensity for particular product sectors across output uses — making it clear that the total technology intensity for output to exports can be expected to differ significantly from that for output to products for domestic use. Also needed from official sources are data on product rather than on industry R&D where such is unavailable, as is the case, currently, in Canada. APPENDIX

Methodology Total technology intensities, the sum of the direct and indirect intensities, can be estimated through the use of input/output techniques. The total of

38 Technology, Energy, International Trade all domestic inputs required to produce each group of products provides a basis for estimating the total requirements for the applied technology embodied in those products. Direct technology intensity can be denned as the ratio of direct R&D expenditures by companies on each group of products. Applying those direct domestic technology intensity ratios (coefficients) to total domestic input/output requirements coefficients yields estimates of total technology requirements coefficients for each input in the output of those products. All of the official input/output tables published by the United States, Canada, and Japan are on a total-requirements basis. Conversion of these tables to a domestic economy requirements basis was necessary before the technology data in this report could be estimated. The INFORUM models used in this paper are all on that basis. The sum of the total technology input requirements coefficients for each input yields the estimated total technology intensity coefficient for each output, that is, the sum of the R&D embodied in each input required to produce each product output. This procedure creates an input/output model of the total R&D requirements of us manufacturing in terms of the total cost of applied R&D per dollar of output. For each output the indirectly required R&D is the total R&D required minus the direct R&D required. For this study, the indirect R&D required is augmented by including in the total the estimated total input-output requirements for capital goods used to produce each output. In addition, for the Canadian technology estimates, the total requirements are based on estimated requirements for only domestically produced inputs and then augmented with the estimated technology embodied in imported goods for which the technology intensities are based on exchange-rate— adjusted us export intensities as a proxy for those for all Canadian imports. To reflect changes in technology, individual input/output coefficients estimated for individual years in the INFORUM models are used to estimate the technology coefficients and values for that year. To reflect the estimated average lag between applied research and development expenditures and the bulk of the production embodying them, the technology embodied in output for a given year is estimated on the basis of R&D expenditures in the prior year. All of the input/output coefficients, output and export data, R&D expenditures data, and resultant technology estimates in this paper are in nominal, no. real, terms. Symbolically, the total embodied technology coefficient for the output of product sectors^ is

39 Technology Intensity of Output and Exports where TTICj

DTICi

TDRCy

is the total (direct plus indirect) embodied technology intensity coefficient for each output, in terms of total cents of embodied technology required in each input ij to produce a dollar of output (shipments) of product^, including as indirect inputs technology embodied in intermediate, capital, and, in the case of Canada, imported inputs. is the direct technology intensity coefficient for each input i required, in terms of cents of R&D expended directly by the manufacturer per dollar of product i shipments. is the total domestic requirements coefficient for inputs ij to produce product j, in terms of cents in input i per dollar of output j, including as indirect inputs the intermediate, capital, in the case of Canada, imported inputs.

In the case of Canada, the total technology coefficient for the output of product sectors j is based on the total domestic technology requirements coefficients plus the import technology requirements coefficients and is TTICj = ^[(DTICi'TDRCtj)

+ (TTIC^:• MRCg)]

where TTIC^j

MRCy

is from the us technology intensity estimates and is used as proxies for total Canadian imports with the us intensities adjusted by the exchange rate from a us-to a Canadian-dollar basis, and is the Canadian import requirements coefficient for inputs.

The exchange rates used to convert us technology intensities, in terms of Canadian dollars per us dollar, are $0.9899 for 1972 and $1.2951 for 1984. The technology intensity of capital inputs is estimated on a similar basis as for intermediate inputs, except that the degree of disaggregation is less. Estimates of physical capital input requirements, on which the estimated technology embodied in capital are based, are based on the INFORUM estimates of capital purchased rather than capital consumed.

40 Technology, Energy, International Trade The value of total technology TTVj of a product sector embodied in a final demand FDj (exports or other uses) for that sector is, for an individual product sector, TTVj = (FDj • TTICj) , and for the total of all product sectors,

TTV = 2,- (FDj • TTICj) . Comparability of the Data A major problem in this type of study is the concording of the raw data. In general, the comparability of the aggregate R&D data used for the United States, Canada, and Japan is fairly close. The us data cover company intramural expenditures, including those on capital, by industry from all funding sources, for applied R&D on manufactured products. The exception is that the Canadian and Japanese data also cover company intramural basic research expenditures on manufactures; however, those basic expenditures account for only a small additional proportion, as they do not cover basic research by government and others and only cover basic research on manufactures. For all three countries, the disaggregated data used are R&D expenditures allocated by product sector, not by individual industry regardless of product. For each country, the TTI estimates are based on the same number of sectors in 1972 and 1984. The number of product sectors differed between countries: thirty us sectors, twenty-three Japanese sectors, and nineteen Canadian sectors. These numbers reflect the maximum disaggregation possible given the inconsistency in classifications for R&D data, producer shipments, and input/output sectors in the INFORUM models. With respect to R&D expenditures on capital, the us and Japanese data include the cost of capital used in the form of charges for depreciation. The data for Canada include capital in the form of capital expenditures. The latter are probably a rough proxy when averaged across companies in a large R&D product group for what might have been their R&D depreciation charges. Because of the erosion of data cells reported by the us National Science Foundation, recent years' data for several us sectors are estimated on the basis of prior years' data and other sources. A major us sector estimated for 1984 is motor vehicles, the data for which are based mainly on data reported by the Motor Vehicle Manufacturers Association of the United States.

41 Technology Intensity of Output and Exports All of the Canadian R&D product data for 1972 and 1984 are estimated by the author using the respective year Canadian input/output "make" tables to convert industry to product R&D data. All of the Canadian TTI coefficients for imports are based on the us export TTI coefficients, which assumes that due to the large share of Canadian manufactures imports in both years supplied by the United States, the us export TTI is an adequate proxy for the TTI of all Canadian imports. The R&D expenditures and producer shipment data were obtained from official national sources in national currencies. The input/output coefficients are those residing in the INFORUM national models of the Interindustry Economic Research Fund, University of Maryland. The us data are calculated using the 484-sector us model; the Japanese and Canadian data are calculated using 84-sector models. Where coverage of output data sectors and R&D expenditure data sectors overlap homogeneity of coefficients is assumed. Also assumed is the proportionality of embodied R&D (R&D expenditures) in the value added for a given type of input, per dollar of input, across all using-product and final demands.

NOTES

1 The input/output computations for this paper were done by Dr Douglas Nyhus using the INFORUM us, Canada, and Japan input/output models of the Interindustry Economic Research Fund at the University of Maryland. The author is also indebted to Judith Schulz for her assistance in producing the report tables; to Humphrey Stead at Statistics Canada and to Jennifer Bond and Melissa Pollack at the National Science Foundation for supplying R&D data; and to Kazuyuki Matsumoto of the Japan Development Bank for providing Japanese input-output data. 2 These high-tech data are reported by the us Department of Commerce using its so-called 0003 definition of high-tech product groups. For the initial report on this definition, see Davis 1982. For more recent discussion of the us high-tech trade performance on this basis through 1985, see us Department of Commerce 1986. 3 See Davis 1982 for the first use of input/output for estimating total technology intensity of us output and exports, subsequently reprinted in OECD 1983. 4 For more on this OECD discussion, see OECD 1985. 5 For more on the use of such multipliers, see O'Conor and Henry 1975 and Miernyk 1965. 6 In 1984 the Canadian import addition was also very large compared

42

Technology, Energy, International Trade

with domestic R&D expenditures, but that import addition was exceeded by the large share of the domestic R&D expenditures embodied in output to final demand for non-manufactures. 7 For a more extensive use of input/output models in the examination of industry's vertical integration and horizontal diversification of embodied technology, using Japan as the example, see Kodama 1986. 8 For the United States, they are the same as those frequently referred to as "high tech" in us Department of Commerce publications using the so-called 000-3 high-tech definition. See, for example, us Department of Commerce 1986.

BIBLIOGRAPHY

Data Sources* UNITED STATES

National Science Foundation. Research and Development in Industry, various issues. Motor Vehicle Manufacturers Association of the United States. Information Handbook, Financial Characteristics of Motor Vehicle and Equipment Manufacturers. Washington, D.C., April 1986. us Department of Commerce, Office of Industry Assessment, Industry Statistics Division, "us Trade Data 1972—1985" us trade by Standard Industrial Classification, unpublished computer printout, December 18, 1986. CANADA

Statistics Canada. Industrial Research and Development Statistics, various issues. Statistics Canada. Science and Technology Indicators, various issues. Statistics Canada. Science Statistics, Service Bulletin, various issues. Statistics Canada. The Input-Output Structure of the Canadian Economy, various issues. JAPAN Statistics Bureau. Management and Coordination Agency, Japan. Report on the Survey of Research and Development, various issues. Economic Planning Agency. Government of Japan. Annual Report on National Accounts, various issues.

*In addition to data supplied by Interindustry Economic Research Fund.

43 Technology Intensity of Output and Exports References Daly, J.D. 1985. "Technology Transfer and Canada's Competitive Performance." In Current Issues in Trade and Investment in Service Industries: U.S.-Canadian Perspectives, edited by Robert Stern, 304—3. Toronto: Ontario Economic Council. Davis, Lester A. 1982. Technology Intensity of U.S. Output and Trade. Staff report. Washington, DC: Office of Trade and Investment Analysis, International Trade Administration, us Department of Commerce, July. Kodama, Fumio. 1986. "Technology Diversification of Japanese Industry." Science 233, no. 4761 (July 18). Miernyk, William H. 1965. The Elements of Input-Output Analysis. New York: Random House, Inc. O'Conor, Robert, and Edmund W. Henry. 1975. Input-Output Analysis and Its Implications. Monograph no. 36, Griffin's Statistical Monographs and Courses. New York: Hafner Press. Organization for Economic Co-operation and Development (OECD). 1985. "Trade in High Technology Products: An Initial Contribution to the Statistical Analysis of Trade Problems in High Technology Products" (DSTI/SPR 84.66, DSTI/IND 84.60). Paris: OECD, January 30. — 1983. "Workshop on Technology Indicators and the Measurement of Performance in International Trade" (DSTI/SPR/83.63). Paris: OECD, August 18. us Department of Commerce, International Trade Administration. 1986. U.S. Trade Performance and Outlook in 1985. Washington, D.C.: us Government Printing Office, October.

CHAPTER THREE

Canadian Industrial Energy Consumption and External Trade K.E. HAMILTON

Canadian exports (dominated by raw and semi-finished goods) were significantly more energy intensive than imports (dominated by manufactured goods) in the 19705 — the net surplus of export energy requirements over import energy requirements totalled 9 percent of domestic use of energy in 1971. Exports of goods and services were 26 percent more energy intensive per dollar than imports in 1976. Canadian intensiveness of energy use, as measured by the aggregate energy gross domestic product (GDP) ratio, changed little from 1971 to 1976. Decomposition of commercial energy requirements per dollar of GDP reveals that final consumption patterns and direct energy use coefficients (the use of energy per constant dollar of production in each commercial sector) moved in the direction of increased energy intensity between 1971 and 1976; these effects were largely counteracted by shifts in the pattern of trade. Canadian energy prices were significantly lower than world prices prior to 1983; there is moderate evidence that this permitted the most energy-intensive secondary sectors to substitute for imports in the domestic market. INTRODUCTION

The hypothesis that the factor endowment of a country will be reflected in its trade (made precise in the Heckscher-Ohlin theorem) was explored empirically by Wassily Leontief (1966) in 1953, with surprising results. In what has come to be known as the "Leontief paradox," his input/output (I/O) analysis of the factor content of American trade revealed that the United States exported goods and services that were more labour intensive and less capital intensive than its imports. Work since that date has focused on the quality of the factors measured (e.g., labour skills) and on different kinds of

45 Industrial Energy Consumption and Trade

factors. The study by Postner (1975), for instance, looked at renewable and non-renewable resource requirements in Canadian trade. The present study examines a resource of great importance in the 19705 and 19805, primary energy. The scale and structure of external trade in goods and services have profound effects on the overall energy intensiveness of any economy. This study is concerned with a comparison of the energy required to produce the bill of goods that Canada exports with the energy savings (or, more accurately, foregone energy requirements) associated with the import bill of goods. There is an intrinsic interest in exploring the extent to which Canada has tended to be a net exporter of energy "embodied" in its goods trade, but there are practical implications as well, particularly with respect to Canada's dealings with the International Energy Agency and to the relationship between Canadian energy policy and industrial development policy. Equally worth investigating are the structural changes, especially in energy consumption patterns, that occurred before and after the oil shock of 1973. To examine this, the present study will look at shifts in the intensity of energy use in the commercial sector between 1971 and 1976, and will decompose these shifts in an effort to compare effects due to trade with other structural changes. It is also worth examining the effects that energy pricing policy may have had on external trade performance, given that Canadian crude oil prices were lower than world prices through the 19705. Both export growth rates and rates of import substitution will be measured. Table 16 provides the necessary background for what follows. In it we see the structure of production and trade for primary energy commodities in 1971 and 1976. The notion of "primary equivalent" employed in reference to primary energy products is one of standard international practice; secondary energy products (such as gasoline and fuel oil) and all electricity are counted as the amount of thermal energy that would be required for their production. "Hydroelectricity" includes nuclear generation in this and the following tables. Aggregate domestic product and trade figures are shown in Table 17, as well as energy indicators such as the energy/GDP ratio, the self-reliance ratio (the ratio of total primary energy production to domestic use), and the trade ratio (the average of exports and imports in proportion to domestic use). These tables display two static elements in the aggregate energy picture: (a) the shares of fuels in domestic energy supply and (b) the energy intensity (as measured by the energy/GDP ratio). The trade-

46 Technology, Energy, International Trade Table 16 Primary Energy Production and Trade 7977 Petajoules Primary Equipment

Coal Nat. gas Hydro-el. Crude oil PLPGa

Total

Share of (DS) (%)

Production

Exports

Imports

Domestic Supply (DS)

405 2404 1732 3298

205 961 76 1670 97 3009

478 15 35 1495 0 2023

678 1458 1691 3123 66 7016

100.0

163

8002

9.7 20.8 24.1 44.5

0.9

1976 Petajoules Primary Equivalent

Coal Nat. gas Hydro-el. Crude oil PLPG" Total

Production

Exports

Imports

Domestic Supply (DS)

Share of (DS) (%)

620 2866 2406 3237 242 9371

343 1006 134 1058 182 2723

439 4 38 1616 0 2097

716 1864 2310 3795 60 8745

8.2 21.3 26.4 43.4 0.7 100.0

" PLPG — primary liquefied petroleum gases. Source: Statistics Canada 1976.

related indicators show more movement: both self-reliance and the proportion of trade to domestic use declined significantly from 1971 to 1976. We see, therefore, both a decline in the surplus of Canadian production over domestic needs and a reduction in the relative size of trade in energy products in comparison with the domestic market. The object of the analysis that follows is to dig beneath the aggregate production and trade figures to reveal how external trade in goods and services influences overall energy requirements, and to examine how these and other influences changed from 1971 to 1976. The methodological basis of the study is input/output energy analysis, first described by Herendeen (1973) and more fully expounded by Flaschel (1982). Working papers produced at Statistics Canada presented Canadian experience in the methodology (McInnis and Hamilton 1975) and its application, including an energy analysis of external trade using 1966 data (Hamilton 1977).

47 Industrial Energy Consumption and Trade Table 17 Aggregate Economic and Energy Indicators Millions of 1971 Dollars

GDP

Exports re-exports

1971

1976

95,635 20,668 18,993

120,113 24,616 27,645

Energy Indicators

Primary energy/ GDP (megajoules/1971 $) Self-reliance ratio Trade ratio

1971

1976

73.4

72.8

1.14 0.36

1.07 0.28

Note: Self-reliance is the ratio of primary energy production to domestic use of primary energy. The trade ratio is the average of exports and imports of primary energy in proportion to domestic use.

The technical details of I/O energy analysis are reserved for the Appendix. All dollar figures in this study are constant 1971 dollars. THE ENERGY ANALYSIS OF EXTERNAL TRADE

Where Table 16 displays the trade in energy products, we now turn our attention to the energy required to produce the goods Canada trades. This will shed new light on Canadian external trade. The power of input/output -based energy analysis is that it permits the estimation of direct and indirect use of energy to produce any of the spectrum of goods and services Canadians consume. As an example, I/O energy analysis permits measurement of both the energy used on the production line in the manufacture of an automobile (the direct energy) and the energy required to produce the steel, glass, plastic, rubber, and so on, of which the auto is constituted (the indirect energy). The total direct and indirect primary energy needed to produce a given product will be referred to as the energy requirements of that product. To estimate the net effect of external trade on Canadian energy consumption, two I/O simulations were carried out, one in which total final consumption, including exports and imports, was main-

48 Technology, Energy, International Trade Table 18 Canada's Total Primary Energy Requirements with and without External Trade Petajoules

With trade No trade % difference0

1971

1976

4607 3947

5860 5449

-14.3%

-7.0%

" Percent difference is with respect to the "with trade" figure in each year.

tained at historical levels, and the second in which there was the same final consumption but exports and imports were set at zero. The "without trade" variant, therefore, simulates a hypothetical closed Canadian economy. In both simulations what was being measured was the energy requirement of total final demand (i.e., direct and indirect energy used to produce the goods and services) rather than the amount of energy actually exported, imported, or consumed by households or governments. These calculations are summarized for 1971 and 1976 in Table 18. We can conclude from Table 18 that Canada's pattern of trade entailed increases in energy requirements in both 1971 and 1976, with the effect being more pronounced in 1971; this is equivalent to saying that export energy requirements were greater than the energy requirements foregone in imported goods. The earlier working paper by Hamilton (1977) reached similar conclusions about the dominance of export energy requirements, although the methodology was different in its detail. Hamilton's paper showed that in 1966 the surplus of export energy requirements over imports amounted to 5.2 percent of domestic use, or 8.5 percent of commercial energy usage - the corresponding 1971 figures derived from Tables 16 and 18 are 9 percent and 14.3 percent. Further light can be shed on these results through a study of the energy intensiveness of exports and imports. We define the energy intensiveness of a bill of goods and services to be the direct and indirect energy required by the commercial sector (i.e., all primary, secondary, and service industry) to deliver one dollar of the product to a final user. Table 19 compares the energy intensiveness of exports and imports. To give a common basis for the comparison, imports are

49 Industrial Energy Consumption and Trade Table 19 Energy Intensiveness of Exports and Imports (megajoules per 1971 dollar)

Exports Imports" % difference*

7977

7976

121.8 94.5

119.8 88.3

-22.4%

-26.3%

" Imports are competitive imports less re-exports. * Percentage difference is with respect to the export figure in each year.

restricted to competitive imports (i.e., those for which there is a competing Canadian producer). This table shows export energy intensity as nearly constant between 1971 and 1976 at about 120 megajoules per 1971 dollar. Import energy intensity declined substantially over these years but remained from 22 percent to 26 percent less than export energy intensiveness. The apparent discrepancy between tables 18 and 19 can be explained with reference to Table 17: from 1971 to 1976 the current-account trade balance in constant dollars changed from a state of modest surplus to one of more substantial deficit. A word needs to be said about the methodology used in measuring import energy intensiveness. Tables 18 and 19 show the result of measuring the energy requirements of imports as if they were produced in Canada with Canadian technology. It is arguable that this overstates the "actual" energy required to produce the goods Canada imports, since Canada's trading partners all experienced sharper energy price increases than Canada did in the 19705. However, it is also arguable that measuring imports according to domestic energy requirements foregone is the only meaningful way to calculate import energy "savings" that are comparable to export energy "costs" and, moreover, that the only relevant savings are those of energy that would be required if production took place in Canada. The reason exports exceed imports in energy intensiveness is not hard to find — Canadian exports are weighted toward raw and semifinished goods, Canadian imports toward manufactures. The two examples in Table 20 are typical: for both paper products and iron and steel products the energy intensity per dollar reaches a peak at the semi-finished stage and declines steeply for the finished product. Processes of extraction, separation, and concentration dominate the production chain up to the point of semi-finished products, and in these stages energy inputs are high, value added low. Beyond the semi-finished state (i.e., through fabrication and assembly), the en-

50 Technology, Energy, International Trade Table 20 Selected Energy Intensities, 1976 (megajoules per 1971 $ Pulpwood & chips Pulp Paper end products

62.4 259.0 105.9

Iron ores & concentrates Pig iron & steel ingots Pipe & sheet metal

198.6 232.3 118.6

ergy inputs are low, value added high. The initial rise and subsequent decline of the energy intensities by the production stage seen in Table 20 are the result. However, it would require a more detailed investigation than that undertaken in the current study to prove this in detail. Having noted the decrease in import energy intensiveness between 1971 and 1976, one might ask to what extent this decrease was owing to changes in techniques of production (and in particular the energyusing characteristics of these techniques) as compared with changes in the mix of goods and services imported. As outlined in the Appendix, I/O energy analysis permits the decomposition of the difference in import energy intensiveness into two terms representing precisely these effects. The results of this decomposition can be seen in Table 21. The "energy intensity term" in Table 21 measures the effects of changes in production techniques, including energy utilization, and shows a minor increase in the energy intensiveness of the average import. The effects owing to shifts in the composition of imports, however, are more pronounced and negative. We see, therefore, that the decrease in energy intensiveness of imports from 1971 to 1976 was almost entirely owing to changes in the mix of imports. STRUCTURAL CHANGES IN ENERGY REQUIREMENTS

The preceding section examined the energy requirements of exports and imports and how these changed from 1971 to 1976. This raises the question of what other changes occurred in the patterns of energy consumption between these two years and how the effects owing to trade compared with them. The starting point in this analysis is to compare each year's total commercial energy use per constant dollar of GDP with that in the other years - dividing by GDP controls for the effects of economic growth. With the use of techniques similar to those employed in the analysis of changes in import energy intensiveness, it is possible to

51 Industrial Energy Consumption and Trade Table 21 Import Energy Intensity Change Decomposition (megajoules per 1971 dollar) 1971 intensity 1976 intensity Difference

94.5 88.3 —6.2

Energy intensity term Import composition term

0.8 - 7.0

decompose the difference in commercial energy use per dollar of GDP between 1971 and 1976 into a series of discrete terms. The terms highlighted here are technology, final demand, imports (both scale and mix), exports (scale and mix), and direct energy-use coefficients. These terms require some explanation: • Technology: this term combines the effects of changing input structures and market shares of all commercial sectors. • Final demand: this term combines the shifts in the quantity and mix of goods and services going to households, governments, investment, and inventory accumulation. • Imports: separate terms for the changing scale of imports (their proportion to GDP) and changes in the mix of imports are employed. • Exports: as for imports, separate terms for scale and composition are used. • Direct energy-use coefficients: a fundamental determinant of energy intensiveness is the quantity of energy consumed per constant dollar of production in each sector. The aggregate effect of changes in these coefficients is measured by this term. Table 22 displays the change in commercial energy use per dollar of GDP between 1971 and 1976 and its decomposition. Negative figures in this table indicate the tendency of the terms to decrease the energy intensity of the commercial sector between 1971 and 1976. Note that the change in the aggregate is negligible, 0.6 percent. However, there are some significant changes in individual components, especially final demand, import scale, and direct energy-use coefficients. Noting that the combined effect of export scale and mix is also significant, we can conclude from this decomposition that there were strong tendencies for the structure of final demand and direct energy use per unit of output in the commercial sector to increase energy

52 Technology, Energy, International Trade Table 22 Decomposition of Commercial Energy Use per Dollar of GDP (megajoules per 1971 dollar) 1971 primary energy requirements 1976 primary energy requirements Change (1976-1971)

48.2 48.4 0.3

(0.6%)

Changes in commercial energy use per dollar of GDP owing to Technology -1.7 Final demand 3.0 Import mix 1.3 Import scale h-3.00 Export mix -1.6 Export scale -1.5 Direct energy coeff. 3.7

(3.6%) (6.2%) (2.6%) (6.2%) (3.2%) (3.0%) (7.7%)

Selected aggregates of these terms Demand, trade, & tech. Demand & trade Imports Exports

(7.2%) (3.6%) (3.6%) (6.2%)

-3.55 - 1.8 -1.7 -3.00

Note: Percentages are absolute values of proportion of 1976 commercial primary energy use.

intensiveness between 1971 and 1976; structural change in external trade largely cancelled these effects. ENERGY INTENSITY AND ECONOMIC PERFORMANCE

The most striking feature of Canadian energy policy in the 19705 was the establishment of a domestic crude oil price significantly lower than the international price. Throughout this decade, oil prices set the pace for other energy prices (the analysis in this section is derived from Hamilton 1988). The extent of the difference in crude oil prices is shown in Table 23. This table displays both nominal prices for crude oil delivered to Montreal and the index of prices relative to GDP. The divergence in nominal prices is marked: by 1980 the international price was over twice the Canadian price. The relative price index for foreign crude took two sharp jumps, in 1974 and 1979, but was fairly constant between those years; for Canadian crude the jumps in relative price were more moderate, while the index increased smoothly during the mid-1970s. The two prices were roughly equal by 1983.

53 Industrial Energy Consumption and Trade Table 23 Domestic and International Oil Prices Nominal Prices ($ per cubic metre)

Relative Price Index (1971-100)

Year

International

Canadian

International

Canadian

1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

19.38 19.33 25.17 69.54 78.73 80.38 93.86 101.14 141.02 233.18 268.67 259.98 229.97 241.03 246.03

22.21 22.21 26.18 40.65 50.59 58.52 69.66 82.99 92.37 114.24 175.82 212.47 229.70 231.52 242.62

87 82 98 238 245 230 253 257 326 487 507 451 380 386 383

100 94 102 139 157 167 187 211 213 238 331 368 379 371 377

Note: The relative price index is derived by dividing nominal prices by the GDP deflator and converting to an index based on the Canadian price. Source: Energy, Mines and Resources Canada 1988.

The economic consequences of this pricing policy should have been fairly obvious: energy-intensive Canadian exports would have enjoyed a price advantage on international markets, while energyintensive imports would have been at a disadvantage in the domestic market. To examine the evidence for the consequences of this policy, tests for correlation of energy intensiveness of industrial sectors against external trade performance were performed for the period 1971—85. The sectoral energy intensities employed were those for 1971.

Table 24 shows the result of two tests of correlation : (a) did more energy intensive sectors exhibit higher export growth rates than less energy intensive sectors? and (b) did domestic sectors succeed in substituting for imports, decreasing the import share in their respective domestic markets, according to their energy intensiveness? The table distinguishes between primary and semi-finished goods sectors and secondary sectors. Energy producers and service sectors are excluded. As seen in Table 24, there is only weak evidence of stronger export growth for more energy intensive primary sectors over the period

54 Technology, Energy, International Trade Table 24 Correlations with Sectoral Energy Intensity Primary and SemiFinished Goods

Export growth rate Import share growth rate

Secondary Goods

0.27

-0.03

-0.42

-0.19

in question, and virtually no evidence of the same correlation for secondary sectors. There is, again, a weak indication that secondary sectors substituted for imports according to their energy intensiveness. The larger negative correlation of primary and semi-finished goods sectors with import shares is likely spurious, being based on only six sectors with significant trends in import shares — in any case, import penetration into these markets was quite small, since energy sectors were excluded. Table 25 shows the results of examining subsets of the secondary goods—producing sectors for correlations of energy intensiveness and import substitution. Rather than correlating all sectors, two subgroups, those with average import penetration greater than 10 percent and those with average import penetration greater than 20 percent (the median import share for all sectors was 18 percent, were tested for the period 1971-85. There is moderate support from Table 25 for the notion that, in sectors where import penetration was significant, lower Canadian energy prices conferred an advantage on secondary sectors that enabled them to supplant imports during the 19708 and early 19805. CONCLUSION

Canada is energy rich and traditionally an energy exporter. This study reveals the degree to which this reality extends to Canada's role as a significant net exporter of indirect energy through its structure of trade. In 1971 energy use inherent in Canadian trade patterns was over 14 percent of that in commercial energy use, or over 9 percent of that in total domestic use. This is a consequence of exports being significantly more energy intensive per dollar than imports, amounting to some 26 percent in 1976. Although there were shifts in the components making up aggregate energy use per dollar of GDP in the commercial sector, the

55 Industrial Energy Consumption and Trade Table 25 Secondary Sectors with High Import Competition Import Share

Correlation against Energy Intensity -0.30 -0.46

> 10% >20%

aggregate changed very little from 1971 to 1976. There are several plausible explanations for this: (a) the energy-using stock items in the economy have long lifetimes (e.g., industrial boilers, private automobiles) ; (b) the development of new energy-conserving technologies takes time (e.g., designing and manufacturing efficient furnaces); or (c) oil prices in Canada, compared with other prices measured relative to GDP, increased smoothly and slowly compared with world prices. Prior to 1983 Canadian oil prices were significantly lower than world prices. This had no apparent impact on export performance. For producers of secondary products, particularly those that faced serious import competition, there is moderate evidence that these sectors successfully increased their domestic market share against imports. APPENDIX

IIO-Based Energy Analysis The Canadian input/output system distinguishes 191 industries and 602 commodities in a rectangular accounting framework. To construct a working model, two assumptions are invoked: (a) industry technology, wherein the inputs of commodities to each industry are held to be fixed in proportion to total industry output, and (b) fixed market shares, whereby the proportion of the total market for a given commodity is assumed to be constant for each producing sector. By convention, the technology matrix is denoted B (and is commodity by industry in dimension); the market share matrix (industry by commodity) is denoted D. For vectors of industry production g (i.e., the sum of all commodities produced by each sector), commodity production q, exports x, imports less re-exports m, and other final demand/, the following hold: q = Bg+f+x-m

(i)

g = Dq.

(2)

56 Technology, Energy, International Trade The first identity says that commodity production is the sum of intermediate demand, final demand (including consumers expenditure, government expenditure, investment, and inventory accumulation), and exports, less imports. The second identity simply states the market-share assumption mathematically. For our purposes, B and D are dimensionless coefficients, and all other vectors are measured in dollars. On this basis, we can determine industry output by substituting (i) into (2): g

= (I-DB)-lD(f+x-m)

.

(3)

With e defined as a row vector of total direct primary-energy equivalents consumed per dollar of output in each commercial sector, energy intensities E (direct plus indirect energy requirements per dollar of demand of each commodity) may be expressed as: E = e (I-DB)-1D .

(4)

Complete specification of the construction of the Canadian I/O tables is given in Statistics Canada 1977. The methodology for constructing the direct energy requirement coefficients is outlined in Deachman and Hamilton 1978. The quality of these coefficients is enhanced by the availability of important energy data on physical quantity in the Canadian statistical system (i.e., for agriculture, mining, manufacturing, utilities, and transportation), eliminating the need to estimate physical quantities. Given the preceding expression for energy intensities, the energy requirements to produce any bill of goods (measured in dollars) can be obtained by taking the inner product of E and the bill of goods. According to this definition, energy requirements are the measure of energy consumption in the commercial sector; in national accounting terms, they represent the intermediate consumption of energy. We should note from expressions (3) and (4) that the technology terms (I-DB)'1D are the equivalent of the Leontief inverse and that they measure the inter-industry relationships with respect to a "pure" Canadian technology, with no leakages of intermediate imports. The net effect of external trade on commercial energy requirements is measured as the difference between the following two expressions: tl = e (I-DB)-1D (f+x-m)

(5)

t2 = e(I-DB)-lDf.

(6)

Expression (5) represents the energy required to deliver demand/to final users combined with exports of goods x and imports of goods m - in this

57 Industrial Energy Consumption and Trade regard goods imports may be viewed as foregone energy requirements. Expression (6) calculates the energy required to deliver the same demand / to final users, but in the absence of trade. The difference in import intensiveness can be decomposed into effects owing to changing commodity energy intensities and changing patterns of imports. Let the superscript * indicate 1971; arrays without this superscript represent 1976 figures. The change in import intensiveness can be represented as

M-M* = Em - E* m* = (E-E*)m + E*(m-m*) .

(7)

Here M is the energy intensity of imports for 1976, and m the pattern or composition vector (i.e., competitive imports less re-exports normalized to $1). The decomposition of aggregate measures employed in expression (7) can be used effectively in examining the components of total commercial energy use per dollar of GDP and how these changed from 1971 to 1976. In what follows, we assume that/, x, m, and g are all measured in constant dollars per constant dollar of GDP. Let S be the total commercial energy required per dollar of GDP, and for convenience of exposition, let R = (I-DB}-1D, d = f+x-m . If, as previously, the superscript * represents values for 1971, then S—S* = ego — e* op~* = e (g-g*) + (e-e*)g* .

(8)

The first term in (8) represents changes in commercial energy intensiveness owing to the structure of demand and production; the second represents changes owing to the structure of commercial energy use, that is, in the direct energy-use coefficients e of expression (4). This expression can be further decomposed as follows:

g-g* = Rd-R* d* = (R-R*)d + R* (d-d*) .

(9)

Changes in commercial energy intensiveness owing to changes in technology (or, more precisely, changes in input coefficients and market shares) are

58 Technology, Energy, International Trade captured by the first term in (9), while the second represents changes owing to the structure of demand. This latter term decomposes in an obvious way into individual terms representing changes in the structure of final demand, exports, and imports. For exports and imports, it is worth detailing the differences owing to scale (i.e., the proportion of GDP) and to composition. If X is the total proportion of exports in GDP for 1976, and t the normalized vector of exports, then x-x* = Xt-X* t* = (X-X*)t + X*(t-t*) .

(10)

Decompositions of the difference of products, such as those appearing in expressions (7) through (10), are not unique; it is easy to prove that if there are n terms in the product, then there are 2n~l different decompositions possible. A check of alternative decompositions for the subject matter of this study yielded similar results.

REFERENCES

Deachman, J., and K. Hamilton. 1978. "Energy Availability, Detailed Disposition, and Industrial Demand Coefficients for Canada, 1971." Statistics Canada, Structural Analysis Division Working Paper 78-08-01. Ottawa. Energy, Mines and Resources Canada. 1988. Energy Statistics Handbook. Ottawa. Flaschel, P. 1982. "Input-Output Technology Assumptions and the Energy Requirements of Commodities." Resources and Energy, no. 4. Hamilton, K. 1977. "An Energy Analysis of Canadian External Trade 1966." Statistics Canada, Structural Analysis Division Working Paper 77-03-03. - 1988. "Energy Intensiveness and Economic Performance Since 1971." Canadian Economic Observer, December. Herendeen, R.A. 1973. The Energy Cost of Goods and Services. Oak Ridge National Laboratory, ORNL-NSF-EP-58. Leontief, W. 1966. Input/Output Economics. New York Oxford University Press. Mclnnis, B., and K. Hamilton. 1975. "Gross Energy Requirements for the Production of Goods - An I/O Baseline." Statistics Canada, Structural Analysis Division Working Paper 75-06-01. Postner, H.H. 1975. Factor Content of Canadian International Trade: An Input/Output Analysis. Ottawa: Economic Council of Canada.

59 Industrial Energy Consumption and Trade Statistics Canada. 1976. Detailed Energy Supply and Demand in Canada. No. 57-207. — 1977. The Input-Output Structure of the Canadian Economy 1961-71. No. 15-506. — 1980. The Input-Output Structure of the Canadian Economy in Constant Prices 7977-76. No. 15-202.

CHAPTER FOUR

Technological Clusters and Competitive Poles: The Case of Canadian Energy CHRISTIAN DEBRESSON

Technological advance affects the trade performance of a country not only by modifying the absolute advantage of one of its industries or the ranking of its industries compared with that of other countries (comparative advantage), but by creating dynamic advantages that go beyond specific industries. This study develops an aspect of Posner's technology gap theory of trade that has received little attention to date. Innovation tends to cluster in related industries and thus confers to a national economy dynamic advantages. This study takes the example of Canada, focusing on the concentration of innovation in and around the production of electrical power and energyintensive transformed primary goods. The innovation and patented-invention clusters in these industries probably sustain Canada's export specialization in these goods — that is, its "competitive pole." It is suggested that a realistic theory of technology's effect on international trade should take into account the various externalities resulting from these innovation clusters.

INTRODUCTION

The previous two chapters in this volume examined, in the Leontief tradition, the effect of factor content on exports. Lester Davis examined the downstream direct and indirect effects of R&D on exports, and Kirk Hamilton the direct and indirect effect of energy on exports. This study examines yet another technological effect on exports. This is not, however, an effect due to the factor content of exports, but rather the effect of the clustering of innovations; that is, it is not an effect of factor content in the traded commodity, but

6i

Technological Clusters

an external effect of the cluster. I will refer to this as the cluster effect of technology on trade. One of the main effects of technology on trade is best conceived as the external effect of technology clusters. The following quotation will set the topic in focus: We should certainly not expect, a priori, the stream of innovations during any particular period to be distributed randomly between industries. There are four reasons for expecting innovations to be concentrated in one industry or group of industries: 1. There is a technical connection between innovation and its successor. A break-through on one front will bring, quite rapidly, associated successes. Thus there will occur "clusters" of innovations. It is impossible to assess a priori whether this effect will produce concentration of dynamism within one industry, or whether, in fact, by a converse effect, the benefits of the original discovery might provoke successors in quite different industries. The associated development of fast engines, speed resistant tires, and disc brakes (all broadly, in the automobile industry), is an example of clustering; the development of pressure die-casting is an example of the converse. But whichever effect predominates, the existence of some such technical correlation is very unlikely to result in an unbiased distribution of innovations between industries (it may be that the existence of correlation will not only redistribute the incidence of innovations, but also increase the number of innovations). 2. There may be a similar connection on the demand side - complementarities in consumption may lead to pressure for innovations in products jointly demanded with that just innovated. 3. It may so happen that one industry, in a particular period, is blessed with an excessive quota of innovating entrepreneurs or research staff. 4. There may be dissimilarities between industries in their expenditure on "innovation-generating" research; i.e. the rate of investment may differ between industries. It is "technical 'clustering'" which is nearest to the pure concept of dynamic economies of scale (as defined ... above), accruing to a firm by virtue of its past experience (or, more precisely, the aggregate of its output in the recent past).... [T]o the extent that the correlation does occur, such foreknowledge can give the original innovator a head start which he need never surrender until the cluster of related innovations comes to an end. By the time the end comes, the unfortunate foreign competitor may have given up the ghost, unless he invests sufficiently (and sufficiently quickly) in research to buy

62 Technology, Energy, International Trade back the opponent's original lead. Conversely, extra investment by the original innovator can help to consolidate and extend his advantage. (Posner 1961) The above excerpt comes from the seminal economic article on the relationship between technology and trade: M.V. Posner's theoretical foundation for technology-gap trade theory. In this article, Posner assumes from the start that technological innovation will necessarily create a temporary monopoly. This technological monopoly will temporarily disrupt the free circulation of factors. It will also disrupt and modify the patterns of trade that have been shaped by factor endowment and previously established absolute and comparative advantages. Furthermore, Posner suggests that such monopolistic advantages may maintain and reproduce themselves through "dynamic scale economies"; that is, cumulative learning processes enable a technological leader to evolve a new technological advance once the previous technological monopoly has wilted away. Some important empirical work has been done on Posner's hypothesis. Gary Hufbauer produced the first, and yet unsurpassed, empirical demonstration of technology-gap trade in his Synthetic Materials and the Theory of International Trade (1966). In the late 19605, the Organization for Economic Co-operation and Development (OECD) produced a series of monographs identifying these technology gaps and the resulting trade patterns (1970). Since then, many of these technology gaps have been bridged, while new ones have appeared. Europe and Japan are assumed to have filled the gap that had existed between them and the United States. Furthermore, it is commonly believed that since the mid19705 Japan has surged forward in creating some new technology gaps of its own, this time in its favour, that put it ahead of other industrial countries — including the United States. The technologygap studies from the late 19608 clearly need to be updated. Luc Soete (1981) argues that the technology-gap theory of trade has a reasonable chance of standing up fairly well to the empirical evidence of more recent years. In a generalized cross-sectional test, he correlated foreign countries' patents in the United States for each industry with their export performance. A recent master's thesis at Concordia University, however, gives rise to some scepticism (Cotsomitis 1989). This study is not concerned with testing the technology-gap theory of trade, although the theory is accepted as a reasonable starting point. I am not so much concerned with the question of whether or

63 Technological Clusters

not technology gaps affect patterns of trade (I take for granted that they do) as with how they affect trade patterns. This chapter will explore yet another trade effect of technology gaps: the cluster effect. Trade analysts have taken little notice of Posner's mention of technological clusters as an element affecting trade patterns. Yet Posner attributed much of the "dynamic scale economies" of a country to the clustering aspects of a technological advantage. The excerpt from his 1961 article, which points to the central place of technological clusters in his technology-gap trade theory, attests to this. TECHNOLOGICAL CLUSTERS

Although the implications of technological clusters for trade have not been drawn out, it is generally accepted that technological innovation is a phenomenon that tends to occur in clusters. J.A. Schumpeter was no doubt the first to have called economists' attention to such clusters: Whenever a new production function has been set up successfully and the trade beholds the new thing done and its major problems solved, it becomes much easier for other people to do the same thing and even to improve upon it. In fact, they are driven to copying it if they can, and some will do so forthwith. It should be observed that it becomes easier not only to do the same thing, but also do similar things in similar lines — either subsidiary or competitive ones - while certain innovations, such as the steam engine, directly affect a wide variety of industries — this seems to offer perfectly simple and realistic interpretations of two outstanding facts of observation: First, that innovations do not remain isolated events, and are not evenly distributed in time, but that on the contrary they tend to cluster, to come out in bunches, simply because, first some then most, firms follow in the wake of successful innovation; second, that innovations are not evenly distributed over the whole economic system at random, but tend to concentrate in certain sectors and their surroundings. (Schumpeter 1939, 100—i)

Schumpeter's bald-faced reference to "outstanding facts of observation" has not gone unchallenged (see, for instance, Rosenberg and Frischtak, 1983). Nonetheless, there are many sound reasons why technology and innovations would cluster. Some are technical; some are economic. A brief summary of these reasons follows (a more complete analysis of the foundations of such clustering appears in DeBresson 1987, 19893, igSgb).

64 Technology, Energy, International Trade First, the emergence of new technical paradigms tends to focus innovative endeavours. Second, system complementarities will focus technical endeavours toward only those technical solutions that are congruent with the system. Third, the transferability of some of the benefits of learning-by-doing to those products at technical proximity that can make use of the same know-how makes innovation more probable in the vicinity. Micro-economic incentives will also induce entrepreneurs to cluster their innovative activity around previous areas of innovation. Fourth, the search for economies of scale will focus technical activity on systems coherence. Fifth, the search for economies of scope and sharing of input costs (labour, equipment) among a number of products will induce technical innovation in the immediate technical proximity of acquired know-how; even pioneering firms will explore the technical frontier by successive but limited incursions into the technical unknown/row their previously acquired base of technical expertise. The incrementalism of exploratory research and development aims at avoiding the potential penalties of dis-economies of scope. Sixth, the search for durable monopolistic advantage will focus firms' technical endeavours toward those areas that present better chances of exclusive technical appropriation. All these factors, be they technical constraints or economic inducements, interact and reinforce one another toward a single outcome: innovation clusters. More than any individual economic factor, it is the synergetic effect of converging factors that leads to the clustering of innovation. Prominent among these synergetic features are inter-firm coalitions - for the common development of a capital goods and engineering capability - and firms' concerted efforts for regional development.

I D E N T I F I C A T I O N OF A COUNTRY'S TECHNOLOGICAL CLUSTERS Innovation clusters can be identified rigorously on an innovation supplier-user matrix (DeBresson 1987, 19890). (Invention clusters can also be identified on patent matrices.) Simple directed-graph analysis enables us to measure the existence of such clusters and their degree of internal integration. Such analysis also provides us with indicators of probable synergy within the clusters. Using the simple notions of reflexivity (aRa), symmetry (if aRb, then bRa), and transitivity (if aRb and bRc, then aRc), one can differentiate between the types of clusters within a matrix. We can thus

65 Technological Clusters

define five main distinct types of clusters (or "directed graphs") that are disjoint - that is, to which one (and only one) matrix corresponds: • Type i — the clique: that is, a connex graph in which all paths (or arcs) between suppliers or users (summits) are taken • Type 2 - the complex: that is, a connex graph in which there exists at least one symmetrical and one non-symmetrical relationship between summits • Type 3 - the simple cluster: a connex graph in which summits have no symmetrical relations with one another • Type 4 — the innovative coupling: two summits with symmetrical relationships • Type 5 - the isolated point of development: one isolated summit with a reflexive arc These five types of clusters correspond to five basic directed graphs as represented in Figure 2. These five types of matrices and directed graphs represent different levels of integration and dynamic interaction. The presence of a clique of innovators (type i), with all suppliers and users interacting with one another, implies more dynamic and rapid technological accumulation than the isolated point of development (type 5). One would thus expect that a group of interrelated industries that have very tight-knit and widespread innovative interactions will accumulate technological know-how faster than will a group of industries that have only a few, unintegrated innovative interactions. In other words, directed-graph analysis enables us to construct indicators of probable synergy in the clusters. These indicators do not relate to the number or relative importance of the innovations but to the interrelations between the innovators. Higher-order clusters of innovation in a country are likely to have a notable and durable impact on international exchanges. The size of directed graphs and intensities of the arcs can also be measured, indicating the number of innovators (summits) and the frequencies or importance of their innovative interactions (weight of the arcs). Each of the first three directed graphs (clique, complex, or simple cluster) may involve different numbers of participants (summits) in the innovation clusters. The intensities of the interaction (arcs) can be measured by the frequences and/or degree of importance of the innovations. This method of analysing clusters in conjunction with the Canadian innovation and patent matrices will be used in a following section.

66 Technology, Energy, International Trade Figure 2 Basic Directed Graph and Types of Clusters

Posner's hypothesis should then be tested not only in relation to individual product (or industry) technology gaps, but also in relation to the effects of domestic clusters of innovation. Innovation clusters can be identified and examined as they relate to trade performance. This is an avenue of analysis we are presently exploring in regard to Canada in order to explain its performance in the energy field.

67 Technological Clusters THE CASE OF C A N A D A : AN ENERGY

CLUSTER

Innovation clusters can be identified in Canada on the supplier-user innovation matrix established in 1980 for the Science Council of Canada (DeBresson and Murray 1984). This innovation matrix represents both the Canadian economy and significant innovations. An innovation survey was statistically representative of the firstorder domestic-requirement input/output matrix: that is, a matrix for all the goods and services supplied by Canadian production for its domestic demand. A triangularized domestic-requirement matrix enabled me to choose seventy-nine industries that represented most of the domestic internal trade. These seventy-nine industries were surveyed for their innovative advances from 1945 to 1979. Furthermore, within each of these industries, a stratified random sample insured that establishments of different sizes and different regions were represented. Care was also taken to ensure that the sample be representative of the significant innovations of the period. Thirty trade journals were scanned for the years 1945-79 for mention of firms' new products and processes. Furthermore, 270 industrial experts in their respective industrial fields of expertise identified significant innovations and innovating firms. All industrial R& D-performing firms and patentors were surveyed. In this way some 4,000 innovations were identified. Innovation was defined as the first introduction in Canada of a new or improved product or process, 1,900 of which were confirmed by the firms. Each respondent was asked to confirm specific pre-identified innovations. Prior identification of the firm's innovations insured a response rate of 70 percent. The respondent was asked to confirm responsibility for the innovation. Furthermore, to enable performance of a. sensitivity analysis, the respondent was asked whether the innovation was a world first or, if not, to identify the year and country of the foreign predecessor. This permitted calculation of an index of novelty according to the time lag of adoption. The originality of the survey lies with the variables examined, and in particular with its inter-industrial conception. For each innovation, the supplier firm and the first user firm(s) and industry(ies) were identified, from which information an innovation matrix was drawn (Table 26). Each cell in the matrix represents the number of innovations that one industry supplied to another. The first-user industries are represented in the columns; the supplier industries, in the rows. In this

Table 26 Innovation Matrix: Canada, 1979 777 100 404 821 013 271 064 059 323 902 572 951 031 378 503 0990 108 909 295 631 251 339 291 806 853 315 864 378 391 335 336 379 339 165 059 308 321 187 309 291 324 271 329 399 273 183 318 304 359 369 162

12 2 13 12 0 0 14 12 16 5 0 0 18 8 0 1 0 13 3 1 0 0 5 0 0 18

51 38 5 9 11 15 16 5 27 1 30 5 15 8 6 4 7 2 0 0 1 4 0 0 5 5 7 5 0 5 1 2 1 2 0 0 3 0 0 0 4 3 3 0 0 2 3 28 0 0 0 0

6 5 12 96 6 0 2 2 11 0 1 0 0 0 0 0 5 5 0 0 0 0 0 0 2 0

4 62 32 26 0 8 17 5 7 12 25 5 1 6 10 5 0 0 12 0 0 8 0 0 3 5 15 2 0 0 0 0 0 0 2 5 10 0 5 110 0 0 5 0 0 0 2 0 5 0 0 3 5 0 2 0 0 0 0 5 0 0 1 2 0 102 0 0 0 0 0 0 0 0 0 0 1 0 2 0 0 0 0 0 0 0 5 0 0 2 2 0 0 0 0 0 0 0 25 0 5 0 2 2

7 11 1 5 5 0 0 7 0 40 5 10 0 1 1 8 0 0 5 1 0 0 0 86 0 0 1 0 0 0 0 0 0 0 0 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 0

28 46 0 10 9 46 0 13 0 0 0 4 0 8 0 0 0 0 0 0 0 0 0 0 0 0

12 58 8 5 0 0 1 0 0 2 5 0 1 0 0 0 2 5 0 0 0 5 0 0 0 5 1 0 0 0 0 0 0 10 5 10 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 9 5 5 81 0 0 2 0 5 0 25 5 0 0 0 1 0 0 0 0 1 0 7 0 0 0 0 0 4 5 0 5 0 0 0 0 0 5 0 5 0 0 0 0 0 0 0 0 0 0 0

0 26 0 46 6 0 5 0 0 3 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0

4 3 0 1 0 0 5 0 0 0 0 0 5 0 0 0 0 0 0 3 0 0 0 0 0 0

0 5 1 31 0 11 26 0 22 0 22 5 5 0 13 0 0 0 0 46 1 0 15 0 0 0 0 0 5 2 0 0 0 0 7 0 0 5 0 0 0 0 0 0 0 0 0 5 0 5 0 0

1 1 0 0 0 0 0 0 2 0 0 0 0 1 0 0 0 0 5 0 0 0 1 0 5 0

18 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 5 5 0 2 0 0 0 0 0 0

0 30 3 5 1 2 0 0 0 3 0 10 0 0 1 4 1 15 0 0 0 0 0 5 0 0 0 0 0 47 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 5 0 28 0 0 2 1 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0

0 1 0 1 1 5 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 9 1 0 0 0

Table 26 (continued) Innovation Matrix: Canada, 1979 311 338 373 302 294 631 374 064 325 375 332 103 295 101

777 100 404 821 013 271 064 059 323 902 572 951 031 378 5030990 108 909 295 631 251 339 291 806 853 0 0 1 1 6 0 32 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 0 0 1 0 0 0 0 0 0 0 0 0 2 0 0 8 0 0 5 4 5 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 10 0 5 0 0 5 0 0 0 0 0 5 0 0 0 0 0 0 2 0 0 1 1 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 1 1 0 0 0 0 1 0 0 3 0 0 0 0 0 0 0 2 0 0 0 0 0 0 1 0 0 51 0 161 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 24 0 0 0 0 0 0 0 0 5 3 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 27 0 0 0 0 0 0 0 0 0 0 3 2 0 0 0 0 0 0 0 0 0 0 9 0 0 0 0 0 0 0 0 0 20 5 0 23 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 5 5 0 0 0 0 0 0 0 0 0 0 0 12 0 0 0 0 0 40 0 0 2 0 0 10 0 0 0 0 0 12 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5

Source: C. DeBresson and B. Murray 1984, 117.

70 Technology, Energy, International Trade Figure 3 The Energy Innovation Cluster

(triangularized) matrix, the user industries (demand vectors) and supplier industries (supply vectors) are ranked. Those industries that are the greatest first users of innovations originating from the greatest number of different industries are in the left-side columns of the matrix; those industries that supply innovations to the greatest number of industries are in the upper rows of the matrix. The innovations can be weighed in accordance with their degree of novelty: a world first is given a weight of 5; an early adoption in Canada (within three years of its first introduction), which will probably require some adaptation, a weight of 3; a later adoption, from four to nine years, a weight of 2; and beyond ten years, a weight of i.

71 Technological Clusters

The matrix represents the innovation interactions in the economy, that is, the activity of those industries that exchange information in order to generate new or improved products and processes. These innovation interactions are the seeds of structural change and growth. The use of a specially designed computer algorithm identifies the innovation clusters. For the Canadian innovation matrix, it yields the following key results. The main concentrations of innovative activity occur in and around the production, distribution, and use of electrical power, the health sector, communications, wood products (including newsprint), and metallurgy. Here we will only concern ourselves with the innovation cluster in the energy sector (Figure 3). Invention clusters can be identified in a similar way thanks to the patent matrices computerized from PATDAT, the Canadian computerized patent database, which lists for each patent up to three possible industries of use and supply (Ellis 1981). The landscape of inventive activity in Canada can thus be represented in matrix form. With the same method as above, patent clusters in Canada have been identified. Inventive activity is less concentrated than innovation. The main areas of activity, however, are communication electronics, metal products, electrical products, and mining. (Different patenting propensities in various industries will dwarf or boost the relative importance of the inventive activity of a given industry.) Here we will concern ouselves only with the patent cluster in the energy sector (Figure 4). THE S I G N I F I C A N C E OF THESE CLUSTERS

What do these directed graphs represent? What are these directed graphs indicators of? Are they adapted for measuring the technological clusters Schumpeter and Posner were talking about? For Schumpeter, innovation clusters were mainly the results of an external technological shock. Elsewhere, I have argued that he was correct in identifying exogenous technical shocks and constraints as a key factor of clustering, but I also argued that there were reasonable micro-economic foundations for the clustering of innovations beyond the shocks of radical innovations (DeBresson 1987, 19893, 19890). In other words, there are many endogenous economic forces that would lead to the clustering of innovations, even without any major external technological shock. As a consequence, these innovation clusters may be more prevalent than Schumpeter himself had

72

Technology, Energy, International Trade

Figure 4 The Energy Patent Cluster

perceived. Not only do radical innovations, usually associated with some major external technological shock, result in the clustering of innovation, but clustering would be a standard feature - a "way of being" - of innovation. Therefore, when we identify the clustering of innovations on the matrices, we can legitimately question what we are measuring. Are we measuring the clusters of innovations due to technological shocks? Or are we simply measuring the reflection on the innovation process of the interdependence of industries in the input/output table? 1 The answer is that we are doing both. We can determine this in the following way. By separating out those innovations that are claimed to be world firsts and running the directed-graph analysis on those only, we find that the energy innovation cluster is significantly more clustered than the innovations

73 Technological Clusters

resulting from adoption and adaptation. In other words, the innovation cluster that is strictly the consequence of the expansion of the technological frontier (i.e., of the creation of new technology)2 is more significant than the one due only to the economic forces of adoption and adaptation. In other words, the bulk of the clustering effect comes from the generation of new technology and not from the adoption of new technology. More generally, the innovation matrix may be nested in the broader domestic-requirement or capital-requirement matrix, but it is not identical to it. We can expect that in a block diagonal matrix some industrial complexes would favour innovation clusters - but not all. We have also found similarities in ranking between the demand for goods and services in the Canadian domestic-requirement matrix of 1976 and the innovation matrix for 1979 - but very dissimilar ranking in their respective supply vectors (DeBresson and Murray 1984). In other words, not all well-integrated industrial complexes give rise to innovation clusters. The innovation clusters are channelled and disciplined by standard input/output relationships, but they are not reducible to the latter. While the structure of demand may discipline the direction of innovation, the clustering of innovation is partly an autonomous phenomenon, as reflected in the innovation matrices. Therefore, the innovation cluster indicates more than the standard interdependence in the exchange of goods. To summarize, the innovation clusters identified by directedgraph analysis are indicators of synergy, of the way the innovations meet, not as such of the level of innovation. The higher-order clusters (the clique) is assumed to indicate more synergy than the lowerorder clusters (simple agglomerations). THE EXPECTED CLUSTER EFFECT ON TRADE

What effect could one expect these innovation and invention clusters to have on Canada's trade in energy technologies and more specifically on energy equipment? (This is one subject of current research at CREDIT.) It is reasonable to believe that clusters of innovation would have a global effect on the competitive pattern of a country. If a country possesses a technological and innovative cluster in a set of related industries (a cluster not possessed by other countries), then it is reasonable to assume that the country will reap a global technological advantage in all these sectors, since, according to Posner's theory, it will be able to reap cumulative dynamic advantages ("dynamic scale

74 Technology, Energy, International Trade

economies") that will enable it to reproduce its technological leadership. This technological advantage due to the cluster can manifest itself in various expressions of competition: control of the domestic market, exports, foreign investment, and/or technology transfer payments. International competitiveness is a multi-faceted phenomenon. A firm may choose one or another strategy (control of domestic market, export, foreign investment, or licensing) or a combination of these. Accordingly, one can only measure the competitiveness of an industry by using these four indicators simultaneously (OECD, 1983, 1984). Conversely, the effect of technological and innovative clusters on competition is structural; that is, it cannot be ascribed to any specific product or industry but acts upon all those in the cluster. The cluster effect on competition is best conceived as a mediumto long-term structural externality. To better highlight this, let me first describe the technological advantage that is internalized by firms. TECHNOLOGY

GAPS: FIRM-

SPECIFIC INTERNAL

EFFECTS

Firms internalize technological advantage through exclusive appropriation. Patent rights can be seen as an indicator of this proprietary technology. Although trade data are essentially at the industry or product level, the economic rationale is better grasped at the firm level. In the production process of each product, a firm will possess a certain proprietary technology. The level of its technological know-how can be measured as a stock variable calculated with weighted patents. A recent patent will, of course, be of more importance than a patent right that has lapsed, but the latter still reflects more technological know-how than the absence of a patent. The firm's patent stock for a given product in a given year will affect its competitiveness in following years.3 More precisely, a firm's increase in technological stock, relative to that of other firms in the same industry in another country, gives it an absolute and durable technological advantage in competition.4 The absolute monopolistic advantage of that firm (or of a country's industry in relation to the same industry in other countries) will eventually affect the absolute advantage of the country. Depending on whether the firm decides to use this durable technical advantage strategically on the domestic market, through foreign exports, or on external markets, through foreign investment or licensing, the firm's technical advantage will translate, with a lag, into an improved per-

75 Technological Clusters

formance in one exchange indicator (domestic sales, exports, foreign investment, licence payments) or in a combination of indicators. Such internalized effects of technology gaps on trade, investment, and technology transfers obviously exist. I am not ignoring them. But they cannot represent the total effect of the technological factor on trade.5 To this internal technological effect on trade, one must add the external effects of a country's advanced technological capability. In other words, the relative technological advantage of a country is not all captured in the firm's internalized advantages in trade. Some of the country's technological advantage is diffused among a number of interrelated industries that together participate in the technological cluster. The more the technological cluster is integrated, the more one can expect the advantages to be shared. For instance, one would expect a dense clique of innovating firms and industries to benefit more from the external effects of their mutual technological capability than a loosely knit group of firms and industries with sparse innovative interactions. E X T E R N A L I T I E S OF TECHNOLOGICAL CLUSTERS' EFFECTS ON TRADE

Labour Externalities

The work force involved in innovative projects is usually highly skilled and experienced. As a result, it possesses certain powers of negotiation in such matters as place and conditions of employment. It is well known that innovative people tend to change employers to achieve their goals. Incubator firms (such as Fairchild in Silicon Valley) only exist because of the extreme mobility of innovative personnel, in particular those who work at the frontier of technological knowledge. Therefore, experience acquired in one firm can be transferred to another. This is the classic labour externality: the benefits of training personnel cannot always be captured by the firm itself. In the case of innovative personnel, this characteristic is even more pronounced. (The external effects should here be attributed to all the firms in the supplying industry.) Scope Externality

Labour is a source of economies of scope; fixed capital is another. Process equipment is more or less specific and dedicated to an output.

76 Technology, Energy, International Trade

Any equipment will possess a range (or scope) of possible outputs. With a specialized machine or an assembly line, this scope may be narrow; with a general-purpose machine in a batch-production mode, the scope of possible outputs may be quite broad. In the latter case, setting up the machine with new and different jigs and fixtures can enable producers to change the output. Although the same equipment will usually be used by the same firm, this is not always the case. For instance, different railway companies share the same tracks, different power companies share transmission lines, and so on. Economies of scope do not automatically imply greater scope of the enterprise (Teece 1981). Part of the economies of scope are internalized, but part are externalized. As a result, neighbouring firms will benefit from scope economies stemming from the scope of fixed capital. (The external effects should here be attributed laterally to all the firms in the supply industry, and downstream to all the user industries.) Component Externalities

The availability of new products from one industry, be they equipment or materials, will make new developments possible in another industry. Each new component has externalities. The benefits are not all captured by the component manufacturer; some are captured by potential users. For instance, composite materials have opened up opportunities for the automobile and aircraft industries, as was previously the case with aluminum and metal alloys. Although these externalities are very diffuse, they will first benefit the user industries that are already in innovative interaction with the supplier. Eventually any user industry in any country will be able to benefit, but initially the learning costs of adoption of a new technique are such that the innovation-user firms and industries are the first to benefit. (The external effects should here be attributed to all the users of the new products.) Complementarities of Demand

Some components are used by a number of industries. Complementarities of demand favour some industries more than others. For example, in Canada's energy system, pumps and valves of various models are used in hydro-electric dams, in petroleum recovery, pipelines, and refining, and in nuclear power. Suppliers of these components can rely on a varied demand. Pipeline building exerted considerable demand on these suppliers for a time. There was also

77 Technological Clusters

a period (before Three Mile Island) when the building of nuclear plants required sophisticated pumps and valves. In the 1973—82 period, high prices led to a boom in the petroleum industry and gave suppliers many orders. In those same years, and also in other periods, hydro-electric dams were a prime source of energy. Complementarity of intermediate demand (or of consumption, as Posner calls it above) for certain components has an effect that cannot easily be attributed to any one industry: each user industry will benefit from the experience the supplier has gained with other client industries and from the fact that the supplier does not have to rely only on the industry's orders but can, in fact, rely on a combination of possible industrial markets. (Those industries that benefit from complementarities of demand are easily identified on the supply vector of an input/output matrix; the attribution of the external effects should in this case be to all the firms in the supplying industry and to all the user industries of these products.) Network/Systems Externalities Technology is systemic. Most technologies present themselves not as isolated artefacts but as systems of interrelated techniques. Many technical systems present themselves as networks (telephone networks, interfacing computers, power grids). Network externalities are well known. The advantage of being a telephone subscriber is proportional to the number of other subscribers whom one can reach in the network. Similar systems externalities exist, to a certain degree, for all technical systems. In other words, there is an economic advantage that stems not only from a product, but from the system coherence. The advantage cannot be captured by the specific contributor of that product rather, it benefits all the participants. Different technologies fit well together, or they don't fit - independently of their individual merits as components. The economic benefits of good fit are to a large extent externalities that benefit all the participants in a system. (These external effects should be attributed to all the member industries in the technical system.) Synergetic Effects Externalities Finally, there is a more difficult, but perhaps even more important, externality: the synergetic effects that stem from the degree of network integration in the innovative interactions. Systemic complementarities in technology allow for a varied set of innovators to interact

78 Technology, Energy, International Trade

dynamically. In doing so, not only do they accumulate technical know-how themselves (a know-how they can appropriate), but they stimulate each other through their varied and iterative interactions — new orders, new requirements, new technical opportunities and experiences, sharing of technical information, and so on. The density of the interactions in the innovative network thus greatly accelerates the process of technological accumulation. (These external effects are to be attributed to all the members of the cluster.) STRUCTURAL EFFECTS

These last four externalities can be called structural; that is, they are the consequence of stable technical interrelationships between industries. The identification of technological clusters makes possible the concrete description of the structural patterns specific to each country. The term structure has to be used with caution. The concept of structure (like that of system) is so general that one must be specific when applying it. One of the analytical powers of the input/output methodology is to enable one to specify what structure is under discussion. Here we specify the structure to be the stable configuration of the clusters. The effects of technological clusters are structural because they are relatively stable and rigid. Innovation clusters do not appear and disappear, or change configuration, in a day. For example, the energy cluster in Canada has had the same basic skeleton since at least 1945. After dividing the postwar years into three periods (1945—60, 1961—69, 1970—79), we found that electrical industrial equipment (Canadian Standard Industrial Classification [sic] 336) and consulting engineering (864), which feed innovations to firms involved in the production and distribution of electricity (572), form the basic skeleton of this cluster in the first and second periods. This cluster, with the same skeleton, is then fleshed out in the latter period with miscellaneous electrical products (339), various machinery (315), communication electronics (335), and instruments (391).6 If we take the corresponding patent clusters for two periods, 1972-77 and 1978-84, we find again the same type of resemblance in basic structure. At the same time, innovation clusters slowly evolve over time. Clusters grow. Consistent with the presumed synergetic effect, the cluster develops around a node of denser technological interactions that integrate new industries and firms into the cluster. (This diachronic dimension is important for our analysis because, as we sug-

79

Technological Clusters

gested, the technological factor on trade patterns is best measured with time series — not cross-sectional analyses.) Let us review some of the probable reasons for this relative stability of clusters. Stability of Partnerships Innovative interactions have initially high transaction costs because of uncertainties in the partnerships between the supplier and first user of an innovation (DeBresson 1987, igSga, 19890). The situation of the two partners in an innovation project can be described as similar to that of a bilateral monopoly. Each of the two partners is usually the only one to possess some of the required knowledge, information, and skills to bring to fruition the innovation. Each of the potential partners are monopolists in their complementary fields: one is a monopolist as the potential first user in search of a new component, the other a monopolist as a potential supplier needing the user's information to develop the innovation. As in all bilateral monopoly nightmares, the outcome is the result of a totally negotiated process. As a consequence, innovative interactions between firms tend to be iterative. Transaction costs diminish dramatically with the second innovative project. It is far easier for the firms to keep innovating with their previous partners than to socially engineer a qualitatively new innovative interaction. Hence innovative interactions tend to be stable. Stability Resulting from Inertia An innovation is irreversible, and a specific innovation can only be made once. The momentum of innovation has irresistible force. The emergence of an innovation will necessarily focus attention on a new set of questions. Compulsive sequences of technical search ensue. Sunk costs in learning and dedicated fixed investments will encourage the exploitation of the previous innovations benefits. Technological development thus seems to have its own predetermined trajectory. One problem leads to the next. The innovation not only opens new opportunities but defines new limits and gives rise to technical problems, system imbalances, and bottlenecks. Innovation slowly constrains and narrows the field of future technical endeavours. The direction of innovative interactions therefore tends to be stable - with the exception of the emergence of radical technological discontinuities (such as superconductors, bio-chemical en-

8o Technology, Energy, International Trade

gineering, or semiconductors). The inertia in the direction of technical improvement contributes to making innovation clusters fairly stable in their basic configuration. Cumulative Nature of Innovations Innovations accumulate. Even minor innovations will cumulatively compound, building toward new performance thresholds and thus making possible the modified technology's application in new fields. The cumulative nature of technical know-how lends a heaviness to its momentum (what the French call une tendance lourde). Because of the above characteristics of technological development, innovation clusters are slow in their changing configuration. As a result, the cells in an innovation matrix (and the arcs in the corresponding directed graph) tend to have some inertia. This is a paradox. Innovative interactions represent a dynamic element in relation to the rest of the economic system (i.e., an element that fosters not only growth but structural change and development - one that will eventually, if the innovation diffuses, transform the input coefficients of an input/output table of goods and services). Nonetheless the direction of these innovative interactions has considerable inertia. The movement of innovation has its own slow momentum, innovation occurring at proximity of previous innovations. As engineers say, "Only one innovation at a time." Innovators are conservative vis-a-vis the choice of areas in which they innovate. Innovative clusters, therefore, represent a fairly stable pattern of technological development. As a result, the structural effect on trade of these technological clusters is durable. It can only be perceived in the medium and long term. Nonetheless, the extension of the technological cluster to new member industries (summets) and its branching out into new directions (new arcs) are of importance. As the synergy in the technological cluster enables it to grow and diversify, one would also expect the cluster effect to extend itself to the new members of the cluster. THE MEASURE OF TECHNOLOGICAL CLUSTERS' EFFECTS

Let us now return to our original question: What effect could we expect technological clusters to have on foreign investment, trade, and transfer? What effect can we expect these clusters to have that is distinct from that of product-specific proprietary technical advance

8i

Technological Clusters

internalized by the firm? How can we measure the effect? The difficulties in measuring it may be similar to those in measuring any external effect. The problem with externalities for econometricians is very much the same as that for economists in general. Insufficient attention to externalities is the rule in economics. Conversely, general references to externalities (without specifying their nature) is the opposite temptation and methodological pitfall. One way of approaching the problem of externalities is to think in terms of the attribution of their effects. Having specified the nature of these "externalities," it may also be possible to inpute the direction of their effects. Input/output matrices have been used (see chapter 2 by L. A. Davis) to inpute the indirect effects of R& D on a country's exports. Obviously, the technological cluster in energy can be expected to have these kinds of downstream effects on Canada's exports. For instance, it is well known that Canada's exports are energy intensive (see chapter 3 by K.E. Hamilton), and it is reasonable to think that accumulated technological know-how in the energy field has contributed to Canada's natural Ricardian advantages in energy endowments. In other words, this country's long-term efforts in the supply of energy technologies has made available and accessible these natural endowments and acquired for Canada some absolute and comparative advantages. Successful technological efforts can be conceived as a modification of natural Ricardian advantages and as the exploitation of favourable factor endowments. The concentration and synergy in technical accumulation in the energy field helped, downstream, Canada's energy-intensive exports. But although I have specified above some external downstream effects, in this case we are more often talking about lateral externalities resulting from the synergy of the cluster. Let me be more specific. If these cluster effects resulting from synergy are external, the technological clusters (the independent variable) will affect all or any one of the exchange indicators — export, domestic sales, foreign investment, and transfer payments - and will do so in a diffuse way. To be more precise, the more integrated the cluster, the more all (or most) of the industries participating in the national cluster will be favourably affected. In other words, the externalities stemming from the technological capability in the cluster should improve the performance of all (or most) of the participants in the cluster. The improved performance of all (or most) members of the cluster, which cannot all be explained by the product-specific advance in technological stock, could be partly attributable to the ex-

82 Technology, Energy, International Trade

ternal effect of the cluster as a whole.7 These would therefore be principally lateral externalities or general spillover effects (for an example of econometric analysis of spillovers, see Jaffe 1986). The structural effect of lateral externalities will be medium to long term. One could not expect the cluster to affect from one year to the next any of the exchange-performance indicators, but the external effect of the cluster will affect the trend over a few years. Cumulatively over a few years (five years or a decade), the existence of a technological cluster will sustain and promote the performance of a number of related products and industries within the cluster. Of course, an external effect will only occur once an industry becomes part of a technological cluster, perhaps with a certain lag. As the dependent variable is diachronic (export-performance time series), it would be useful to devise a diachronic indicator for the clusters. From the innovation matrix (Table 26), we can calculate transition matrices for various periods. The transition matrices, in which the total of the lines are equal to i, indicate the conditional probability that if an industry i (on the row) is innovating, it will sell its new product innovations to industry j (on the column). Then the data in the transition matrices may be used to impute the external technological effect and be matched with the trade-performance data for, with a lag, a given period. Furthermore, the level of synergy displayed by the directed graphs can be accounted for with a dummy variable. The technological cluster in energy has probably a favourable effect on Canada's foreign trade in energy technologies, bearing out the reasonableness of our hypothesis. Innovation is concentrated in this area. The performance trend for Canadian exports of energy equipment is much more favourable than the general trend for Canada's other equipment exports and all manufactured exports. POLICY IMPLICATIONS

For a long time, provincial politicians in Canada have thought that energy initiatives, in particular hydro-electric projects, present an opportunity for a strong natural advantage (in the Ricardian sense). This natural advantage (dam sites) is seen as a potential comparative advantage that would sustain the industrial development of the country. In the early twentieth century, hydro-power and its decentralized use constituted the backbone of industrialization in Ontario (Nelles 1974). In the Ontario industrial saga, Adam Beck was the populist champion of a development strategy based on a decentralized electrical power network. In contemporary Quebec, Premier Robert Bourassa represents another advocate of this line of development,

83 Technological Clusters

although he is not always explicit about how industrial externalities and development will follow from such development. Hydro-Quebec's contracting out of hydro-electric engineering work has, however, resulted in significant industrial spin-offs and externalities. The low industrial electrical costs have attracted many industries, even those that do not process local raw materials (e.g., aluminum). As Hamilton pointed out in chapter 3, Canadian exports are energy intensive. This development of energy-intensive industrialization may present Canada with an opportunity to escape from its traditional dependence on the export of raw materials. More generally, if developing countries could identify, as Francois Perroux suggested, their "poles of development," then they would be able to build development strategies and mobilize resources to dynamically transform their economies in a durable way. THEORETICAL IMPLICATIONS

The French school of economics has for a long time called attention to the fact that economic structures affect the patterns of a nation's trade. The economic structures they refer to are the stable interindustrial relationships that support (or impede) the relative competitiveness of industries. Perroux introduced fourth-year courses on economic structures and systems into the French university curriculum (these are not to be confused with comparative systems courses in North America). His concepts of development poles and growth poles are related to these structural effects. International economics specialist Jean Weiller (1989) suggested that different countries have various "structural preferences"; that is, given that national communities have different collective consumer preferences and industrial strengths, governments favour the development of certain identified structures as a means of sustaining national specificity, originality, and economic power. Specific collective preferences ("non-homotheticity," in the trade theory jargon) result in structural preferences. The new French school of regulation has also taken up similar ideas. Aglieta and Boyer (1985) coined the term "competitive poles." In another vein, Francois Chesnais (1986), writing a committee document on science, technology, and competitiveness for the Organization for Economic Co-operation and Development (OECD), called attention to the structural effects of technology on competition. If one accepts the simple premise that technology is systemic that is, that technology is what determines which products are complements and which are substitutes - one cannot look at the effect of technology on international exchanges without looking at struc-

84 Technology, Energy, International Trade

tural aspects. The dominant technical systems in a country will affect long-term dynamic patterns of trade. Innovation clusters will affect medium- to long-term change in the direction of international exchanges. In order to appreciate the effects of technology on trade, one must look at trade (and other forms of international exchange) in new ways. Since technological know-how is a stock that will upset trading patterns (Johnson 1975), it is insufficient merely to consider the free flow of commodities and factor inputs into these commodities in order to understand the effect of technology on trade. Technology cannot simply be considered as a new factor. Competitive poles, hinging on innovation clusters, may be a useful way to examine the medium- to long-term effects on technology on trade. IMPLICATIONS FOR CANADA

Writing in light of the Depression of the 19305 and more specifically the collapse of the multilateral trading pattern in 1931, Folke Hilgert illustrated how in 1870 trilateral trade networks merged into six or seven exchange poles. Each group of countries had positive exchange balances in certain commodity groups with other groups of countries, but deficits for other products with other sets of countries (Hilgert 1941, 1942). Each group represented a distinct pole of specialization within a multipolar, multilateral world market. When all debts converged on the United Kingdom, it was forced to cease converting the pound; thereafter no currency played the role of international equivalent for exchange until 1948. After the Second World War Bretton Woods, the creation of the International Monetary Fund and the European Payments Union (with its required revolving credits), and an acute postwar "dollar shortage," the us dollar was established as the international equivalent for exchange. With this there evolved a new multilateral exchange pattern.8 This multilateral trading network must have been modified again with the emergence of the OPEC petroleum cartel and new industrial countries. One suspects, however - and this is the relevant point here - that there still exist groups of countries with distinct specialization and exchange patterns. These multipolar relationships may rest on medium- to long-term structural effects, including technological specialization. Conversely, the occasionally shifting specialization, may be a consequence of the emergence of new structural advantages in a country. In Hilgert's The Network of World Trade, Canada (along with Australia, Argentina, Uruguay, New Zealand, and South Africa) was

85 Technological Clusters

classified in a group of countries of recent settlement in a temperate climate that exported raw material staples to Europe and the United States. The United States had emerged out of this group at the turn of the century as a distinct industrial power that exported both raw material staples and transport equipment. Simultaneously, Germany had emerged as an exporter of mechanical engineering know-how and chemicals, along with its more traditional coal staple. Today Canada, perhaps at the end of its specialization in staple trade, may be emerging as a country that exports energy-intensive goods (be they based on domestic or foreign natural resources). This new pattern of Canadian trade appears to be sustained by a long-term tech nological specialization in energy technologies. NOTES

1 I thank Francois Chesnais for calling my attention to this. 2 A more exact measure of the technological innovation matrix would be to separate out those innovations that correspond to patents and inventions from those that do not. 3 As we are comparing firms' (or different countries' industries') patent stocks in identical patent classes, we can assume that identical patenting propensities exist within the same industries and technologies in different countries. The well-known variations of patenting propensities in different industries should not affect the analysis. 4 The economic rationale of Posner's technology-gap theory of trade implies that the variation in the stock of technological knowledge affects the variation in the absolute advantage. Therefore, it is preferable that the theory be tested with time series rather than by cross-sectional analysis. One of the limitations of Posner's theory is that it is not framed with respect to comparative advantage. 5 Posner seems to assume that all the "dynamic scale economies" would be internalized and "accrue to the firm." This is probably just an oversight he made in the course of formulating a new theoretical framework. 6 The innovation matrix was built to reflect the stock of innovation in use in 1979. Because innovations emerging long before 1979 (e.g., the CANDU) were still in use and being improved, it covered a long historical period. The end of the Second World War was chosen, as for the SPRU (Science Policy Research Unit) innovation database, as the natural starting point. But this does not imply that the matrix is amenable to a diachronic analysis. The memories of respondents privilege the recent period. As a consequence, the sparse matrices are denser for the 19705

86 Technology, Energy, International Trade and 19605, but very sparse for the earlier period. A diachronic analysis on the shifting configuration of technological clusters would be possible only with PATDAT or the SPRU innovation database for the United Kingdom. SPRU repeated its survey of 1969 in 1979 and again in 1983, and therefore the time bias of respondents can be assumed to be the same as for the 19605 and 19705 or for the first two four-year periods (1976-79 and 1980-83). 7 By noting an effect that is not accounted for by the internalized stock of technological advance in the industry, I am in fact ascribing the effect to a residual. I am assuming hypothetically that this residual factor could be a technological factor not yet captured, and in order to specify the nature of this technological factor, am assuming further that it is the effect of technological clusters as a structural externality. 8 Unfortunately, since it would provide a useful tool for identifying the tension points in the world system, this new network, in spite of the computerization of trade data, has not been adequately described. REFERENCES

Aglietta, M., and R. Boyer. 1985. "Poles de competitivite, strategic industrielle et politique macro-economique." No. 8222, CEPI-CEPREMAP, Paris. Chesnais, Francois. 1986. "Science, Technology and Competitiveness." Science, Technology & Industry Review (Paris: OECD) 1185-129. Cotsomitis, John. 1989. "Bilateral Trade in R&D Intensive Industries in the OECD = A Test of the Posner-Hufbauer Technology Gap Theory of Trade." Master's thesis, Concordia University, Montreal. DeBresson, C. 1986a. "Conceptual Notes on the Measurement of Innovation." (Paris: OECD.) Mimeo. — 1987. "Poli tecnologici di sviluppo: verso un concetto operative." L'lndustria VIII, no. 3 (June). 301—35. - 1989a. "Breeding Innovation Clusters: A Source of Dynamic Development." World Development 17, no. 1:1—16. - 1989b. "Les poles technologiques du development: vers un concept operationel." Revue Tiers Monde 30, no. 118:245-70. DeBresson, C. and B. Murray. 1984. Innovation in Canada. 2 vols. Reprint. New Westminster Cooperative Research Unit on Science and Technol-

ogyDeBresson, C., B. Murray, and L. Brodeur. 1986b. L'innovation au Quebec. Quebec: Les Publications du Quebec. Ellis., E.D. 1981. "Canadian Patent Data Base - The Philosophy, Construction and Uses of the Canadian Patent Data Base, PATDAT." World Patent Information 3:13-18.

87 Technological Clusters Hilgert, F. 1941. Europe's Trade. Geneva: League of Nations, Economic Intelligence Unit. — 1942. The Network of World Trade. Geneva: League of Nations, Economic Intelligence Unit. Hufbauer, Gary. 1966. Synthetic Materials and the Theory of International Trade. Cambridge, Mass.: Harvard University Press. Jaffe, A.B. 1986. "Technological Opportunity and Spillovers of R&D: Evidence from Firms' Patents, Profits and Market Value." American Economic Review 76, no. 5:984—1001. Johnson, Harry G. 1975. Technology and Economic Interdependence. London: Macmillan for the Trade Policy Research Centre. — 1975. "Technological Change and Comparative Advantage: An Advanced Country's Viewpoint. "Journal of World Trade Law 9:1—14. Nelles, H.V. 1974. The Politics of Development: Forests, Mines and HydroElectric Power in Ontario 1849—1941. Toronto: Macmillan of Canada. Organization for Economic Co-operation and Development (OECD). 1983. "Technological Indicators and the Measurement of Performance in International Trade." DSTi/spR/83-45 A, B, C. — 1984. "Indicators of the Technological Position and Performance in OECD Member Countries during the Seventies." DSTi/SPR/84.43. — 1970. Technology Gaps. Paris. Perroux, F. 1961. L'economie du XXeme siecle. Paris: Presses Universitaires de France. — 1955. "Note sur la notion de pole de croissance." Economic appliquee VIII, nos. 1—2:307—20. Posner, M.V. 1961. "International Trade and Technical Change." Oxford Economic Papers New Series 13:323—41. Rosenberg, N., and Claudio R. Frischtak. 1983. "Long Waves and Economic Growth: A Critical Appraisal." American Economic Review, May, 146-51. Schumpeter, J.A. 1939. Business Cycles. 2 vols. New York: McGraw-Hill. Soete, L. 1981. "A Generalized Test of the Technology Gap Theory of Trade." Welwirtschaftliches Archiv 117, no. 4. Soete, L.G., and S.M.E. Wyatt. 1985. "The Use of Foreign Patenting as an International Comparable Science and Technology Output Indicator." Scientometrics 5, no. i: 31—54. Teece, D. 1980. "Economies of Scope and the Scope of the Enterprise." Journal of Economic Behaviour & Organization i: 2 23—45. Vernon, Raymond, ed. 1970. The Technology Factor in International Trade. New York: National Bureau of Economic Research. Weiller, Jean. 1989. Economic Internationale: Hier et aujourd'hui. Grenoble: Presses Universitaires de Grenoble.

This page intentionally left blank

PART TWO

State, Technology, and Competitiveness

This page intentionally left blank

CHAPTER FIVE

Technological Innovation and International Competitiveness G I O V A N N I DOSI AND LUC SOETE

Soete and Dosi provide a critical examination of the role of technology in current theories of international trade. They reject the dominant neoclassical paradigm for its unrealistic assumptions of identical production functions and tastes among nations, compensating adjustment mechanisms, and perfect competition. They classify the heterodox theories in three major groups: (i) those that accept all but one or two of the assumptions of the neoclassical theories, (2) those whose starting points are international technical disparities among nations, and (3) those that emphasize the international diffusion of technology. While these theories provide a slightly more realistic framework for the analysis of international trade and technological change, they are still far from offering a coherent and integrated theory to replace the dominant neoclassical perspective.

INTRODUCTION

In contrast to many other fields of economic theory, international trade theory has traditionally kept the importance of technical change in explaining international trade flows and the international "competitiveness" of a country or an industry at the centre of much economic debate. This can be explained to a large extent by the almost unique influence of "classical" thinking in the area of international trade, with many contemporary trade theorists expressing, particularly with regard to the technology assumption, strong doubts as to the relevance of the "pure" neoclassical case. The fact that pure trade theory is still so prominent in international trade textbooks and is still held in such esteem by policy makers (at least until recently) has indeed little to do with the way "factor en-

92

State, Technology, and Competitiveness

dowments" (pure Heckscher-Ohlin-Samuelson) trade theory explains international trade flows. Its value as a descriptive theory that is, national differences in endowments of productive factors form the basis for trade - is reputedly very limited. The strength of the pure orthodox theoretical framework lies primarily in the relatively straightforward normative implications (in terms of the gains from trade for both trading partners, as well as international factor price equalization) that can be built around the model. The fact that pure trade theory must therefore rely on a set of "heroic" assumptions is then generally justified in terms of cost/benefit analysis: the insights gained by such a simple but complete trade/welfare picture outstrip by far the disadvantages of more realistic but more complex and less clear analyses. Such a view requires, of course, that the distortions and imperfections of the real world lead only to minor or short-lived aberrations with relatively little consequence for the normative or policy conclusions of the theory. Nowhere is this more clearly illustrated than in the seminal review Hufbauer (1970) presented of the emerging and growing evidence in support of the so-called neo-technology accounts of international trade flows. In interpreting his neo-technology results, Hufbauer, himself author of one of the most detailed technology-gap trade studies on synthetic materials (1966), remained, if anything, rather ambiguous. His neo-technology results, while powerful in explaining the actual trade flows and admittedly closer to the real world, represented an approach which, in his words, was "not geared in answering the traditional questions of economic inquiry." And Hufbauer added with some irony: "It can as yet offer to compare with Samuelson's magnificient (if misleading) factor-price equalisation theorem" (Hufbauer 1970, 197). While Hufbauer's contribution was exceptional in its frankness, it was in no way exceptional in bringing out the dilemma between relevance to and consistency with a general and established theoretical framework that has characterized the analysis of technical change in economic theory. Some authors privilege the first criterion (relevance) and find in the evidence on technological change a powerful challenge pushing toward a search for a radically different theory. As Rosenberg puts it, "in a world where rapid technological change is taking place we may need an analytical apparatus which focusses in a central way upon the process of technological change itself, rather than treating it simply as an exogenous force which leads to disturbances from equilibrium situations and thereby sets in motion an adjustment process leading to a new equilibrium" (Rosenberg 1970, 69—70).

93 Technological Innovation

Conversely, other economists stress as a necessary condition for the theoretical consideration of the phenomena related to technological change precisely their tractability within the traditional model, or they simply consider the absence of any alternative as a sufficient condition for their neglect. In Bhagwati's words, "the realistic phe nomena ... such as the development of new technologies in consumption and production involve essentially phenomena of imperfect competition for which, despite Chamberlin and Joan Robinson, we still do not have today any serious theories of general equilibrium. ... Unless therefore we have a new powerful theoretic system ... we cannot really hope to make a dent in the traditional frame of analysis" (Bhagwati 1970, 23). In these few pages we aim to illustrate that such dents appear more and more clearly. One dent is from within trade analysis itself, another from within diffusion theory, particularly as it applies to international diffusion. TECHNOLOGY AND TRADE THEORY

Consider first the neoclassical pure theory of trade in its simplest textbook form. There are generally four fundamental assumptions: (1) On technology: Differences in technologies can be adequately represented by production functions. The latter are assumed to represent the real world, are well behaved, continuous, differentiable, exhibit non-increasing returns to scale, and so on. Moreover, they are assumed to be identical across countries. (2) On behaviours: Perfect competition prevails throughout. Agents are maximizers under budget constraints. (3) On demand: Tastes are identical across countries and utility functions are well behaved. (4) On adjustment mechanism: Adjustments are such as to guarantee ex hypothesi the clearing of all commodity and factor markets. It is then assumed that hypotheses (i) to (4) offer a reasonably accurate description of the prevailing "state of the world" and the main interdependencies in the international arena. Accordingly, any possible distortions or imperfections of the real world lead only to minor or short-lived aberrations with relatively little consequence for the interpretative and normative conclusions of the theory. In its simplest form, the pure theory of international trade then goes on to prove some the most "classic" theorems of economic the-

94 State, Technology, and Competitiveness

ory: on relative specialization determined by relative factor endowments (Heckscher-Ohlin-Samuelson theorem) 1 ; on factor-price equalization; and on the effects of changing prices on factors' returns (Stolper-Samuelson theorem of comparative statics) and of changing endowments upon commodity outputs (Rybczynski theorem of comparative statics).2 We will not consider here the developments and refinements of all four above hypotheses3 but will limit our review to some of those contributions which have tried to modify some of the assumptions (i) to (4), particularly assumption (i). One way of relaxing the simplest technological assumptions has been to allow production functions to be different between countries. Jones analyses some of the implications: factor-price equalization does not occur any longer, "differential rates of technical differences between countries come to dominate the determination of comparative advantages" (Jones 1970, 84), but the Heckscher-Ohlin theorem on specialization still applies in a modified form. Berglas and Jones (1977) embody in their model a mechanism of learning-by-doing characterized by "local learning" (Atkinson and Stiglitz 1969) of the techniques effectively in use. Findlay (1978) develops a steady-state dynamic model that includes technology transfers between an "advanced" country and a "backward" one. Chipman (1970) considers the case of moving production functions whereby technical progress is itself endogenous along Kennedy—von Weizsacker—Samuelson lines (cf. Kennedy 1964; Von Weizsacker 1965; Samuelson 1965). Purvis (1972) presents a model with international technological differences and capital mobility, illustrating that in this case contrary to the standard model, factor mobility and trade may be complementary. The issue of capital mobility is also considered by Ferguson (1978) and Jones (1980): the patterns of trade now turn out to be essentially determined by technology gaps and relative labour costs. Another way of relaxing the standard assumption with regard to the production function is to introduce economies of scale. Since the analysis of the latter must be generally associated with behavioural assumptions different from the pure competitive model,4 one may consider these two variations on the standard model together.5 First, as Dreze (1960, 1961) and Ohlin (1933) himself, more than fifty years ago, point out, economies of scale taken on their own can be an explanatory variable of trade patterns. Second, from a more normative point of view, they may well influence the welfare effects of trade so that a country may even lose from trade, as suggested originally by Graham (1923).

95 Technological Innovation

More recently, several interesting theoretical developments have been produced in this area (Dixit and Norman 1980; Krugman 19793, 1980). Either (1979, 19823) has explored the conditions under which Graham's arguments hold: they depend on the nature of the increasing returns (which are either "national" or "international") and on the pattern of change in relative prices due to the transition from autarky to trade. "Imperfect competition" due to increasing returns may imply gains from trade for both trading partners (cf. Melvin 1969; Krugman 19793) but may also imply losses. In the case of imperfect competition, a large number of conclusions emerge that may be diametrically in conflict with the standard Heckscher-OhlinSamuelson model.6 For example, factor prices will not be equalized, but on the contrary, the price of the factor used intensively in the production of the export good may actually be high in each country (cf. Markusen and Melvin 1980, 3). Similarly, factor mobility, instead of substituting for trade (trade in factors as opposed to trade in commodities) as in the standard model, will be complementary to trade, with each country achieving an equilibrium where it is well endowed with the factor used intensively in the production of its export good. As Markusen and Melvin note, "In the HeckscherOhlin model this is, of course, the basis for trade whereas in the present model it is the result of trade" (Markusen and Melvin 1980, 3)In general, as shown by Markusen and Melvin (1984) sufficient conditions for the gains-from-trade theorems to hold are, first, on the behavioural side, marginal pricing and, second, on the technological side, the convexity of the production possibility sets. The analysis of differentiated products, on the other hand, has led to attempts to synthesize theories of monopolistic competition and of intra- and inter-industry trade.7 Differentiation is supposed to come from a demand for a variety of product characteristics (cf. Barker 1977; Dixit and Stiglitz 1977; Krugman 19793, 1980, 1981) or from different combinations of some fundamental attributes (cf. Lancaster 1979) embodied in each product. Thus, whereas intraindustry trade is explained on the grounds of monopolistic competition (Grubel and Lloyd 1975), the explanation for the interindustry trade flows is left to the traditional Heckscher-Ohlin model. These models predict that intra-industry trade will be highest between similar countries (in terms of per capita income and patterns of demand [Linder 1961]) whereas inter-industry trade flows will be more important the greater the difference between countries in terms of their "endowments."8 An alternative (Ricardian) model of

96 State, Technology, and Competitiveness

intra-industry trade is provided by Petri (1980), where intra-industrial specialization for any given pattern of demand is determined by relative labour productivities and cost conditions within sectorspecific and country-specific structures of production. Another line of analysis of those market structures different from pure competition has been pioneered by Caves (1971, 1974) in an attempt to link instruments and concepts of industrial organization (multinational corporation, oligopolistic competition, strategic behaviours) with a general equilibrium trade model. A growing literature on industrial organization and international trade has since emerged.9 While some of the results can be formally represented in terms of the traditional model with specific factors,10 this line of enquiry has more clearly drawn attention to the significance of the link between industrial structures and trade flows (given whatever "endowments") and to a different adjustment mechanism (international capital mobility in the form of multinational investment rather than intra-national, intersectoral mobility). This line of analysis therefore allows, at least in principle, the considerations of countryspecific variables, both institutional and economic in nature, which as such represent absolute advantages/disadvantages of, and thus also incentives/obstacles to, the location of international capital (Jones 1980). Under the broad heading of "industrial organization and international trade," one must also mention parts of the vast literature on the origins and effects of multinational corporations. Some of the studies are quite far in spirit and construction from the neoclassical assumptions listed above (e.g., Hymer 1976): technological differences between companies and countries, country-specific absolute advantages, and high degrees of "imperfection" of markets in general and of the market for technology in particular are implicit from the start. These features of the world are indeed the necessary structural conditions for the existence of multinationals. Other interpretative models also try to incorporate some neoclassical elements. This appears to be the case in regard to Dunning's eclectic theory (Dunning 1977, igSia, and igSib; Buckley and Casson 1976) whereby Heckscher-Ohlin mechanisms of adjustment in prices, quantities, and relative specialization are considered as one of the processes at work whose relative importance depends on the sectors, the degrees of development of the countries, and the nature of the technology. Finally, other interpretations - such as that of Rugman (1980) - try to reconcile the existence of multinationals, intra-firm trade, and so on with traditional analysis. Rugman recognizes the widespread existence of imperfections - and thus the limited validity

97 Technological Innovation

of assumptions (i) and (2) above. However, he assumes that companies face and overcome these imperfections by internalizing the relevant transactions. Therefore, multinationals theory become a kind of "second-best approximation" of the standard model. Some models adopt "Ricardian" hypotheses on technology — with coefficients of productions fixed and different between countries while generally retaining general equilibrium assumptions on prices determined through a market-clearing process. Dornbush, Fisher, and Samuelson (1977) present a two-country Ricardian model with a "continuum" of commodities and the patterns of specialization determined by relative wages and relative productivities. Wilson (1980) extends the model to many countries and non-homotetic demand schedules. Jones (1979) considers the conditions under which technical progress may produce "immiserizing growth" for either of the trade partners. A simple but illuminating picture of the technology-trade relationship emerges from Krugman's North-South trade model (19793, 1982). Starting from an innovative North and a non-innovative South, where the North's innovations take the form only of new products produced from the first in the North, but only after a lag in the South, Krugman (19793) shows how new industries have to emerge constantly in the North in order to maintain the North's living standards, since the new industries decline and disappear sooner or later in the face of low-wage competition from the South. In Krugman's model, the reason for this is that the North's wages reflect the rent of the North's monopoly on new technology. "This monopoly is continually eroded by technological borrowing and must be maintained by constant innovation of new products. Like Alice and the Red Queen, the developed region must keep running to stay in the same place" (Krugman 19793, 262). In other words, while the North will be able to achieve some "moving equilibrium" through a rate of innovation large enough to maintain its living standards, any slowing of innovation or acceleration of technology transfer will narrow the wage differentials between North and South and might even lead to an absolute decline in living standards in the North. The most interesting aspect of Krugman's model is, maybe paradoxically, the set of simplistic and, from a traditional trade point of view, totally unrealistic assumptions behind the model: there are no differences in factor endowments, since there is only one factor of production (labour), and all goods, old and new, are produced with the same cost function, leaving no room for differences in labour productivity. Neither neoclassical nor Ricardian trade explanations are relevant, there is no fixed pattern of trade, but trade is deter-

98 State, Technology, and Competitiveness

mined by a continuing process of innovation in the North and technology transfer to the South. Yet, despite these simplifications, some of the conclusions that emerge from the model are very appealing, not in the least because, as Krugman observes, "the picture of trade seems in some ways more like that of businessmen or economic historians than that of trade theorists" (Krugman 19793, 265). It is obviously very difficult to provide a synthetic assessment of all these quite heterogeneous streams of literature, characterized as they are by very different directions and degrees of "revisionism." Three general conclusions, however, may be drawn. First, there is probably little disagreement, even among neoclassical trade theorists, about the inadequacy of the "canonic" factor proportions theory to explain by itself international trade flows. As Krugman puts it, "casual observation seems to militate against a simple factor proportions theory. The emphasis on factor proportions in international trade is ... not the result of an empirical judgement" (Krugman 19790). Second, most of the studies implicitly highlight the lack of robustness of the major Heckscher-Ohlin-Samuelson results in terms of both predictions and welfare implications. Relaxation of the least realistic assumptions (i.e., perfect competition, constant returns to scale, factor immobility, immediate and free diffusion of technology, existence of well-behaved production functions) leads, more often than not, to indeterminate predictions in relation to the direction and volume of trade. Moreover, the factor-price equalization theorem does not generally follow. In terms of welfare implications, depending on which assumption is relaxed, conclusions on the "gains from trade" are sometimes in accordance and sometimes at variance with the orthodox model. Third, and notwithstanding the above, quite interesting results sometimes emerge, despite the continuing presence of highly restrictive assumptions. This set of conclusions might well provide the first pieces of a more substantial dent in the traditional trade theoretical framework: for example, the role of technology gaps; country-specific absolute advantages; different forms of industrial organizations; the importance of economies of scale and of various types of learning; and the absence of any general tendency toward factor price equalization. T E C H N O L O G Y AND THE LESS PURE TRADE THEORY

The discussion so far has focused upon that stream of economic analysis concerned primarily with one theoretical question, namely

99 Technological Innovation

the determinants of specialization, and one functional mechanism, namely the adjustment processes induced in the latter by the interdependencies between markets, both within each country and between countries. It is a line of enquiry that — despite the great differences in the assumptions on technology, demand, and nature of the markets — links Ricardo, the neoclassical school, and all those more recent contributions based on a general equilibrium framework. One of the fundamental premises of such a stream of thought is that trade (or the notional transition from autarky to trade) affects the intersectoral (and, sometimes, international) allocation of inputs, quantities, and prices but does not affect the rate of utilization of the stocks of inputs themselves (and thus the rates of macro-economic activity).11 This is straightforward in modern general equilibrium analysis, where, as already discussed, full employment of all factors is assumed by hypothesis. It is equally true for that part of Ricardo's Principles (Ricardo 1966) concerned with international trade that is based on the assumption that no extension of foreign trade will immediately increase the amount of value in a country, although it will very powerfully contribute to increase the mass of commodities, and therefore the sum of enjoyments. As the value of all foreign goods is measured by the quantity of the produce of our land and labour, which is given in exchange for them, we should have no greater value if, by the discovery of new markets, we obtained double the quantity of foreign goods in exchange of a given quantity of our's. (Ricardo 1966, 128)

Since production techniques are given in Ricardo's model, the assumption concerning an unchanged "amount of value of a country" is precisely equivalent to an assumption of constancy of the rates of macro-economic activity throughout the notional transition from autarky to trade. In the history of economic thought, however, one can also identify another group of contributions, highly heterogeneous in scope and nature, seldom thoroughly formalized, heretic in spirit, and often produced by theorists outside the dominant economic tradition. In this composite group, one may include early economists from the eighteenth and nineteenth centuries, such as Ferrier, List, and Hamilton, as well as Adam Smith in respect to parts of his analysis. In more recent times, one finds an equally heterogeneous set of writers, ranging from some technology-gap and product-cycle authors (Posner, Freeman, Vernon, Hirsch) to Kaldor, Cornwall, and Thirlwall - broadly, in the post-Keynesian tradition, "structuralist" writers in development economics, especially within the Latin American tradition; economic historians, such

ioo State, Technology, and Competitiveness

as Gerschenkron, Kuznets, and Balogh; and some modern French writers such as Bye, de Bernis, Lafay, and Mistral. Obviously, these contributions are highly different in nature and scope. However, one may state that they have in common, explicitly or implicitly, one or several of the following assumptions: • International differences in technological levels and innovative ca pabilities are a fundamental factor in explaining the differences in both levels and trends in export, imports, and income among countries. • General equilibrium mechanisms of international and intersectoral adjustment are relatively weak, and thus trade has important effects upon the rates of macro-economic activity of each economy. Putting it another way, the growth of each economy is often balance-of-payment constrained and this constraint becomes tighter or looser according to the levels and composition of the participation of each country in world trade flows. The weakness of price/quantity adjustments between sectors and between countries has to do partly with the nature of technology (fixed coefficients, irreversibilities, etc.) and partly with the nature of demand (sticky baskets of consumption, etc.). As a result, what adjusts in the international arena is world market shares within each sector and, through that, the levels of macro-economic activity generated by foreign demand. • That same weakness of general equilibrium adjustments is such that the intra-sectoral distribution of trade shares between countries and their evolution through time can be explained by a set of country-specific absolute advantages and, without explicit reference, at least in a first approximation, by price/quantity adjustment between sectors and between factors' returns. • Technology is not a free good. • The allocative patterns induced by international trade have dynamic implications that may yield either "virtuous" or "perverse" feedbacks in the long term. These assumptions have generally been stated in a rather confused way by the early writers, who did not share the rigour and depth of a Ricardo or Samuelson and who were often motivated simply by policy issues, such as protection versus free trade. Nonetheless, these writers had valuable, if confused, insights into complex problems of economic dynamics that were later neglected in the cleaner but more restrictive formalizations of modern theory. For example, Tucker (1774), assuming that there is a macro-economic link between tech-

ioi Technological Innovation

nological advantages, international competitiveness, and incomes, discusses whether the product-cycle effects induced by the lower wages of the "poor country" will eventually reverse the competitive position of the "rich" vis-a-vis the "poor." His answer is reassuring for England: continuous technical progress, higher capabilities of accumulation, and institutional factors will keep an absolute advantage there, despite the lower wages of the more backward countries. Ferrier (1805) deals with the relationship between trade and rates of macro-economic activity in the light of the historical experience of the Continental Blockade, arguing that there is a direct negative link between import penetration and employment levels in a relatively backward country due to a generalized technological disadvantage and to the long-term effects that de-specialization in the most advanced products (in this case, manufactures) exerts upon the capability of a country to progress and accumulate: "I compare a nation which with its money buys abroad commodities it can make itself, although of a poorer quality, with a gardener who, dissatisfied with the fruits he gathers, would buy juicier fruits from his neighbours, giving them his gardening tools in exchange" (Ferrier 1805, 288). Interestingly, Adam Smith (1961, vol. i) was equally aware of the dynamic implications of trade, and his and Ferrier's positions appear almost symmetrical. First, they agree that trade has a beneficial effect upon the rates of macro-economic activity and employment because, in contemporary words, exports increase aggregate demand. This is close to what Myint (1958) later defined as a "vent-for-suplus" trade model. Second, according to both Smith and Ferrier, the enlargement of the market due to international trade feeds back to the domestic division of labour and thus to the trends in productive efficiency. The argument of List (1904) is directly against Ricardo and Say. The practical matter at stake was the political advocacy of protectionism and industrialization. In List's view, there is nothing in the adjustment mechanisms in the international market (in List's terminology, the adjustments "based on the theory of exchange values") that guarantees dynamic convergence between nations in terms of productive capabilities and incomes (the "growth of productive forces of a Nation"). In several respects, this view involves much more than an "infant industry argument," the idea being that the long-term position of each country depends jointly on its degree of capital accumulation, its global technical and learning capabilities,12 and a set of institutional factors (social consensus, factory discipline, political conditions). According to List, the adjustment processes set

102 State, Technology, and Competitiveness

in motion by international trade might well be detrimental to the development of these aspects of the "national production forces." Putting it in modern words, static and dynamic economies of scale and differing income elasticities of the various commodities will lead under free-trade conditions to divergence rather than to factor-price equalization, and to growth polarization with the concentration of production in one country rather than to welfare gains for both partners. In a similar perspective, these points have been emphasized in much of the early development/trade/dependency literature (Prebisch, 1950) and in the historical analysis of the early industrialization/opening-of-trade process in the United Kingdom. More recently and along the lines suggested by Kaldor (1970, 1975, 1980), Thirlwall and Vines (1983) have formalized such views in a multi-sector North-South model and have studied the "consistency conditions" between two countries and the various sectors. The Kaldor-Thirlwall-Vines approach, while incorporating some ideas similar to earlier "two-gap" models of development (whereby the growth of the industrializing countries is shown to be constrained by either saving/investment capacity or by the foreign exchange requirements) 13 embodies the general hypothesis that world growth is determined by "asymmetrical" patterns of change in technical coefficients and demand composition. In this view, processes of interfactoral and intercommodity substitution in response to relative prices and excess factor supplies are of minor importance. Instead, adjustments are made in the level of sectoral and macro-economic activity. An ambitious multi-sector model along similar lines is that of Pasinetti (1981), whose open-economy version determines the difference between the relative rates of growth of economies in terms of the evolution of relative productivities and the income elasticities of the commodities each country produces. In all these models, the difference in the income elasticities of the various commodities plays a fundamental role and is assumed to dominate upon the price/quantity adjustments in consumption baskets. Thus, as Thirlwall (1980) shows, the income elasticities enter into the determination of the foreign-trade multiplier of each economy (via import propensities and export elasticities of world income). The other factor is obviously technology. Polarization in innovativeness is shown to imply polarization in growth. Interestingly, while both the Ricardian and neoclassical perspectives focus upon the determinants of the patterns of specialization, the set of contributions reviewed above focuses on the relationship between trade, levels of activity, and growth. In terms of adjustment mechanisms, both Ricardo and the neoclassical school hold the rates

103 Technological Innovation

of activity constant and study trade-induced changes in relative prices and relative quantities; conversely, the "heretic" stream often assumes away price/quantity adjustments and studies the link between trade and rates of activity in both the short and long term. A major factor counteracting this link between polarization in technology and polarization in income levels is, of course, the international diffusion of technology. Indeed, most modern technology-gap models focus on the crucial time element between innovation and imitation abroad - the trade-and-income—polarizing reversal factor. It is to this issue that we now turn. INTERNATIONAL TECHNOLOGY DIFFUSION

From a totally different theoretical angle - diffusion analysis - a number of contributions have highlighted the relationships between innovativeness, technology diffusion, and international competitiveness. The importance of foreign technology and its international diffusion is undoubtedly a historically well-recognized factor in the industrialization of both Europe and the United States in the nineteenth century, and even more strikingly of Japan in the twentieth century. That importance becomes more evident every day with regard to the rapid industrialization of the so-called newly industrializing countries, such as South Korea, over the last two decades. There exists, of course, a voluminous literature on this subject that has been a focal point of research for economic historians (Landes 1969; Hobsbawn, 1970; Rosenberg 1970). We do not intend to review this literature here. Suffice to say that some convergence between these latter contributions (based on in-depth case studies of the growth and domination in certain countries of the production and use of particular technologies)14 and the more recent international trade and growth models (based on imitation and "catching up")15 appears to emerge. Indeed, the basic assumption of modern technology-gap trade accounts is that technology is not a freely^ instantaneously, and universally available good, but that there are substantial advantages in being first. Thus, in Posner's seminal model, it is suggested that while technical changes and developments may influence some industries and not others, it is the technical change originating in one country and not in others that will induce trade "during the lapse of time taken for the rest of the world to imitate one country's innovation" (Posner 1961, 323). A similar point is made in Freeman's case study of the plastics industry: "Technical progress results in leadership in production in this industry, because patents and commercial secrecy

104 State, Technology, and Competitiveness

together can give the innovator a head start of as much as 10 15 years" (Freeman 1963, 22). Once imitation has taken place, the more traditional factors of adjustment and specialization would again determine trade flows. In Hufbauer's words, "Technology gap trade is ... the impermanent commerce which initially arises from the exporting nation's industrial breakthrough and which is prolonged by static and dynamic scale economies flowing from the breakthrough" (Hufbauer 1966, 23). There is, of course, nothing necessarily "impermanent" about these static and dynamic scale economies. Coupled with new or improved product innovations, they might well lead to a more or less continuous trade flow. Product-life-cycle theories (Hirsch 1965; Vernon 1966) provide an articulated trade picture along similar lines. They also integrate foreign direct investment and view technology as part of a wider set of market-structure factors, including entry, product differentiation/ standardization, and nature of demand. Vernon's original model is primarily demand determined: high levels of income and sophisticated demand patterns induce innovative responses from domestic firms. More recently, the introduction of supply factors has dealt with some of the weaknesses of the original model (see, for a critical assessment, Walker 1979). The contributions here relate primarily to theories of innovation and can be seen to extend post-Schumpeterian "evolutionary" models to the international field, with the focus now on the international technology-diffusion process. This convergence between technology trade theory and diffusion models clearly puts the emphasis back on the historical institutional framework within which the process of imitation/technological catching up takes place; significant factors include the role of historical accidents, the importance of development constraints (be they primarily economic, such as the lack of natural resources, or more political in nature), the role of immigration (Scoville 1951) and other "germ carriers," and the crucial role of governments.16 As a parenthesis, we will first turn to the relevance of the theoretical debate around diffusion models for industrial growth and catching-up arguments, and then, in the following section, we will consider the international dimension of this issue. TECHNOLOGY DIFFUSION MODELS AND INDUSTRIAL GROWTH

There is a striking level of methodological similarity between the traditional epidemic innovation-diffusion model and some of the

105 Technological Innovation

models of industrial growth and economic development developed in the 19305 by Kusnetz (1930) and Schumpeter (1934), among others. This is in many ways not surprising. The concepts of "imitation" and "bandwagons," so crucial to the diffusion literature, have been and still are central in many of the more structural accounts of economic growth, where the S-shaped diffusion pattern is similar to the emergence and long-term rise and fall of industries. An attempt to link the two theories is made in Freeman, Clark, and Soete 1982. They examine the notion of "clusters" of innovations, including the follow-up innovations made during the diffusion period. These clusters are linked to the rapid growth of new industries and will in the extreme case even provide the ingredients for an upswing in overall economic growth. In the more restrictive diffusion terminology, this phenomenon could be viewed as an "envelope" of diffusion curves of a set of closely interrelated clusters of innovations which, occurring within a limited time span, might tilt the economy in the early diffusion phase toward a higher rate of economic performance. Another similarity with diffusion models can be found in Rostow's theory (1960) of the stages of economic growth, again with a distinct S-shaped pattern of take-off, rapid growth in the "drive to maturity," and slower growth in the "age of high mass-consumption" and standardization. Rostow's phases contain many of the S-shaped development patterns assumed to exist for new products, as typified in the marketing and subsequent international trade literature on the product life cycle. Such an argument was also put forward in the mid-19605 by Hirsch (1965), who showed how the relative importance of certain production factors would change over the different phases of the product cycle. Hirsch and after him Vernon (1966) and many other proponents of the product—life-cycle trade theory illustrated how such changes would shift comparative advantage in favour of less-developed countries as products reached the maturity phase. Within the development literature, particularly that of the "dependencia school," this view and especially Rostow's theory were heavily criticized; the mechanistic, quasi-autonomous nature of the process of economic growth assumed by Rostow was even brandnamed as "ahistorical." Interestingly though, the critique on the mechanistic nature of Rostow's growth model finds its reflection in much of the recent diffusion literature that criticizes the "mechanistic, atheoretical" nature of the S-shaped "epidemic" technologydiffusion models. These recent diffusion contributions also provide a number of interesting insights into some of the broader industrial growth theories mentioned earlier. The first area of critique on the standard

io6 State, Technology, and Competitiveness

diffusion model has led to the application of "probit analysis" to develop a new model of inter-firm diffusion. Probit analysis was already a well-established technique in the study of the diffusion of new products among individuals. The central assumption underlying the probit model is that individual consumers (or firms) will be found to own the new product (or adopt the new innovation) at a time when their income (size) exceeds some "critical" level. This critical, or tolerance, income (or size) level represents the actual tastes of the consumers (the receptiveness of the firms), and these tastes can be related to any number of personal or economic characteristics. Over time, however, with an increase in income and assuming an unchanged income distribution, the critical income will fall with an across-the-board change in taste in favour of the new product due to imitation, greater and better information, bandwagon effects, and so forth. The relevance of the probit model for industrial growth theory is self-evident. The concept of a "critical" income per capita level can be directly applied in Rostow's theory of the stages of economic growth by substituting "individuals" with "countries." Different behaviour between countries in their growth performance can thus be explained and expected. Given both the extreme variation in different countries' ability to take risks and "assess new innovations" (the variation in consumer tastes in the probit model) and the extreme degree of income inequality at the world level, it should come as no surprise that industrialization at the world-wide level (diffusion) has been slow and that many poor countries, even with the fall over time in the critical income industrialization level, have not reached the stage of take-off. The second major critique of the standard diffusion model relates primarily to its static nature and the way the diffusion process is reduced to a pure demand-induced phenomenon. Metcalfe (1981) in particular has emphasized the limits of the standard model in this area. There are, as many detailed studies of the innovation process have indicated, plenty of reasons for expecting both innovation and its surrounding economic environment to change as diffusion proceeds. At the technological end, one may expect to see significant improvements in the innovation as diffusion proceeds further. These incremental developments can either be more or less autonomous or be induced by the diffusion process - for example, through user feedback information as well as through the wider application of the innovation to new users requiring better performance and/or more precisely defined quality characteristics. As diffusion proceeds and the specific users' demands become more stringent, the effective use

107 Technological Innovation

of scientific knowledge in improving the performance, quality, and reliability of the innovation can be expected to increase substantially. From an economic perspective, too, the static, demand-focused nature of the standard diffusion model is questionable. In the model developed by Metcalfe, the price of the new innovation is no longer a constant and does not evolve along a particular time path, but is itself determined by the process of diffusion. In addition, the supply of the innovation is limited by productive activity, the rate of increase of which depends on the profitability of producing the innovation. A typical Schumpeterian scenario of an entrepreneurial innovation emerges, the temporary reward consisting of the initial monopoly profits, which are gradually diminished through competition as imitation takes place and "the innovative potential" is exhausted. At the same time though, as the rate of return of the innovation supplier^) falls, "the associated reductions in price increase the profitability of adopting the new innovation," which is thus further diffused. Once the importance of these supply factors is fully recognized, it becomes more apparent how past investment in the old, established technology can slow down the diffusion of the new innovation — past investment not just in physical capital but also in human, even "intellectual," capital. The importance of past investment in, and existing commitment to, the established technology to the slowing down of new technology diffusion points also toward the phenomenon of inter-technology competition and introduces questions of nonergodicity and "locked-in" technological development. We will not enter this discussion here (see Arthur 1985, 1988; DeBresson 1987). However, it should be clear that new technology will compete on disadvantageous terms against existing technology. As Rosenberg (1976) has observed, the diffusion of steam power in the last century was significantly retarded by a series of improvements to the existing water-power technology that prolonged the economic life of the old technology. The "dying" process of a technology is indeed a slow one, with the old-technology firms often living off fully recovered past investments and sometimes being able to underprice the innovation-adopting firms. TECHNOLOGICAL CATCHING UP AND INTERNATIONAL TECHNOLOGY DIFFUSION

The implications for the international diffusion of technology and the potential for technological catching up are straightforward but far-reaching. There is no reason not to expect that the vast majority

io8 State, Technology, and Competitiveness

of new technologies will originate primarily within the technologically most-advanced countries. There are, however, good reasons to expect that the internal national diffusion of such major new technology will be hampered by the various factors mentioned earlier, the new technology competing (in its diffusion) on disadvantageous terms. Thus, the possible previous investment outlays in the existing technology, the commitment to the latter of both management and the skilled labour force, and even the "development" research geared toward improving existing technology might all hamper the diffusion of the new technology to such an extent that it will diffuse more quickly elsewhere, that is, in a country uncommitted, both in terms of actual production and investment, to the old technology. At the same time, as diffusion proceeds, some of the crucial, incremental innovation, resulting, for example, from user-feedback information, will further shift the technological advantage to the country in which the new technology is diffusing more rapidly. The nineteenth-century industrialization of Germany, France, the United States, and a number of smaller European countries, provides ample support for this view. The dramatic change in fortune in the United Kingdom's position — from absolute technological leadership, producing in the mid-nineteenth century more steam engines than the whole of the rest of the world put together - is a powerful illustration of this phenomenon. This then underlines the significant advantages of "late industrializers," both in terms of catching up with present technological leaders and in terms of acquiring foreign technology at a more competitive price. In recent times, this has been most obvious in the case of Japan in the 19608 and 19705 and South Korea in the igSos: through their rapid industrialization, largely on the basis of initially imported technology, world "best-practice" productivity levels were achieved over a very short time in steel, cars, electronics, numerically controlled machine tools, and, most recently computers. The scarcity of such successful examples, however, illustrates how non-automatic and exceptional such processes of effective technological catching up are. As straightforward as might appear the use of foreign, imported technology as a short cut to industrialization, as difficult and complex is the effective assimilation of foreign technology. As already hinted at above, the country's, and the domestic firms', absorptive capacity is crucial. As illustrated in Perez and Soete (1988), a central problem in the previous discussion on the international diffusion of technology is the almost automatic interuse of the words "use," in the narrow diffusion sense, and "entry," in the broader technological sense. Here

iog Technological Innovation

we are interested in a country's technological catching up, in its effective use of foreign technology with the aim of producing it one day. We are interested in the acquisition and assimilation of technology. From a less-developed country or from any other lagging position, this can in the first instance only be done through the use of foreign technology. Diffusion theory gives us hints as to why such international diffusion is not automatic and why there are delays in the adoption of new technology. As we saw in the previous section, the probit model emphasizes the threshold level below which no adoption will take place. Below this level, there will, almost by definition, be no acquisition of technology, learning, or catching up. CONCLUSION

In conclusion, it might be useful to make a brief synthesis of the major gaps in our understanding of the role of technological innovation in international competitiveness that are evident from the critical literature survey undertaken here. First, there are the numerous theoretical issues, some of which seem to have come to the attention of trade theorists in recent years. Among the crucial issues, there are, in our view, • the determinants of different national capabilities to innovate and imitate and the need to account for such continuous but internationally uneven flows of product and process innovations; • the analysis of the adjustment mechanisms within and between countries following such innovative processes; • the relationship between sector-specific patterns of competitiveness and "general equilibrium" factors in the broader sense that are linked to relative prices, intersectoral capital mobility, etc; • the analysis of economies of scale, oligopolistic forms of market organization, international investment, and all other factors that generally go under the heading "imperfect competition" within an international framework; and • the long-term relationships between innovation, trade, and growth and, more specifically, the possible extensions of dynamic Schumpeterian evolutionary models to the international trade field. Second, there are the no-less-crucial normative implications of the various alternative trade theories, which have so far only rarely been analysed. This is certainly the case for both the technology-gap and Schumpeterian accounts of international trade flows. This is surprising in view of the emerging agreement among several economists

no

State, Technology, and Competitiveness

and policy makers about the crucial advantages of innovation for international competitiveness. These normative issues are also of importance in assessing the objective grounds for the worries expressed within several OECD countries about the increased diffusion of technology to less- and not-so-less-developed countries, or, conversely, in assessing the importance of the polarization effect induced by new technologies on international production and trade patterns. Among the major questions that have direct normative implications are the following: • the relationship between present patterns of international competitiveness and long-term innovative capabilities; • the impact of innovation-based trade upon rates of macroeconomic activity and levels of per capita income; and • the notion of international efficiency in an ever-changing and imperfect world. Finally, there are the major empirical gaps that remain in our understanding of the importance and role of innovation in presentday international trade flows, which have not been discussed here. NOTES

1 That is, the Heckscher-Ohlin theorem, which states that the relative specialization of each country is in those commodities that use intensively those factors that are relatively abundant in that same country. 2 The so-called Stolper-Samuelson and Rybczynski theorems. 3 Such as, for example, the analytical treatment of those cases with more commodities than factors, etc. 4 Of course, this is necessarily so if the economies of scale are internal to each firm. 5 For a thorough review, see Helpman 1984. An interesting collection of some of the state-of-the-art contributions in the field is in Kierzkowski 1984. 6 For "imperfect competition" models, among others, see Markusen 1980, Lancaster 1980, Helpman 1981, Helpman and Razin 1980, Melvin and Warne 1973. The implications of economies of scale in a neoclassical open-economy growth model are analysed in Krugman 19843. 7 For a recent overview, see Greenaway and Milner 1986. 8 See Helpman 1984. This line of enquiry is in many ways an attempt to synthesize the Heckscher-Ohlin-Samuelson model and Linder's

111

9

10 11 12 13

14 15 16

Technological Innovation

model (cf. Linder 1961). For a model accounting also for multinational investment, see Helpman 1984. Cf. the special issue of Journal of Industrial Economics edited by Caves (1980), Brander 1981, Jacquemin 1982, Soete 1978, Brander and Krugman 1983, Dosi 1984, Momigliano and Dosi 1983, and Caves, Porter, and Spence 1980. That is, a general equilibrium model with sector-specific and intersectoral immobile factors (see Jones and Neary 1984). Obviously, this assumption is necessary in order to base the analysis on unit functions, indifference curves, isoquants, etc. For a reappraisal of List's view on the importance of the national techno-scientific system, see Freeman 1982. See Chenery and Bruno 1962, Chenery and Strout 1966, and Findlay 1973. For a thorough critical analysis of the debate on North-South differences, terms of trade, and development, see Bacha 1978. For a review of the trade/development literature, see Findlay 1984. See in particular Ames and Rosenberg 1963, Habakkuk 1962, and von Tunzelmann 1978. See in particular Posner 1961; Freeman 1963, 1965; Gomulka 1971; and Cornwall 1977. For a broad overview, see Yakushiji 1986. REFERENCES

Ames, E., and N. Rosenberg. 1963. "Changing Technological Leadership and Industrial Growth." Economic Journal 73:214-236. Arthur, W.B. 1985. "Competing Technologies and Lock-in by Historical Events." Center for Economic Policy Research Paper 43, Stanford University. - 1988. "Competing Technologies: An Overview." In Technical Change and Economic Theory, edited by G. Dosi et al. London: Pinter. Atkinson, A. and J. Stiglitz. 1969. "A New View of Technological Change." Economic Journal 79:573—8. Bacha, E. 1978. "An Interpretation of Unequal Exchange from PrebischSinger to Emmanuel. "Journal of Development Economics 5:319—38. Barker, T. 1977. "International Trade and Economic Growth: An Alternative to the Neoclassical Approach." Cambridge Journal of Economics 1:153-72. Berglas, E., and R.W. Jones. 1977. "The Export of Technology." In Optimal Policies, Control Theory and Technology Exports, edited by K. Brunner and A. Meltzer. Carnegie-Rochester Conference on Public Policy. Bhagwati, J.N. 1970. "Comment." In The Technology Factor in International Trade, edited by R. Vernon. New York: Columbia University Press.

112

State, Technology, and Competitiveness

Brander, J.A. 1981. "Intra-Industry Trade in Identical Commodities." Journal of International Economics 11:1—14. Brander, J.A., and P.R. Krugman. 1983. "A Reciprocal Dumping Model of International Trade." Journal of International Economics i3:3 1 3-2iBranson, W., and N. Monoyios. 1977. "Factor Inputs in us Trade."Journal of International Economics 7:111—31. Buckley, P., and M. Casson. 1976. The Future of the Multinational Enterprise. London: Macmillan. — 1981. "The Optimal Timing of a Foreign Direct Investment." Economic Journal 91:75~87. Caves, R.E. 1971. "International Corporations: The Industrial Economics of Foreign Investment." Economica 38:1-27. - 1974. "International Trade, International Investment and Imperfect Markets." Special Papers in International Economics, no 10, International Finance Section, Princeton University. - 1980. "International Trade and Industrial Organisation Introduction." Journal of Industrial Economics 29:113—19. Caves, R.E., M.E. Porter, and A.M. Spence. 1980. Competition in the Open Economy: A Model Applied to Canada. Cambridge, Mass.: Harvard University Press. Chenery, H.B., and M. Bruno. 1962. "Development Alternatives in an Open Economy: The Case of Israel." Economic Journal 72:79—103. Chenery, H.B., and A. Strout. 1966. "Foreign Assistance and Economic Development." American Economic Review 56:679-733. Chipman, J.S. 1970. "Induced Technical Change and Patterns of International Trade." In The Technology Factor in International Trade, edited by R. Vernon. New York: Columbia University Press. Cornwall, J. 1977- Modern Capitalism: Its Growth and Transformation. London: Martin Robertson. DeBresson, C. 1987. "The Evolutionary Paradigm and the Economics of Technological Change." Journal of Economic Issues 21, no. 2:751—62. Dixit, A., and J.E. Stiglitz. 1977. "Monopolistic Competition and Optimum Product Diversity." American Economic Review 67:297—308. Dixit, A., and V. Norman. 1980. The Theory of International Trade. Cambridge: Cambridge University Press. Dornbusch, R., S. Fisher, and P.A. Samuelson. 1977. "Comparative Advantage, Trade and Payments in a Ricardian Model with a Continuum of Goods." American Economic Review 67, no. 5:823—839. Dosi, G. 1984. Technical Change and Industrial Transformation. London: Macmillan. Dosi, G., K. Pavitt, and L. Soete. 1990. The Economics of Technical Change and International Trade. Brighton: Wheatsheaf.

113 Technological Innovation Dreze, J. 1960. "Quelques reflexions sereines sur 1'adaptation de 1'industrie beige au Marche commun." Comptes rendus des travaux de la Societe royale d'economie politique de Belgique, no. 275 (December). - 1961. "Les exportations intra-CEE en 1958 et la position beige." Recherches economiques de Louvain 27:717—38. Dunning, J.H. 1977. "Trade, Location at Economic Activity and Multinational Enterprises: A Search for an Eclectic Theory." In The International Allocation of Economic Activity, edited by B. Ohlin et al. London: Macmillan. - 19813. "Explaining the International Direct Investment Position by Countries: Toward a Dynamic or Development Approach." Welwirtschaftliches Archiv \ 17:30—64. — igSib. International Production and the Multinational Enterprise. London: Allen and Unwin. Ethier, W. 1979. "Internationally Decreasing Costs and World Trade." Journal of International Economics 9:1—24. - 1981. "A Reply to Professors Metcalfe and Steedman",/OMrna/ of International Economics 11:273—7. — 19823. "Decreasing Costs in International Trade and Frank Graham's Argument for Protection." Economica 72. - ig82b. "National and International Returns to Scale in the Modern Theory of International Trade." American Economic Review 72:389-405. Ferguson, D. 1978. "International Capital Mobility and Comparative Advantage. "Journal of International Economics 8:373—96. Ferrier, F. 1805. Du gouvernement considere dans ses rapports avec le commerce. Paris. Findlay, R. 1973. International Trade and Development Theory. New York: Columbia University Press. — 1984. "Growth and Development in Trade Models." In Handbook of International Economics, edited by R.W. Jones and P.B. Kenen. Amsterdam: Elsevier-North-Holland. Freeman, C. 1963. "The Plastics Industry: A Comparative Study of Research and Innovation." National Institute Economic Review, no. 26:22— 62. — 1965. "Research and Development in Electronic Capital Goods." National Institute Economic Review, no. 34:40—97. — 1982. The Economics of Industrial Innovation. London: Pinter. Freeman, C., J. Clark, and L. Soete. 1982. Unemployment and Technological Innovation: A Study of the Long Waves and Economic Development. London: Pinter. Gomulka, S. 1971. Inventive Activity, Diffusion and the Stages of Economic Growth. Working papers of the Institute of Economies, Aarhus University, no. 24.

114 State, Technology, and Competitiveness — 1978. "Growth and the Import of Technology: Poland 1971—1980." Cambridge Journal of Economics 2:1—16. Graham, E. 1979. "Technological Innovation and the Dynamics of the us Comparative Advantage." In Technological Innovation for a Dynamic Economy, edited by C. Hill and J. Utterback. New York: Pergamon Press. Graham, F.D. 1923. "Some Aspects of Protection Further Considered." Quarterly Journal of Economics 37:199—227. Greenaway, D., and C. Milner. 1986. The Economics of Intra-Industry Trade. Oxford: Blackwell. Grubel, H.G., and P.J. Lloyd. 1975. Intra-industry Trade: The Theory and Measurement of International Trade in Different Products. London: Macmillan. Habakkuk, H. 1962. American and British Technology in the Nineteenth Century. Cambridge: Cambridge University Press. Helpman, E. 1981. "International Trade in the Presence of Product Differentiation, Economies of Scale and Monopolistic Competition: A Chamberlain-Heckscher-Ohlin Approach. "Journal of International Economics 11:305—40. - 1984. "Increasing Returns, Imperfect Markets, and Trade Theory." In Handbook of International Economics, edited by R.W. Jones and P.B. Kenen. Amsterdam: Elsevier-North-Holland. Helpman, E., and P. Krugman. 1985. Market Structure and Foreign Trade: Increasing Returns, Imperfect Competition and the International Economy, Brighton: Wheatsheaf. Helpman, E., and A. Razin. 1980. "Monopolistic Competition and Factor Movements." Seminar paper, no. 155, Institute for International Economic Studies, University of Stockholm. Hirsch, S. 1965. "The us Electronics Industry in International Trade." National Institute Economic Review, no. 34:92—7. Hobsbawn, E. 1970. Industry and Empire. London: Penguin. Hufbauer, G. 1966. Synthetic Materials and the Theory of International Trade. Cambridge, Mass.: Harvard University Press. — 1970. "The Impact of National Characteristics and Technology on the Commodity Composition of Trade in Manufactured Goods." In The Technology Factor in International Trade, edited by R. Vernon. New York: Columbia University Press. Hymer, S.H. 1976. The International Operations of National Firms: A Study of Direct Foreign Investment. Cambridge, Mass.: MIT Press. Jacquemin, A. 1982. "Imperfect Market Structure and International Trade - Some Recent Research." Kyklos 35:75-93. Jones, R.W. 1970. "The Role of Technology in the Theory of International Trade." In The Technology Factor in International Trade, edited by R. Vernon. New York: Columbia University Press.

115 Technological Innovation — 1979. International Trade: Essays in Theory. Amsterdam: Elsevier-NorthHolland. — 1980. "Comparative and Absolute Advantage." Schweizerische Zeitzchrift fur Volkswirtschaft und Statistik. Jones, R.W., and J.P. Neary. 1984. "The Positive Theory of International Trade." In Handbook of International Economics, edited by R.W. Jones and P.B. Kenen. Amsterdam: Elsevier-North-Holland. Kaldor, N. 1970. "The Case for Regional Policies." Scottish Journal of Political Economy 17:337-48. - 1975. "What Is Wrong with Economic Theory." Quarterly Journal of Economics. Vol. 89:337—348. - 1980. "The Role of Increasing Returns, Technical Progress and Cumulative Causation in the Theory of International Trade." Paris: Institut de sciences mathematiques et d'economie appliquees (ISMEA). Mimeo. Kemp, M. 1969. The Pure Theory of International Trade and Investment. Englewood Cliffs, N.J.: Prentice-Hall. Kennedy, C. 1964. "Induced Buyers in Innovation and the Theory of Distribution." Economic Journal 74:541—47. Kierzkowski, H., ed. 1984. Monopolistic Competition and International Trade. Oxford: Clarendon Press. Klein, B. 1979. Dynamic Competition. Cambridge, Mass.: Harvard University Press. — 1979. "The Slowdown in Productivity Advances: A Dynamic Explanation." In Technological Innovation for a Dynamic Economy, edited by C. Hill and J. Utterback. New York: Pergamon Press. Koizumi, I., and Kopecky, K. 1980. "Foreign Direct Investment, Technology Transfer and Domestic Employment Effects. "Journal of International Economics 10:1—20. Krugman, P. 19793. "A Model of Innovation, Technology Transfer and the World Distribution of Income." Journal of Political Economy 87:253— 66. — i97gb. "Increasing Returns, Monopolistic Competition and International Trade." Journal of International Economics 9:469—79. — 1980. "Scale Economies, Product Differentiation and the Pattern of Trade." American Economic Review 70:950—9. - 1981. "Intra-Industry Specialization and the Gains from Trade."Journal of Political Economy 89:959—73. - 1982. "A Technology Gap Model of International Trade." Paper presented at the International Economic Association Conference on Structural Adjustment in Trade-Dependent Advanced Economies, Yxtahohn, Sweden. — 1984. "Import Protection as Export Promotion: International Competition in the Presence of Oligopoly and Economies of Scale." In Monopo-

116 State, Technology, and Competitiveness listic Competition and International Trade, edited by H. Kierzkowski. Oxford: Clarendon Press. Kuznets, S. 1930. Secular Movements in Production and Prices. Boston: Houghton Mifflin. Lancaster, K. 1979. Variety, Equity and Efficiency. New York: Columbia University Press. - 1980. "Inter-Industry Trade under Monopolistic Competition."Journal of International Economics 10:151—75. Landes, D. 1969. The Unbound Prometheus. Cambridge: Cambridge University Press. Linder, S.B. 1961. An Essay on Trade and Transformation. New York: Wiley. List, F. 1904. The National System of Political Economy. London: Longmans. English translation from German original, 1844. Markusen, J. 1980. "Trade and the Gains from Trade with Imperfect Competition." Seminar paper, no 153, Institute for International Economic Studies, University of Stockholm. Mimeo. Markusen, J., and J. Melvin. 1980. "Trade, Factor-Prices and the Gains from Trade with Increasing Returns to Scale." Institute of International Economic Studies, Seminar Paper 153, University of Stockholm, Stockholm. Mimeo. — 1984. "The Gains-from-trade Theorem with Increasing Returns to Scale." In Monopolistic Competition and International Trade edited by H. Kierzkowski. Oxford: Clarendon Press. Melvin, J., and R.D. Warne. 1973. "Monopoly and the Theory of International Trade." Journal of International Economics 3:45—72. Melvin, J. 1969. "Increasing Returns to Scale as a Determinant of Trade." Canadian Journal of Economics 3:389—402. Metcalfe, J.S. 1981. "Impulse and Diffusion in the Study of Technological Change." Futures 43:347—59Metcalfe, J.S., and L. Soete. 1984. "Notes on the Evolution of Technology and International Competition." In Science and Technology Policy in the icjSo's and Beyond, edited by M. Gibbons et al. London: Longmans. Momigliano, F., and G. Dosi. 1983. Tecnologia e Organizzazione Industrial Internazionale. Bologna: II Mulino. Myint, H. 1958. "The Classical Theory of International Trade and the Underdeveloped countries". Economic Journal 68:317—37. Nelson, R.R. 1968. "A 'Diffusion' Model of International Productivity Differences in Manufacturing Industry." American Economic Review 58, no. 5:1219-48. Nelson, R.R., and S. Winter. 1982. An Evolutionary Theory of Economic Change. Cambridge, Mass.: Harvard University Press, Belknap Press. Nelson, R.R., S. Winter, and H. Schuette. 1976. "Technical Change in an Evolutionary Model." Quarterly Journal of Economics 90:90-118.

117 Technological Innovation Ohlin, B. 1933. Interregional and International Trade. Rev. ed. Cambridge: Cambridge University Press, 1967. Pasinetti, L.L. 1981. Structural Change and Economic Growth. Cambridge: Cambridge University Press. Perez, C., and L. Soete. 1988. "Catching Up in Technology: Entry Barriers and Windows of Opportunity." In Technical Change and Economic Theory, edited by G. Dosi et al. London: Pinter. Petri, P. A. 1980. "A Ricardian Model of Market Sharing." Journal of International Economics 10:201—11. Posner, M. 1961. "International Trade and Technical Change." Oxford Economic Papers 13:32 3—41. Prebisch, R. 1950. The Economic Development of Latin America and Its Principal Problems. New York: Economic Commission for Latin America, United Nations. Purvis, D.D. 1972. "Technology, Trade and Factor Mobility." Economic Journal 82:991—9. Ricardo, D. 1966. On the Principles of Political Economy and Taxation. Edited by P. Sraffa. Cambridge: Cambridge University Press. Rosenberg, N. 1970. "Comments." In The Technology Factor in International Trade, edited by R. Vernon. New York: Columbia University Press. — 1976. Perspectives on Technology. Cambridge: Cambridge University Press. Rostow, W.W. 1960. The Stages of Economic Growth. London: Cambridge University Press. Rugman, A.M. 1980. "Internationalization as a General Theory of Foreign Direct Investment: A Reappraisal of the Literature." Weltwirtschcftliches Archiv 112:210—234. Samuelson, P.A. 1965, "A Theory of Induced Innovation Along Kennedy-Weizsacker Lines." Review of Economies and Statistics 47:343~ 56. Schumpeter, J. 1934. The Theory of Economic Development. Cambridge, Mass.: Harvard College. — 1939. Business Cycles: A Theoretical, Historical and Statistical Analysis of the Capitalist Process. 2 vols. New York: McGraw-Hill. Scoville, W. 1951. "Minority Mitigation and the Diffusion of Technology." Journal of Economic History 11 ^47—60. Smith, A. 1961. The Wealth of Nations, London: Methuen. Soete, L. 1978. Inventive Activity, Industrial Organization and International Trade, Brighton, sn. - "Technological Dependency: A Critical View." In Dependency Theory: A Critical Reassessment, edited by D. Seers. London: Pinter. Thirlwall, A.P. 1980. Balance-of-Payment Theory and the United Kingdom Experience. London: Macmillan.

n8 State, Technology, and Competitiveness Thirlwall, A.P., and D. Vines. 1983. "A General Model of Growth and Development on Kaldorian Lines." Paper presented at the conference on "The Dynamics of Employment and Technology," Udine, Italy, 1—3 September. Mimeo. Tucker, J. 1774. Four Tracts, Together with Two Sermons on Political and Commercial Subjects. Gloucester. Vernon, R. 1966. "International Investment and International Trade in the Product Cycle." Quarterly Journal of Economics 80:190—207. von Tunzelmann, N. 1978. Steam Power and British Industrialization to 1860. Oxford: Clarendon Press. Von Weizsacker, C.C. 1965. "Tentative Notes on a Two-Sector Model with Induced Technical Progress." Review of Economic Studies 32:85— 104. Walker, W. 1979. Industrial Innovation and International Trading Performance. Greenwich, Conn.: JAI Press. Wilson, C. 1980. "On the General Structure of Ricardian Models with a Continuum of Goods." Econometrica 48:1675—702. Yakushiji, T. 1986. "Technological Emulation and Industrial Development." Paper presented at the Conference on Innovation Diffusion, Venice, 17—21 March.

CHAPTER SIX

The State and International Trade: Technology and Competitiveness JORGE NIOSI AND PHILIPPE FAUCHER

Because international trade has evolved since the nineteenth century, so should neoclassical theory. Since the Second World War, industrial development has rapidly evolved from labour-intensive activities to capital-intensive and more rfcently to knowledge-intensive activities. As the weight of static natural advantage has diminished, that of the dynamic socially produced advantages has grown. As convincingly demonstrated by the booming industrial latecomers, the state is a key actor in the creation of dynamic comparative advantages. The state's role includes the active promotion of technological change and its diffusion in the productive structure. Theories of international trade, as well as those of international investment, are mostly theories of pure economics. In other words, they consider exclusively, and take as granted, market forces as they express themselves through supply and demand. Other factors, such as institutional actors, are treated as exogenous variables in the models. Because of their unpredictable behaviour, such factors are considered residual and ultimately a hindrance to the genuine expression of demand and supply. The state is certainly the single most important exogenous actor that affects trade. Still influenced by nineteenth-century reality, economists concentrate on commercial policy where state intervention is associated with the medieval habit of erecting huge walls (Chinese style) to prevent foreign invasions. Barriers meant to prevent foreign competition result in modifying trade flows in a "negative" sense through preferential tariffs, free-trade zones, quotas, and non-tariff protectionism. The state also manipulates trade by modifying the exchange rates of its own currency, thus changing the relative prices of the goods and services produced in the country compared with foreign goods and services.

120 State, Technology, and Competitiveness International trade theory most often assumes a given endowment of factors for each country, perfect competition, and international immobility of factors. Given these assumptions, the state can certainly not be seen as a producer of comparative advantages in itself, an independent actor capable, and willing, to modify the offer of productive factors in the national economy and, in this way, change, in the long run, the flow of a country's exports and imports. Similarly, no market failure can induce government intervention on the factors' markets. The whole issue of the regulation of international technology trade and multinational corporations as well as national policies on labour migration therefore becomes irrelevant for trade theory. We believe that the current framework in which trade theory is being conceptualized fundamentally limits its capacity to contribute to the understanding and shaping of our international economic environment. The central role of the state and state policy in modifying the availability of factors, and in framing the institutional context in which both international trade and the national production of factors take place, should not only be acknowledged, but be incorporated as a central part of the theory. This chapter is organized in three sections. The first enumerates some key trends of postwar international trade that run against the received international trade theory. The second recalls briefly some elements in the development of this theory. The third points out some of the economic policies of the state that have a crucial influence on the technology endowment of modern societies. We conclude on the necessity of "bringing the state back" into economic analysis; while this would mean some degree of mathematical sophistication would eventually be lost, both predictive power and theoretical strength would be gained.

WHEN THEORY FLIRTS WITH DOGMATISM

Basic epistemology teaches that, in empirical science, theories are abandoned when they are unable to integrate a disproportionate number of observations (Bunge 1967). This is what normally happens in physics, biology, or chemistry. Unfortunately, it does not happen in the social sciences, beginning with economics. International trade theory is not an exception to this unhappy situation: mounting massive evidence shows the inadequacy of received generalizations. Nevertheless, academic theory seems generally unaffected by this empirical literature based on observation. Let us summarize some major trends of the postwar period that can not adequately find their place in international trade theory.

121 The State and International Trade Table 27 Quantum Index of Exports by Large Groups of Products: Market Economies, 1948-1983 (selected years; 1970 = 100)

1948

1958

1970

1975

1980

1983

All products

22

38

100

132

164

164

Food & raw materials Fuel Manufactured goods

36 18 16

55 39 32

100 100 100

120 112 139

161 123 196

166 90 210

Exports

Source: UN, Statistical Yearbook, 1977, 1984.

Any basic analysis of the evolution of international trade in the last forty years shows the declining share of commodities and the increasing proportion of manufactured goods in world commerce. Commodities have declined relatively in volume, in price (with the exception of fuels since 1973), and in aggregate value (see tables 27, 28, and 29). In manufactured goods, capital- and knowledge-intensive products show the most rapid rates of growth in world trade (see Table 30). These trends, which recall Prebisch's analysis (1957) on low revenue elasticity of food and raw materials, strongly point to the fact that countries specializing in commodities (and, therefore, making intensive use of their natural, static advantages) are bound to decline in world trade. On the other hand, countries making use of their dynamic, variable comparative advantages (capital and technology) will represent an increasing share of world commerce. During the whole postwar period, trade was rapidly liberalized in the capitalist world. Tariff barriers were reduced through GATT (General Agreement on Tariffs and Trade) negotiations, trade zones were abolished, and colonial empires and preferences disappeared almost entirely. The creation and the constant strengthening of the European Economic Community (EEC) is certainly the most spectacular illustration of this new emerging order. It is true that the trade negotiations produced some unequal results; the most important changes took place during the late 19608 and early 19708 as a direct consequence of the Kennedy Round. Protection rates for the manufacturing industry in Britain fell from 15.8 to 9.2 percent between 1968 and 1972 (Oulton 1976, 66). For Germany it went from 14.8 to 10 percent between 1964 and 1972 (Hiemens and Rabenau 1976, 23-5), and for Canada the decline has been from 17 to 12.7 percent between 1961 and 1970 (Wilkinson and Norrie 1975, 49).

122

State, Technology, and Competitiveness

Table 28 Value of Exports in Current us Billion Dollars by Large Groups of Products: Market Economies, 1948-1983 (selected years)

1948

1958

1970

1975

1980

1983

All products

54

96

279

788

1,826

1,609

Food and raw materials Fuels Manufactured goods

25 5 24

34 11 50

67 26 182

155 155 470

309 434 1,047

272 322 992

Exports

Source: UN, Statistical Yearbook, 1977, 1984.

The trend certainly continued with the conclusion of the Tokyo Round treaty. As a consequence, the weighed average tariff for all industrial products for seventeen OECD (Organization for Economic Co-operation and Development) countries dropped from about 7 percent to 4.5 percent (OECD 1983, 101). Parallel to the dismantling of the traditional methods of state intervention in trade was the rise of new types of government intervention, called "positive adjustment policies" (OECD 1983). These were designed to improve the competitiveness of the industrialized capitalist economies in the newly liberalized world market (Blais and Faucher 1981). The postwar period was one of sweeping growth in foreign direct investment (FDI) by corporations of the industrial (and, recently, the newly industrializing) countries. FDI rose from $25 billion in 1951 to $549 billion in 1984. Not only were multinational corporations (MNCS) moving factors (capital, technology, and skilled labour) from country to country, they were also responsible for an overwhelming part of the exports of any market economy. An increasing proportion of world trade is intra-firm; intra-firm prices for technology, products, and services are administered but not market prices (Murray 1981). In 1977, for example, 48 percent of American imports and 44 percent of American exports were done through intra-firm transactions, both by us corporations with foreign affiliates and by foreign transnational with us subsidiaries (UN 1983, 423). This figure is higher for commodity exports (most of which are made through intra-firm trade). Since the Second World War, FDI by the multinational corporations of developed countries is mostly directed to other industrialized countries. Resource-rich and labour-rich in less-developed countries

123 The State and International Trade Table 29 Unit Value Index of Exports: Market Economies, 1948—1983 (selected years) Exports All products Food and raw mat. Fuels Manufactured goods

1980

1983

1948

1958

1970

1975

90

90

100

213

400

102 104 84

91 111 86

100 100 100

191 541 185

286

243

1,429

1,343

294

259

356

Source: UN, Statistical Yearbook, 1977, 1984.

only attract a small (and decreasing) proportion of this massive FBI. Instead, MNCS invest in search of markets. No natural advantage (with the exception of a small number of strategic raw materials) attracts international investment in sufficient amounts to these lessdeveloped countries. More than 70 percent of FDI of all market economies, except for Japan, was located in other developed countries (OECD 1981; UN 1983). But even for Japan, the figure is increasingly similar to that of other industrialized countries, since the recent Japanese surge in FDI is essentially located in North America and Western Europe (Keizai 1986). Since the 19505, the world has witnessed a reduction in the international gaps in industrial and technological capabilities. The overwhelming American leadership of the 19405 and 19505 is now history. Not only have Europe and Japan caught up with the United States in most industries, but some resource-poor newly industrializing countries (NICS) have taken advantage of their cheap labour force and, through a massive effort of industrial planning and technical education, have mastered Western technology in several key manufacturing sectors (Bellon and Niosi 1987). These NICS are rapidly reducing their industrial backwardness through the careful implementation of industrial policies that are designed to optimally allocate their scarce resources and avoid capital flight and brain drain, the common afflictions of most capitalist underdeveloped countries. The implementation of these industrial policies by Japan and the most successful NICS in the postwar period, and the adjustment policies that older advanced countries had to apply in order to compete, has brought to the forefront the inadequacy of laissezfaire theories and recipes. Not only are industrially successful countries not free-traders, but in the strict sense of the term, no country is a free-trader any more.

124 State, Technology, and Competitiveness Table 30 Recent Trends in High-Technology World Trade: Market Economies, 1970-1984 (billions of current us dollars) Export

1970

1980

1984

Telecommunications equipment Aircraft Automatic data-processing equipment Electronic microcircuits Medical and pharmaceutical products

4,712 4,128 n.a. n.a. 2,686

9,768 25,042 12,557 4,591 13,918

22,802 24,992 24,206 1,371 14,716

Source: UN, Yearbook of International Trade Statistics, 1972, 1982, 1986.

THEORIES ON INTERNATIONAL TRADE: THE M A R G I N A L ROLE OF THE STATE Arghiri Emmanuel has pointed out that the classical and neoclassical theory of international trade (the Ricardo-Mill-Marshall-HeckscherOhlin doctrine) easily survived through more than 150 years, while most other aspects of the received economic doctrine were objects of intense dispute and debate (Emmanuel 1969, 23). This is probably less true today than it was in the 19605, when Emmanuel wrote his book on unequal exchange. Stimulated by the "Leontief paradox" and by the emerging theory on international investment and multinational corporations (associated with, among others, Raymond Vernon, Stephen Hymer, and Charles Kindleberger), extensive theoretical and empirical research took place in the United States and elsewhere as economists sought to explain the patterns of foreign trade in both developed and developing nations. This extensive literature overwhelmingly accepted the main conditions of the Heckscher-Ohlin theorem; these include, among others, perfect markets, free trade, complete international immobility of productive factors, identical production functions and consumption patterns throughout the world, and qualitatively indentical productive factors (Bhagwati 1964). Another theoretical condition went unnoticed by and large in the economic literature: the state was limited in its economic functions to two types of intervention - tariffs and non-tariff barriers, and exchange-rate manipulations. It is true that most of the economic functions that states perform in the real world (upgrading the skill of the labour force, performing research and development, financing exports of goods and services, funding selected industries to overcome barriers to entry) were theoretically

125 The State and International Trade excluded from the model by the underlying economic assumptions already mentioned. Generally speaking then, in international trade theory, governments act only as impediments to the resource optimization that the economic forces would produce in the absence of these governmental barriers and currency manipulations. It is more than time to change this simplistic and basically ideological view of market dynamics. We defend the position not only that government intervention is much broader than depicted in international trade theory, but also that it has its own rationale. State policies in the fields of industry, technology, and finance are often oriented and have demonstrated their capacity to strengthen the national factor endowment and to provide an orderly market framework in which national resources can be managed so as to improve the international competitiveness of the economy. To understand the role of the state with regard to the international trade position of a given economy, we will have to abandon most if not all of the conditions of the Heckscher-Ohlin theorem. Recent empirical literature on the structure of international trade, however, even if it has been theoretically cast in the neoclassical framework, can be used in our own discussion on the role of the state. There is no point here in providing a summary of the theories of international trade from a historical perspective. Others have done it before us (Chipman 1964; Bhagwati 1964), and we do not claim any particular expertise in the history of economic analysis. In this section we shall try to explain, both on theoretical and historical grounds, why the state plays such a marginal role in international trade doctrines when it is such an important element in real world commerce. The Ricardian Legacy Originally exposed by Ricardo (1817), the theory of comparative advantages (or comparative costs) was in fact a powerful attack on British mercantilism in that it favoured free trade. The main target of the theory was, precisely, the British state, which protected local agriculture from low-priced imports through an array of tariffs. The policy goal of the theory was to reduce the intervention of the (British) state in order to permit the import of cheap food and raw materials. Theory, then, had the political or welfare goal of minimizing the role of the state. The British states in Ricardian times was devoid of the multifarious functions that it performs in modern capitalism. It obtained most of its revenues from taxes on foreign trade and had the modest

126 State, Technology, and Competitiveness

duties of regulating currency, improving domestic transportation, and taking responsibility for defence and police (including, of course, defending the Empire from rival powers). No public enterprise, no massive education system, no governmentally sponsored research and development, and no state development bank or export credit device was there to complicate the picture. A marginal state was not only a desired state of things, but also a relatively realistic assumption. The main conditions of this theory were already present in the Ricardian legacy as well. Perfect competition, international immobility of factors, similar productive conditions, no transportation costs, two countries, and two commodities were involved in trade. It was left to nineteenth-century followers (mainly J.S. Mills and A. Marshall) to give the theory a more explicit form. Marshall added the mathematical modelling that, from this point on, covered the assumptions and gave academic credibility to the theory (Schumpeter 1954, 609). The historically correct assumption regarding an economically weak state remained confined to matters involving tariff management and currency manipulations. The Heckscher-Ohlin Theorem

While some authors consider the Heckscher-Ohlin theorem of the 19305 and 19408 to be a radical departure from the Ricardian tradition (Bhagwati 1964, 17—18), it is more often seen as a refinement of the previous theory of comparative costs. While Ricardo postulated one factor (labour), the new theorem assumes two (capital and labour). Ricardo postulated different production functions in the trading countries; the new theorem assumes that they are identical throughout the world. All other postulates remain the same for both including, as previously noted, a given supply of factors, constant returns to scale, and perfect competition. Under these conditions, the nation-state of international trade theory has essentially remained a producer of barriers to trade (tariff and non-tariff) and currency manipulations. It is curious to observe that while in the real world of the interwar and postwar periods government intervention expanded in very different ways (public enterprises, government-sponsored laboratories, a vast taxation revolution, mass education, trade theory concerning the role of the state remained fixed to the original Ricardian themes of Victorian Britain. It is also remarkable that after the Second World War, tariffs (which were declining in the world trade through the GATT negotiations) and exchange rates (which until 1971 were relatively fixed in ac-

127 The State and International Trade

cordance with the Bretton Woods Agreements) remained, according to international trade theory, the most important instruments of the state. Leontiefs Paradox When Wassily Leontief (1954) tried to estimate the factor intensities of the average exports and competitive imports of the United States, he inaugurated an era in which logical coherence was no longer the sole criterion of truth for international trade theory. From then on, trade theory also had to pass the test of facts. Empirical studies multiplied, as did the number of paradoxes. It was ascertained not only that us exports were labour intensive, while us imports were capital intensive, as demonstrated by Leontief, but also that Indian exports to the United States were capital intensive, while Indian imports from the United States were labour intensive (Bharadwaj 1962). The same paradox holds true for Canadian-American trade (Wahl 1961). The number of productive factors increased rapidly as explanations for these empirical results were sought. Labour was soon divided into the skilled and the unskilled (Keesing 1965, 1966). Research and development was identified as a key factor in international trade (Gruber, Mehta, and Vernon 1967), and its importance grew compared with that of other factors (Sveikauskas 1983). Lindert and Kindleberger (1982, chap. 4) suggested that at least five productive factors must be distinguished in the analysis of foreign trade: crop land, minerals, unskilled labour force, skilled labour force, and non-human capital. Present-day studies allow for many more factors. In spite of the fact that science, technology, and the upgrading of the labour force are often governmental endeavours, all these debates and verifications have not destroyed the universal credibility of the theory, and its assumptions, including that of the marginal role of the state. Oligopoly and International Investment From the 19608 on, the accepted theory of international trade came under fire from another angle. Empirical studies showed that oligopoly was the prevailing market structure in capitalist countries, both developed and underdeveloped (Bain 1966). Multinational corporations were shown to have an increasing proportion of international trade. The link between oligopoly and its transnational by-

128 State, Technology, and Competitiveness

product, the multinational corporation, was definitively established (Hymer 1960). And equally the links between technology (and particularly the international flow of technology), oligopoly, international investment, and international trade became the subjects of the (then) new theory of the product life cycle (Vernon 1966, 1971; Wells 1972). Oligopoly was first theorized in the 19308 by J. Robinson and E. Chamberlain, but their contribution was stunted by the lack of empirical proof, as well as by the contemporary Keynesian revolution. Empirical studies had to wait thirty years before they showed in the works of, among others, J. Bain, J. Blair, F.M. Scherer, GJ. Stigler, K.D. George, and P. Sylos Labini) that perfect competition was a very unrealistic assumption. The massive empirical literature on the multinational corporation dealt a severe blow to the concept of international immobility of factors. The empirical analysis of the international transfer of technology had become, in itself, a branch of economic and policy studies. If a small number of large corporations are responsible for most of the production and exports of any industrial or newly industrializing country, the question arose as to whether a firm's competitiveness was not at least as important as country-specific factor endowments. And if this was the case, why could the state not recognize and pick the winners through selective aid in order to boost exports ? In this new world plagued by market imperfections of different kinds, with a variable mobility of productive factors across borders, with barriers to entry, and with international information and technology gaps, the state would normally have plenty to do. However, international trade theory left the state confined to its traditional role in the tariff and currency fields. The neoclassical division of economics from other fields of social enquiry prevailed. Recent Developments in the Neoclassical Theory

The Heckscher-Ohlin theorem was extended in a variety of ways since the 19605. Some models assumed several goods and/or several countries instead of the "two countries/two goods" of the original model. Others allowed for growth. Some included imperfect competition. Several new models allowed for international capital and/or labour mobility (Bhagwati and Srinivasan 1983). In every case, however, only one or two of the assumptions of the HeckscherOhlin theorem were modified, while the others were retained. In all cases, the state continued to be perceived as a nuisance to international trade, and its interventions were viewed as essentially

i2g The State and International Trade

confined to tariffs. This point of view is well summarized in the literature on international trade contributed by the new public choice school, led by James M. Buchanan. This school of thought is particularly important because it intends to bridge the gap between economics and political theory by applying neoclassical principles to governmental processes. The public choice description of tariffs on imports and their effects is as follows: Tariffs: (i) raise the price of foreign goods and domestic substitutes; (ii) make some domestic producers competitive in local markets; (iii) may raise the earnings of labour in the protected sector; (iv) redistribute income in favour of certain groups; (v) make it easier for the inefficient and the mediocre to thrive; (vi) may permit the support of a larger total population; (vii) will usually stimulate domestic investment in the protected sector; (viii) add to public revenues; (ix) require the creation of jobs in the public sector; (x) give discretionary powers to senior bureaucrats who have to administer the tariff structure; (xi) make it possible to discreetly reward friends and punish foes either by a change in the tariff rate, in the tariff list or in the mode of administration of the tariff law; (xii) must be co-ordinated with the tariff structure of other countries and therefore necessitate international contacts and travel abroad, with all the amenities that such travel necessarily entails. (Breton 1978)

In this new (old) theory, government policies are confined to tariffs, and such policies are seen as a sort of plot worked out by politicians and bureaucrats, together with businesspeople and workers in inefficient industries — a plot that reduces the economic welfare of the general population. New theories often look very much like Ricardian or Marshallian statements.l THE S T A T E AS A P R O D U C E R OF PRODUCTIVE FACTORS AND AS AN INSTITUTIONAL FRAMEWORK OF THE FACTOR

MARKET

The demise of the international trade theory is taking place through a two-way process. In the first place, from a theoretical and academic point of view, the neighbouring social sciences (among them political science, public administration, and business studies) are bringing to view fresh new data and theoretical insights on industrial and tech-

130 State, Technology, and Competitiveness

nology policy, as well as on multinational corporations and their links with the state (see Reich, Roth well, Johnson, Vernon, Murray, Tyson and Zysman, Bonin, and Harris). In the second place, from a more practical point of view, the rise of Japan and of the newly industrializing countries, all of them having adopted strategies involving some sort of industrial and technological planning, show that the international competitiveness of a national economy can be improved through the active intervention of the state. In fact, the "planned market economy" (as the Japanese call their own economic system) posits itself as a model, not only for the neighbouring Asian countries, but for the advanced capitalist nations as well. In the debate that took place in the United States over the question of implementing an industrial policy, the "Japanese model" gained political (though not yet academic) credibility (Norton 1986). We believe that it is of major theoretical significance that an attempt be made to systematize the analysis of the role of the contemporary state in international trade. Public intervention goes far beyond the "infant industry" argument advanced by List in the nineteenth century and generally conceded later to be the exception to the general rule of "free trade/no public intervention. Our arguments are either purely theoretical nor normative. We argue that most present-day states actively intervene to increase the international competitiveness of their economy with far more sophisticated devices than those depicted by the current international trade theory. We acknowledge that policies may disrupt markets, inducing inefficiencies in the allocation of resources, and that programs, at great cost to the taxpayers, may not produce the results envisioned. Our presentation is confined to state intervention on the technology factor. We point out what we believe are the most important types of economic intervention by the state in industrial and technological fields. Technology and the State: Where "Market Failures" Are the Norm

Technology is now recognized as a key factor in international trade in most or all empirical studies on the factor intensity of the exports or imports of any country. It is widely acknowledged by mainstream economists that technology plays a key role in the determination of trade flows. A country will undertake massive public subsidization of R&D in order to contribute to the creation of national champions and to defend the international competitiveness of its national pro-

131 The State and International Trade

duction. For those not comfortable with this political perspective, a market justification can also be found; in its most straightforward form, it reads as follows: In a market economy in which a given commodity usually sells at a uniform price, firms cannot appropriate unto themselves all of the benefits generated by their products. For that reason, they may not undertake certain innovative projects because their prospective private returns are inadequate. If this inadequacy were offset by subsidies, these projects would likely produce social returns in excess of the subsidies plus the costs of delivering them. Thus government subsidization of certain innovative projects undertaken by private firms is necessary. Without it, the overall level of R&D and related activity forthcoming from the private sector would be less than socially optimal. (ECC 1983, 75)

The growth of international trade and foreign competition has made countries more aware of their relative strengths and weaknesses. In the igGos the United States, then the hegemonic economic power, was both a threat and a challenge to its trading partners. For reasons of both economic survival and political sovereignty, governments had little choice but to enter the technological race (Blais 1985, 117; OECD 1983, 63). More precisely, still following the factor-proportion version of comparative advantage (the Heckscher and Ohlin approach), technology has entered trade theory through the distinction that has been drawn between skilled and unskilled labour (Keesing 1966). It is interesting to note that the major challenge suffered by the theory of comparative advantage, the Leontief paradox,2 was "resolved" when it was shown that us exports were intensive in the use of skilled labour, while us imports were intensive in the use of unskilled labour (Harris 1985, 41). Saving the theory was enough for most economists, while others, more policy prone, got the idea that a country's specialization could be changed through education and job training. Later on, it was noted that technology was also involved in innovation, in R&D and in the knowledge composition of different industries. Factors such as technical competence were considered as givens. A country's specialization was always the result of the optimal combination of factor endowments. But the conceptual framework didn't allow for the causal relationship to be considered in the other direction. This is probably the major weakness of the classical theory. While the technology-based specialization was fully acknowledged, no causal pattern was identified. Logically, any intervention that makes a productive factor more available (i.e., less expensive) will

132 State, Technology, and Competitiveness

potentially affect the comparative advantage position of this market. As this argument was pursued, those responsible for the management of economic policies soon discovered that competitive advantages could be produced and the resulting specialization tampered with. Still, it remained unnoticed by trade theorists that the state had been deeply involved in the production of technology. Technology first appeared in empirical, post-Leontief discussions on international trade through the analysis of labour skills. It was observed that developed countries export products that are intensive in high-skilled labour (Keesing 1965, 1966). But those who introduced this technology factor explained national skill differences through differences in culture and history, and cumulative economic factors, including migration induced by international wage differences. The skill composition of any labour force, as Keesing has correctly pointed out, can, however, be modified through training. Training is generally a governmental service. Any industrial, semi-industrial, or newly industrializing country, capitalist or socialist, has a masseducation system, which includes high school and university levels. The state controls education either entirely or partially and, in most countries, funds it. In some countries, like Japan, South Korea, or Taiwan, public education (and especially postsecondary education) is part of the government's industrial policy (Ozawa 1974; Botkin 1982): the education system has to provide the large skilled labour force required by the industrial sectors. Consequently, the university attendance rates are very high, and a significant percentage of the graduates earn diplomas in the engineering and technical fields. In Japan, the engineering, technology, and physical sciences faculties account for nearly 50 percent of the university graduates (Okimoto et al. 1984, 24—32). In Korea, education was one of the main factors in that country's industrialization and international competitiveness. As the Korea Development Bank ably puts it: [T]he nation's manpower with its managerial and technical skills has been a great asset in the course of industrialization, enabling industrial development despite scarce natural resources and low capital formation in the early ig6o's. Increasing demand for skilled manpower in line with the industrial development has been met by increased investment in education and various manpower development programs. Emphasis has been placed on the acquisition of skilled manpower resources, such as technicians and engineers, needed to sustain industrial development and economic growth ... The development of manpower resources contributed to the improvement of labour productivity which has been a fundamental factor in Korea's development process. (KDB 1984, 20—21)

133 The State and International Trade In other words, the entry of a country into capital- and knowledgeintensive industries requires an abundant supply of skilled labour that only the state can provide in adequate amounts. In-company training, although increasingly common, is not a sufficient substitute for this massive amount of state upskilling, since it can only provide a complement for the sometimes unsatisfactorily constituted work force, a specially trained labour force that suits the particular technological context and technical resources of enterprise. Industrial production and export competitiveness are directly affected by the national education strategy (or the lack of it). However, producing skilled labour is not the only way in which the state intervenes in the technology factor. Gruber, Mehta, and Vernon (1967) showed the positive correlation that existed between us research-effort measures and us export performance. But again, the key role of the state went unnoticed in their seminal contribution. In fact, the state funds R&D directly (through public R&D laboratories and public enterprise research, as in Canada), indirectly (through subsidies to private laboratories, as in the United States and Japan), or through tax credits and other tax rebates to the private sector for research expenses undertaken. During the postwar period and up to the igSos, the state was funding more than 50 percent of all R&D in the OECD area. Today the figure is still close to the 50 percent mark (OECD 1984). It is widely known that the American export balance shows a surplus in industries with high R&D intensity, largely funded by the state (aircraft production, electronics, and defence material). This is also true for Britain and France. The market mechanism always works within a given institutional environment in which the state plays a major role. This environment may be more or less conducive to technological learning, adaptation, and innovation. Until recently it has remained unnoticed in trade theory (Dosi 1988), as economic institutions are usually taken for granted and excluded from economic theory. A brief comparison between the United States and Japan may, at this point, be useful. The United States has no policy concerning technology imports from abroad. In contrast, Japan carefully monitors the entry of foreign technology through the MITI in order to maximize the returns the country gets from any private transaction between a foreign licensor and a Japanese licensee. No restrictive clauses concerning adaptation, innovation, or exports by the licensee are permitted in the contracts. Thus, Japan imports "hands-free" technology (Ozawa 1974)In the field of innovation, again, American policy has until recently markedly differed from Japan's, us antitrust policy forbade association between corporations during the phase of initial R&D. Under

134 State, Technology, and Competitiveness

increased pressure from foreign competition, American authorities have allowed coalitions in high-tech ventures. Sematech, a joint venture involving the largest American electronics firms, is a good case in point. This new research corporation, in operation since 1988, receives $125 million a year from thirteen partners and $100 million from the us federal government in order to rescue American semiconductors from the Japanese competition. Japanese antitrust policy has always permitted the association of companies in the expensive and risky stage of pioneering innovation. Japanese corporations are, however, forced to compete once the product or the process is developed and patented by the corporate pool. The computer industry deserves a brief mention in this regard. In 1981 Japan's government launched a plan for the production of fifth generation computers with an initial subsidy of $450 million. Six months later, the government created, with public funds, the Institute for New Generation Computer Technology. And the National Electrotechnical Laboratory of Japan, another governmental endeavour, organized a pool of the six largest Japanese electronics producers (Toshiba, NEC, Fujitsu, Hitachi, Mitsubishi, and Oki) to pursue collective R&D on a very rapid conventional computer. These and other initiatives of the Japanese government, similar to those launched fifteen years before in consumer electronics, show in a more precise way what the concept of a "planned market economy" means (Feigenbaum and McCorduck 1983; Okimoto et al. 1984). Through its massive presence as a buyer of goods and services, the modern state gives direction, purposely or not, to economic activity. It is well known, for example, that the United States' enormous defence budget explains the us postwar world leadership in the electronics, aircraft, and aerospace industries. The fact that the Japanese are catching up in electronics is closely related to the buying policies of the Nippon Telegraph and Telephone Corporation (NTT), a public enterprise whose goal was to create a competitive semiconductor and telecommunications equipment industry in Japan (Okimoto et al. 1984; OECD 1985). In most industrially advanced economies, with the familiar exception of the United States, and in all of the newly industrializing ones, there is a definitive relationship between technologically advanced activities - such as aeronautics, electronics, communications, transports, energy production — and direct state intervention. The state is, either directly or through its agencies, the major user of the products and even, in many instances, the nation's main producer. Through placing orders and deciding upon product specifications and performance, the state sets national production goals. Thus, national champions are created through a market reserve that gives

135 The State and International Trade selected firms a definite advantage for sliding down the learning curve. It was noted above that differences in institutional framework affect the R&D capacity of industrial sectors, but more needs to be said about the relationship between economic concentration and technological development. Influenced by the Schumpeterian hypothesis which relates monopoly power and technological innovation, many governments have actively encouraged mergers (Japan, Britain, and France). Their purpose was to assist in the creation of very large enterprises that would have both the resources to get involved in R&D and the capacity to be active in foreign markets. The public promotion of technology was thus an indirect consequence of the active support given to corporate concentration. A whole new generation of multinational enterprises is the result of such policies (Faucher and Niosi 1985; Anastassopoulos, Blanc, and Dussauge 1985)Policies do fail, however, and it is not clear to what extent public intervention has contributed to improved markets. We are neither presenting a plea on behalf of public intervention nor endorsing all programs that are regrouped under the general heading of industrial policy. Now that these specific interventions have been thoroughly experienced and placed under close scrutiny, we know more of their relative effectiveness. Technology policies have been precisely formulated to correct market imperfections and, more precisely, to increase positive externalities from R&D investments. Interventions have had more to do with creating the proper environment for firms to invest in R&D than with providing direct subsidies or resorting to massive fiscal incentives. Japan is known for programs that facilitate the flow of information, pooling both private and public expertise (thus contributing to the reduction of risk and uncertainties) while contributing relatively few resources for the realization of the project. This approach has inspired the creation of pre-competitive research projects in Europe such as EUREKA, ESPRIT and FAST since 1984. Unlike industrial policy, which is often undiscriminating, technology policy has in many countries been limited to specific types of technology and/or projects. In Japan only eight projects were financed in 1987 in the large-scale project program, and fourteen projects, which were distributed in only three sectors (new materials, biotechnology, and advanced electronics), were supported by the Basic Technologies for Future Industries program (OECD 1989). The importance of the diffusion process has been recognized, and there have been attempts to make technology available to sectors traditionally less R&D intensive to small enterprises, and to regions.

136 State, Technology, and Competitiveness

Finally, it has been acknowledged that the production and generation of technology is an international process. Instead of traditional protectionism, what is found is an increasing amount of international co-operation. The result is the growing importance of an institutional framework in the promotion of technological change and dynamic adjustment. Obviously, given the very mixed results achieved, particularly the dismal situation of many heavy interventionist Third World countries, technological development is far from being only a matter of public policy. In fact, considering the average rate of success, the probable outcome of public intervention is negative. If this were not so, governments would be quick to adopt this successful course of action in order to promote economic and technological growth. On the other hand, as we have tried to show, a few countries have managed to actively promote and facilitate technological development. As a general rule, recent examples show that interactive policies combining private initiative with a limited amount of public resources can achieve more than what can be accomplished through cumbersome state planning and public ownership. Thus, the outcome of public intervention depends largely on the intrinsic quality and dynamism of the country's entrepreneurship. CONCLUSION

International trade theory is based on neoclassical perfect and pure competition, international immobility of factors, and given endowment assumptions that keep the role of the state in world commerce from being considered in anything but a marginal way. Once these assumptions are abandoned and more dynamic assumptions and a more realistic oligopolistic market are introduced, the role of the public sector becomes more understandable. Through different policies, all nation-states try (and some have succeeded) to improve the international competitiveness of their national economies by overcoming market failures, mainly in technology and finance, that impede either the proper combination of existing factors or the growth of the national endowment of factors. In this way, the state's goal is to create dynamic comparative advantages for local industries and firms. We suggest that with industrial development, with its evolution from labour-intensive industries to capital-intensive and then knowledge-intensive industries, the weight of static natural advantages diminishes and that of dynamic socially produced advantages grows. In the more rapidly growing economies of the industrial latecomers,

137 The State and International Trade

the role of the state, a key actor in the creation of dynamic comparative advantages, is bound to develop. Even the older industrial powers could, through state intervention, slow the decline of their international competitiveness. Trade theory should integrate the fundamental and systematic quest in which governments of all major countries are involved — the quest for comparative advantage. Intervention is certainly not limited to the promotion of R& D, but involves a whole array of policies. Borrus, Millstein, and Zysman synthesize the Japanese approach: the more general national goal of creating comparative advantage, rests on a conscious government and industry strategy of controlling access to the domestic Japanese market, structuring the terms of domestic competition, making available stable sources of cheap capital, and using the controlled and structured domestic market as a secure base from which to gain entry and competitiveness in international markets. (Borrus, Millstein, and Zysman, 1983, 183)

It is not sufficient to claim that these policies have no other rationale than the nationalistic manipulation of resources and markets. In order for some theoretical progress to take place, it must be acknowledged that the creation of comparative advantage results in a different, but no less efficient, allocation of resources. What is left of the neoclassical theory in the present international market? Because technical factors have become increasingly important in the composition of trade, and to the extent that these factors are the result of public decisions, the accepted model's explanatory capacity diminishes. State intervention increases the degree of uncertainty of the system because the factor endowment of many countries can be rapidly changed through appropriate policies. The static character of the model is a major hindrance to the understanding of reality — even more so in a world market that has been changing swiftly because of the increased liberalization that took place in the post war period and the resulting necessity of industrial adaption in advanced countries. Protectionism by foreign partners can temporarily freeze the composition of trade of a particular country and slow down the impact of the changes in the internal factor endowment. It follows that the more protectionism there is, the more a static theory of international trade would apply. Neoclassical theory has put the emphasis on the defensive, protectionist interventions of the states. However, the word "defensive" implies that there is also an offensive intervention. It is quite true that government support for modernization and industrial

138 State, Technology, and Competitiveness

restructuring may result in a waste of public funds. In such an occurrence, the collectivity has to pay the costs of the public intervention. But it serves governments no purpose to dwell on failure, since the only consequence is that nothing fundamental has changed in trade flows. On the other hand, when state intervention succeeds, there is a change in the factor composition of the economy that may have a positive impact on international trade. Impacts, as just noted, may not always be positive, but if only the failures are focused upon, change brought about by state intervention will forever remain unexplained. It is, as already suggested, successful public intervention that brings about change. Further studies could determine what types of public intervention produce the most significant changes in the trade competitiveness of industrial countries. NOTES

1 It should not be concluded that all economists are happy with the neoclassical theory as it stands today. Richard G. Harris introduces the right nuance in this debate: "Every economist, of whatever methodological bent, admits that the theoretical structure underlying the neoclassical theory of resource allocation and its open economy extension do not give a fully accurate picture of the real world. Nevertheless, its defenders regard it as useful, approximately true, and both elegant and logically consistent in its internal structure" (Harris 1985, 12). 2 Being labelled a paradox is a kind of curse, or as more ably formulated, "Taradoxe"' c'est 1'etiquette, sous laquelle la science economique range avec interet, amusement et courtoisie distante, toutes les choses qui sont trop solides pour etre rejetees purement et simplement, trop contrariantes pour etre adoptees" (Emmanuel 1969, 33). REFERENCES

Anastassopoulos, Jean-Pierre, Georges Blanc, and Pierre Dussauge. 1985. Les multinationales publiques. Geneva: Institut de recherche sur les multinationales, Presses Universitaires de France. Bain, J. 1966. International Differences in Industrial Structure. New Haven and London: Yale University Press. Bellon B., and J. Niosi. 1987. L'industrie americaine: Fin de siecle. Paris and Montreal: Seuil and Boreal. Bhagwati, J. 1964. "The Pure Theory of International Trade: A Survey." Economic Journal 74, no. 293 (March): 1—84.

139 The State and International Trade Bhagwati, J., and T. Srinivasan. 1983. Lectures on International Trade: Cambridge, Mass.: MIT Press. Bharadwaj, R. 1962. Structural Basis of Indian Foreign Trade. Bombay: University of Bombay. Blais, Andre. 1985. Une sociologie politique de I'aide a I'industrie. Vol. 45. Study prepared for the Royal Commission on the Economic Union and Development Prospects for Canada. Toronto: University of Toronto Press. Blais, Andre, and Philippe Faucher. 1981. "La politique industrielle dans les economies capitalistes avancees." Canadian Journal of Political Science 14, no. i (March): 3-35. Borrus, Michael, James E. Millstein, and John Zysman. 1983. "Trade Development in the Semiconductor Industry: Japanese Challenge and American Response." In American Industry in International Competition, edited by John Zysman and Laura Tyson. Ithaca: Cornell University Press. Botkin, J., et al. 1982. Global Stakes: The Future of High Technology in America. Cambridge, Mass.: Ballinger. Breton, A. 1978. "Economics of Representative Democracy." In The Economics of Politics, edited by J.M. Buchanan. London: Institute of Economic Affairs. Bunge, M. 1967. Scientific Research. 2 vols. New York: Springer-Verlag. Chipman, J.S. 1965. "A Survey of the Theory of International Trade." Econometrica 33, nos. 3, 4: 477—519, 685—760. Dosi, G., et al. 1988. Technical Change and Economic Theory. London: Pinter. Economic Council of Canada (ECC). 1983. The Bottom Line: Technology, Trade and Income Growth. Ottawa: Minister of Supply and Services. Emmanuel, A. 1969. L'echange inegal. Paris: Maspero. Evans, P., et al. 1985, Bringing the State Back In. Princeton: Princeton University Press. Faucher, P., and J. Niosi. 1985. "L'etat et les firmes multinationales." Etudes Internationales 16, no. 2: 239—59. Feigenbaum, E., and P. McCorduck. 1983. The Fifth Generation. Reading, Mass.: Addison-Wesley. Gruber, W., D. Mehta, and R. Vernon. 1967. "The R&D Factor in International Trade and International Investment of United States Industries." Journal of Political Economy 75, no. i :2O—37. Harris, Richard G. 1985. Trade, Industrial Policy and International Competition. Vol. 13. Study prepared for the Royal Commission on the Economic Union and Development Prospects for Canada. Toronto: University of Toronto Press. Harrod, R.F. 1958. International Economics. Chicago: University of Chicago Press.

140 State, Technology, and Competitiveness Hiemens, Ulrick, and Kurt V. Rabenau. 1976. "Effective Protection of German Industry." Public Assistance to Industry, edited by W.M. Corden and Gerhard Pels, 7—45. Boulder, Col.: Westview Press. Hymer, S. 1960. The International Operations of National Firms. Reprint. Cambridge, Mass.: MIT Press, 1976. Keesing, D.B. 1965. "Labor Skills and International Trade: Evaluating Many Trade Flows with a Single Measuring Device." Review of Economics and Statistics 47, no. 3:287—94. — 1966. "Labor Skills and Comparative Advantage." American Economic Review 56, no. 2:249—58. Keizai. 1986. Japan 1985: An International Comparison. Tokyo: Keizai. Korea Development Bank (KDB). (1984). Industry in Korea. Seoul: Korea Development Bank. Leontief, W. 1954. "Domestic Production and Foreign Trade: The American Capital Position Reexamined." Economia Internationale 7. Lindert, P.H., and C. Kindleberger. 1982. International Economics. 7th ed. Homewood, 111.: Irwin. Murray, R. 1981. Multinationals beyond the Market. London: Harvester. Norton, R.D. 1986. "Industrial Policy and American Renewal."Journal of Economic Literature 24, no. i (March): 1—40. Organization for Economic Co-operation and Development (OECD). 1981. Recent International Direct Investment Trends. Paris. - 1983. Positive Adjustment Policies. Paris. — 1984. "Resources for Science Newsletter." Newsletter, Paris. — 1985. L'industrie des semi-conducteurs. Paris. — 1989. Evolution des politiques industrielles dans les pays de /'OCDE. Paris. Okimoto, D., et al. 1984. Competitive Edge: The Semi-conductor Industry in the us and Japan. Stanford: Stanford University Press. Oulton, Nicholas. 1976. "Effective Protection of British Industry." In Public Assistance to Industry, edited by W.M. Corden and Gerhard Fels. Boulder, Col.: Westview Press. Ozawa, T. 1974. Japan's Technological Challenge to the West, 7950—7974. Cambridge, Mass.: MIT Press. Porter, M. 1980. Competitive Strategy. New York and London: Macmillan. — 1985. Competitive Advantage. New York and London: Macmillan. Prebisch, R. 1957. Hacia una dindmica del desarrollo latinoamericano. Mexico: Fondo de cultura economica. Ricardo, D. 1817. On the Principles of Political Economy and Taxation. Reprint. Cambridge: Cambridge University Press, 1966. Schumpeter, J. 1954. History of Economic Analysis. London: Allen and Unwin. Sveikauskas, L. 1983. "Science and Technology in United States Foreign Trade." Economic Journal 93, no. 37:542—54.

141 The State and International Trade United Nations (UN). 1983. Transnational Corporations in World Development, Third Study. New York. — Statistical Yearbook. Annual. - Yearbook of International Trade Statistics. Annual. Vernon, R. 1966. "International Investment and International Trade in the Product Cycle." Quarterly Journal of Economics So, no. 2:190—207. - 1971. Sovereignty at Bay. New York: Basic Books. Wahl, D.F. 1961. "Capital and Labour Requirements for Canada's Foreign Trade." Canadian Journal of Economics and Political Science 27 (August): 349-58. Wells, L.T., Jr. 1972. The Product Life Cycle and International Trade. Boston: Harvard University Press. Wilkinson, B.W., and Kenneth Norrie. 1975. Effective Protection and the Return to Capital. Ottawa: Economic Council of Canada.

CHAPTER SEVEN

Technological Competitiveness Considered as a Form of Structural Competitiveness FRANCOIS

CHESNAIS

The issue of international competitiveness — and the links competitiveness might have with a country's research and development and its capacity to innovate — became central to the igSos debate concerning the conceptual research on innovation and the elaboration and implementation of national industrial and technological policies (of course, these policies include the decision to abandon such policies, to stop "intervention" and to deregulate and privatize the public sector). The aim of this chapter is to assist and to clarify the debate. It begins with the assumption that the international competitiveness of a national economy depends on the competitiveness of the firms within its boundaries, but it identifies a national economy's competitiveness as being something more than a simple expression of the collective or "average" competitiveness of its firms. I adopt the notion of "structural competitiveness" as a way of expressing the fact that, while the competitiveness of firms obviously reflects successful management practices, it also stems from the strength and efficiency of a national economy's productive structure, its technical infrastructure, and other factors determining the "externalities" on which firms can build. This holistic approach, although considered rather "heretical" until recently, has received the support in the United States of MIT'S interdisciplinary group, which was commissioned to study the causes of the decline in us industrial productivity and thus of the American economy (Dertouzos et al. 1988).' This approach relates to the work carried out in the 19805 in the field of innovation and technological change, the results of which give particular importance to the question posed by T. Hatzichronoglou and the author (1985) at the Organization for Economic Co-operation and Development (OECD): Can technological competitiveness depend solely on a few

143 Technological Competitiveness very R&o-intensive sectors (known as "high-tech" sectors), or on the other hand, must it be supported by a much wider industrial base, as well as by considerable mechanisms for intersectoral diffusion of new technologies? (One of the most convincing effects of new technologies is the importance of exports in sectors of average R&D intensity, where countries such as Germany and Japan excel.) Two definitions of high technology, one of American origin and the other of Japanese, both proposed in 1984, can throw light on the significance of this question as well as on the different responses that can be made to it in terms of international competitiveness.

TWO D E F I N I T I O N S OF HIGH TECHNOLOGY

The American definition of high technology was proposed by the OECD in its work on the obstacles to trade in high-technology products (Hatzichronoglou 1985). It lays down five major characteristics: (a) high dependence on a strong technology base and a vigorous research effort; (b) considerable strategic significance for governments; (c) long lead-times from basic research to industrial application, short lead-times in commercialization, and accelerated obsolescence under the competitive pressure of new product and process introductions; (d) high risks and large capital investments; and (e) high degree of international co-operation and competition in R&D, production, and world-wide marketing. The Japanese definition was proposed by Ken-ichi Irnai, who uses the twofold criteria of technology: that is, high technology has a "high intensity of research and development" as well as "the characteristics of a system." This means that "individual technological elements are utilised in the economy combined into a system, and that each technological element is evaluated according to the function it fulfils within the entire system." Imai continues: "[Cjlearly, the focus is on the combination of technologies and not on each separate technology.... [T]he reason high technology is made an issue in connection with industrial policy ... is probably strongly attributable to its systemic characteristic" (Imai 1984). The difference between these two definitions, which are both interesting on the conceptual level, is striking. The American definition has the twofold perspective of the "strategic" requirements (i.e., military) of the federal government and of the company strategies that concentrate on the individual high-tech product and its life cycle in isolation. The Japanese definition, on the other hand, stresses two

144 State, Technology, and Competitiveness

notions, that of the combination of technologies and that of their systemic interaction with the larger economic. The American definition corresponds to an international trade approach, with competitiveness based on comparative advantage and the choice of specialization based on products known as "niche specializations." In contrast, the Japanese definition assumes the elaboration of an approach in terms of "structural competitiveness" and "pole specialization."2 The emphasis is placed on the functioning of the national economy and its component sectors as a whole, which presupposes the recognition in that economy of more or less systemic characteristics that allow the optimal exploitation of the technology's own systemic dimensions. The term "structural competitiveness" will be defined later. It refers to competitiveness that manifests itself both on the level of costs and prices and on the level of "non-price" factors. 3 Within the framework of the French theoretical debates, structural competitiveness is an extension of what J. Weiller (1945) defined as "the national preference regarding economic structures" in view of a country's role in the international division of labour, as well as an extension of the work of F. Perroux (1962). It is, of course, also linked with J. Mistral's work (1978) on "long-term competitiveness." The concept of structural competitiveness has the advantage (or perhaps the inconvenience) of pointing out very sharply the great ambiguity in the notion of international competitiveness. Competitiveness implies the existence of an active economic agent (a "subject" of the economic process) that makes choices, defines strategies, and seeks to control variables; it also assumes the possibility that clear-cut indicators of performance can be established and measureable results achieved with some degree of agreement (Michalet 1984). In the case of the firm, and therefore of business management economies, which gave rise to the concept of competitiveness, the situation is clear: competitiveness falls within the realm of "the dynamic entrepreneur" of the Schumpeterian type or of Porter's "competitive advantage." However, when one moves from the level of the firm to the macroeconomic level of a country's economy, the existence of an active economic agent presupposes an active governmental industrial and technological policy (with a view to identifying the factors that constitute the national preference regarding structure - for instance, the choice of specialization by pole), which depends on two-way cooperation between the state and the business sector. Within the capitalist analytical framework, the essential territorial base required for the formation of structural interdependence has been the economy of the nation-state. It is this economy that since

145 Technological Competitiveness

the early ig6os has been the framework of what the governance school calls "monopolist governance" and the sphere, above all, of industrial and technological policy. Today, however, in many countries, but not to the same extent nor in the same shape in all, the very idea of an industrial and technological policy is being questioned at a time when the basis of the economy of the nation-state seems to be disintegrating under the force of internationalization. The approach taken here is as follows: I begin with a reminder of the significance of an aspect of the theoretical debates which have taken place since the early igSos on the relationship between technology and the international competitiveness of national economies. Then I shall analyse the notion of structural competitiveness in relation to recent developments in the theory of innovation and technological change. I conclude by briefly discussing internationalization, or more exactly, the double process - overlapping but nonetheless distinct from internationalization - namely, transnationalization. This discussion is all the more necessary because technology has been assigned a central role by many, both in globalization and in the process of the disintegration of the economy of nationstates, particularly in Europe. However, it is not certain that the essential causes of the economic crisis of the nation-state should be sought in the characteristics of the constraints on new technologies. TECHNOLOGY AND INTERNATIONAL COMPETITIVENESS: THE AMERICAN DEBATE The discussion on the relationship between technology and international competitiveness is particularly lively in the United States. The deterioration of the American trade balance in high-tech sectors, as well as in others, calls for a re-examination of the assumptions used vis-a-vis the possibility of building up a comparative advantage in limited to highly R&o-intensive sectors. These assumptions, based on numerous neoclassical trade studies, attempted to account for the division of labour between countries and for international trade in terms of comparative advantage arising from the specialization made possible by "endowed factors": natural resources of one kind or another, abundant and therefore cheap manpower, and a large stock of capital (the notion of capital being extremely difficult to measure). This approach was based, and is still based (neoclassical theory is still being taught), on a group of hypotheses that the orthodox economists must then, to use J.H.

146 State, Technology, and Competitiveness

Dunning's descriptive expression, "loosen." These hypotheses, which together would suggest pure and perfect competition, are the following: the existence of homogeneous production functions and thus of identical costs from one country to another; the immobility across national boundaries of factors of production; and the absence of government intervention in areas directly linked to foreign trade, apart from government-imposed tariffs. In a way, neoclassical theory favours the nation-state economy to the detriment of firms. In fact, as C.A. Michalet notes, "the only agents taken into consideration by neo-classic theoreticians are the countries. However these nation-states are reduced to a blueprint or abstraction. They are merely the containers of given combinations of factors. Their only reality is to constitute a pretendedly insuperable barrier to the movement of productive factors" (Michalet 1976, 42) (emphasis added). The assumption of the immobility of factors beyond frontiers is overlapped by the hypothesis, or better still the prescription (which today represents one of the bases for "structural adjustment" policies), to the effect that domestic factors may be moved, if not at will, at least quite easily from one activity to another. A complementary policy prescription according to neoclassical theory deals with the need to adjust "price factors" downwards, and another recommends breaking all "rigidities," particularly those originating from workers arid trade unions. When an activity or an industrial sector loses its comparative advantage the neoclassical approach does not take account of the interdependent structural relations that could link the activity or sector to other parts of the national productive system. According to neoclassical writers, declining sectors can be considered in isolation and their productive resources "displaced" (moved) to other activities or sectors that seem to benefit from comparative advantage. In the American case, this approach has been taken a long way.4 According to some theoreticians, since the importing of Japanese goods signifies a loss of comparative advantage in the us manufacturing sector, the United States should move its specialization toward services to firms. All through the 19708, the dominant thesis was that the United States could find its comparative advantage, its international trade specialization, and its foreign trade balance strictly by servicing hightechnology industries. The past quarter-century in the history of dominant international trade theory has consisted mainly of a series of attempts to improve the theory of specialization by factor endowment without abandoning the basic theory. The main attempts are those that issued from

147 Technological Competitiveness

the work of Leontief on the structure of us trade by product and, later, those that grew out of research undertaken on the role of the technological factor. They have resulted in an enlargement of the Heckscher-Ohlin-Samuelson model, which was constructed on only two factors (labour and capital), with the introduction of other econometric variables: first, the occupational qualifications of the work force (or human capital) and later R&D expenditures and stock of scientific personnel. The purpose was to allow technology to be taken into account as a specific productive factor, while according it the status and analytical treatment identical to those of labour, capital, or one raw material or another. Besides econometric tests, other American work has effected a division of industries according to the intensity level of their R&D expenditures (high, medium, or weak R&D intensity) in order to study the comparative performances of the three groups. Undoubtedly, those who consider technology to be important in the contemporary economic process will be satisfied to note that within the framework of the neoclassical theory of international trade, numerous econometric empirical tests carried out in the 19708 have shown that innovation activity — measured by R&D expenditure, the number of researchers and engineers, or the number of patents - has effectively constituted, alongside human capital (i.e., qualified manpower), one of the explanatory factors of comparative advantage in an economy such as that of the United States.5 Nevertheless, the attempt to understand the role of technology in international trade simply by submitting it to the same tests as labour, capital, or natural resources — that is, simply as a stock that may be correlated to performance indicators (production or exportation) - results in the expulsion from the analysis of those intermediary trends that lie between R&D expenditure and measures of performance.6 The attempt also assumes a linear and direct relationship between the two orders of grandeur that leads to a mutilation of the main specific characteristic of technology. The conceptual weakness of work of neoclassical inspiration explains the paradoxical findings to which it gives rise. The last set of American studies that sought to correlate growth rates of industrial production or exports of the manufacturing sector with indicators representing the endowment of the country in R&D and qualified manpower (or in human capital) led their authors to observe that in the early 19808, although R&D expenditures had not fallen (and had even begun to grow rapidly with the launching of new military programs), the basis of the us comparative advantage in high-tech

148 State, Technology, and Competitiveness

products began to disappear in relation to Japan, and also in relation to several European countries (us Congress Joint Economic Committee 1980; us Department of Commerce 1981). The dead end reached through explanations of neoclassical origin, combined with the rapid fall in the competitiveness of sections of important sectors of the us manufacturing industry, the rise of a structural deficit in the trade balance, the commercial and industrial Japanese penetration, and the first studies of the de-industrialization of America (Business Week, 30 June 1980), has brought forth a new series of studies in which the reasoning is in structural terms. Mellman (1983) places the accent particularly on what he calls "the deterioration of the industrial system" and "the erosion of the infrastructure which supports production" (production support base), linking these phenomena to the overwhelming weight of the military sector, to the very specific process of selection of innovation that this causes, and to the progressive proliferation in the competitive sector of the "cost-plus" — accounting management methods used in military procurement. Mellman is a "heretic" in America, the oldest and most vigorous critic of military expenditure and all that is linked to it. In consequence, his support for a structural approach could be considered as marginal. This could not be said of the book Manufacturing Matters by S. Cohen and J. Zysman of the Berkeley Roundtable on International Economy, as they are respected members of the Californian intellectual establishment and can hardly be suspected of leftish views (Cohen and Zysman 1987). The book contains one of the first, or in any case one of the most complete and systematic, defences of the analysis of the competitiveness of national economies in structural terms yet published by American economists. It is all the more interesting because it is not supported by any strong indigenous school of thought or national theoretical tradition from within the United States. It engages in straightforward debate with the neoclassical school, for whom the only linkages between branches of industry worthy of interest in economics are of the Walrasian type, that is, those established through the market and the price system. In contrast to the "loose linkages" of the price system and the structural adjustment policy supported in America by economists like R.Z. Lawrence of the Brookings Institution, Cohen and Zysman describe both "tight linkages" (i.e., intra- and inter-industrial relations of a structural nature, which arise from the technical interconnections between industries) and "medium linkages" (i.e., those associated with processes of a cumulative type that characterize investment and growth). According to Cohen and Zysman, both of the

149 Technological Competitiveness

latter types of linkages control the different national economies' capacity to exploit technological change to their advantage. They summarize their most important conclusions in this respect as follows : (i) Technological developments can provoke rapid market shifts, (ii) Technologies are shaped by the needs and arrangements that exist in the nations from which they emerge, (iii) Some critical technologies can affect the competitive position of a whole range of industries; and if one nation uses these technologies to gain a lead in a vital product, it can forge an important trade advantage for itself. These are strategic transformative industries characterized by imperfect competition and with powerful interindustry spillovers, (iv) Continued technological development depends heavily on the connections between producing firms, their suppliers, and their customers. A web of structural and operating arrangements supports technological development, and that web can unravel, (v) This reshuffling of market position in a period in which important new strategic transformative sectors are emerging is powerfully influenced by government policy, (vi) The reshuffling can result in new international hierarchies of wealth but also of power. (Cohen and Zysman 1987, 109-10) (Italics in original)

Cohen and Zysman recognize several non-American influences on the elaboration of their analytical positions. Their research included, for example, a careful examination of the Japanese economy and Japanese industrial theory. Also important were their readings of work by and discussions with members of the French "regulation" economic and social governance school. Their position has also been nourished by progress made in the analysis of technological change to which certain "heterodox" Americans - principally, N. Rosenberg, R. Mowery, and R. Nelson — have made important contributions, but of which the centre of gravity in the Anglo-Saxon world was the Science Policy Research Unit (SPRU) under the direction of C. Freeman and K. Pavitt. The influence of the French governance school — M. Aglietta, R. Boyer, J. Mistral — on Cohen and Zysman is significant. Ironically, just when the American authors, realizing the limits of neoclassical theory, had come closer to an analysis that took structural competitiveness into account, this approach was abandoned in France due to the impact of what was above all a political failure. Those who wanted to put France "in time with California" did not seem to realize that in California, under the threat of Japanese competition, an increasing number of economists were reconsidering the primacy of the market and enterprise culture and were moving toward an ap-

150 State, Technology, and Competitiveness

proach favouring interdependence and cumulative processes — precisely those processes that the French school in the Perroux tradition had analysed. The studies by J. Zysman, S. Cohen, and M. Borrus at Berkeley were followed by the publication in Boston of an important MIT report, Made in America (Dertouzos et al. 1988). This work gives considerable support to the American school, which is as interested in interconnections and structural relations in economics (as well as in the large range of factors flowing from the "social relations of production" in the wide sense) at least as much as it is interested in markets and pricing systems. Indeed, the factors to which the MIT report attributes the loss of industrial competitiveness are of the structural type. These factors include the conservatism of American industrialists with regard to the organization of manufacturing production (i.e., their attachment to Taylorism and Fordism); their constantly shortening temporal horizons, determined by the criteria and demands of short-term profit, the fear of stock-exchange raids, and fluctuations in the financial and money markets; their redundancy policy and their lack of knowledge of the essentials of long-term manpower-resources management; the difficulty that these independent firms have co-operating with each other; and finally, the contradictions between the objectives and vision of the federal state and those of the private sector (one of the principal problems on the state side being the weight of military priorities at a time when completely different measures are required for the restoration of industrial competitiveness). STRUCTURAL COMPETITIVENESS AND NATIONAL SYSTEMS OF PRODUCTION

In a 1983 report on science, technology, and competitiveness, the Science Policy and Technology Committee (SPTC) of the OECD accepted an approach based on the following hypothesis: The international competitiveness of a national economy is built on the competitiveness of the firms which operate within, and export from, its boundaries, and is, to a large extent, an expression of the will to compete and the dynamism of firms, their capacity to invest, to innovate both as a consequence of their own R&D and of successful appropriation of external technologies; but the competitiveness of a national economy is also more than the simple outcome of the collective or "average" competitiveness of

151 Technological Competitiveness its firms; there are many ways in which the features and performance of a domestic economy viewed as an entity with characteristics of its own, will affect, in turn, the competitiveness of firms.

The report recognizes that the analysis of these overall macroeconomic features or structural factors is complex. It embraces a Xvide variety of economic and institutional phenomena that relate to the way economies function, and that represent for firms either "externalities" or "dis-economies" (in other words, factors that either stimulate or, on the contrary, constitute a brake on progressiveness and competitiveness). Nevertheless, the report adds, the term "structural competitiveness" has been coined to express the idea that while the competitiveness of firms obviously reflects successful management practices in the private sector, it also reflects the strength and efficiency of a national economy's productive structure, the corresponding long-term trends in the rate and structure of capital investment, the economy's technical infrastructure, and other factors determining the externalities on which firms can lean. The reference to economic externalities is, of course, a reference to Alfred Marshall, the first economist to demonstrate in a systematic way that the competitive qualities of a firm are determined by its environment (Marshall 1929). Before examining the support this approach receives from contemporary theory of technological change and innovation, attention must be drawn to the contribution of J. Mistral. In the approach he proposed in 1978 and subsequently long-term competitiveness is related to capital formation (or rate of growth over a long period of investment), considerable allowance being made for the size of the capital-goods sector and the role of the internal market. In one of his recent and most complete formulations, Mistral — starting from the polarisation of world trade between industrialized countries, the increase in intra-branch trade, and its concentration in the electromechanical industries - states that it is just as necessary to consider international trade between industrial countries in terms of "intraindustry competition" as it is to consider it in terms of specialization. In this context, he writes (i) competitive capacity depends to a large extent on the cohesion of the productive system: this property reflects, first, the ability of downsteam sectors to develop a range of products and a continuing flow of innovations closely adapted to trends in demand, since that is a prerequisite if an industry is to play an active role in the dynamic redeployment of consumer patterns;

152 State, Technology, and Competitiveness moreover, it is a property which reflects the integration of industries downstream with the capital goods industries, which can enable them to improve their production techniques and increase their productivity; (ii) the above considerations make it clear that control over the domestic market (which will, of course, be selective) and a large measure of international specialisation are not irreconcilable but rather go hand in hand. The former ensures sound management of technologies and provides a launching pad for new products; the latter results from a judicious selection of the activities for which production for the world market is justified by the extent of the economies of scale to be achieved. (Mistral 1983)

Mistral considers that only with this approach will it be possible to establish "national preferences regarding economic structures" in Weiller's sense. Indeed, the approach puts "on the same level the relationship of each productive system with its domestic market and with the world market, as the development of exports cannot be seen as submission to a pre-established specialisation, any more than the sharing of the domestic market is in terms of a technical or natural constraint" (ibid). The structural competitiveness approach leads implicitly to a theory of the capitalist nation-state economy. However, this represents one of the large blank areas in economics, except for the well-known, but little read, work of F. List. In the neoclassical theory of international trade, the nation-state is presented as constituting "a reservoir of factors of production." Those who conceptually oppose the neoclassical economists have usually not presented a more satisfactory alternative theory. For most economists, the reduction of the national economy to something "taken for granted" ("un fait d'evidence" as M. Beaud has said in a pertinent way [Beaud 1987], or better, to a "bunch of industrial sectors," to employ B.A. Lundvall's expression [Lundvall 1988]), historically has been part of a wider attempt to expel from the field of economic analysis the state, social classes, and technological change — in brief, in the name of a search for laws of general equilibrium to expel all the elements that link economics with history and society. From this point of view, one of the ways of understanding the limits of the "Keynesian revolution" is to note that neither the general theory nor the state and the interventionist economic policies to which it gave rise, brought forth any serious work on the economy of the nation-state, particularly in the United Kingdom. In Keynesian analysis, the nation-state undoubtedly ceases to be a simple "reservoir of factors." Rather, it becomes an area where a body

153 Technological Competitiveness

of macro-economic equilibriums are set in place with the help of active state intervention, with macro-economic policy and national accounts serving as the state's instruments. However, the Keynesians (whether of the Labour Party or more recently the neo-Cambridge school) have not produced a valid theory that posits the nation-state as the necessary framework, albeit a historically determined one, for the accumulation of capital and the place for the formation of structural interdependencies. The defeat of Keynesianism by Thatcherite liberalism, the re-affirmation by the Thatcherites of comparative advantage, and the implementation of structural adjustment policies hence derived, have been greatly facilitated by the Keynsians inability to come up with such a theory. With reference, again, to Mistral, it can be seen that two strong ideas of a complementary nature come out of his positions: first, the cohesion of a country's productive fabric depends both on certain properties of downstream industries that allow these industries to adapt to market trends, and on the integration of these branches with capital-goods industries situated upstream; and second, solid forms of international specialization depend on the selective mastery of the domestic market. The important role played by the capital-goods sector in the formation of hierarchical positions within world trade, in both the duration and adaptive capacity of certain specializations that are based on upstream machine production, is confirmed by historical and contemporary analysis. The growth and sophistication of the capitalgoods sector, the diversification and specialization among firms producing different types of machinery, as well as the purchasing price and the quality of intermediary goods necessary for their manufacture, automatically become strong factors of differentiation among the productive structures of the different countries or regions that participate (some actively, others by coercion) in the capitalist international division of labour. These factors create hierarchies among these countries, determining the way they are situated in that international division. England's international dominance until about 1880 was due less to its textile industry than to its metallurgical and mechanical industries, the technology these industries incorporated, and also England's military capacity, which allowed the industries to be established. The same would apply later to Germany and the United States and again today to Japan. It seems that the hypothesis concerning the particular functions and attributes of the capital-goods sector should also apply outside the large economies. The long-term performances of small economies would also benefit from internal effort and a domestic market,

154 State, Technology, and Competitiveness

however limited, if there existed specialized domestic capacities to produce machines. For example, the prosperity of the Swiss economy was founded not only on the success of its banking activity and its food and pharmaceutical multinationals, but also on an. important industrial sector that produced and exported machines (large electrical equipment, machine tools, and precision instruments). In conjunction with its agro-food industry - particularly its milk-products industry — Denmark has built up a specialized capital-goods industry, the role of which has been important in maintaining over a long period the structural cohesion of that country's industrial specialization. Even the remarkable adaptive capacity of the downstream branches of the Swedish economy is partly dependent on the strength of Sweden's upstream capital-goods industry (Andersen et al. 1981). On the other hand, although Belgium had an iron and steel industry, its capital-goods industries were very weak, which contributed to the industrial dislocation it suffered due to transnationalization (Coomans 1986). A similar observation emerges from a comparative examination of the capacity for adaptation and survival of textile industries in the 19705. Benefiting from equal protection against offshore producers with low-price labour (GATT [General Agreement on Tariffs and Trade] agreement on fibres), the textile industry of different advanced capitalist countries has managed to survive or not, depending on whether it has a thriving textile machinery industry. A recent Brookings Institution study that compared the evolution of the productivity and exports of a series of American industries showed the favourable performance of the textile industry in comparison to many others; this was attributed in large part to the existence of a dynamic, specialized machine sector upstream (Baily and Chakrabarti 1988). The second part of the Mistral approach to long-term competitiveness is related to what he calls "selective domination (or better control) over one's domestic market": a relative control, which does not exclude control over intra-branch international trade even for the industries where such a control has already been established; domination or control over the external market, seen as a springboard for exports, the character of which would be founded on an important element of "non-price" competitiveness; and domination or control over areas that do not depend over very long periods on external protection (now essentially non-tariff) but rather on the solidity of the productive fabric and the complementarity of the competitive poles they include. Now we come to what seemed, not long ago, to be the second distinctive contribution of the French industrial economy — that is,

155 Technological Competitiveness

an analysis in terms of "filieres," of which J. de Bandt has no doubt provided the most complete version with his notion of "mesosystems." Contrary to "filieres," understood as a series of stages in manufacturing linked by the flow of trade (trade between suppliers and customers, or when these are integrated, the transfer of products between affiliates), meso-systems, de Bandt says, stress the modalities of the organization of overall relations - market and non-market between agents. Further, meso-systems provide a strategic space where the strategies of the actors are affirmed and confronted (and, we would add, where they can be co-ordinated) (de Bandt 1988). This approach is interesting for several reasons. It emphasizes the market and non-market relationships that form a meso-system; further, it underlines the organizational element therein. This approach leads straight toward the contemporary analysis of technology that we shall now examine, but it also reminds us of the necessity of strategic intervention in providing both the goals and shape of an industrial meso-system. THE CONTEMPORARY THEORY OF I N N O V A T I O N AND ITS IMPLICATIONS FOR COMPETITIVENESS

In the Perroux-Mistral school of thought, long-term competitiveness depends mainly on recognition of the effects of the cumulative process relating to the rate of continuous investment on a controlled interior market and on phenomena of "virtuous" cumulative interaction between different industrial branches and economic sectors. These particular attributes occupy little space therein. However, when they are introduced to the analysis, they strengthen the theory that it is essential to adopt the technological competitiveness approach, which pays particular attention to interdependence between branches and sectors and is especially relevant for industries with a strong "core" or "modal" character, which facilitates technological diffusion and training. Since the late 19708, the economic theory of innovation and technical change, after considerable difficulty due to the dominant neoclassical economic paradigm, has made qualitative steps in the account it gives of technical change. Several aspects of recent advances interest us here in a direct way: • the recognition of generic features of some technologies, which, being based on key basic scientific knowledge, have great pervasiveness and a large capacity for inter-industry diffusion;

156 State, Technology, and Competitiveness

• the recognition of the role played in innovation and productive technical change by the learning process and the long-run accumulation of productive technological knowledge; • analysis of the modality of the flows and diffusion of intersectoral technologies, which includes the central role diffusion plays in the structuring of flows of certain industries, especially capital goods; and finally, more generally, • the necessity of realizing that new technologies are not exogenous elements coming from outside the economic sphere, but are, on the contrary, essentially social products within the dynamic process, influenced by the play between interactive procedures and mechanisms and between research organizations, firms, and the state (firms themselves being considered not as "individual" economic agents, but as a particular form of organization).7 These concepts have important implications for the establishment and implementation of technological policies at the level of firms and states. Here my comments deal essentially with the second. With regard to firms, I shall limit myself to two points. The first point deals with what M. Giget, in several reports and articles, and LA RE A/ CEREM (Laboratoire de recherche en economic appliquee/Centre d'etudes et de recherches sur les entreprises multinationals) and GEST (Groupe d'etude en science et technologic), in a collective book, have defined as firm strategies of the "arborescent" type, of which figure 5 provides a graphic representation. In Giget's view (illustrated particularly by numerous Japanese examples), it is the firm's own technological and industrial potential, and not financial assets, that directs the financial policy and that should be the main focus of the firm's strategy. It is here that a decisive part of the competence that gives the firm its specificity is to be found. Starting with the generic technologies situated upstream of product lines, these capacities should have the ability and the means necessary to evolve into competitive activities. At the prescriptive and normative level, the arborescent strategy assumes that no firm with a large technological potential would specialize in only a few market products. Its growth perspectives depend on its capacity to develop a technology-based polyvalency (Giget 1988). The second point concerns the importance of co-operative agreements and inter-firm alliances on technology (see, in particular, Teece 1986, 1989; Mytelkaand Delapierre 1987; and Chesnais 1987, 1988). Such agreements and alliances today offer firms a wide range of means of access to complementary or pioneering technologies of relevance to their own potential technology. In a strategy for the

157 Technological Competitiveness Figure 5 Tree Structure Showing the Technological Function of the Company

Source: Futures, 1988.

enhancement of technology, a twofold function - interface upstream in relation to scientific knowledge, and interface downstream in relation to the market - is decisive. Ideally, it is this, and not the financial function, that should dictate the diversification, purchase, and connections of the firm. At a time of almost complete domination of corporate strategies by the criteria and priorities of stock marketdominated money capital, we have a long way to go. Nevertheless, it is the price that has to be paid for the full exploitation of the generic technology potential of the firm. When one moves from the level of the firm to that of the economies of nation-states, the elements brought to light by the contemporary

158 State, Technology, and Competitiveness

theory of innovation and technical change all tend to strengthen the idea that the technological competitiveness of firms depends, to a greater or lesser degree, on the existence of "external economies" (still Marshall's notion), which together add up to what C. Freeman and R. Nelson (also Henry Ergas, although he is close to the neoclassical economists on many subjects) have named the "national system of innovation" (Dosi et al. 1988). Here again one must emphasize the progress made by British and American theorists in recent years, and also the fact that French researchers, who had in the mid-19705 evolved a similar notion ("national scientific and technical potential"), have let it lie fallow since then. The central idea here, to employ an expression used by Andersen and Lundvall, is that "the relevant context of an individual innovation is (part of) the total 'system of innovation' which ought to be modelled explicitly." And the same authors explain that "the system of innovation is not simply another name for the 'research system' of industrial R&D, public research institutes and universities. On the contrary, we think that an analysis of the system of innovation is greatly strengthened when it takes as its starting-point an analysis of the system of production and consumption" (Andersen and Lundvall 1988). The national system of innovation would therefore be a part of what the French school in the Perroux tradition calls the national productive system. "NATIONAL SYSTEMS OF INNOVATION" AND THE EFFECTS OF L E A R N I N G AND TECHNOLOGICAL ACCUMULATION The work that led to the formulation of the idea of "national systems of innovation" first focused upon the effects of technological accumulation and learning. The conclusions of this work can be summerized as follows: innovation is the result of long processes of accumulation and appropriation of technology, whether it is situated at the end of the R&D chain (where the capacity to identify the economic possibilities offered by progress in fundamental research and by the ability to transpose this knowledge to new products and processes of production has a decisive importance) or at the level of industrial production (where the industrial know-how and long processes of assimilation by experience have had important roles). These innovation processes determine, in particular the conditions that allow firms to assimilate and successfully appropriate exogenous

159 Technological Competitiveness

knowledge, that is, knowledge resulting from scientific and technical activities carried out elsewhere in the system. In the case of scientific knowledge, the capacity for successful assimilation generally implies that firms have previously participated in similar or closely related fields of knowledge. In the case of industrial technologies, there is a similar implication, since an accumulation of industrial know-how on the part of firms can in no way be improvised. Some people have tried to distinguish practical learning (learning by doing) from learning through use (learning by using). Learning by doing (Arrow 1962) consists of developing abilities during production and, generally, leads to a fall in unit costs. Learning by using (Rosenberg 1982) is the result of using the product: it concerns, particularly, the capital goods studied by Rosenberg, but now covers complex products of software and also data-processing materials, which are progressively brought up-to-date by the users. The weight of learning by using is due to the complexity of the interaction of the product's components; this method results in a progressive differentiation of the product, which makes it possible to undertake utilizations that themselves become differentiated through use. The framework in which the effects of learning have most often be studied is that of the firm in isolation. Andersen and Lundvall (1988) have, however, shown that it would be interesting to consider the relationships among a few firms, or even among a large set, in order to fully analyse the effects of "learning by interaction," a kind of learning whose importance in the national framework they stress. One of the most important dimensions would be the relationships among firms in regard to technical learning, which takes place between the producers and users of advanced technological goods or new production processes. Andersen and Lundvall consider that the experience of Nordic countries, because of the density and the quality of their "interactive" relations, provides one of the best frameworks at present for an analysis of both learning by interaction and accumulation of technology. The learning-by-interaction process (i.e., collective learning through the implicit or explicit co-operation of numerous firms and organizations) has also attracted the attention of Japanese economists; Imai states that an analysis of new "high technologies" and of the corresponding industrial and technological strategies and policies must take two aspects into consideration: the high R&D intensity of the high technologies as well as their interdisciplinary and systemic character. In this approach, Imai says that "the various technical elements in the economy make up a system, and each technical

160 State, Technology, and Competitiveness

element should be evaluated according to the functions it fulfils within the global system (so that) attention is drawn to the whole constituted by the association of different technologies and not to technologies taken separately." Imai thinks the debate opened by Schumpeter, in which the scale of enterprises has been a point of dispute in discussions concerning the efficiency of technological innovation, should be pursued further. "It is more pertinent to consider the size of the system within which the learning process for the development of technological innovation takes place as representing the real factor influencing the results of innovation.... Account must be also taken of the efficiency of the diffusion and transfer of knowledge and information within this system" (Imai 1984, 33)The framework of such an interactive process may be a large industrial group, a group combining partners and/or satellites (firms or research teams), or firms and laboratories constituting a national productive meso-system. In the second situation, the interactive process may consist of a whole range of formal or informal cooperative agreements. Such agreements, even though they are quite common, are not absolutely critical for production. There may be "collective" production without some of the participating units being completely conscious of the fact. This is especially true of public or university laboratories, whose research and findings may be used by firms without the former's knowledge and without their receiving financial reward. In the third situation, such an occurrence could be more easily avoided. THE INTERSECTORAL DIFFUSION OF TECHNOLOGY AND THE INTERACTIVE LINKS BETWEEN SECTORS

Another area where the theory of innovation has helped to nourish the idea of national systems of innovation is that of the process of intersectoral diffusion of knowledge and learning. All sectors of the economy and all industrial branches call on technology to some degree. However, as industrial R&D statistics show, the use of technology is extremely concentrated within the economy. It is carried out essentially in four industrial branches, grouped in two large sectors: mechanical and electrical construction (including data processing at the heart of this sector) and the chemical and related-industries sector.

161 Technological Competitiveness Table 31 Estimated Breakdown of Industrial R&D in the OECD Area Industrial R&D As % of Total Average Annual Growth Expenditure, 1983 Expenditure Rate, 1979-83 (%) ENGINEERING BRANCH

Electrical/Electronics group Machinery group Aerospace group Other transportation equipment group BASIC METALS GROUP CHEMICALS BRANCH

Chemicals group Chemicals-linked group

81.2

65.7

6.7

29.4 21.0 17.8

23.8 17.0 14.4

8.0 7.9 6.3

13.0 5.1 26.3 21.6 4.7

10.5 4.1 21.3 17.5 3.8

2.7 6.4

3.8

3.1

5.0

116.4

94.2

6.5

1.3 5.9

1.0 4.8

5.4

123.6

100.0

6.6

5.9

6.7 3.2

OTHER MANUFACTURING INDUSTRIES GROUP

Sub-total manufacturing industries Sub-total mining and agriculture Subtotal services Total BERD in the OECD area COMPOSITION OF INDUSTRY GROUPS

Electrical/Electronics group Machinery Aerospace group Other transportation equipment group Basic Metals group

Electrical machinery: electronic equipment and componenets (excluding computers) Instruments, office equipment, and computers, machinery NEC Aerospace (including missiles) Motor vehicles, ships, other transportation equipment Ferrous metals, non-ferrous metals, fabricated metal products Chemicals, drugs, petroleum refining Food and beverages, textiles and clothing, rubber and plastic

Chemicals group Industries related to the chemicals industry (or chemicals-linked group) Other manufacturing Stone, clay, and glass; paper and printing; furniture; industries group wood; cork; and other manufacturing industries Services group Utilities, construction, transportation and storage, communications, commercial and engineering services, other services Source: OECD, STUD Data Bank, February 1988.

162

State, Technology, and Competitiveness

In other studies, the extent to which the structure of R&D expenditure is influenced by the weight of military R&D has been emphasized (Chesnais 1990). Here, in a complementary way, other dimensions of a particular configuration of research expenditure must be stressed. The great disparity in R&D between different branches of activity cannot simply be interpreted as indicating that some branches need little technology while others need a great deal. Although it is true that such an interpretation would apply in some cases (e.g., low R&D intensity leather industries vs. high R&D intensity semiconductor industries, the majority of sectors combine their own R&D expenditure with acquisitions of technology from other sectors. Some industries constitute a sort of "reservoir" from which there are various flows and intersectoral transfers of technologies. The British data bank on innovations at the Science Policy Research Unit of Sussex University, Brighton, and the relevant analyses referring to the data bank (DeBresson and Townsend 1978; Pavitt 1986; Robson et al. 1986) at this stage provide the best descriptions of the structure of flows and transfers. The SPRU identified the sectoral origin and the first branch or sector of utilization for 4,378 innovations (identified forms and easily limited technical changes in products and processes, patented and non-patented). The SPRU is careful to point out that the data are only indicative, since they do not include improvements of a secondary nature made to the original innovation. They cover only the first known sector of application, and sometimes other sectors turn out to be much more important users. Finally, the data need to be completed by other complementary indicators, such as flows of licences and qualified personnel. While taking account of its limitations, one should nevertheless, consider the results of the SPRU'S work carefully. Nearly 65 percent of all innovations come from five sectors: chemicals, machinery, mechanical construction, precision instruments, and electronics. Six other sectors - metallurgy, electric construction, engineering for drilling at sea, vehicles, construction materials, and rubber and plastic products - take up another 23 percent of the total, leaving little for other branches and sectors. Although it may be considered that the SPRU models express certain particularities of British economic history, they clearly have a wider significance. Two aspects are especially interesting: • the share taken up by machines and mechanical construction, as well as the interdependence of precision implements and electronics; and

163 Technological Competitiveness • the fact, nevertheless, that the main destinations of innovations coming from the electronics sector are situated outside the manufacturing sector in the military sector, in the business sector, and in the management and execution of R&D. T H E "CORE" C H A R A C T E R O F MACHINE PRODUCTION

It is difficult not to relate the SPRU'S work to that done in 1978 by B. Gille (see Figure 6). It is to Gille that we owe the first formulation of what he called, rather deceptively, "a technical sector" (it is, in fact, a "techno-industrial" system): As a rule all techniques differ to some degree, but they are interdependent and some coherence is necessary among them: this body of coherence, of different levels of all structures of all bodies and all filieres, composes what can be called a technical system" (Gille 1978, 19). This hypothesis depends on research carried out since P. Nantoux's work on the technology that accompanied and accelerated the growth of capitalist industry in the first half of the nineteenth century. The most significant aspects of Figure 6 are, first, the fact that Gille includes both technologies and branches of industry, the coherence of which depends on their interdependence, and second, the fact that the two big technological products belonging to the investment-goods sector (steam engines and machine tools) occupy central points in the system. The objection that some could make to the effect that the second feature, though perhaps valid in the past, is not valid today, is contradicted by the work of another important name in the field, D. Sahal. The conclusions of his book Patterns of Technological Innovation are worth quoting: • ... technical progress is first and foremost a matter of learning or the accumulation of practical experience in design and production processes; • ... the crucial factor in the phenomenon under consideration is learning in the capital-producing rather than in the capital-using sector. This conclusion from quantitative analyses is also supported by the qualitative evidence indicating that industries in the former sector (e.g. machine tools and electrical equipment) tend to be far more interconnected than industries in the latter sector (e.g. paper products and textiles); • ... learning in the capital-producing sector is especially important because of its synergistic effect on what is demonstrably a highly integrated system of industries; •... one consistant result that emerges from a wide variety of case studies

164 State, Technology, and Competitiveness Figure 6 Simplified Diagram of the Technological System of the First Half of the Nineteenth Century"

" The dates correspond to the first recorded innovation, the generalized diffusion conies later. Source: Gille 1978, Figure 4.

presented here is that (cumulated) gross investment is a key determinant of technical progress. That is, all investment has ipso facto the character of investment in R&D activity.... [Technical progress is not just a matter of conscious R&D activity. It is also crucially dependent on the growth of the capital-producing sector of the economy. (Sahal 1981) Even if one believes, like J.L. Gaffard, that it is now necessary to go beyond the idea of "the factory as the centre of productive activity and so of machines as the sole expression of technology" and also

165 Technological Competitiveness

"to present productive activity as an activity which poses questions and which solves specific problems rather than which provides products" (Gaffard 1988, 691), the relationship noted by Sahal, between the rate of formation of capital and the rate of technical progress remains true. The precise composition of industries and technology, which together constitute the investment/goods-production sector, is, at present, in rapid evolution, as is the internal hierarchy within the sector, but this does not seem to change the sector substantially. EFFECTS OF P R O X I M I T Y AND B A R R I E R S ON THE D I F F U S I O N OF TECHNOLOGY

It would be a great mistake to think that the intersectoral diffusion of technologies resulting from the purchase of equipment or semifinished products with high-technological content happens automatically. This was, however, Pavitt's opinion when he put forward the thesis that industries and sectors based on science would become motors of diffusion and give impulse to growth (Pavitt 1984). In a recent critical work on the SPRU'S analysis, B. Quelin expressed the opinion that this thesis should be "more solidly supported" (Quelin 1988). His work on Pavitt's results shows that although R&D-intensive industries and sectors constantly encourage and feed the propensity toward or even the constraints on innovation in other sectors, the heart of intersectoral diffusion remains in the capitalistic industries ("production-intensive industries," in Pavitt's terminology). One of the merits of Quelin's analysis is that it was an attempt to rationalize Pavitt's view of the problem with the classical work of Perroux, as well as to take account of the research conducted by ISMA (Institut de sciences mathematiques et d'economie appliquee) on inter-branch relations and on the effects, in terms of investment and productivity, of the influence exerted by some industries on others. Thus, Quelin brings us right back to the analytical work conducted by the French on long-term competitiveness and the national productive fabric. One of Quelin's conclusions is that two main channels of transmission of innovations emerge from all the varieties of industrial relations: economic proximity and technological and organizational proximity. The first covers, of course, the market relations between branches, which indicate the intensity of their reciprocal trade. The second type of proximity involves the transmission from one branch to another of a procedure or technique, or the transmission of an innovation from one organization to another on the basis of a high

i66 State, Technology, and Competitiveness

degree of similarity in technologies or productive processes, without their necessarily having close market relations. Under this latter form of diffusion, cost considerations, as well as the firms' strategies, play a determinative role, either allowing for the intersectoral transmission of an innovation or, on the contrary, blocking its diffusion. The latter aspect is the subject of an interesting concept developed by C. Passadeos (1987) in his study of the diffusion of new composite materials from the military sector. Passadeos starts from the idea that the sectors constituting the national productive fabric will often be characterized by large differences in their propensity to undertake R&D and to innovate, but also in their dominant objectives, which govern the R&D of firms. If the first aspect is widely recognized, the second is far less so. Differences in objectives do represent, however, significant brakes on intersectoral diffusion that could in certain circumstances become real barriers. This is the case for R&D and innovative activities with final military-strategic aims. Such activities undergo a process of selection filtering in two phases: first, the passage from high technology with strategic priority to high technology with commercial priority, then, when passing to the sectors, to "traditional technology" or "mature technology." Passadeos suggests a list of the main factors that constitute this twofold process of selection filtering (the F i and F 2 blocks in Figure 7). All this shows the importance of moderating the idea of easy and rapid intersectoral diffusion, which could be suggested by the models proposed by the SPRU. TECHNOLOGY IN RELATION TO "INTERNATIONALIZATION" AND "TRANSNATIONALIZATION" To end this rapid survey of some aspects of the contemporary theory of innovation, I will answer the question posed in the introduction. The technological competitiveness of a country cannot depend simply on a few isolated high technologies. The complementarities, interfaces, and synergies associated with contemporary technology can only flourish if the organizational forms - market and non-market - in which they emerge constitute an interconnected whole. The only known models of this are (a) the industrial group of the Japanese "keiretsu" type and (b) the nation-state economy when it is based on solid, interconnected industrial meso-systems (of which Japan is also the best example at present). As long as the qualitative differences in the objectives of innovation do not erect high barriers to transfers and to the technological im-

167 Technological Competitiveness Figure 7 The State/Research/Industry System: Transfer and Diffusion of Advanced Technologies

petus resulting from investment, the fabric of inter-industrial relations characteristic of the economy of an industrialized nation-state will also represent a framework in which the intersectoral flows of technology can operate most effectively, while benefiting from cumulative processes stemming from investment made in complementary sectors and branches. In the same way, it is the national economic base (plus the regional base in large states or countries with diversified geographical structure) that provides the required foundation for learning by interaction and for the special needs of the technological accumulation process. In contrast to what constituted, in the mid-19708, the experience of the French approach to long-term competitiveness, these elements are capable of founding a "good international specialization," that is, an industrial specialization capable of directing, if not forming,

i68 State, Technology, and Competitiveness

the evolution of world demand for certain products and of following it very closely for others. In France, in the wake of the work of Aglietta and Mistral, specialists in international trade agree that the strong economies in international trade are those that have strong poles of competitiveness.8 However, these poles are nothing more than productive meso-systems whose strength is founded mainly on the quality of the economies' relations, interactions, and accumulations. The almost complete abandonment in many countries (including France) of the structural competitiveness approach, as well as the industrial and technological policies likely to achieve it, seems to be a fait accompli at least for the time being. That abandonment is increasingly blamed on the inexorable process of internationalization. Two questions should be raised in this regard. Does the degree reached in the process of internationalization still allow for analysis and policy formation in terms of structural competitiveness? Do the most recent technological developments create a particularly decisive obstacle to any approach that continues to refer to a national framework. For decades, the idea of internationalization has been ignored by the dominant currents in economic thought. The first systematic study of the subject is, no doubt, that of N. Boukharine in 1915. He would certainly be surprised to see who has rallied recently to this idea. The term "internationalization" is simultaneously evocative and vague, as it covers a very wide range of processes and relationships by which national economies, previously fairly distinct from one another, are linked — the primary linkage being international trade. These national economies have become increasingly interconnected and will henceforth be interdependent to a degree never reached before. The process of internationalization is contained in the movement of capital insofar as the pursuit of constantly higher labour productivity is concerned, as well as in the successful termination of the cycle of capital enhancement, which must necessarily lead to foreign outlets for goods and for capital. To that extent, the process is irreversible or "inexorable" and suggests the development of different forms of organization of productive forces. But in a discussion of international competitiveness, where the existence of a nation-state is assumed and the main objectives are perceived as the loosening of external constraints and the search for equilibrium in the foreigntrade balance, it seems permissible and indispensable to ask how far internationalization can go without destroying the basis of competitiveness. National competitiveness viewed from within a country

169 Technological Competitiveness

never consisted of scatterings of isolated capitalist firms, but always of firms that depend on external economies, that is, those that make up the coherent productive fabric and provide solid national technological structures. From this point of view, consideration should be given to M. Beaud's idea that behind the word "international" there are two distinct concepts: "internationalization" (identifiable economic relations between two or more states) and "transnationalization" (relations rising out of the formation of transnational areas within multinational and transnational firms and banks active simultaneously in a number of countries) (Beaud 1987). Although there is considerable overlapping between the two, the distinction has great analytical importance. In a structural competitiveness approach, "international" refers to a problem of development of productive forces. This can include the growth in international trade, the increasing complexity of schemas of international trade specialization, trade in technology, and the constitution of a common pool of scientific knowledge at the international level open to all those with the level of competence required to understand it. Unlike internationalization, transnationalization relates to movements to increase the value of capital, which has become highly concentrated and above all strongly centralized (particularly in the form of money capital). Today transnationalization relates essentially, on the one hand, to the transnationalization and autonomization of monetary and financial markets (i.e., the place operations are carried out to enhance the value of financial assets or so-called financial assets) and, on the other hand, to the globalized activity of large industrial and financial groups (transnational corporations) and to the oligopolistic interdependencies woven between them on a world level within concentrated transnational structures. Under the impacts of the world economic crisis and the political pressures of Thatcherism and Reaganism, the advanced industrialized economies have been brought under the "sway of the market" (de Brunhoff 1987). Most developed-country governments have abandoned their earlier attempts to preserve the few instruments of economic policy that could direct the use of disposable investment assets to targeted productive objectives. At the same time, they have more or less left it to the big groups in the private sector to define on their own their financial, industrial, and technological strategies; these governments have ceased giving any consideration to industrial policy on the national level except in regard to the military sector. Within industries that are highly concentrated on a world level and

170 State, Technology, and Competitiveness

strongly internationalized, the only references of the big groups have become the world-wide financial markets and the complex relations of competition, mutual recognition, and co-operation among members of each world-wide oligopoly. Technology belongs to the realm of productive forces. Even though it is and will continue to be at the centre of the relationship that links members of the various oligopolies, it is not technology that is responsible for the process of dislocation in the productive fabric. It is certain, for example, that the techno-industrial and Taylorist-Fordist paradigm has forced a number of growing industries to enlarge their markets by means of exports. The influence of technology can be seen more generally in the increasing complexity of industrial products and consumer goods, as well as in the procuring processes necessary for their manufacture. Technology's influence is marked by the complexity and the increasing cost of scientific, technical, and industrial "inputs" and by the processes required for development of these inputs. Thus, technology leads to ever finer distinctions in the division of international labour and specialization, which has led to special interdependencies in intra-industry trade. Undoubtedly, this marks considerable change in the form of the division of labour that links advanced capitalist countries together. It suggests a very important change in the international division of labour within the manufacturing sector, particularly the chemical, mechanical, and electric sectors. These sectors, however, are not themselves responsible for the dislocation of factors that ensure the cohesion of national productive sectors. The same can be said of numerous very recent technological developments. Thus, the flexibility associated with variety 9 certainly suggests an accentuation on specialization and an even greater development in some fields of intra-industry trade, as well as intensive scientific and technological exchange and co-operation between firms. All these elements have very new aspects, but are not in themselves responsible for dislocative effects on the fabric of national production. To attribute a major role in the dislocation of national production to technology — as many politicians and trade unionists are increasingly inclined to do — is to commit, I believe, a serious error of judgment. In an approach based on the notion of structural competitiveness, industrial micro-electronics and the technology of new materials certainly demand a remodelling of the national productive fabric as well as the technological renewal of many branches of industry. But the characteristics of these technologies and their com-

171 Technological Competitiveness

patibility with small firms, universally recognized, would by nature allow considerable mastery of these changes on the national level (Perez 1987). Equally, biotechnology originated in the universities and large public research laboratories and was then developed in small firms; it does not call for gigantic industries either. The hold of multinationals active in pharmaceuticals and phyto-sanitary products depends above all on their mastery and control of market entry. The upstream implementation of fundamental and applied scientific research involving the most modern technologies demands considerable investment and recourse to increasingly varied disciplines and, therefore, leads to the call for joint funding and for the co-operation of laboratories belonging to scientific and technical bodies situated in several countries. At present, this international cooperation takes place far less between states and more and more among big industrial groups. While belonging to the field of productive capital, these groups submit to a logic often more "transnational" than "international." Agreements are often established between large groups, or there are inter-firm agreements where one big firm plays a pivotal role within a network of agreements (involving small firms and university laboratories) in which it has the initiative and control. In the majority of cases, these agreements have the support and authorization of the government administration, particularly within the Common Market, and are shaped by considerations which above all relate to the strategies of groups within world-wide oligopolies (Chesnais 1987; Mytelka and Delapierre 1987). Therefore, they tend to represent a collective defence of the status quo, directing knowledge towards and centralizing it within the strongest firms and those best organised to use it. Indeed, such agreements represent the subordination, if not the dislocation, of the scientific and technical potential of the weaker partners (countries, universities, firms). However, this cannot be attributed to the failure of scientific and technical co-operation itself, but to the form it takes, which suggests that cooperation in R&D should be established on different bases, more strictly inter-state, which would preserve national scientific and technical potential. It will be said, no doubt, that there is at least one area where technology demands or even causes the breakup of national frameworks. This field is the world-wide system of telecommunications and satellites, the emergence of which radically modifies the previous relative autonomy of national systems in relation to the world system. Here again one should have a closer look at the situation. The technological transformation taking place certainly implies new relation-

172 State, Technology, and Competitiveness

ships between national telecommunications systems as well as the establishment of a supranational management system. It does not, however, imply the deregulation, privatization, and attack on the public sector in the field of telecommunications and the media that is now taking place. The origin of this attack has nothing to do with technology as such. It is the result of the methods used to increase the value of financial capital (which is, at present, one of the main clients and the main beneficiary of transnational systems). This kind of attack is also a consequence of the rivalry between and international battles of different fractions of capital, which seek to exploit without hindrance the fruitful market of the media. Leaving everything to the logic of transnational financiers is to resign oneself to the hypothesis of an "international division of labour generated strictly on the world level by the choice made by industrial groups" (Michalet 1989). But on this level, it should be noted that faced with the crisis, with the formation of world-wide international oligopolies, and with the attraction of monetary operations, corporate planning is often short term and increasingly subordinate to purely financial strategies (including strategies formulated as a defence against stock-exchange raids). Corporate planning is characterized by rapid movements between engagement and disengagement of capital, and gives priority to "new forms of investment" (Oman 1980, 1989), showing a growing preference for the international exchange of industrial assets between big groups that are members of the same world-wide oligopoly. Apart from the Boetskys of this world, the present game is "I shall give you my medical electronics sector and you will give me your public electronics," of which the 1987 Thomson/General Electric agreement is a good example. On such foundations there can be no mastery of technological change any more than there can be a durable response in terms of the competitiveness of the national economy. It is therefore at this level, in relation to financial capital and its operations, that the real problem of competitiveness is situated - nowhere else. In the framework of this discussion on competitiveness, the conclusion now being reached is close to that of G. Destanne de Bernis when he writes: "[AJdaptation to what does not exist [i.e., a world-wide productive system as the result of internationalization] is a myth and an illusion; the external constraint is above all the result of a destruction of productive systems, from inside countries," resulting from Thatchertype government policies; and finally, "the crisis strikes simultaneously [more exactly, most strongly, because no country escapes

173 Technological Competitiveness

completely] all the countries that have allowed themselves to permit this destruction" (Destanne de Bermis 1987). Having established this, the following question, of course, arises: What conclusion can be drawn on the policy level? But that question lies outside the framework of this chapter. NOTES

1 The implications of this study for us competitiveness were subsequently drawn by K. Bozdogan (1989). 2 The terms "niche-specialization" and "pole-specialization" are borrowed from Turpin 1989. 3 Our definition of structural competitiveness differs from the one sometimes used by the CEPII (Centre d'etudes prospectives et d'informations internationales), where the term is synonymous with non-price competitiveness. See Fouquin 1986, chap. 2. 4 A good illustration of this type of approach to comparative advantage and structural adjustment is found in Lawrence 1984. For a presentation of the thesis of the United States' supposed comparative advantage in services, see inter alia us Congress, Office of Technology Assessment 1987. 5 For a summary of this literature, see Mucchielli and Sollogoub 1980 and Graham 1979. 6 For an earlier critique of this approach, see Chesnais and MichonSavarit 1980. 7 This summary of recent advances in the theory of technical change is based on my own reading of the collective volume edited by G. Dosi et al. (1988). See also Chesnais 1990. 8 See the special issue of Economic et statistique, no. 217—8 (JanuaryFebruary 1989). 9 See the studies by the BETA at the Louis Pasteur University in Strasbourg, in particular, the articles by E. Zuscovitch. REFERENCES

Andersen, E.S., B. Dalum, and M. Villumsen. 1981. "The Importance of the Home Market for the Technological Development and Export Specialisation of Manufacturing Industry." In Technical Innovation and National Economic Performance. IKE monography, Aalborg University, Denmark. Andersen, E.S., and B.A. Lundvall. 1988. "Small National Systems of In-

174 State, Technology, and Competitiveness novation Facing Technological Revolutions: An Analytical Framework." In Small Countries Facing the Technological Revolution, edited by C. Freeman and B.A. Lundvall. London: Pinter. Arrow, K. 1962. "The Economic Implications of Learning by Doing." Review of Economic Studies, no 29:155—73Baily, M.N., and A.K. Chakrabarti. 1988. Innovation and the Productivity Crisis. Washington, DC: Brookings Institution. Beaud, M. 1987. Le systeme national mondial hierarchise. Paris: La Decouverte. Boyer, R. 1986. La theorie de la regulation: une analyse critique. Paris: La Decouverte. Bozdogan, K. 1989. "American Industrial Competitiveness." MIT. Mimeo. Chesnais, F. 1987. "Les accords de cooperation technologique et les choix des entreprises europeennes." Communication au colloque Europrospective, CPE, Forecasting Agency for Science and Technology (FAST), Commissariat general au plan, La Villette, Paris, April. - 1988. "Technical Cooperation Agreement Between Independent Firms." STI Revue (OECD, December), no. 4:55-132. - iggoa. "Technological Cumulativeness, the Appropriation of Technology and Technological Progressiveness in Concentrated Market Structures." In Frontiers in Technology Diffusion, edited by F. Arcangeli, P. David, and G. Dosi. London: Oxford University Press. Chesnais, F., ed. iggob. Competitivite international et depenses militaires. Paris: Economica. Chesnais, F., and C. Michon-Savarit. 1980. "Some Observations on Alternative Approaches to the Analysis of International Competitiveness and the Role of the Technology." Paper presented at the first OECD Conference on Science and Technology Indicators, Paris, September. Cohen, S., and J. Zysman. 1987. Manufacturing Matters: The Myth of the Post-Industrial Economy. New York: Basic Books. Coomans, G. 1986. "Systeme productif et petites nations." Economies et societes, Cahiers de /'ISMEA 20, no. 5:49—67. de Bandt, J. 1988. "La filiere comme mesosysteme." In Traite d'economie industrielle, edited by R. Arena et al. Paris: Economica. DeBresson, C., and J. Townsend. 1978. "Notes on the Inter-industrial Flow of Technology in Post-War Britain." Research Policy 7:48—61. De Brunhoff, S. 1987. L'heure du marche. Paris: Presses Universitaires de France (PUF). Dertouzos, M.L., R.K. Lester, and R.M. Solow, eds. 1988. Made in America: Regaining the Productive Edge. Findings of the MIT Commission on us Industrial Productivity. Cambridge, Mass.: MIT Press. Destanne de Bernis, G. 1987. "Observations sur la contrainte exterieure." Economies et societes, series HS, no. 29, vol. 19, no 14.

175 Technological Competitiveness Dosi, G., et al. 1988. Technical Change and Economic Theory. London: Pinter. Fouquin, M., ed. 1986. Industrie mondiale: La competitivite a tout prix. Paris: Economica. Gaffard, J.L. 1988. "Mutations technologiques et choix strategiques des entreprises." In Traite d'economie industrielle, edited by R. Arena et al. Paris: Economica. Giget, M. 1988. "The Bonsai Trees of Japanese Industry." Futures, 20, 2:147-154. April. Gille, B. 1978. Histoire des techniques. Paris: La Plei'ade. Graham, E.M. 1979. "Technological Innovation and the Dynamics of us Comparative Advantage." In Technological Innovation for a Dynamic Economy, edited by C.T. Hill and J.M. Utterback. New York: Praeger. Hatzichronoglou, T. 1985. "An Initial Contribution to the Statistical Analysis of Trade in High Technology Products." OECD. Mimeo. Imai, K. 1984. "Japan's Industrial Policy for High Technology Industries, Conference on Japanese Industrial Policy in Comparative Perspective. New York University. Mimeo. LAREA/CEREM at Paris X, inter alia see Mytelka and Delapierre 1987. Lawrence, R.Z. 1984. Can America Compete! New York: Praeger. Lundvall, B.A. 1988. "Innovation as an Interactive Process: From UserProducer Interaction to the National System of Innovation." In Technical Change and Economic Theory, edited by G. Dosi et al. London: Pinter. Marshall, A. 1929. Principles of Economics. 8th ed. London: Macmillan. Melman, S. 1983. Profits without Production. New York: Alfred Knopf. Michalet, C.A. 1976. Le capitalisme mondial. Paris: PUF. — 1984. "Introduction." In L'integration de Veconomie franfaise dans Veconomie mondiale. A collective volume by LAREA/CEREM. Paris: Economica. Mistral, J. 1978. "Competitivite et formation du capital en longue periode." Economic et statistique 97 (February)^—23. — 1983. "Competitivite du systeme productif et specialisation internationale." Paris: OECD. Mimeo. Mucchielli, J.L., and M. Sollogoub. 1980. L'echange international fondements theoriques et analyses empiriques. Paris: Economica. Mytelka, L.K., and M. Delapierre. 1987. "The Alliance Strategies of European Firms and the Role of ESPRIT. "Journal of Common Market Studies 26, no. 2 (December). Noble, D.F. 1984. Forces of Production: A Social History of Industrial Automation. New York: Alfred Knopf. Oman, C. 1980. New Forms of International Investment. OECD Development Center Studies. Paris: OECD. — 1989. New Forms of Investment in Selected Industries in Developing Countries. OECD Development Center Studies. Paris: OECD.

176 State, Technology, and Competitiveness Passadeos, C. 1987. Recherche militaire et industrielle civile: elements de methode et remarques sur le cas de materiaux composites en France. Paris: Fondation pour les etudes de defense nationaie. Pavitt, K. 1984. "Sectoral Patterns of Technical Change." Research Policy 13, no. 6:343-74. Perez, C. 1987. "New Technologies and Development." In Small Countries Facing the Technological Revolution, edited by C. Freeman and B.A. Lundvall. London: Pinter. Perroux, F. 1962. L'economic des jeunes nations, industrialisation et groupement de nations. Paris: PUF. Quelin, B. 1988. "La diffusion des innovations: Une analyse interindustrielle." In Traite d'economie industrielle, edited by R. Arena et al. Paris: Economica. Robson, M., J. Townsend, and K. Pavitt. 1988. "Sectoral Patterns of Production and Use of Innovations in the U.K. 1945-1983." Research Policy 17, no. i: 1—14. Rosenberg, N. 1982. Inside the Black Box: Technology and Economics, especially chap. 6. Cambridge, Mass.: Cambridge University Press. Sahal, D. 1981. Patterns of Technological Innovation. New York: AddisonWesley. Teece, D.J. 1986. "Profiting from Technological Innovation." Research Policy, no. 15, 6:285—305. — 1989. "Technological Development and the Organisation of Industry." International Seminar on Science, Technology and Growth (Paris: OECD) June. Turpin, E. 1989. "Le commerce exterieur francais: Une specialisation industrielle fragile." Economie et statistique, nos. 217—8. JanuaryFebruary .-51-6 2. us Congress, Joint Economic Committee. 1981. The International Economy: U.S. Role in a World Market. Washington, DC: us Government Printing Office, December. us Congress, Office of Technology Assessment. 1987. International Competition in Service. Washington, DC us Department of Commerce. 1981. An Assessment of U.S. Competitiveness in High Technology Industries. Washington, DC: us Government Printing Office. Weiller, J. 1945. Problemes d'economie Internationale. Paris: PUF.

CHAPTER EIGHT

Indicators of Industrial Competitiveness: Results and Limitations THOMAS HATZICHRONOGLOU

In the context of a global economy, the indicators that measure the results of industrial competition, especially those indicators based on international commerce, increasingly lose their value if they are not linked to other variables, especially foreign direct investment (FDI) and technological exchange. Also, these indicators should not be isolated from the totality of the determinant factors if they are to be interpreted accurately. These factors can be directly tied to relative prices or to other structural causes. In this regard, we will point out the importance of the role of technology in determining competitive advantage. We will also examine the process of specialization followed by the major OECD (Organization for Economic Cooperation and Development) countries, since this process was the key element in the improvement or weakening of their competitive position.

THE P R I N C I P A L INDICATORS OF COMPETITIVENESS

The Concept of Competitiveness and Indicators of Competitive Position The term "competitiveness" is often employed by economists and public officials even though no consensus exists as to its definition. Certain authors use the term in different and at times contradictory senses, while others make no distinction between causal and effect variables. Thus, the terms "comparative advantage," "factor productivity," and "unit labour costs" are often used as synonyms for competitiveness.

178 State, Technology, and Competitiveness

This lack of clarity is due in part to the fact that competitiveness is a multidimensional concept that cannot be reduced to a single variable or indicator, no matter what its importance. A second difficulty arises from the fact that it is difficult to consider indicators of competitive position separately, without reference to underlying causal variables. As will be shown below, utilizing indicators of competitive position without referring at the same time to the variables directly or indirectly linked to those indicators may lead to errors in interpretation. These complications make it difficult to arrive at a commonly accepted definition of competitiveness. It seems, however, that there is a growing consensus among economists that an industry or firm is competitive if it is capable of maintaining its share of existing markets or of conquering new ones. This definition suggests that we can analyse competitiveness at either a macro-economic, meso-economic (i.e., sectorial) or micro-economic (i.e., firm) level. The indicator that is most frequently used to measure the conquest of a market is, of course, market shares: exports in the case of foreign markets, and the weight of imports in domestic demand in the case of the home market (rate of import penetration). Because the conquest of a market can be achieved not only through the export of goods and services but also through direct investment or the sale of technology (patents and licences), the concept of market share should be expanded. These three modes of market penetration or conquest can be used in unison or as complements or substitutes according to the characteristics of the markets in question, the goods and services being exchanged, and the technology gaps that exist between partner countries. Accordingly, indicators of market shares can be defined as follows: Foreign Markets

where

MSy MPy Xy My Dy

Domestic Markets

= market shares of country i for product j, = rate of import penetration, = exports of goods, technology, or capital, = imports of goods, technology, or capital, and = domestic demand for goods, technology, or capital.

179 Indicators of Industrial Competitiveness

Although the notion of market share is frequently used in the case of the export of goods and services, this is not the case for direct investments or for the sale of technology. In fact, an increase in direct investment is not, a priori, proof of improved competitiveness on the part of a firm. The reasons that a firm is driven to invest abroad are varied and at times may even be associated with a decrease in competitiveness. Thus, a firm whose costs of production are increasing more rapidly than those of its competitors may relocate a part of its production in order to benefit from lower labour costs. Similarly, a loss in technological advantage, which leads to decreased export returns for a given firm, may make it necessary for that firm to invest directly in countries where its competitors produce. However, very competitive firms, in order to consolidate their competitive position, also relocate part of their production in order to benefit from more favourable foreign conditions (i.e., lower labour costs, fiscal advantages, proximity of raw materials, more lenient labour legislation). In the face of the globalization of markets, such direct investment allows firms to acquire an international dimension. This strategy can gain impetus when the exports of the firms in question reach such a high level as to elicit protectionist pressures on the part of importing nations. Nevertheless, experience shows that few firms faced with a loss of competitiveness will invest in foreign countries. In the majority of cases, such direct investments indicate a certain dynamism and momentum geared toward the conquest of foreign markets. In order to take into account direct investment as an indicator of market conquest, we have the choice of the following dimensions: • sums invested abroad; • the turnover of subsidiaries weighted by the percentage of foreign control of capital; • realized profits; • repatriated profits. Each of these four indicators has its advantages and disadvantages. In practice, one must choose the one that can be most likened to exports. It would seem that repatriated profits most closely play such a role. Although firms can repatriate profits arising from their direct investment in the same manner that they can repatriate foreign exchange arising from their exports, the legislation on the repatriation of profits differs from country to country. On the other hand,

i8o State, Technology, and Competitiveness

numerous firms prefer to reinvest their profits, whereas others choose to forgo profits for a more or less extended period in order to strengthen their market shares. Other factors, such as the appreciation or depreciation of the host country's currency, can also encourage foreign investors not to repatriate their profits. However, the ultimate goal of direct investment is the repatriation of the profits that it generates. It is for this reason the most appropriate choice as an indicator of the return on investments appears to be repatriated profits. In regard to the sale of technologies as a means of market control, it is reasonable to choose returns coming from patents, licences, know-how, and technical assistance as an indicator of performance. l As far as foreign markets are concerned, repatriated profits and receipts from technologies can be likened to a flow of exports. On the other hand, in the case of the domestic market, comparable indicators include export profits of foreign firms and technological payments, which can be likened to a flow of imports. The simultaneous comparison of exports, repatriated profits from direct investment, and returns from technology payments raises numerous technical problems that are linked to temporal lags (CNRS 1983). At the same time, they raise another difficulty. While exports of goods and technologies are identical at the conceptual level, this does not seem to be the case for profits arising from direct investments, whether they be repatriated or not. In order to compare these three types of indicators, one would require data on profits arising from the export of goods and services as well as on profits accruing from the sale of technology. It is difficult, however, to define and measure profits derived from the sale of technologies if one takes into account the methodological difficulties encountered in trying to identify the production and real transfer costs of a given technology. The preceding highlights the gap that exists between ideal indicators of competitiveness and those presently available. Main Limitations The preceding indicators, regardless of the type of trade involved (goods and services, technology, or capital), only reflect export gains or losses in market share. We will see that further information is required in order to fully appreciate the significance of these indicators. Export market share. The fact that a country's share of industrial exports in a given market is increasing or decreasing with respect to

181 Indicators of Industrial Competitiveness

other competing countries is not in itself sufficient in determining the evolution of its competitive position. A host of other factors may well obscure the true meaning of observed variations. Changes in specialization. Losses in market share for a particular product can be the result of a change in industrial specialization. In fact, in numerous OECD (Organization for Economic Co-operation and Development) countries, certain categories of mature (i.e., standardized) products are successively abandoned in favour of new products that are more in demand. This apparent loss in competitiveness with regard to standardized products cannot be properly understood unless one takes into account such structural transformations. In order to overcome this problem, the constant market share technique has been developed (Learner and Stern 1970). This method consists of breaking down the gap in growth between world exports of manufactured goods and the exports of a given country into three components: • an effect due to the geographic location of the export market; • an effect due to the categories of products exported (categories of products in which the given country has a comparative advantage); and • a residual factor. Although this method provides a better mode of analysis, it suffers from two main drawbacks. First, it implicitly assumes that a country has the same ability to penetrate across different markets. However, it is well known that for various reasons (e.g., institutional barriers to entry, geographic barriers) the elasticity of substitution of exports of a given country with respect to each foreign market is not the same. Second, the residual factor component, which often groups together all other factors of competitiveness, frequently carries a weight greater than the other two components. The term "residual factor" may be misleading, since a country may, for economic reasons linked to its industrial strategy, choose to reduce its share of a given market, thereby seeming to become "uncompetitive." Moreover, it may export products to regions of rapidly increasing demand that do not have a sufficient internal capacity to maintain their market share. Under such a hypothesis, the country in question may raise its export prices while at the same time reducing its market share. Nonetheless, a poor geographic distribution of export markets that persists for an extended period of time generally reflects a lack of

182 State, Technology, and Competitiveness

ability on the part of the industry in question to capture expanding markets. It also puts into doubts its capacity to penetrate such markets. More generally, any analysis of a positive or negative change in the market share of an industrial sector must take into account the evolution of the industry as a whole, including services, in order to properly measure the impact that a significant change in specialization between different industrial sectors, or between industry and services (i.e., industrial losses compensated for by gains from services and vice-versa) may have on the competitiveness of a given sector. Strategic choices of firms. One can also find oneself in the situation where an increase in market share reflects a strategic choice on the part of firms. Japanese firms, for example, tend, more so than their competitors, to favour gains in market share over maximizing profits in the short term. Other firms, which maximize immediate profits over gains in market share, have an improvement in their financial position as their short-term objective. The successful implementation of these strategies depends on the initial level of the firm's production costs and on the influence that these exert on competitor's prices. If these costs (expressed in a common currency) are relatively low, exporting firms are freer to increase their prices. Under such a hypothesis, these firms can take advantage of their favourable position by increasing their profit margins if foreign prices happen to rise. Conversely, if production costs are high and deteriorating, the firms in question will accept a decrease in their profit margins in order to prevent a worsening of relative prices. These different possibilities, as a function of the initial level of a country's relative costs and of their evolution, are summarized in Table 32. Differentiated growth rates between domestic and foreign demand. Differentiated growth rates in domestic and foreign demand can bias the interpretation of changes in market shares. In effect, when the growth of domestic demand in a given country exceeds the growth of demand in its export markets, then there is a tendency for firms to use a portion of the production that would have gone into exports to satisfy the excess domestic demand. This phenomenon makes the interpretation of the above indicators even more difficult in that the fall in export market share that follows may be accompanied by a rise in the rate of import penetration. In nearly all countries, an increase in domestic demand is almost always accompanied by an increase in imports of differentiated prod-

183 Indicators of Industrial Competitiveness Table 32 Reaction of Export Prices to Variations in Competitors' Prices Initial Position

Relative Costs Variation

Reaction Prices

1. Low

Improvement

2. Low

Deteriorating

3. High

Improving

4. High

Deteriorating

Relative prices: good Profits: increasing Relative prices: falling Profits: maintain Relative prices: improving Profits: stable Accept a decrease in profits to avoid a deterioration in relative prices

Observed Effect on Competitiveness Strong Weak Weak Strong

ucts. In order to correctly measure the impact of different rates of growth in domestic and foreign demand on competitiveness, it is necessary to evaluate the difference between gains realized by national producers in the domestic market due to excess demand and the gains that they would have realized had their means of production allowed them to satisfy foreign demand. Symmetrically, if domestic demand in a country grows more slowly than foreign demand, then the rate of import penetration will fall while at the same time the market share of exports may rise. The real competitiveness of the country in question thus remains unchanged. Similar phenomena can have an impact on the indicators of third countries. If, for example, European exports for various reasons fall, the market shares of the United States and Japan will automatically increase, without the competitiveness of these two countries necessarily having improved. In order to avoid such errors of interpretation, it is important to highlight a certain asymmetry with respect to the competitive ability of firms vis-a-vis a collapse of demand on the one hand and an increase in demand that is difficult to foresee on the other. In the first case, we can conclude that the competitiveness of the firms should not be put into question. In the second case, however, the reply is more subtle. The absence of any leeway in a firm's capacity to respond to a sudden increase in demand may be due to insufficient investment or weak management. A firm is all the more competitive if it is able to adjust its capacity rapidly in order to capture a new market. Changes in the exchange rate. Movements in the exchange rate can influence the interpretation of market shares to the extent that they

184 State, Technology, and Competitiveness

affect the structure of relative prices. In reality, changes in export and import prices may not reflect changes occurring at the level of the exchange rate. A change in the relation between foreign prices and domestic prices (when the difference in these prices is due to differing inflation rates or a change in the exchange rate) will have repercussions on the competitive position of firms and will modify their price strategy. Supposing that profit margins remain constant, we can ask to what extent changes in monetary parity are reflected by changes in relative prices. It all depends on the elasticity of supply and demand of imports and exports. In the case of a two-country general equilibrium model, the variation in export prices with respect to a change in the exchange rate is given by the ratio

where:

Px = export prices expressed in local currency, R = the exchange rate, nx>(i = elasticity of exports with respect to foreign demand, nx>s = elasticity of exports with respect to domestic supply, and where the hat corresponds to a proportional rate of change. If national supply is perfectly elastic, that is to say the elasticity of exports with respect to domestic supply tends to converge toward infinity (nx>s —> +°°), export prices (in local currency) will not be affected by the exchange rate because PJR —> 0. When foreign demand is perfectly elastic (i.e., exports of raw materials are sold at world prices, since the country's market is too small to influence them), then the elasticity of export prices (in local currency) with respect to foreign demand will converge toward negative infinity (nx,d ~* ~°°)> and the variation in prices (in local currency) will be proportional to the exchange rate while at the same time remaining constant in foreign currency given that thatPx/R —> 1. The same reasoning may be applied in the case of imports. However, these considerations are valid only in the short run. In the long run, all will depend on the links that are established among prices, wages, and the profit rate. Substitution and complementarity between different modes of market penetration. Another phenomenon that renders the interpretation of mar-

185 Indicators of Industrial Competitiveness

ket shares difficult is the partial or total substitution or complementarity of different modes of market penetration. The relocation of production by means of direct investment can create over a certain period of time new exports of goods and services and has a complementary effect on trade in products. However, after a certain threshold is reached, relocalized production may be substituted for exports and may even be transformed into a flow of exports to the country of origin.2 Under these conditions, the market share of exports may diminish in favour of gains arising from direct investment. As far as investment strategy is concerned, difficulties in interpretation can occur around the real meaning of the term "repatriated profits." It often happens that declared profits, notably those of multinational firms, do not correspond to reality. For fiscal reasons, certain subsidiaries, following the global strategy of the group to which they belong, may artificially report a loss so as to minimize their taxes and maximize overall profits. Most often it is the price of services that facilitates such operations because of increased verification problems in this regard. This highlights the magnitude of the problems that the interpretation of indicators of competitiveness conceived at the national level can give rise to when the principal actors involved are internationally integrated groups who formulate their strategies at a global level. The difficulties in properly interpreting receipts from technology are similar, and at times even closely linked, to problems in interpreting repatriated profits. It is not unreasonable to formulate a hypothesis whereby firms offset a part of their profits coming from other activities against the transfer cost of know-how and of technical assistance, costs whose level is more difficult to determine. Under this hypothesis, it would be necessary to add the two entries together in order to capture any compensation effects that may occur. Conversely, exports of technology (patents or licences), although they represent only a small percentage of the export of goods and services, may have a direct impact both on trade and on direct investment. The links between the flow of export goods, capital, and technology are complex and up to now have not been the subject of extensive empirical study. Import penetration rate. Competitiveness in the domestic market, as measured by the rate of penetration of imports, is based on the idea that a national industry tries to increase, or at least maintain, its share of the domestic market. However, if we recognize that certain national producers may be foreign firms and that a certain

186 State, Technology, and Competitiveness

proportion of imports may come from subsidiaries of national firms, then a proper evaluation of market penetration by foreign firms must take into account production under foreign control that is intended for the local market (assimilated under imports). It must also exclude that portion of imports produced by subsidiaries controlled by national producers. Unfortunately, the majority of available data make no distinction in regard to the ownership of capital, thus reducing the significance of foreign penetration indicators to a considerable degree. As we will see later on, the gap between apparent and effective penetration can be important in the case of certain sectors. Analogous reasoning can be applied to export market shares. In effect, the exports of domestic firms controlled by foreign capital can be replaced in calculations by exports of national firms situated abroad. It is evident in the final analysis that what is lacking is a clear definition of the term "national firm." The preceding considerations highlight the need for a proper interpretation of the most frequently used indicators of competitiveness (export market share and the rate of import penetration) and for complementary information that is often missing. On the other hand, the indicators in question only reflect past or static competitive positions. In order to envisage a more predictive approach toward competitive tendencies, it is necessary to identify the principal factors that contribute to changes in the preceding indicators as well as to establish the causal relations between these two categories of variables. In the next section, we will briefly examine the changes that have occurred from 1975 to 1987 in each country's strategy for conquering domestic and foreign markets. The Three Modes of Market Penetration and the Results of Principal Competitive Positions Foreign markets. Up to the first oil shock (1974-75), the conquest of foreign markets in most OECD countries took place primarily through exports. Direct investment, which actually had increased during the 19605, occupied only a secondary position relative to exports, save for certain countries such as the United States and to a lesser degree the United Kingdom and Canada, where for historical reasons industrial development coincided with the introduction of numerous production units into foreign countries. Table 33 presents changes in the relative shares of exports, direct investment, and technological receipts of each of the seven principal

187 Indicators of Industrial Competitiveness Table 33 Relative Weights with Respect to Totals of the Seven Countries (percentages rounded off)

1975 Flow of

1985 Flow of Direct Inv. Abroad

Technological Receipts

52

70

9 8 4

4 5

Italy Canada

26 17 22 9 13 8 5

Total

100

Export Manuf.

us Japan W. Germany France UK

Export Manuf.

22

11

1 4

2 1

20 22 23 10 10 8 7

100

100

100

7

Direct Inv. Abroad

Technological Receipts

31

68

17 11 6

7 5 8

21 7 7

8 2 2

100

100

Source: Calculations based on various OECD sources.

OECD countries for 1975-85. During the whole of this period, the United States continued to be the principal seller of technology (selling more than two-thirds of the total of the seven countries). This dominant position is in sharp contrast to its weakening position in the share of exports and flows of direct investment. Between 1975 and 1985, the weight of Japan in all three modes of market penetration increased, more so in the case of exports and direct investment. A similar phenomenon occurred in nearly all the other countries, with the exception of the United Kingdom, whose market share shows a slight decline both in the export of goods and in the sale of technology, and the United States, as described above. The evolution that we observe during the 19808 reflects certain changes in industrial strategy at a global level. The internationalization of markets, rapid changes in technology, and the appearance of new industrializing countries (NIGS) as competitors forced firms to move off shore a growing part of their production. It also forced them to find a new equilibrium between strategies based on exports and those based on direct investment. The decline in the us market share of manufactured exports (see Table 35) as of 1980 is not attributable to an increased substitution of direct investment for exports or to a decline in the growth of export markets (see Table 43). Nor does it correspond to an improvement in the market shares of exports in other categories, such

i88 State, Technology, and Competitiveness Table 34 Export Market Shares of Services (1982 prices) 1975

us Japan W. Germany France UK

Italy Canada Total 7 Countries

25.9 7.6 15.0 17.6 18.9 10.2 4.5 100.0

1980

1982

26.5 9.8 14.9 18.1 15.9 10.2 4.4

25.8 10.8 17.2 17.9 14.5 9.8 3.8

100.0

100.0

1985

23.3 10.4 17.6 19.0 15.6 9.7 4.1 100.0

1986

24.9 10.0 17.3 18.1 15.9 8.9 4.6 100.0

1987

27.3 10.7 15.6 16.9 16.0 9.1 4.2 100.0

Source: OECD, Industry Division.

as services (see Table 34). On the contrary, it corresponds to the rise in the us dollar during this period. If this rise in the dollar played a fundamental role in the decline in the market shares of manufactured exports during the first half of the igSos, its predictive power remains limited, since, conversely, market share did not improve during the 19705 when the dollar was weak. The principal change observed since the dollar began to depreciate in 1985 is that exports have again become the primary American strategic instrument for the conquest of foreign markets. Japan, until the mid-19805, followed a different path than that of the United States. During its period of restructuring and technological catching up, exports were the primary route to its conquest of foreign markets. For many years, helped by a weak yen, this route continued to be the main mode of penetrating foreign markets. However, after the second oil shock and especially since 1985 (a period corresponding to a strong appreciation of the yen with respect to the us dollar and to the proliferation of protectionist measures), Japan embraced a new strategy, choosing direct investment and the sale of technology as equally important vectors for the conquest of foreign markets. Japanese exports may have even suffered a slight decline as reflected by a fall in Japanese export market shares commencing in 1986. It would be erroneous, however, to see this relative decline of Japan in foreign markets as a sign of weakened competitiveness. On the contrary, the rapidity with which Japanese firms successfully directed an important part of their production toward interior markets so as to compensate for the decline in the volume

189 Indicators of Industrial Competitiveness Table 35 Export Market Shares of Manufactured Products" (1982 dollars)

us Japan W. Germany France UK

Italy Canada

Total

1975

1980

1982

1985

1986

1987

25.7 16.6 21.8 9.6 12.8 8.0 5.3

24.8 18.7- . 21.3 10.6 10.7 7.9 5.9

20.8 20.4 23.7 10.6 10.2 8.3 5.8

20.0 22.0 23.3 9.7 9.8 8.2 6.9

21.0 21.4 23.3 9.4 9.8 8.1 7.1

23.0 20.2 22.7 9.0 10.1 7.9 6.9

100.0

100.0

100.0

100.0

100.0

100.0

" Classification SITC 5 + 6 + 7 + 8 + 9. Source: OECD, Industry Division.

of their exports shows a great competitive dynamism and a strong capacity to adapt. As to the strategy of transferring and selling technology, we can say in general terms that this is linked to two factors: the simultaneous presence of a high level of research and development and a strong network of foreign subsidiaries. We find these two conditions present primarily in the United States and to a lesser extent in the United Kingdom. These are the only large countries whose balances of technological payments are in surplus. The domestic market. An increase in the rate of penetration of manufactured imports as well as an increase in the purchase of foreign technology cannot be systematically seen as a sign of weakening competitiveness in the domestic market. The modernization of productive equipment can justify these types of purchases, and such modernization in principle signals an improvement in competitiveness. The United States, since the start of the 19608, has seen its position weaken in both the domestic market and foreign markets. Since the early 19705, while its export market share of manufactured goods has diminished, falling from almost 26 percent in 1975 to 23 percent in 1987 (see Table 35), the share of its domestic demand satisfied by imports has more than doubled, rising from 5.5 percent in 1970 to 12.2 percent in 1984 (see Table 39). In the domestic market, the falling rate of penetration relates to products of low to medium R&D intensity. While the penetration of imports, as well as their elasticity with respect to domestic demand, has been greater in products of

igo

State, Technology, and Competitiveness

Table 36 Export Market Shares in Manufacturing Industries as a Function of Their Technological Content (current prices) Total Ind. Manufactures

1970

High-Intensity

Medium-Intensity

Low-Intensity

R&D

R&D

R&D

1986 1970 1986 1970 1986 1970

17.8 9.7 16.6 8.2

13.3 16.2 18.0 8.4

28.3 12.0 16.1 7.3

9.2

6.9

9.6

8.8

Italy Canada Other OECD

6.4 6.4 130

7.5 5.3 24.4

5.0 3.5 18.2

4.3 2.3 16.7

19.5 7.6 20.7 7.7 10.7 6.4 8.0 19.4

Total OECD

100.0

100.0

100.0

100.0

100.0

us Japan W. Germany France UK

22.1 24.1 14.6 7.1

1986

12.2 17.6 21.7 8.2

11.3 11.1 12.7 9.1

6.7

7.5

6.7 6.7 20.2

7.1 5.9 35.3

10.8 5.5 35.2

100.0

100.0

100.0

8.5

8.9 15.6 9.6 5.9

Source: OECD, Industry Division.

Table 37 Relative Weights in the Total Flow of Outward Foreign Direct Investment of the Seven Countries

1965 us Japan W. Germany France UK

Italy Canada Total

1973

1978

1980

73.5

51.9

48.9

43.8

1.1 4.6 3.8

8.7 7.5 4.2

7.2

5.4 9.5 7.1

12.6

2.6 1.7

100.0

10.9 5.4

22.8

20.6

1.1 3.5

0.5 6.2

1.7 6.1

100.0

100.0

100.0

26.1

1985

1986

31.0 17.0 11.0

33.7 17.4 10.8

21.0

21.4"

6.0

7.0 7.0

100.0

6.3 3.2

7.0°

100.0

" Estimation. Source: OECD, Industry Division.

high technological content, the trade deficit of recent years is essentially attributable to products of low to medium R&D intensity, which have become a more important part of trade, especially in imports. The most notable structural change regarding American imports of manufactures is the increased importance of Japan and NICS in American imports. Since 1965, the weight of Japanese products in total American imports of manufactures has doubled (increasing

igi

Indicators of Industrial Competitiveness

Table 38 Relative Weights of Technological Receipts with Respect to Total Receipts of the Seven Countries (current dollars)

us Japan W. Germany France UK

Italy Canada Total

7972

7975

7975

1980

1982

1985

68.6 4.5 5.7 6.4 11.9 1.7 1.0

71.1 4.1 4.2 6.9 10.5 1.6 0.9

68.0 5.6 4.2 8.1 10.9 2.1 0.9

66.8 6.6 5.7 8.9 8.9 2.1 0.9

67.3 7.1 5.4 8.2 8.4 1.5 1.8

70.3 7.9 5.0 7.3 6.3 1.1 1.9

100.0

100.0

100.0

100.0

100.0

100.0

Sources: OECD, Industry Division, and DISTI.

Table 39 Apparent Rate of Import Penetration for Total Imports of Manufactured Goods

us Japan EEC

W. Germany France UK

Italy Canada

7970

7975

7950

7952

7954

7955

5.5 4.5 20.6

7.0 5.8 25.0

9.3 5.7 30.0

9.6 5.6 31.6

12.2 5.8 34.8

11.7 5.4 35.7

19.2 16.2 14.2 15.8 26.9

24.1 17.9 19.4 21.4 29.7

30.6 22.8 25.3 29.2 31.9

33.1 25.3 26.2 27.8 26.5

37.9 26.6 31.2 28.2 33.1

39.2 27.5 31.3 30.0 33.8

"Twelve countries. Source: OECD, Industry Division.

from n in 1965 to 26 percent in 1987), mostly to the detriment of Europe and Latin America. At the same time, the growth in the share of imports coming from the NICS has been even more spectacular (increasing from 8.7 percent in 1963 to 20.2 percent in 1984). However, this increase involves primarily large-scale consumer goods that have become standardized. Starting in 1980, there has been a resumption of direct investment in the American domestic market. Notwithstanding Japan's increased importance, Europe remains the United States' main foreign investor. However, during recent years the growth in Japanese investments has been four times that of Europe. It is not easy to interpret such movements in terms

192

State, Technology, and Competitiveness

Table 40 Apparent Penetration of Manufactured Imports as a Function of Their Technological Content Total Ind. Manufactures

1970 us Japan W. Germany France UK

Italy Canada

5.5 4.5 19.2 16.2 14.2 15.8 27.0

1985

11.7 5.3 39.0 27.5 31.3 29.8 33.9

High-Intensity R&D

Medium-Intensity R&D

Low-Intensity R&D

1970 1985 1970 1985 1970 4.9 6.2 22.6 24.0 17.3 19.4 43.5

7.0 5.3 22.5 23.0 19.0 22.5 52.7

17.8 7.4 64.6 40.0 55.0 45.8 66.9

16.3 5.6 41.9 34.6 44.1 33.9 61.7

4.8 6.6 17.0 11.8 11.6 12.0 12.8

1985 7.7 4.7 32.0 21.0 20.3 24.1 13.3

Source: OECD, Industry Division.

Table 41 Inward Direct Investment Flows (millions of current dollars)

us Japan W. Germany France UK

Italy Canada

1965

1970

1975

1980

1982

1984

7956

415 47 915 237 551 286 496

1464 94 597 622 870 -5 867

2620

16920

13800

25390

25050

225 687 1457 3361 631 713

278 424

278 825 1563

-10 740 2198 -241 1290 1312

226 136

3327

10124 597 684

5286

636 -831

2749 7960

-168 1116

Source: OECD, Industry Division.

of competitiveness, at least in the short run. However, this buying up of American enterprises by foreign firms, particularly in sensitive high-tech sectors, may have consequences on industrial competitiveness that are difficult to foresee. The strengthening of Japan's competitive position involves exterior markets as much as it does that country's domestic markets. While Japan's share of export markets has increased greatly, the apparent penetration of imports has shown a remarkable overall stability. Import penetration has even been reduced by half in the computer industry (from 22.5 percent in 1970 to 12.6 percent in 1985) and has barely been maintained at i percent in the automobile sector. The sectors that show a net increase in the rate of import penetration are those that plan a reduction in capacity and that, in

193

Indicators of Industrial Competitiveness

Table 42 Technological Payments (millions of constant dollars)

us Japan W. Germany France UK

Italy Canada

1972

1975

1978

294 705 499 355 405 376 121

473 592 629 471 578 480 190

610 687 851 686 714 827 324

1980

725 1057 1449 1105 820 635 372«>

1982

215 1134 1116 995 725 597 419

1985

196 1231 1202 1064 597 566 464

" 1971; * 1979;r 1981. Sources: OECD, DISTI.

Table 43 Real Growth Rates (average annual rate) Export Markets for Manufactured Products

Domestic Demand

us Japan W. Germany France UK

Italy Canada

1976-80

1981-85

1985-87

1976-80

1981-85

1985-87

2.6 4.3 3.3 2.5 1.1 3.9 3.1

4.0 3.5 0.8 1.3 3.2 1.8 1.7

3.0 4.5 1.5 1.6 3.9 4.0 4.1

2.6 2.5 4.1 3.5 3.6 3.8 5.9

3.1 5.0 4.6 3.7 2.6 3.3 12.2

1.8 3.4 3.6 2.3 2.0 3.7 7.3

Source'. OECD.

a first stage, substitute imports for national production. The slight increase in the penetration of imports in high-technology sectors is imputable to American products, while that of low-technology products is due in large part to imports from South-East Asia. Europe has lost ground since the early 19705 both in its domestic and foreign markets.3 In the domestic market, the high rate of import penetration reflects above all the opening up of the European Community (EEC) to international trade. After the second oil shock, the growth in the rate of import penetration was as rapid as that in the United States. There, however, the growth of domestic demand was two times more rapid than in Europe. These changes indicate a certain loss of European competitiveness. It is especially noticeable in leading sectors where, since 1970, the penetration of imports from

Figure 8

Rate of Apparent Penetration of Imports of Manufactured Products for the us, Japan, and the EEC

Source: ?: OECD,

Statistiques du Commerce Exterieur et Banque de donnees COMTAP.

195 Indicators of Industrial Competitiveness Table 44 United States: Foreign Penetration, 1977-1982 Portion of Production under Foreign Control

Rate of Effective Penetration

Rate of Apparent Penetration

Sectors

1977

1982

1977

1982

1977

7952

Chemicals Construction materials Iron, non-ferrous metals Electrical/electronic good Wood, paper, printing Food products Transport materials Metal products Textiles, clothing

9.7 5.0 5.5 4.9 2.7 3.7 0.1

24.1 10.7 10.3

5.6 4.2 8.2

14.7

30.9 14.8 17.9 19.3 12.8

2.2 1.2

6.7 5.3 5.1 4.3 2.9 2.1

5.2 3.3 5.9 9.9 6.5 3.7

Total

3.7

8.2

8.3

12.4

11.2 14.9

3.9 8.9

16.7

11.1

20.9

13.0

10.0

14.8

6.7

8.8

10.5

17.1

11.0

7.6 4.1 5.1

9.1 7.5

6.1

9.1 8.0

Source: us Department of Commerce, "Survey of Current Business.'

the United States and Japan has increased i .6 and 5 times respectively. Apparent and effective penetration of imported manufactured goods. In the case of the United States, the comparison of apparent and effective import penetration at the sectorial level allows us to measure, beyond traditional trade relationships, the impact of multinationalization on the penetration of the us domestic market. Production under foreign control was relatively small in the United States until 1977 (about 3.7 percent). Since then, however, it has shown remarkable growth, especially after the 1986-87 fall in the dollar (more than 10 percent). The results of certain OECD calculations (see Table 44) show that effective penetration4 in the United States is much stronger than traditional calculations would lead us to believe. This rapid increase confirms the priority given by many foreign firms, which aspire to achieve a significant share of the world market, to setting up in the American market. Comparisons between apparent and effective penetration in the United States show that penetration is doubled (and even quintupled in the case of the chemical industry) if production and exports under

196

State, Technology, and Competitiveness

Table 45 Canada: Foreign Penetration, 1972-1980 (%)

Sectors Machine and metal products Construction materials Chemicals Paper, printing Other products Iron Textiles, clothing Wood, furniture Food products

Portion of Production under Foreign Control

Rate of Apparent Penetration

1972

1980

1972

1980

75 54 92 35 48 19 31 24 39

66 60 81 32 42 15 32 16 30

43 13 16 10 41 22 19 5 7

47 16 15 11 45 22 20 5 7

Source: CEPII (France).

foreign control are taken into account. In the case of the chemical sector, this result may be due to the pharmaceutical industry, where American health policies encourage the reduction of imports and promote production in situ. In Canada, more than 50 percent of the production of manufactured goods is controlled by foreign capital. If in the past Canada relied on foreign capital, it now does so less and less: in 1960 it received 13.7 percent of international investment, 16.8 percent in 1971, and 10.4 percent in 1983. Most of these investments were American and British in origin. In 1980, 42 percent of foreign subsidiaries in Canada were American in origin and 28 percent were British (if we consider foreign subsidiaries as firms in which more than 50 percent of their capital is of foreign origin). In general, as can be seen from Table 45, foreign control is higher in those sectors where Canada has a trade deficit: chemicals, machinery, and construction materials. It is weaker in those sectors that are in surplus: wood, paper, steel, non-ferrous metals, and food products. THE P R I N C I P A L FACTORS OF COMPETITIVENESS

The analysis of competitiveness often rests on the distinction between price and structural competitiveness. Structural competitiveness is

197 Indicators of Industrial Competitiveness

often defined in terms of those factors that do not relate to price. In particular, it includes economic specialization, technological innovation, the extent of distribution networks, and a host of other factors that represent the state of supply. These two aspects of competitiveness are in fact closely linked. A favourable change in competitiveness can result from a price advantage but may also result from a judicious specialization that is the result of investment decisions or an important technological innovation. Price Competitiveness

Price competitiveness can be evaluated by measuring price differentials (production, export and import prices) between different producers and exporters. Export prices are in reality indices of export unit values recorded by customs. This is an advantage to the extent that this method assures the representativeness of goods that are in real competition on export markets. However, these prices do not take into account losses in competitiveness of potentially exportable goods that are not exported owing to their high prices. Equally, cost indicators could be a better measure of competitiveness than export prices if we consider that the latter also reflect changes in profit margins. Cost indicators in general use today relate primarily to labour costs (in particular, wages). A better comparison involving costs would instead involve reasoning in terms of net hourly costs (i.e., after taxes). In effect, differences between countries in hours of work and in taxation (professional taxes, social charges, etc.) can lead to distortions in international comparisons. At the same time, in order for changes in costs to be a more revealing indicator, variations in labour productivity must be taken into account. The comparison of costs can just as easily be affected by a certain disequilibrium linked to the geographic location of activities. For example, in the case of industrial restructuring, the closure of a production unit in a region with low labour costs can have a negative impact on the structure of costs for the overall production of an industry. On the other hand, the majority of these costs are calculated on products sold on the national market, whereas more and more firms subcontract out a growing part of their production to firms in countries with low labour costs. The transition in international comparisons from global wage costs to unit wage costs raises other difficulties because of the absence of specific exchange rates reflecting purchasing-power parities that correspond to each industrial product.

198 State, Technology, and Competitiveness

The major drawback to using costs as a measure of competitiveness comes from the fact that they refer only to labour costs. It is not an easy matter to correctly calculate the cost of capital. It is also difficult to find comparable international data (Bernheim and Shoven 1986) to the extent that it is necessary to take into account the rate of interest and taxation system of each country. If one assumes that these factors, once correctly measured, are found not to differ substantially in any two given countries, a difference in costs attributed to capital may have as its origin a difference in the lifespan of equipment. Unfortunately, other important categories of costs are also not taken into account, neither in the cost of labour nor in that of capital. These include R&D costs, costs of distribution (agent costs), negotiation costs (purchasing-agent costs), and various other categories of financial charges. All these drawbacks significantly limit the relevance of cost indicators. Insofar as production prices are concerned, they have the drawback of involving a whole range of products and services that are not subject to international competition. To the extent that international comparisons between prices or costs are carried out by means of a common currency, competitiveness, as represented by price differentials, will be a real effective exchange rate (Durand and Gorno 1987). The advantage of the latter with respect to a nominal effective exchange rate lies in the fact that it takes into account changes in real prices in different markets. In effect, experience shows that exporters often prefer to lower their prices in certain markets in order to maintain their price competitiveness. In summary, for any given country, price competitiveness is the difference between its price and a weighted average of competing prices. Price competitiveness in the domestic market (import prices). In any given market, we assume that for a particular product there is an equalization of the national producers' prices. Competitiveness will then be measured by the gap between the prices of national producers and those of their competitors. For a given product in the domestic market of a country this gap can be defined as follows:

where

PR? = the relative price between the domestic price and the prices of the competitors on the domestic market, of a country X,

199 Indicators of Industrial Competitiveness

Pd Pi\ W2A

= national producers' prices in country A's market, ~ prices of country is exports to country A, = the market share held by country i in the total imports of country A.

The denominator of this relative price is nothing but an approximation of the import price of all the competitors in market A. On the other hand, because of the lack of homogeneity in production prices, we can take the price of domestic demand as an approximation of the price of production. Price competitiveness in foreign markets (export prices). This section deals with how to measure the price competitiveness of a country in foreign markets (other than its own). In reality, measuring a country's price competitiveness means being able to situate the export prices of that country with respect to all markets. For example, in the Japanese market an exporting country is in competition not only with national producers but also with other non-Japanese competitors. The prices of that country's competitors in the Japanese market will be determined as a function of the structure of Japanese supply (production and imports). The price of the country's competitors with respect to all markets is obtained by aggregating their prices in each market. Assuming that export prices of each country do not depend on the country of destination (which is a strong hypothesis), then the relative price of the country in question with respect to those of its export competitors in a given market A will be equal to

where

P/?iA PXi PCSix PA PXfl W^

= the relative price of country i with respect to its competitors in foreign market A, = export price of country i, = the export price of country i's competitors in market A, = the production price in market A, = the export price of country /A, = the share of imports from country /u, in market A in the total supply (production and imports) of A,

200 State, Technology, and Competitiveness W AA

= the share of production of country A in the

total supply of market A, and

= the share of imports coming from country fJL in the total supply in market A with the exception of imports from country i. This is justified to the extent that we want to measure the export prices of country i's competitors and by the fact that country i cannot be in competition with itself. The preceding approach involves the measurement of the relative export prices of a country with respect to its competitors in a particular market A. In order to proceed to the total of all markets it is necessary to aggregate the prices of competitors in each market according to the export structure of the country under consideration. Thus,

where

PCXi = the aggregated price of country i's competitors on all markets, PCXfo = the export price of country i's competitors in market A, and XJA = the share of country i's exports to market A in the total exports of country i.

Therefore, the relative export prices of country i with respect to those of its competitors in all markets (PR*) is given by

where again the price differential between country i and its competitors is equal to PXZ - PCX, . The Role of Factors of Competitiveness Linked to Relative Prices

The evolution of relative export prices, whose principal indicators were described above, primarily reflects changes in costs; effective exchange rates, and policies regarding profit margins, notwithstand-

2oi

Indicators of Industrial Competitiveness

ing the fact that variations in profit margins (sometimes involving more than 30 percent of cases considered) can be linked to gains or losses in financial markets rather than to industrial activity. The simultaneous comparison of these indicators allows us to identify the periods during which non-price factors (structural factors) have played a more important role in the evolution of industrial competitiveness. Thus, in the United States, during the period 1976-80, we observe a slow progression in labour costs, denominated in local currency, which, under the effect of a fall in the effective exchange rate, show a decrease that is also reflected by relative export prices. If we take into account the evolution of costs and prices during this period, recorded losses on export markets (see tables 35 and 46) do not seem to be due to a weakening of growth in these markets or to a more rapid growth of the domestic market (see Table 42). We can therefore suppose that during the period 1976-80 the role of prices was less determinant in the loss of foreign markets. On the other hand, the period 1981—85 is marked by a strong rise in the dollar and a deterioration of costs and relative prices denominated in common currency. Clearly, during the course of this period, the fall in export market shares, the increase in the penetration of imports, and the deterioration of the trade deficit in manufactured products were linked to price rather than to structural factors. This period is characterized by a strong growth in domestic demand, while at the same time import demand more than doubled. According to a study of the FBR of New York (see Quarterly Review, Spring 1986), between the end of 1980 and the start of 1985 this single factor explains 40 percent of the deterioration in the trade balance of the United States. The growth in the American budgetary deficit and the appreciation of the dollar, notably vis-a-vis the yen, account for another 30 percent of the deterioration. We can observe that during the course of the first subperiod (1967—80), competitiveness decreased in the majority of sectors, including leading industries, while from 1980 the most affected sectors are mature ones, with the sole exception of the computer industry (see Table 46). The rise of the dollar had two structural consequences. The dropping competitiveness of American industry as a whole (particularly in mature sectors) accelerated the tertiarization of the American economy, notably at the employment level. At the same time, numerous segments belonging to traditional industries that had become non-competitive disappeared from the national market and were replaced by imports. The recent fall in the dollar, starting in 1985, has been unable to revive those industrial segments

Table 46 United States: Trade Balance (millions of dollars) World

1985

OECD

1986"

1985

Japan

EEC

1986"

1985

1986"

1985

France

Germany

1986

1985

1986

1985

Canada.

1986

1985

1986

HIGH R&D INTENSITY

Aerospace Office machinery, computers Electronic components Drugs, medicine Instruments Electrical machinery Subtotal

10743

1 906

699

911

-673

-523

-581

-829

4879 -3 449 -4 809 319 -13355 -14620 149 394 350 974 2718 -3374 -1 411 -2 436 -2 972 5968 -20013 -23522

866 -97 -31 -47 -536 851

922 -278 -32 -165 -833 522

738 15 93 290 -54 408

754 69 70 326 -121 575

1 469 -163 142 727 373 1 968

1 096 120 151 686 192 1417

10578

3638

4294

952

3889 1433 -17274 -19027 884 892 477 -634 -1 748 -4723 -6003 -13085

3749 -12603 500 -656 -2972 -8344

2337 -13870 592 -1 682 -4006 -12334

4539 656 38 1 162 -776 6573

-39 820 -51 804 4081 46322 -7229 -8471 2737 -3007 -5530 -6373 -5064 -6037 -1 524 -1 568 -52351 -72631

-41 341 -1 272 -2924 -7087 -2 174 -3037 -1 338 -59 176

-51 802 -9464 -11 401 -24253 -916 -971 -1 329 499 -3200 -1 900 -2028 -867 -11353 -3607 -5831 -5 544 -2499 -850 -906 -999 -793 -93 -3406 922 -1 326 -174 -190 -1 185 -74917 -17 708 -22251 -32443

1 058

1 551

MEDIUM R&D INTENSITY

Motor vehicles Chemical Other manufacturing ind. Non-electrical machinery Rubber, plastics Non-ferrous metals Other transport Subtotal

-31 780 -7509 -9301 -722 -637 -6337 -7094 -694 -526 -641 -901 -1 036 -731 368 30 45 -111 -122 -87 -90 -926 1 659 2 144 -6 746 -2609 -3 807 -26 -203 -329 -532 -243 -265 -166 -182 -1 028 -128 -225 -86 -102 -1 658 -2019 -256 -27 -44 -20 -81 11 -55 -1 103 -41 345 -11 428 -14 668 -2 146 -2 029 -6817 -8512

LOW R&D INTENSITY

Ceramics, clay, glass Food, beverages, tobacco Shipbuilding Petrol refineries Ferrous metals Fabricated metal products Paper & printing Wood, cork, furniture Apparel, shoes, leather Subtotal TOTAL MANUFACTURING IND.

" Excluding Spain. Source: OECD.

-280 -309 -189 -240 -3 109 -3952 -2764 -1 562 -1 895 -2255 -732 -732 -308 -889 -875 -6966 -7006 -2969 -2489 -1 917 -3688 1 093 1 491 -290 -174 -63 -39 -114 -141 -336 -465 -185 -137 150 -252 2 -67 -50 -81 -49 -12923 -8878 -1 206 -1 657 -1 114 709 520 -2332 -907 -759 -614 -10526 -8983 -7 191 -3463 -2966 -3 497 -2 565 -1 076 -8530 -817 -443 -103 -99 -3981 -4940 -3 131 -2707 -930 -1 508 -1 489 -348 -14 26 14 -3901 -4004 77 -5308 -5050 22 314 389 -35 -56 -6813 -7469 -755 -119 -172 -48 -4792 -4612 -713 172 245 -277 -298 -486 -477 -25 121 -28 294 -5933 -4 188 -4321 -1 122 -1 231 -5572 -73 192 -73 643 -34 747 -33 550 -15012 -14 009 -4 906 -3 836 -2634 -2 679 -2 595 -2437 -130 154 -157915 -101 449 -120 122 -25 859 -29 793 -56868 -68261 -13 149 -16796 -4330-3891

89 -804 393 -1 358 -830 -227 -4981 -4037 22 -11 733 -16644

-56 -1 005 311 -662 -1 010 -479 -5332 -4207 23 -12419 -19619

203 Indicators of Industrial Competitiveness

and reduce imports. The rate of penetration of imports has continued to increase, especially in the structurally weak sectors. The gains in export markets realized after 1985 are the result of a fall of 14 percent in relative prices and of traditional industries having again become competitive. However, when one takes into account the magnitude of the fall in relative prices, these gains appear relatively weak. They remain fragile to the extent that many American industrialists expect a new appreciation of the dollar and therefore are hesitant for the moment to initiate new investments in productive capacity. In Japan during the period in question, costs have progressed more slowly than in most other countries, but prices, with the exception of the period 1981-85, have increased more rapidly than in the United States. When the us dollar was rising, Japan's costs and relative prices remained low, which explains in part its good performance in export markets as well as in its domestic market. It is only during the period 1985-87 that the gap between costs and relative prices expressed in a common currency widened. The latter increased by only 9 percent, while relative costs and the effective exchange rate increased 18 and 20 percent annually. This difference can only be explained by the fact that Japanese firms preferred to compress their profit margins rather than increase their prices in proportion to the exchange rate. However, the explanation of the performance of Japanese industry in terms of market shares, rate of penetration, and trade surpluses is not restricted solely to costs or relative prices. Other structural factors played an equally important role (as will be outlined in the next section). In Europe this evolution varies between countries. In Germany, until 1985, the weak progression and even the fall in export prices were obtained only with a moderation of labour costs and profits — this, despite successive re-evaluations of the Deutschmark. In France, and to a lesser extent in Italy, the rapid increase in relative costs (except for the period 1985-87) could only be sustained thanks to monetary devaluations that allowed industries to limit the worsening of relative prices and to avoid the compression of their profit margins, the latter having recently improved, notably in France as of 1985. In the United Kingdom, relative costs that had shown a strong progression between 1970 and 1980 had a negative impact on prices and on foreign market shares. However, as of 1981, price competitiveness seems to have contributed to a lesser degree than formerly to the worsening of the trade balance and to import penetration. More generally, the weakening competitiveness of the whole of the European Economic Community is centred in high-technology

Table 47 The Evolution of Factors of Competitiveness Linked to Price in Manufacturing Industries with Respect to the Twenty-four Countries of the OECD" (average annual rate increase) Relative Export Prices (common currency)

Relative Costs (common currency)

Relative Costs (domestic currency)

Effective Rates

Exchange Profit Rateb

1976-80 1981-85 1985-87 1976-8( 1 1981-85 1985-87 1976-8'0 1981-85 1985-87 1976-80 1981-85 1985-87 1976-80 1981-85 1985-87

us Japan W. Germany France UK

Italy Canada

-1.3 + 0.5 + 1.3 -2.7 + 7.5 + 2.0 -3.1

+2.1 -0.7 -0.5 +1.0 -2.3 -0.1 +1.0

-14.0 + 9.0 + 6.5 + 1.5 -1.1 + .1.5 -2.0

-1.1 -2.8 +2.1 -0.4 +12.0 -1.1 -5.6

+ 4.0 -2.7 -0.4 -0.8 -5.1 + 2.8 + 2.0

Export Market Shares in 1982 Prices (variation in percentage points)

us Japan W. Germany France UK

Italy Canada

-18.6 + 18.0 + 8.1 + 0.5 -4.0 + 2.8 + 0.0

+1.1 -8.6 -2.3 +1.0 +9.0 +4.9 -0.1

-1.5 -5.6 -2.4 + 3.2 -1.1 + 5.5 + 3.6

-2.5 -1.2 +0.7 +0.8 +0.9 +1.8 +3.1

-2.1 +5.8 +2.8 -1.4 +2.7 -5.9 -5.5

Rate of Import Penetration (variations in percentage points)

7977-50

1981-85

1985-87

1976-80

1981-85

-2.6 + 0.0 -1.1 -0.1 -1.3 -0.1 -0.1

-4.7 + 1.7 + 1.2 -0.6 + 0.1 + 0.2 + 0.8

+ 1.0 -1.1 -0.4 -0.1 + 0.3 -0.1 + 0.4

+ 1.8 + 0.7 + 3.9 + 2.4 + 4.6 + 5.1 + 2.1

+ 2.2 + 0.1 + 6.5 + 3.9 + 6.8 + 2.4 + 1.9

" To the 24 OECD countries are added: Hong Kong, Taiwan, Korea, and Singapore. '' Gross operating surplus on value added of manufactures. Source: Calculations are based on various OECD sources.

+ 5.5 + 2.8 + 1.8 -4.2 -4.1 -5.4 -1.5

-15.5 +19.6 +7.3 +1.0 -5.0 +0.1 -3.0

-4.3 + 0.4 -4.6 -0.7 + 1.1 + 5.9 + 3.5

+2.2 +0.8 +6.9 +1.3 +9.3 -1.0 -0.2

+ 5.0 + 0.6 + 1.9 + 2.7 + 3.8 + 3.3 + 0.1

Rate of Coverage (exports/imports) (annual average values) 1985-87

1976-80

1981-85

1985-86

1.00 4.30 1.70 1.16 1.17 1.59 0.73

0.82 4.19 1.67 1.10 0.95 1.54 0.82

0.56 4.08 1.67 1.06 0.89 1.42 0.81

205 Indicators of Industrial Competitiveness Trade 48 Trade Balance with the World in Manufactured Products" (billions of current dollars)

1970

us Japan W. Germany France UK

Italy Canada

3.7 12.6 12.9 1.6 5.6 3.4 -1.4

1976

12.7 51.3 41.3 5.4 8.4 10.7 -8.7

1980

19.9 94.2 61.9 7.0 8.6 17.2 -9.7

1981

12.9 113.5 62.0 8.9 6.2 22.9 -12.4

1985

1986

-101.5 132.5 67.2 7.9 -6.3 22.2 -11.1

-136.3 152.8 88.9 1.9 -1.1 23.5 -13.1

" Classification SITC 5 + 6 + 7 + 8 + 9. Source: OECD, Industry Division.

industries and more specifically in the electronics and computer industries, where the rate of import penetration has been between two and three times higher than the average rate for all industries. Conversely, low-technology industries showed a greater dynamism despite stiff competition from developing countries. In the domestic market, the deterioration during the period 1980-85 could have been somewhat overestimated. It could reflect to a certain degree the strong appreciation in the dollar and the weak price elasticity of high-technology goods, which are in net progression in the import demand of the EEC. Notwithstanding their strong growth, the weakness of these sectors is also confirmed by losses registered on foreign markets, which suggests the presence of structural deficiencies. In Canada the good performance of relative export prices was obtained in part by a fall in domestic currency, notably vis-a-vis the American dollar. The strong growth in American demand, combined with an improvement in the price competitiveness of Canadian industry, allowed a growth in market shares, particularly in sectors of low to medium technological intensity, where Canada has certain comparative advantages, as well as in certain sectors of the aerospace industry. The preceding results speak of the absence of a negative and systematic correlation between variations in export market shares and relative prices. On the contrary, during certain years (taking into account some temporary time lags), we observe that export market shares progress in the same sense as do relative prices. The importance of structural factors, at least during subperiods, has been key, notably in high-tech industries or in industries where demand is strong.

206 State, Technology, and Competitiveness Table 49 Commercial Trade Balance with the World in Services

us Japan W. Germany France UK

Italy Canada '

1970

1976

1980

1981

1985

1986

3.0

-1.8 -2.1

18.7 -5.8 -6.7

42.3 -13.5 -12.2

21.0 -5.1 -3.5

18.6 -4.9 -6.5

0.6 2.5 0.9

3.7 6.8 1.1

34.4 -11.3 -13.4 13.0

-2.1

-6.5

8.2 5.5

-9.8

9.6 9.3 1.7

-12.8

7.6

9.1

10.7

14.2

1.5

0.5

-14.1

-15.1

Source: OECD, Industry Division.

Table 50 Trade Balance with the World inAll Goods and Services (billions of current dollars) 7970

1976

5.6 2.1 3.5 0.9 2.4 0.5 0.8

9.2 4.0

us

Japan W. Germany France UK

Italy Canada

10.5 -0.9 -0.2 -3.1 -4.6

1980

1981

8.9

14.3

-9.2 -3.2 -0.0 11.4 -11.5 -1.8

6.3 5.5

-0.5 16.0 -9.9 -6.4

1985

1986

-101.1 50.8 25.2

-125.7 87.8 49.8

2.2 7.9

7.2 1.7 5.3

-4.7 -1.2

-7.2

Source: OECD, Industry Division.

Aspects of Structural Competitiveness

Until recently, we have characterized as structural those competitive factors that do not relate to price. It would be convenient, however, to more precisely attribute to the term "structural competitiveness" those factors that, as long as they play an important role, cannot easily be changed in the short run. In the context of the present article, we will limit our discussion to two structural factors that seem of great importance: technology and industrial specialization. Technology. Among supply factors, technology holds a predominant position to the extent that it exercises either a direct or indirect effect on other determinants of supply (capital, labour) as well as on demand, trade, and direct investment. It constitutes a permanent disturbance factor in the creation of comparative and competitive advantages held respectively by countries and firms. Technological

207 Indicators of Industrial Competitiveness

development has, above all, radically transformed modes of production (new processes but also new work organization) and has accelerated the decline of certain mature sectors while favouring the emergence of new products. The diffusion of technology has allowed the rapid development and industrialization of some Third World countries, and these newly industrializing countries now occupy a growing position in world trade. The overcapacity thus created in many traditional sectors has heightened tensions between old and newly industrializing countries in the sharing of world markets. At the same time, competition between member countries of the OECD has intensified in leading sectors where acquired positions can, at any moment, be put at risk by the technological effort of firms and by the capacity of the industries of each country to absorb and to draw the maximum profit from new and generic technologies. The rapidity of technological change has altered the nature of international competition itself. No longer is such competition limited to the industrial sector. It now involves the whole of society. The quality of the educational system, the development of an extensive scientific and technological culture, and the collective effort each society is willing to put forth to maintain its position and rank in the world will, from now on, determine the success or failure of industrial competition. With the second oil shock in 1979 - when changes in productive equipment started to accelerate - industrial R&D expenditures began to increase more rapidly than production or material investments in the manufacturing sector. It is not surprising that during almost the whole of this period it is in Japan that this growth in R&D expenditures has been the greatest, to such an extent that Japan has been responsible for a quarter of the increase in R&D expenditures for all OECD countries (see Table 51). In the United States, the slowing down in the growth of R&D expenditures during the period 197075, even though largely imputable to a reduction in the space program, may, if time lags are taken into account, be connected to a loss in export markets during the period 1976-81. In effect, between 1970 and 1975, expenditures on R&D in electronic components fell by 4.2 percent per year in real terms; those of the aerospace industry by 4.6 percent; and those of the automobile industry by 2.3 percent. At the same time, between 1980 and 1985, there has been an increased effort toward improvement, as is evidenced by the evolution of the ratio of R&D expenditures to value added. In comparison, the growth of R&D expenditures in the European Community is modest. We even see a slight fall in the period between the two oil shocks.

208 State, Technology, and Competitiveness Table 51 Growth of Industrial R&D Expenditures and GDP (annual average rate) 1970-75 R&D

us Japan EEC

W. Germany France UK

Italy Canada

-0.3 4.6 3.0 2.3 4.8

-0.4 4.0

-1.1

1975-80

1980-85

GDP

R&D

GDP

2.4 4.3 2.7 2.0 4.0 2.2 2.4 5.0

4.6 7.0 5.0 7.1 4.0 4.5 2.8 8.1

3.3 4.9 2.9 3.4 3.2 1.6 3.8 3.3

R&D

GDP

9.4

3.0 3.8 0.6 1.2 1.1 1.8 1.7 2.5

11.9 4.1 7.1 5.0 2.0

11.1 8.2

Sources: OECD, DISTI, and national accounts.

With respect to physical investments, investment in R&D in the United States and most European countries is at a much higher level than that in Japan. A first explanation may lie in the importance of military R&D in total R&D, notably in the United States and the United Kingdom.5 We can assume then that military R&D is far more costly and risky than civilian R&D and that, by its very nature, it generates less physical investment than is the case for civilian mass production. It is not surprising, therefore, that in the United States the ratio of R&D to physical investment is so high and that it is inversely correspondingly low in Japan. (See Table 53.) At the same time, a failure to translate on a commercial level the comparative advantage the United States enjoys in a whole range of advanced technologies has been a source of preoccupation for several years. This question has become more and more pressing with the worsening of the us trade balance and its resistance to classical remedies. The unfavourable evolution of trade in high-technology products is particularly noteworthy. In terms of the implications for government policy, the debate has tended to crystallize around the impact of military research, especially in view of the latter's weight in the total federal budget for R&D (75 percent of public funds are directed to military programs). Though large industrial enterprises appear to benefit from such programs, the technologies developed in this context are neither easily usable nor marketable. The problem has two dimensions. First, certain experts think that military research, without turning from its primary objective, can be modified in such a way as to increase its civilian applications and

2og

Indicators of Industrial Competitiveness

Table 52 R&D Expenditures of Manufacturing Industries as a Percentage of Value Added

us Japan W. Germany France UK

Italy Canada

7977

7975

7979

7957

7955

6.5 2.8 3.4 3.5 4.5 1.5 2.1

6.5 3.4 3.7 3.6 4.3 1.5 1.7

6.4 3.7 4.5 3.7 5.4 1.3 1.9

7.6 4.5 4.9 4.3 6.3 1.6 2.7

7.7 5.3" 5.4 4.5 5.3" 2.2" ?

" 1984. Sources: OECD, DISTI, and national accounts.

Table 53 R&D Expenditures of Firms with Respect to the Gross Fixed-Capital Formation of the Manufacturing Sector

us Japan W. Germany France UK

Italy Canada

7977

7975

7979

7957

1984

77.8 7.0 21.9 20.6 38.8" 10.0 14.3

60.9 9.5 35.7 25.0 35.2 10.2 10.7

54.5 10.7 40.1 35.7 37.3fc 11.3 13.9

60.5 12.9 40.1 31.6 66.7 13.5 13.7

90.3 13.3C 47.7 37.7 66.3 21.9 25.6

" 1972: '' 1978; r Estimation. Sources: OECD, DISTI, and national accounts.

make them more accessible to a larger array of industrial enterprises. Second, while civilian applications are often considered marginal, they are at times suspected of disproportionately favouring foreign firms that are more able to appropriate the commercial applications inherent in these types of technologies. It is the national consequence of this increased international appropriation of the fruits of innovation that is questioned here. It may even be that an additional dimension of protectionist temptation is thus revealed. In a general fashion, R&D expenditures are only an approximate indicator of the technological effort of each country. On the other hand, the technical and conceptual difficulties inherent in quantitatively isolating, among a host of other factors, the influence of research effort on the main indicators of competitiveness has, for the moment, led to findings of limited application.

2io State, Technology, and Competitiveness Figure 9 R&D Expenditures of Manufacturing Industries in the us, EEC, and Japan as a Percentage of All Such Expenditures of OECD Countries

Source: OECD.

If R&D is determinant in creating a technological advantage, the presence of other factors is necessary in order for R&D to be transformed into both a competitive advantage and a gain from trade. The adaptation of industrial specialization to domestic and world demand. To be effective, industrial specialization must, in large part, satisfy the needs of domestic demand and must be aimed at those sectors and products that are most in demand in foreign markets. The quality of specialization will be improved to the extent that the adjustment of supply, in the face of changes in demand, takes place rapidly, facilitated by a flexible production system and a judicious choice in the branch of specialization. In order to capture these qualities, we use three indicators: specialization by niche, which takes into account the adaptation of the means of production to changes in domestic demand; revealed comparative advantage, which reflects the structure of exports of a country with respect to that of a reference group (in this case the OECD

211 Table 54

Specialization by Niche

(

production \ x 100 I domestic demand /

Strong Demand"

us Japan W. Germany France UK

Italy Canada

Indicators of Industrial Competitiveness

Medium Demand"

Weak Demand"

1970

1980

1985

1970

1980

1985

1970

1980

1985

102 109 126 105 110 102 94

103 121 128 105 106 99 84

94 132 140 107 97 94 91

100 104 108 99 99 102 102

102 107 113 101 101 111 98

97 122 120 102 96 110 99

97 105 98 101 98 105 107

95 102 93 98 95 122 111

92 102 94 96 91 115 104

" Strong-demand industries correspond to industries with high R&D levels, except for the electrical machinery (see Table 46), chemical, and automobile industries. Medium-demand industries include the electrical machinery, agriculture and food, paper, mechanical, stone, clay, and glass industries. Weak-demand industries include metal, wood, furniture, refineries, textile, leather, and footwear industries. Source'. OECD.

countries; and finally, contribution to the trade balance, which measures the weight occupied by each industry in the total trade deficit or surplus of manufactured products (see the Appendix for definitions and modes of interpretation of these indicators). We will briefly present the trajectory followed by each country in inserting itself in the international division of labour, and we will see to what extent such choices have allowed it to assure itself a stable industrial growth. United States: The adaptation of America's industrial structures has followed a pattern that is in conformity with the "product life cycle." As we see in tables 54 and 55, this specialization has been directed toward sectors with a high technological content in which the United States continues to have the overall highest comparative advantage within the OECD. Though these sectors have helped to restoring the trade equilibrium, their weight is yet insufficient to resolve it (see Table 56). The structural erosion of America's industrial competitiveness during the 19708 in traditional or mass consumption sectors took place in the context of increasing international competition and accelerated the redeployment of resources toward the high-technology and service sectors. The weakening of traditional industries, in terms

212

State, Technology, and Competitiveness

Table 55 Revealed Comparative Advantage" (average OECD = 100) High R&D Intensity

1970

us Japan W. Germany France UK

Italy Canada

158 123 97 86 104 78 55

Medium R&D Intensity

Low R&D Intensity

1980 1985 1970 1980 1985 1970 1980 1985

156 141 93 87 121 62 48

166 148 81 84 126 57 43

109 78 124 94 117 99 124

106 105 117 96 108 91 108

91 108 120 97 96 89 126

63 114 76 110 81 111 91

64 75 83 108 80 128 112

64 55 86 114 86 144 103

" See definition in the Appendix. Source: OECD. Table 56 Contribution to the Trade Balance" Medium R&D Intensity

High R&D Intensity

us

Japan W. Germany France UK

Italy Canada

Low R&D Intensity

1975

1980 1986 1975 1980 1986 1975

11.7

10.3

12.5

10.8

6.1

1.3 0.6

7.4

9.6

7.6

12.8 19.1

-2.2 18.6 18.8

4.4

3.2 7.4 1.2

4.2 3.8 0.9

-1.3

-0.1 -0.6

4.3

1.9

-3.8 -6.9

-2.6 -0.3

18.9

4.9

13.6

-3.7 -5.9 -9.9 -10.0

6.3 5.1

-7.2

-0.4

1980 1986

-23.6 -17.6 -11.1 -7.2 -18.6 -20.7 -19.3 -18.8 -16.0 -3.0 -4.3 -3.6 -18.1 -9.9 -9.3 0.1

11.9

4.5

16.9

6.0

10.3

" See definition in the Appendix. Source: OECD.

of foreign trade, not only has occurred at the level of price competitiveness but also involves a disequilibrium in investment. For more than twenty years — with the exception of 1987 — the productivity gains of America's competitors have been more rapid than that of the United States. As a result, certain American products have progressively disappeared from the marketplace. American producers and investors have, it seems, preferred to seek higher returns either in new sectors or in service industries that are less open to competition. This corresponds to a certain preference on the part

213 Indicators of Industrial Competitiveness

of us industry to favour activities that generate short-term profits. However, this specialization in high-tech industries has not yet been able to compensate for the weakening of American competitiveness in sectors where technology is standardized or where diffusion is rapid. American specialization has thus taken place in the context of a deterioration of price competitiveness, and preservation of competitive advantage on a structural level. The recent growth of the trade deficit, even in leading-edge sectors, reflects to a large extent the strength of domestic demand and the structure of import elasticities. Notwithstanding the fall in the dollar, capital goods imports (which are less sensitive to price variation) have increased more than other categories of goods (see Table 57). On the other hand, the fall in the dollar favoured exports in traditional sectors where the United States does not have a lasting comparative advantage unless there is a repeated devaluation of the dollar. However, the rehabilitation of industrial structures that has been witnessed since 1985 and that has been the result of a trend toward greater concentration (takeovers, mergers, joint ventures) offers the opportunity for a rebound based on a strengthened and more dynamic industrial base. However, the growing buyout of American enterprises by foreign firms, if it contributes to reduction in the foreign deficit, will lessen effective penetration and will transfer problems on the trade front to the direct investment front. Japan: Japan has shown a very strong capacity to adapt to changes in the world economy, entering expanding sectors as soon as possible and exiting progressively from sectors in decline. Japan presents us with the unique example of a country that, within fifteen years, has doubled its export market shares in leading-edge sectors (see Table 34) while at the same time maintaining the rate of import penetration of manufacturing sector at a very low level (two times less than that of the United States and seven times less than that of the European Community). To achieve this, Japan has ceased to specialize in those declining sectors in which it is in a surplus position so that domestic demand can be satisfied without turning massively to imports. Conversely, in those sectors in which it is in a deficit position, its disengagement has been slow to moderate (non-ferrous metals, industrial chemicals, wood furniture). The aerospace industry appears to be an exception to this tendency to the extent that the increased growth in the rate of penetration here is a reflection of Japan's effort and ambition to maintain a presence in this highly strategic sector. All such successes have occurred as a result of creation of comparative advantages

214

State, Technology, and Competitiveness

Table 57 Import Elasticities with Respect to Global Domestic Demand, 1974—1985 United States Originating from Japan

Originating from the EEC

Elasticities

Imports (billions of dollars)

Elasticities

Imports (billions of dollars)

1974-1985

1985

1986

1974-1985

1985

1986

2.05 0.96 0.93 1.73

30.2 21.8 11.8 63.8

37.1 21.9 12.7 71.6

1.33 1.14 1.20 0.95

58.1 11.1 2.1 71.3

71.2 10.7 2.4 84.3

Capital goods" Intermediate goods Consumption goods Total manufactured goods

Japan Originating from the us

Originating from the EEC

Elasticities

Imports (billions of$)

Elasticities

1974-1985

1985

1986

1974-85

1985

1986

1.14 1.28 0.97 1.15

8.3 6.0 2.9 17.3

8.7 6.2 3.5 18.5

0.85 1.08 0.97 0.95

2.5 3.2 2.4 8.1

3.9 4.3 3.3 11.4

Capital goods" Intermediate goods Consumption goods Total manufactured goods

Imports (billions of $)

EEC

Capital goods" Intermediate goods Consumption goods Total manufactured goods

Originating from the us

Originating from Japan

Elasticities

Elasticities

Imports (billions ojf $ )

Imports (billions of $)

1974-85

1985

1986

1974-85

1985

1986

1.97 0.72 0.95 1.62

26.2 11.7 3.9 41.8

27.9 11.9 4.6 44.5

1.32 1.17 1.04 1.24

18.3 3.1 0.9 22.3

26.5 4.1 1.1 31.8

" Including consumer electronics. Source: Calculations are based on various OECD sources.

215 Indicators of Industrial Competitiveness Table 58 Contribution of Technical Progress and the Substitution of Capital/Labour to the Apparent Labour Productivity, 1973-1983 High-technology Industrie!? Contribution (%) Rate of Technical Progressb

Substitution Capital! Labour

2.2 11.2 4.0

55.0 72.3 65.0

45.0 27.6 35.0

4.0 4.2

67.5 57.1 53.1 77.5

32.5 42.9 46.9 22.5

Labour Productivity

us Japan EEC

W. Germany France UK

Italy

3.2 4.9

" These include the following industries: aerospace, electronics, telecommunications, computers, scientific instruments, pharmaceutical products, and electrical machines. * Total factor productivity. Source: EEC, Commission Services.

based both on structural factors of competitiveness (advanced technology, quality of products, distribution, trade networks) and on price factors involving a rationalization across all stages of production and commercialization. Japan's success conforms to the attitude of Japanese businessleaders who systematically adopt medium- to long-term strategies of action that are compatible with an industrial logic, in contrast to the strategies adopted by those who are tempted by short-term gains, which are more compatible with a financial logic. This demonstrates the extraordinary efficiency of Japanese management, whose success is being felt more and more outside Japan. The performance of Japanese industry in high-technology sectors (see Table 55) can also be explained in terms of a greater contribution by technical progress to labour productivity than is the case in other countries (see Table 58). This implies a very great ability on the part of Japanese productive capacity to efficiently assimilate new technological innovations in these sectors. Moreover, technical progress (if we look at the distribution of expenditures on R&D in different industries) seems to have been better spread across Japanese industrial sectors than across those of other countries (see Table 59). As well, strong growth, the low rate of unemployment, social consensus

216 State, Technology, and Competitiveness Table 59 Variation Coefficient0 of Industrial Expenditures on R&D

us Japan W. Germany France UK

Italy

1973

1983

1.22 1.00 1.41 1.34 1.32 1.41

1.16* 1.10 1.38 1.39 1.64 1.26

" Mean to standard deviation. In the present case, twenty-three manufacturing industries were considered, which sum corresponds to the total manufacturing sector. * 1980. Sources: OECD database and DISTI database.

(notably in the area of job security), as well as the high level of work force training, have facilitated industrial restructuring and have occurred sooner in Japan than in other countries and under better economic conditions. The European Economic Community: The trade surpluses and comparative advantages of the EEC'S industries have, since 1970 and in particular after the second oil shock, been in decline in leading-edge sectors. On the other hand, sectors of weak growth have shown some improvement. As has been noted above, the difficulties of adaptation of productive capacity in the face of domestic demand seem to be concentrated in a few high-tech industries such as computers and electronics. Conversely, in foreign markets, the EEC has maintained its market shares in industries where traditionally it has had a strong, often German, performance (chemicals, machinery, agro-food). Germany. Germany is notable among other OECD member countries by its dominant position within the EEC and by the strength of its specialization (see Table 52). The majority of its trade surpluses come from other European countries, whereas its position weakens slightly in certain foreign markets (Japanese and American). The strategy that German firms seem to have adopted has been to strengthen their acquired position in those sectors that, for a long time, have been the source of their industrial competitiveness. At the same time, German firms have improved their specialization in the electronics industry, which alone in the EEC has a surplus position. Growth sectors, in terms of specialization and trade, rep-

217

Indicators of Industrial Competitiveness

resent 40 percent of manufacturing production, which approaches the structural specialization of Japan. Despite the continued appreciation of their currency, German firms, which have benefited from the reputation of their products, have been able to adapt to recent changes in the world economy through increased investment and, above all, through a strengthening of the diffusion of technical progress in more traditional industries. France: France, like other OECD-member countries, has had difficulty in competing with Germany inside the European Community and in inserting itself solidly in the non-EEC economy. The problems raised by the evolution of its trade balance have less to do with relative prices than with the weakness of its structure of specialization. The few strong points that characterize this specialization make France's industrial performance fragile and heavily dependent on international circumstances. Moreover, these strengths are dispersed throughout its industrial sectors. Consequently, the small trade surpluses that they engender do not have the same chain-reaction effect that they have in Germany or Japan. Equally, the presence of large export contracts, which are too often influenced by circumstances and the time gap between growth in domestic and foreign demand, frequently aggravates trade disequilibria. United Kingdom: The structure of British specialization raises a number of questions given the United Kingdom's weak competitiveness in recent years and the resulting deficits. The deterioration in the competitive position of traditional industries in the United Kingdom has been more pronounced than in other countries. The UK deficit in traditional industries is six times higher in absolute terms than that in other categories of industries. The loss of competitiveness in these industries, both in the domestic and foreign markets, is linked to the prior appreciation of the pound sterling following the exploitation of North Sea oil reserves. This made numerous mature industries non-competitive and forced specialization toward certain high-technology and service industries. Without the benefit of American advantages, British industry has been obliged to turn more and more toward tertiary activities in which it already has a solid base. For historical reasons, large British firms have, for a long time, acted as financial holding companies, investing massively abroad. British assets abroad (direct investments and portfolio investments) are actually two to five times higher than those of Germany or France. Even if British industrial restructuring, already under way for several years, has substantially improved productivity

218 State, Technology, and Competitiveness

and industrial growth, the persistence of trade deficits shows that wealth coming from foreign investments can only partially compensate for a weak industrial base. Italy: The specialization of Italian industry with respect to other European countries presents two peculiarities. On the commercial level, even though Italy is more dependent on traditional sectors than are its partners (clothing, textiles, shoes, leather, wood, furniture, metallic ornaments), it is nonetheless in an overall surplus position. At the same time, during the period 1970—85, other sectors have improved their specialization, such as mechanical engineering and the electrical and electronics industries. Equally, the automobile industry, which up to 1984-85 had been restructuring, is now beginning to recapture its market share after having been in a deficit situation. This specialization, unique in Europe, has allowed Italy to generate substantial trade surpluses, especially within the European Community where it is the only country with which Germany has registered a deficit. The capacity of Italian industry to occupy markets where other countries are less present is a reflection of the dynamism of its entrepreneurs, especially after 1980. To maintain its advantage in traditional sectors, Italian industry will without doubt need to continue to innovate in terms of concepts, design, and automated production. It must do this in order to meet a more specialized demand and competition from NICS, notably those of South-East Asia. Canada: The structure of Canadian industry resembles that of Italy in that trade surpluses come primarily from sectors of low technological intensity. In effect, the industries that constitute the source of Canada's competitiveness are non-ferrous metals, wood, paper, petroleum refineries, and certain industrial chemicals. There are two primary reasons for this specialization: the presence of numerous natural resources and the geographic proximity of American markets, the destination of 50 percent of Canadian exports. The combination of these two factors has caused Canadian industry to adopt a structure of specialization that is complementary to that of the United States, especially in certain sectors of high technology such as aerospace and telecommunications. The control of more than 40 percent of Canadian industry by American capital has acted to lessen the export to the United States and to third markets of products produced under American licence. On the other hand, during the 19805 Canadian industry has benefited more than other coun-

2ig Indicators of Industrial Competitiveness

tries from the strong growth in the American domestic market. The recent free-trade agreement between Canada and the United States should reinforce its complementarity vis-a-vis the United States and no doubt will offer new opportunities for Canadian industry. CONCLUSION

In this article I have presented the main indicators of industrial competitiveness and have underlined their limitations. I have especially highlighted the technical and conceptual difficulties involved in developing a set of coherent indicators that allow for a good and comparative understanding of industrial performances. However, a number of problems remain little explored. These include the empirical analysis of the complementarity and substitutability of exports and direct investment; the construction of cost indicators that take into account factors other than labour; the absence of exchange rates (purchasing-power parities) specific to each product and to each industrial sector; the non-distinction in flows of imports and exports of the ownership of capital; and the absence of technological variables in empirical studies on competitiveness. Greater knowledge in these areas seems to be a precondition to improving overall indicators of industrial competitiveness and to developing measures of analysis that are more relevant for decision making. APPENDIX

i. Specialization by niche: Sy

The question concerns surplus production in relation to internal demand. Thus, for country i and product j, specialization by niche S y is equal to

where Yy; Xy, and My- represent, respectively, production, exports, and imports. It is evident that if Sjj = 1 we are in a balanced position because either Xy= My * 0 Or Xy = My = O.

If Sy > 1 we are in a surplus position (Xy > My), where the more Sy is increased, the more the greater part of production will be exported.

220 State, Technology, and Competitiveness If Sy < 1 we are in a deficit position (My < Xy). If specialization is very weak or non-existent, then production will be less than or equal to imports. 2. Revealed comparative advantage: CA~ For country i and product J, the revealed comparative advantage is equal to CAg = 100

where Xy- are the exports of product j to country i and Xim are the exported manufactures of country i. 3. Contribution to the trade balance: C,y This indicator serves to identify for a country those sectors for which it is in a relative surplus position and those for which it is in a relative deficit position. Thus, for country i, the contribution to the trade balance Cy of sector j will equal

where X^ and My are, respectively, the exports and imports of country i and sector j to/from a particular geographic area; and X and M are the exports and imports of manufactures from all sectors j. Applied to the different industrial sectors that make up a country's productive fabric, this indicator provides a profile. We can easily show that when indicator Cy has a value greater than 2, it denotes strength. When it is less than — 2 , it suggests weakness, and when it is between — 2 and +2, it denotes a position of "relative indifference."

NOTES

Author's Note: The ideas expressed in this paper are the author's and do not necessarily reflect those of the Organization for Economic Co-operation and Development. 1 Technical assistance differs from know-how to the extent that it reflects the level of development and the industrial structures of the host country. From this point of view, it can be excluded from technological receipts. 2 The complementarity and substitutability of flows of capital and commodities and their influence on the trade balance have been the object

221

Indicators of Industrial Competitiveness

of numerous debates. Essentially, it is a question of whether exports of capital have a negative impact on the commercial exports of a country. Most empirical studies have attempted to show that a correlation exists between flows of exports and direct investment. They show that, for the most part, these two flows are complementary and that a growth in investments does not inhibit that of exports. However, the existence of a positive correlation between these two flows does not constitute proof of causality. The fact that the two flows move in the same direction does not necessarily prove the existence of a true complementarity, especially given that the majority of calculations relate to flows destined to all countries of the world. 3 On the level of exports, this weakening essentially involves hightechnology industries (notably the computer industries) whose market share, after having stabilized during the 19705, declined following the second oil shock. Conversely, sectors of low R&D intensity fared much better, notwithstanding stiff foreign competition from developing countries. All the same, three-quarters of the European Community's trade surplus comes from sectors of low and especially medium technological intensity, regardless of the reference period. 4 (Imports + national production under foreign control less national exports controlled by foreign firms)/domestic demand. 5 According to some estimates, military research accounts for about 28 percent of all R&D in the United States, 27 percent in the United Kingdom, 22 percent in France, 4 percent in Germany, and 0.35 in Japan.

REFERENCES

Bernheim, Douglas, and John Shoven. 1986. "Taxation and the Cost of Capital." Paper presented to the American Council for Capital Formation, September. CNRS. 1983. Commissariat general du plan, Institut national de statistiques et d'etudes economiques. "Technological Indicators and Measures of International Trade Performance." Actes du collogue sur I'Econometrie de la recherche. Paris, September 9—10. Durand, Marline, and Claude Giorno. 1987. "Indicators of International Competitiveness: Conceptual,Aspects and Evaluation." OECD Economic Studies, no. 9 (Paris). Learner, Edward E., and Robert M. Stern. 1970. Quantitative International Economics. Boston: Allyn and Bacon.

This page intentionally left blank

PART THREE

Technological Development and Government

Strategy

This page intentionally left blank

CHAPTER NINE

New Modes of Competition in the Textile and Clothing Industry: Some Consequences for Third World Exporters LYNN KRIEGER MYTELKA

The textile and clothing industry is currently undergoing a series of dramatic changes, perhaps the most fundamental of which is the increasingly important role that knowledge is coming to play in the production process. This is reflected less in the classical sense of research and development and more in the design, management, and marketing innovations that are exercising a profound impact on the industry's structure and its mode of competition. Combined with growing protectionism, in the form of increasingly more restrictive Multifibre Arrangements (MFAS), these technological and organizational changes have significantly altered the environment within which the newly industrializing countries [NICS] and those aspiring to this status are forced to compete. After analysing this new mode of international competition, this paper will turn to an examination of the way in which some Third World textile and clothing exporters have begun to adjust to it.'

THE CRISIS AND THE SEGMENTATION OF TEXTILE AND CLOTHING MARKETS

From the earliest beginnings of modern mill production through the interwar years, the development of textile and later clothing production was predicated upon the growth of a mass market (Piore and Sabel 1984). As a result, technological change in these industries focused primarily on the lengthening of production runs for standardized products and to a lesser extent on labour-reduction strategies associated with increased machine speeds, reduced machine down-

226 Technological Development

time, and the elimination of steps in the production process. The latter, along with international subcontracting, began to gain particular importance in the 19605 when trade liberalization within the context of the General Agreement on Tariffs and Trade (GATT) and the creation of the European Economic Community (EEC) coincided with rising exports from low-cost Asian producers, notably Japan, Hong Kong, China, Korea, Taiwan, India, and Pakistan. Although they were far more productive than their antecedents, the major innovations of the 19508 and 19608 - open-end spinning, shuttleless looms, and circular knitting machines - were initially quite limited in the number of applications and range of products for which they were suited. In the case of open-end spinning, for example, this meant a restriction to coarser counts, and in knitting, to synthetic fibres. With the relatively limited reduction in costs obtainable from modernization at that time and under pressure from rising imports, many countries opted for protectionism. Resort to protection through the Short Term Cotton Arrangement in 1961 and the Long Term Arrangement a year later, however, slowed the diffusion of new spinning and weaving techniques in many of the industrialized countries during the 19605 (CGP 1986, 99—110). This further opened the possibility of competition from lower-cost producers, but not all of these were NIGS. Within Europe, Germany and Italy emerged as formidable competitors following the adoption of a dual strategy of textile modernization coupled with, in Germany, the internationalization of production by German clothing firms and, in Italy, domestic subcontracting (Frobel, Heinrichs, and Kreye 1980; Dubois and Barisi 1982; CETIH 1984). So successful was this two-pronged strategy that by the early 19705 Germany had catapulted into first place among the world's top textile exporters, while Italy retained its position as the world's third major textile and second leading clothing exporter (GATT 1984). The vulnerability of those textile and clothing firms that failed to pursue a continuous process of adjustment in the postwar period would increase as the very nature of the competitive process itself changed in the 19705 under the effects of the global economic crisis. We can look at these changes in terms of the intersection between two dynamics, the first having to do with the nature of demand and the second with the intensive mode of capital accumulation common to the advanced industrial economies and characterized by an accelerating rate of technological change (Boyer 1987; Lipietz 1982, 1986). As incomes rose in the advanced industrial countries, a secular decline in the rate of growth of consumer spending on clothing

227 Textile and Clothing Industry

became evident. In the 19705, the economic crisis accelerated this decline. In Japan, for example, the annual average percentage rate of change in consumer spending on clothing fell from 6.9 percent in 1963—73 to 0.3 percent in 1973-82, declining further into a negative rate of growth in the early 19805. In the EEC, the growth of consumer spending on clothing fell from 3.9 percent in 1963—73 to 0.9 percent in 1973-83, declining to a negative rate of growth of0.2 percent during the early 19805. As in the Japanese case, consumer spending on clothing lagged behind total consumer spending.2 Only in the United States, the United Kingdom, and Sweden, where the relative price of clothing remained well below that of other consumer goods, did the rate of increase in consumer expenditure on clothing in the period of 1973-82 remain above the annual average percentage rate of change in consumer expenditure as a whole (UNCTAD 1984, 120). In addition to a growing sensitivity to price changes, price/quality relationships resulting from the differential impact of the crisis on incomes (intermediated by government policies) became increasingly important. This emerged most clearly from a series of standardized surveys of household consumer spending patterns in Germany, France, the United Kingdom, and Italy (Eurostat 1985, 25-7) and pointed in particular to the differences in price elasticities of demand across income categories. The segmentation of the market into income categories having different price and income elasticities of demand suggests that in the advanced industrial countries, in spite of the secular decline in the demand for clothing, the market for cheaper textile and clothing products — those, for example, sold in large chain and discount stores - will continue to grow, as will, however, upscale segments of the market where demand appears far less affected by rising prices. On the whole, however, current slower growth in Third World consumption resulting from high levels of indebtedness, the imposition of austerity measures, and the effects of economic crisis on overall consumption in the advanced industrial countries is likely to prolong the period of declining rates of growth in world textile and clothing production that began in the 19705. The increasingly zero-sum nature of international competition in this industry resulting from slower growth in the consumption of textile and clothing world-wide, coupled with the market-segmenting effect of the crisis, has put into question earlier growth strategies based exclusively on mass-produced, standardized products at the same time as it has focused attention on the industry's need to maintain market shares. This has led to a change in the mode of com-

228 Technological Development

petition itself, from one based primarily on price to one based simultaneously on price and creativity. Because cost-reduction and product-differentiation strategies require innovation in products, processes, marketing techniques, and organizational forms, this double competitive dynamic has further stimulated the growing knowledge intensity of production in the textile and clothing industry. A C C E L E R A T I N G THE RATE OF TECHNOLOGICAL INNOVATION AND DIFFUSION

The late 19705 and early igSos thus witnessed a tremendous acceleration in the demand for new innovations in the textile and clothing industries. States also increased their subsidies for the modernization of "traditional" industries in this period (Mytelka 1982; Mytelka and Mahon 1983). The pull of textile demand, moreover, was accompanied by a push from the electronics, computer, and some metalworking industries than facing a slowdown in the growth of demand from established markets. Efforts were undertaken to apply these advanced techniques in new ways, and in the textile and clothing industry, such efforts are reflected in the development of automated pattern cutting (the American Gerber method finding its inspiration in metal-cutting techniques developed for the aerospace industry), the French Lectra Systems technique drawing upon laser technology, in the incorporation of electronic devices into spinning, weaving, and knitting machines, and in the computerization of the design and management processes. Already highly automated, electronic controls in the spinning, weaving, and dyeing branches of the industry improved diagnostic and monitoring capabilities and reduced downtime when models, patterns, or colours changed or threads broke. This in turn provided major cost savings through reduced labour time and energy consumption, materials savings, and improved product quality. In Japan, operating manpower requirements in spinning, for example, were reduced from 76.6 workers per 10,000 spindles in 1975 to 43.5 workers in 1982 (Fukuda 1982, 11). In Germany, the productivity of capital in the spinning industry, measured in terms of kilograms of yarn per spindle, rose from 62.8 kg in 1960 to 122 kg in 1983, while the productivity of labour increased from 6,100 kg per employee to 16,000 kg over the same period (Hartmann 1985, 11—12). In weaving, the gains were even more remarkable - an increase from 14,300 m 2 of cloth per loom in 1960 to 43,700 m 2 in 1982 (Hartmann 1985, 12), and this does not take into consideration newer, more

229 Textile and Clothing Industry

productive weaving techniques such as the air-jet loom recently developed and diffused in Japan. These looms, whose speed, measured in terms of the weft insertion rate, are 4.5 times faster than traditional looms and half again as fast as earlier shuttleless looms of the projectile and lance type. Because their efficiency depends upon careful maintenance and fine adjustments, Japan's traditional concern for quality control has produced additional productivity gains here. Thus, in comparison with the United States, where air-jet looms are stopped on average 1.5 to 2 times an hour, in Japan the average has fallen to one stop per hour (Kuramoto 1985, 13). In knitting, techniques that both were labour saving and offered flexibility in model changes and design only became available in the 19808. Their appearance revolutionized the knitwear industry, as the time needed to change over to a new model fell from three hours to thirteen minutes. The newest of these machines, by coupling computerized pattern design with the electronic selection of needles and knitting to form, has significantly reduced materials wastage and dramatically increased throughput from design to production, thus acceierating the speed of response to changes in demand for seasonal merchandise. For a number of knitwear products (notably socks, certain types of pullovers, and T-shirts), the amount of assembly (sewing) time - the only labour-intensive segment of the production process left - has been reduced to under five minutes. Because of the increased flexibility of these new knitting machines, the gains from economies of scale are limited. This opens new possibilities for small, modernized firms in the advanced industrial countries; such firms are particularly well placed to play the role of domestic subcontractors in those markets where segmentation favours the production of more diversified and design-intensive products. The effects of these changes on the price competitiveness of European industry has been notable. In 1983, in a study that disaggregates the clothing imports of the European Economic Community [EEC] to the six-digit level, Nicolas Marian, former principal economist of GATT and a member of the International Textiles and Clothing Bureau (ITCB) in Geneva, identified approximately fifty articles, including stockings, socks, children's clothes, T-shirts, and pullovers, for which the unit value in ECU per kilogram of imports coming from other EEC countries was below that of imports from the Third World (ITCB 1985, 11-15). The same, however, is not true for clothing assembled from woven fabrics. With computer-aided design, collections are becoming larger and model changes more frequent. The integration of design, pattern grading, marking, and cutting, the principal preparatory stages

230 Technological Development

in the clothing industry, has reduced costs and increased flexibility still further. In addition to materials savings of some 10 percent, such integrated systems have resulted in a 40 percent reduction in the number of persons previously performing these operations, a reduction in the time needed to gradate and make patterns by a factor of from two to six (that is, from roughly three days to one hour), and a 50 percent cut in the time it takes to move a product from the decision stage to production (Hoffman and Rush 1985, 4.4-4; Clauzel 1985, 42—7). By turning cash registers into terminals that collect data on sales, such as Benetton has done (Belussi 1987; Bruce 1987; Heskett and Signorelli 1985), production is also linked more closely to distribution, thus significantly reducing the time elapsed from product conception to the point of sale. These process, product, organizational, and marketing innovations, by increasing the speed of response to changes in demand, give a competitive edge to firms that are in close contact with their market. In most instances, this favours firms located in the advanced industrial countries. Competition has also stimulated the development of collaborative R&D ventures between clothing, electronics, and robotics firms, accelerating still further the speed of technological change in these industries. In the making-up stage of the clothing industry, which accounts for nearly 90 percent of the labour costs, major innovations are now closer to commercialization than one would have imagined in the early 19805. In the United States, TCS, for example, is currently testing programmable robotic sewing machines with prehensile arms that are capable of such complex sewing operations as sleeve and pocket making (Confection 2000, January 1987, 77). In Japan, the development of three-dimensional sewing techniques by a consortium of robots firms and sewing-machine manufacturers (Brother, Juki, and Mitsubishi), sponsored by the Agency for Industrial Science and Technology, has also advanced (Escande 1985, 43; Tuloup 1986). Not to be outdone, PFAFF, the German sewingmachine manufacturer, has joined forces with the United Kingdom's largest clothing firm, Courtaulds, and largest electronics firm, General Electric Company, to develop automated sewing frames under the European Community's BRITE (Basic Research in Industrial Technologies for Europe) program (Commission of the European Communities 1986). There is no doubt that the clothing industry is poised for the same type of dramatic transformation that the textile industry underwent over the 19808. More than a decade of slow growth and high unemployment in the advanced industrial countries, thus, has had two major consequences for the textile and clothing industry. On the demand side,

231 Textile and Clothing Industry

it has radically altered consumption patterns in the clothing industry, leading to the development of a market segmented by both income and product. These changes in demand, as well as in the nature of international competition, have induced firms in the advanced industrial countries to pursue a dual strategy of modernization and delocalization, emphasizing through each an increase in the knowledge intensity of production from product conception to distribution. On the supply side, innovations that increase the design intensity of products (making possible more numerous and more frequent model changes), that enhance flexibility (permitting the production of more diversified articles in shorter and more efficient production runs), and that decrease the elapsed time between conception and distribution (thereby introducing new competitive advantages for those firms in close contact with their markets) are diffusing rapidly. THIRD WORLD RESPONSES TO PROTECTIONISM, DELOCALIZATION, AND TECHNOLOGICAL CHANGE

With the rapidly changing competitive environment in the 19708 and 19805, Third World textile and clothing exporters were obliged to restructure as well. Yet, the flexibility with which these firms were able to respond to the constraints and opportunities posed by protectionism, delocalization, and technological change in the advanced industrial countries, was in large part a function of the degree to which the firms in question had become autonomous decisionmaking centres. This, in turn, depended upon the relationship between local firms and foreign parents or subcontractors and the extent to which the local firm had consciously sought to develop indigenous technological capabilities, notably in product design, production engineering, marketing, and management. Thus, a local firm's autonomy was weaker to the extent that crucial decisions concerning product range and design, process choice, and marketing strategies, including the decision to export, were taken by a parent firm, licensor, or subcontractor and to the extent that the firm itself had not mastered technology sufficiently to go beyond the stage of machine operation and maintenance (Mytelka 1985). Using this notion of autonomy as a guide, we can classify Third World textile/ clothing exporters along a rough continuum from the most autonomous (e.g., firms in Korea and Hong Kong), to those in an intermediate position (of which Moroccan firms are an example), to the

232

Technological Development

least autonomous (of which firms in the Ivory Coast are typical). Given space limitations here and the availability of a more detailed analysis of the African experience elsewhere (Mytelka 1981, 1985, 1986), the following discussion will focus primarily on firms at the more autonomous end of the spectrum. The first Multifibre Arrangement, negotiated in 1973, was designed to limit the growth of imports of textile products from Third World countries to about 6 percent over the period 1974—77. Compared with MFA ii and in, MFA i was far less restrictive, primarily because of the flexibility with which quotas could be managed unused quotas could be shifted to other product categories or carried over into the following year — and because of the fewer bilateral agreements signed within its framework - seventeen by the United States with its Third World suppliers and, somewhat belatedly, fourteen by the European Community with its suppliers. This flexibility, coupled with the limited modernization of textile and clothing firms in a number of the advanced industrial countries, meant that the NICS were able to increase their exports by shifting unused quotas to high-demand product categories. Over the four years covered by MFA i, imports from the Third World thus continued to increase; they rose especially rapidly in the EC because restrictions were slower to be applied there than in the United States. With an MFA that extended the voluntary export restraints imposed by the United States in 1971-72 to 1977 and that enlarged the terrain of their application to the EC, independent textile manufacturers in Taiwan, Korea, and Hong Kong, however, were faced with a difficult choice: they either had to watch the growth possibilities of their exports diminish or they had to alter the range of products then being manufactured (Toyne et al. 1984, 121—3; Jang 1984; Yoffie 1983). The flexibility of MFA i encouraged them to select the second option, but this required major new investments in machinery and equipment capable of producing more sophisticated and higher-quality products. Tables 60 and 61 provide data on the extensiveness of these investments. During the 19705, as these data show, deliveries of new textile machinery kept pace with the modernization efforts then under way in the advanced industrial countries. With respect to open-end spinning, for example, cumulative deliveries for the years 1974-80 totalled 61,276 rotors in Taiwan, 58,656 in Hong Kong, and 21,088 in Korea as compared with 11,716 in Belgium, 20,442 in the United Kingdom, 58,650 in France, and 60,978 in Germany (ITMF 1981, 23). In the case of shuttleless looms, cumulative deliveries over the same period totalled 6,908 in Taiwan, 5,986 in Korea, and 2,527 in

233 Textile and Clothing Industry Table 60 Rate of Modernization in Third World Spinning Industries Cumulative Deliveries

Cumulative Deliveries 1984"

1981-84

688

-

960 172

564

2,236 18,000 1,920 3,840

1974-80

1981

1982

1983°

38,582

1,248 18,000

1,400

960

AFRICA

Egypt Sudan Tunisia Morocco

11,268

672

9,804 4,404

4,640

840

2,432

-

ASIA

Philippines Indonesia Thailand Malaysia Singapore Hong Kong Korea Taiwan China India

300

600 6,376 58,656 21,088 61,276 20,152

840

540

7,968 1,008 9,408 2,168 1,880

32,232 41,936

3,760 5,276

168

732

9,752 728 14,360

1,636 504 4,924 1,032

2,792

512

1,380 5,260 4,284 2,100 1,800

1,108 7,416 3,704 0 0 20,736 7,500 32,976 5,700 6,984

704

940 532

564 564

5,968 8,388

1,108 416

400

—

2,360 1,964

LATIN AMERICA

Mexico Brazil

2,016

" In 1983 and 1984 the Czechoslovakian firm Investa did not provide data on its deliveries. The figures for these two years thus underestimate cumulative deliveries over the period 1981-84. Source: International Textile Manufacturers Federation (ITMF).

Hong Kong as compared with 2,770 in Belgium, 2,836 in the United Kingdom, 8,781 in France, and 9,812 in Germany (ITMF: 1981, 32). In addition to the growing capital intensity of production in the textile industry resulting from these investments, the new competitive strategy significantly shifted employment away from the textile and into the clothing industry. In Hong Kong, for example, the number of employees in the textile industry fell from 106,463 in 1973 to 67,226 in 1984, whereas the number of employees in the clothing industry rose from 174,622 to 301,545 in the same period (GATT 1985, 88). In the Korean garment industry, employment rose by 4.5 percent between 1980 and 1983, while it fell 5.9 percent in the textile industry during the same period (GATT 1985, no). For the Third World as a whole, a decline in the share of textiles in net manufactured output, from 11.3 percent in 1963 to 10.8 percent in 1978, was accompanied by a rise in the share of

234

Technological Development

Table 61 Rate of Modernization in Third World Weaving Industries Cumulative Deliveries

Cumulative Deliveries 1981

1982

1983"

1984a

285

481

1,363

557

117 375 -

126 171 -

102 123 -

196 193 -

245 5 59 219 316

2,654 5 483 707 316

800 1,267 240 67 122 2,527 5,986 6,908 603

1 1,563 161 61 95 551 1,018 1,705 348 129

22 488 244

54 181 148 1

18 73 468 1,391 343 245

434 1,264 2,317 265 224

20 262 45 84 4 149 2,427 5,687 772 573

97 2,494 598 146 117 1,207 5,177 11,100 1,728 1,172

4,382 2,086

1,460 353

869 240

47 233

106 147

2,482 973

1974-80

1981-84

AFRICA

Egypt Sudan Tunisia Morocco Mauritius ASIA

Philippines Indonesia Thailand Malaysia Singapore Hong Kong Korea Taiwan China India LATIN AMERICA

Mexico Brazil

" In 1983 and 1984 the Czechoslovakian firm Investa did not provide data on its deliveries. The figures for these two years thus underestimate cumulative deliveries in this period. Source: ITMF.

clothing in net manufactured output, from 3.6 to 4.1 percent over the same period (UNIDO 1982, 19). In Asia, where the shift had first begun, the result was a dramatic rise in clothing exports. Thus, from 1970 to 1975, when Asia's share in world textile exports rose from 11.5 to 13.8 percent its share in world clothing exports soared from 19.1 to 29.1 percent (CEPII, Data Bank). Successive renewals of the MFA in 1978 and 1982 progressively enlarged the scope of the agreement to include all textile and clothing products that were or were likely to become (through the surge mechanism) sensitive (Choi et al. 1985; Dolan 1983). The advantages obtained through diversification were progressively whittled away, while efforts to fill all quotas to capacity brought rewards. The drive to modernize the textile sector was thus given an added stimulus, leading in the 19803 to considerable overcapacity in synthetic yarns and cloth among Asian Nics. 3

235 Textile and Clothing Industry

Accelerated rates of modernization in the 19805, combined with a far higher annual rate of machine utilization in the Asian NICS than in the advanced industrial countries (8,496 hours per year in Hong Kong and 8,472 in Korea compared with 6,868 in the United States, 5,520 in Germany, and 5,312 in France [Werner Associates]) and considerably lower salaries and longer working weeks (work weeks of 54.5 hours in Korea and 45.5 in Hong Kong compared with 41.1 in Japan and 40.1 in the United States [International Herald Tribune, 9 April 1986, 11]), have enabled the Asian NICS to successfully compete in price terms with the advanced industrial countries over a wide range of standardized products that benefit from economies of scale through mass production. But the protection afforded by the MFA quota system to the NICS as opposed to would-be NICS with still lower salaries was beginning to erode as the NICS reached the limits of their now barely growing quotas. Competing successfully with the advanced industrial countries in highervalue-added diversified production, however, implies increasing the knowledge intensity of textile and clothing production. In this the NICS have been encouraged by an MFA system in which quotas are fixed by volume rather than by value. In the case of Korea, far from abandoning the drive toward modernization in its textile and clothing industries and despite some earlier errors, the government in its current five-year plan anticipates major improvements in product quality, design, and variety, thereby imitating with a several year lag the move toward competition on the basis of both price and creativity initiated in the advanced industrial countries and followed early in the 19805 by Hong Kong and Taiwan. The adoption of a dual strategy of competition on the basis of price and creativity places those firms in the NICS that are able to make this transition in direct competition with firms in Europe, Japan, and the United States. At the same time, the shift from pure price to a form of competition that stresses both price and creativity enables firms in the NICS to maintain their dominance in these faster-growing segments of the market. For them to do so, however, presupposes that the diversified markets of the advanced industrial countries remain open to their products. The rising levels of protectionism that characterized the mid-19805 would not appear to auger well for such a strategy. Since they have pushed to the top of their quota ceilings in the EC, further expansion there seems unlikely. Insofar as the United States is concerned, the number of calls (restrictions) on products coming from Hong Kong rose from thirty-two in 1982 to sixty in 1984, and the percentage of imports subject to restrictions increased from 60 per-

236 Technological Development

cent in 1982 to 70 percent in 1984 (ITCB 1985, 24, 25). Because of such restrictions, American subcontractors of standardized, lowpriced, classic articles whose production can be planned well ahead have increasingly sought out suppliers in other low-wage countries - Indonesia, Malaysia, Sri Lanka, and Bangladesh — where quotas are not yet filled (us Senate 1984). For firms from the NIGS that have already attained an international size, the option to delocalize production as a complement to a modernization strategy at home has thus gained in attractiveness. Of particular interest as production sites are countries that are in close proximity to major advanced industrial country markets and have either few restrictions on their exports to those markets, as, for example, the Lome Accord countries, or have relatively unused quotas. The list of delocalizations undertaken by the NICS to such countries is already extensive. Three Taiwanese firms have established joint ventures in Panama to produce pullovers, jeans, and slacks destined for the American market (USITC 1985, 424); others have invested in Macao and Malaysia (UNCTC 1984, Annex 5.32). One Korean firm has delocalized a part of its shirt production to the Dominican Republic (USITC 1985, 424), and another has announced its intention to build a spinning plant in Tunisia (L'industrie textile, February 1985). Many Hong Kong firms have invested in Macao, Mauritius, Thailand, and Malaysia. Many more have subcontracted spinning and weaving operations to China (USITC 1985, 187; Goodstadt 1985, 64). Four Hong Kong firms have even jumped the EC'S protective barriers and established factories in Ireland, France, and Italy (Business Week, 26 August 1985; ITGLWF 1984, 110-3). Fang Bros Knitting Ltd has a plant in Ireland, and Peninsula Knitters Ltd, another major Hong Kong knitwear firm, located a plant in northern England (Lardner 1988). Although Henry Y.Y. Tang, managing director of Peninsula, acknowledged that it costs 30 percent more to make sweaters in the United Kingdom, the investment is regarded as an important safety valve, since, despite us restrictions on imports of Asian textiles, the United States continues to allow European manufacturers to ship unlimited numbers of garments (Business Week 1985).4 Whether subcontracting is done by firms from the NICS or from the advanced industrial countries, the local plant frequently tends to produce a very narrow range of finished products, lacks forward and backward integration, receiving most of its inputs from the parent firm or subcontractor, has no autonomous design and limited marketing capacity, and because of a lack of qualified textile engineers, often follows poor maintenance schedules. Such firms are

237 Textile and Clothing Industry

highly vulnerable to changes in the strategies of the foreign firms on which they depend and to variations in the level of their export quotas. The most extreme cases here are firms located in the exportfree zones of countries like Mauritius where no effort is made to articulate those activities with production in the wider economy or in countries like the Ivory Coast where export-oriented manufacturing was grafted onto an import-substituting textile industry dominated by foreign owners and technology suppliers. Moroccan firms are in a more intermediary position. In 1973 the institution of a system of temporary import licences enabled local firms to import duty-free all raw materials and intermediates used in the manufacture of an exported finished product. Since this system was put into place, subcontracting by European, especially French, clothing firms has risen dramatically. By the mid-19805, the clothing sector accounted for 28 percent of Morocco's employment in manufacturing and one-third of its manufactured exports (Lanone 1985, 46). Most of the cloth used in the preparation of finished garments for export,to Europe, however, is furnished by subcontractors, since spinning and weaving firms cannot produce complex products. In spinning, for example, a recent mission from the French Fashion Institute found no innovation in the type of yarns being produced and a slower rate of machine replacement compared, for example, with what is found in the Asian NICS (see Table 58). In weaving, the use of classical looms has limited the extent to which Moroccan firms can create new, multicoloured woven goods (IFM 1988, 16) and the rate of modernization has been even slower (see Table 59). With regard to the clothing sector itself, the use of second-hand machinery is common and the organization of production remains rigidly Taylorist. Under these conditions, only relatively long production runs have proven profitable in all but the newest factories, where an effort has been made to invest in newer equipment, including automatic pattern-grading machinery (IFM 1988, 17). Products that require considerable labour inputs and are destined for the middle and upper ranges of the price/creativity scale are also able to compete with Asian imports in the European market. Because of its weak technological base and dependence on imported inputs, the Moroccan industry has been highly vulnerable to shifts in demand, in the strategy of European subcontracting firms, and in levels of protection on goods entering the European Community, which to date remains its most important market. Thus, when French quotas on imported slacks were slashed from 5.3 million pairs in 1982 to less than 4 million in 1983, many small Moroccan

238 Technological Development

subcontractors went bankrupt (Lanone 1985, 46). Larger firms, having learned this lesson, are now seeking to follow the path charted by the NICS, especially insofar as developing their own design and marketing capabilities is concerned. The number of firms capable of doing so, however, is quite small. C O N S E Q U E N C E S OF THE NEW MODE OF INTERNATIONAL COMPETITION FOR THE THIRD WORLD IN THE COMING DECADE

Several additional consequences for the pattern of international competition flow from the changes in demand for textile and clothing products and from the accelerated diffusion of new techniques that have marked the late 19705 and early 19808. These are likely to further shape the ability of Third World firms to compete. Foremost among these consequences is the fact that slow growth coupled with market segmentation in the advanced industrial countries has permitted both mass-market and more specialized, highervalue-added textile and clothing production to survive. This has had an effect on both production and distribution, as well as on the relationship between producers and distributors in the advanced industrial countries, and in turn is giving rise to a number of important consequences for the Third World. With regard to production, in the case of mass-market textiles, to compensate for slower growth, marketing strategies based on product differentiation and continuous fashion changes, once associated with more specialized production, have become common. The sheer size of this market, however, makes longer production runs possible, and price elasticities of demand in this market favour efficiency over flexibility in many product categories. Where it has been possible to significantly reduce labour time through technological innovation, firms in the advanced industrial countries have regained competitiveness. This, we saw above, is true for a number of important massmarket textile, knitwear, and clothing products, such as pullovers, socks, and men's shirts. To the extent that Third World countries introduce similar technology and are able to use it efficiently,5 they, too, will be able to compete. In the period 1981-84, Taiwan and Korea made many such investments. The alternative is to move into higher—value-added products as Hong Kong has done. As new techniques diffuse more rapidly, Third World countries that are highly indebted, as, for example, Brazil, may not be able to keep up. International production-cost comparisons in spinning and

239 Textile and Clothing Industry Table 62 A Comparison of Manufacturing Costs in Spinning and Weaving, 1979, 1983, and 1987 (in us dollars) Spinning

1979 Cost per kg of yarn Index 1983 Cost per kg of yarn Index 1987 Cost per kg of yarn Index

Brazil

Germany

India

Japan

Korea

us

0.909

1.229

0.841

n.d.

0.949

0.071

(73.9)

(100)

(68.4)

n.d.

(77.2)

(79.0)

1.134

1.014

0.765

0.962

0.700

0.867

(111.8)

(100)

(75.4)

(94.9)

(69.0)

(85.5)

1.353

1.839

1.227

1.677

0.972

1.169

(74)

(100)

(68)

(91)

(53)

(64)

Brazil

Germany

India

Japan

Korea

us

0.297

0.421

0.289

n.d.

0.337

0.305

(70.5)

(100)

(68.7)

n.d.

(80.1)

(72.6)

0.366

0.374

0.275

0.296

0.236

0.335

(97.9)

(100)

(73.5)

(79.1)

(63.1)

(89.6)

0.340

0.562

0.324

0.474

0.271

0.344

(60)

(100)

(58)

(84)

(48)

(61)

Weaving

1979 Cost per yd. of fabric Index 1983 Costs per yd. of fabric Index 1987 Costs per yd. of fabric Index

Source: ITMF, 1979, 1983, 1987.

weaving prepared by the International Textile Manufacturers Federation (ITMF) in Zurich, for example, reveal a considerable decline in Brazil's advantage over the years 1979 to 1987 and in comparison with other Third World countries (Table 62). State policies that assure credit to medium- and large-sized independent firms in Brazil and encourage an interface between Brazil's growing computer industry and the textile and clothing sector could, however, alleviate some of the pressures generated by austerity. For a large number of mass-market products, countries with more labour-intensive techniques do remain competitive at particular wage

240 Technological Development

rates, assuming a reasonable level of efficiency in the utilization of these techniques. Even this, however, may not help potential newcomers. This is primarily due to the fact that the X-efficiency of existing plants (i.e., the productivity of existing factors of production) is low in many of these countries. Machine efficiencies in Pakistan, the Philippines, Colombia, and the Ivory Coast, for example, are often less than 75 percent of comparable international standards, and unit costs are further reduced through underutilization of capacity, poor quality control, and inadequate attention to maintenance (de Vries and Brakel 1983, 13-43; Mytelka 1985; UNIDO 1986, 5164). Firms in such countries will have even more difficulty running the newest machines as productively as in the advanced industrial countries. This is particularly true of techniques such as the new airjet looms now being diffused in Japan, whose efficiency depends upon a continuous and accurate maintenance of the pressure and rate of rotation of the jet and of operating conditions for each part of the equipment, including temperature and humidity control. Yet these are among the looms included for purchase under the Ivory Coast's structural adjustment program! With regard to distribution, large retail establishments - such as Marks and Spencers, Sears, Wards, J.C. Penny, C&A, and Carrefour — have played an important role in organizing the trade in massmarket textile and clothing products through subcontracting to suppliers at home and abroad for merchandise that they market under their own labels. In the past, this has provided many aspiring NICS with access to markets in Europe and North America. As competition rose, especially in countries where retailing of textile and clothing products is relatively concentrated (e.g., in the United States, where three firms account for nearly 20 percent of textile and clothing sales, or in England, where a single firm accounts for nearly one-third of the sale of textile and clothing products), smaller manufacturers were gradually placed at a clear bargaining disadvantage and even larger companies have found it necessary to merge with or take over competitors in order to increase their market power. In France, for example, DMC took over Herwiller and Anny Blatt in order to add long fibres (synthetics) to the short fibres and the "fantasy" yarns it already dominated. Prouvost (La Laniere de Roubaix) increased its points of sale by adding to its Rodier franchise that of Vitos. In Italy, Marzotto bought Lanerossi, a subsidiary of the state-owned ENI petroleum corporation, thereby replacing Benetton as Italy's premier textile and clothing firm. In England, Vantona Viyella, having merged with Nottingham Manufacturing, merged again, this time with Coats Paton. The new firm replaced

241 Textile and Clothing Industry

Courtaulds as England's largest textile and clothing complex and has moved into third place after Burlington Mills (us) and Milliken (us) and ahead of J.P. Stevens (us), Kanebo (Japan), Toyobo (Japan), West Point Pepperell (us), Courtaulds (UK), Prouvost (France), Collins & Aikmann (us), DMC (France), and Marzotto-Lanerossi (Italy) among the world's top textile and clothing firms (Textile Wirtschaft, no. 52 [25 December 1986], Le Figaro, 24 July 1987). Among the medium-sized enterprises, intra-European mergers have accelerated as 1992, the year in which a full common market is to take affect, approaches. Thus, Pepe (UK) took over Buffalo (France), Triumph (West Germany) took over Horn (France), Schiesser (Switzerland) took over Eminence (France), and Riorda (Italy) took over Rica Levy (France) (Journal du textile, no. 1063 [9 March 1987]). Larger firms have also expanded the use of designer labels and developed franchising as a means to capture a greater share of the gains from innovation in design and to enlarge their market shares in both specialized and mass-produced textile and clothing products. This process appears to have accelerated in the 19805 as the Multifibre Arrangements stabilized the growth of imports from the newly industrializing countries and led to intensified competition among advanced industrial countries within, for example, the EC market. Both of these factors place smaller Third World firms at a disadvantage, since many of these firms have relied almost exclusively on subcontracting from European and American manufacturers to acquire product designs and to secure market access. In a new development, the French knitwear industry has recently sponsored a special salon, "Interselection," expressly designed to permit small local firms to exhibit their superior design capabilities and establish direct sales links with the large chain store. Firms in the Mediterranean basin that develop their own design capacity and/or can obtain designs from large retailers that, seeking to bypass local manufacturers, are working more closely with independent fashion designers, may also benefit from the opportunity to make contacts that this new biannual fair represents. Lastly, insofar as the higher-value-added products are concerned - men's and women's suits, coats, dresses, and skirts - fashion considerations have led to shorter production runs, which provide a further impetus to the development of flexible production techniques, and to the rapid diffusion of computer-assisted design, pattern-marking, grading, and cutting systems. The spread of computer-assisted inventory and management systems, moreover, has enabled larger firms to analyse their cost structures, varying the combination of fibres, fabrics, and colours used and optimizing the

242 Technological Development composition of their collections according to the price and creativity elasticities of different market segments. The decision as to whether to delocalize all or a part of the production for any given article can thus be integrated into an overall optimizing strategy, a strategy that can change with changes in the availability of new, more efficient techniques, and in the characteristics of final markets of sale and production sites. While the MFA encourages the continued growth of international subcontracting, the integration of design, production, and commercialization tends to favour those sites nearest to major consuming markets. The use of designer labels, moreover, creates additional barriers to entry to independent producers from the less-developed countries. Current efforts by Hong Kong and more recently Korea to develop training programs for local clothing designers will permit firms in these countries to diminish their reliance on imported designs and subcontracting relationships in the near future. Low wages, however, will remain the key to their competitiveness in these products, especially as firms in the advanced industrial countries move toward computer-integrated manufacturing in the clothing industry. Progress in developing design capabilities will likely be much slower in intermediate countries like Morocco, while design autonomy will not be achievable over the coming decade in those countries where little effort has been made to date to build indigenous technological capabilities. In each of the latter cases, export-oriented manufacturing will largely be possible only within a narrow range of product/ price market segments and competition among Third World countries within those market segments will be fierce. NOTES 1 Research for this paper was partially funded by the Commissariat general du plan (CGP), Paris. 2 These data were compiled from a variety of sources, including OECD, National Accounts, vol. 2, various years; Eurostat, National Accounts, various years; UN, Yearbook of National Accounts Statistics, various years; and GATT 1984. 3 The crisis in the Korean polyester weaving sector is a good example. In 1978-79 polyester fabric weavers in Taegu, the capital of Korea's textile industry, began investing heavily in water-jet looms. These looms cannot be used to weave cotton cloth because the cotton absorbs water. "By 1982, Korea had 3,000 of them and in the next two years, despite government urging to slow down their investment, added 2,000 more.

243 Textile and Clothing Industry What these firms overlooked ... was that Japan and Europe had abandoned the water jet loom ... [for] the more expensive, but even more efficient, air jet loom. More seriously, Taegu's weavers had failed to realize that the world textile market was shifting away from polyester and other man-made fibers.... Taegu's polyester weavers are saddled with an overcapacity that has created a financial crisis in the sector. There are enough looms to produce 100% of Korea's domestic demand for polyester fabric in just two months and exports are declining" (Jang 1984, 24). 4 In the case of the Yangtzekiang Garment Manufacturing Co. Ltd of Hong Kong, production facilities have been established in the United States and garment wholesale and retailing companies have been created in the United Kingdom, Italy, France, and the United States (ITGLWF 1984, 112). 5 It is important to recognize that costs may be reduced not only by increasing output relative to costly factors of production such as capital and labour but also by paying greater attention to quality and by using machinery and equipment efficiently. The latter strategies imply a pattern of work organization that ensures careful machine maintenance and calibration, and this in turn presupposes both worker-training and management skills frequently lacking in the Third World. REFERENCES

Belussi, Fiorenza. 1987. Benetton: Information Technology in Production and Distribution: A Case Study of the Innovative Potential of Traditional Sectors. SPRU Occasional Paper Series, no. 25. Brighton: University of Sussex. Boyer, Robert. 1987. "Les economies au milieu du gue, changements techniques et interventions publiques depuis une decennie." In R. Boyer et al., Aspects de la crise. Vol. 3. Paris: CEPREMAP. Bruce, Leight. 1987. "The Bright New Worlds of Benetton." International Management, November, 24—35. Business Week. 1985. "Hong Kong's End Run Around us Protectionism," 26 August. Centre d'etudes techniques des industries de 1'habillement (CETIH). 1984. Rapport de la mission de veille technologique franqaise en Italic. Paris: CETIH. Centre d'etudes techniques des industries de Phabillement CETIH. 1984. Rapport de la mission de veille technologique frangaise en Italie. Paris: CETIH. Choi, Y-P, H.S. Chung, and N. Marian. 1985. The Multi-fibre Arrangement in Theory and Practice. London: Pinter. Clauzel, F. 1985. "La robotisation du pret-a-porter. Journal du textile, 8 October, 44-8. Commissariat general du Plan (CGP). 1986. L'enjeu du textile franqais: Le

244

Technological Development

marche mondial. CGP. Report prepared by Lynn Krieger Mytelka for the Commission "Prospective des echanges internationaux". Paris: CGP. Commission of the European Communities. 1986. BRITE Project Synopses: Projects Supported under the First Call for Proposals — 1985. Brussels: Basic Research in Industrial Technologies for Europe (BRITE), November. de Vries, B., and W. Brakel. 1983. Restructuring of Manufacturing Industry: The Experience of the Textile Industry in Pakistan, Philippines, Portugal and Turkey. World Bank Staff Working Papers no. 558. Dolan, Michael. 1983. "European Restructuring and Import Policies for a Textile Industry in Crisis." International Organization vol. 37, no. 4 (Autumn): 583-616. Dubois, Pierre, and Giusto Barisi. 1982. Le defi technologique dans I'industrie de I'habillement: Les strategies des entrepreneurs Franfais et Italiens. ATP Internationale — 1981. Paris: Centre national de recherche scientifique (CNRS).

Escande, P. 1985. "Les 'High-Tech' envahissent la confection." Usine nouvelle, no. 28 (i i July): 41—3. European Economic Communities (EEC). 1985. Situation de I'industrie Textile. Working document of the Services de la commission, Brussels. Eurostat. National Accounts. Various years. — Comite des textiles. 1985. "Resumes des renseignements recus des pays participants." Doc. No. Com. Tex/42, 27 November. Frobel, V., J. Heinrichs, and O. Kreye. 1980. The New International Division of Labour: Structural Unemployment in Industrialised Countries and Industrialisation in Developing Countries. Cambridge: Cambridge University Press. Fukuda, K. 1982. "Technological Development Trends in Textile Machinery." Digest of Japanese Industry and Technology. No. 172: 11—14. General Agreement Tariffs and Trade (GATT). 1984. Textiles and Clothing in the World Economy. Background study prepared by the GATT Secretariat to assist work undertaken by the contracting parties in pursuance of the decision on textiles and clothing taken at the November 1982 Ministerial Meeting. Geneva: July. Goodstadt, L. 1985. "Textile House of the Rising Son." South, March, 64. Hartmann, U. 1985. "Structural Adjustment: The West German Experience." Paper presented at the CEPS/EPL Conference, 9-10 December, Centre Borschette, Brussels. Heskett, James, and Sergio Signorelli. 1985. "How Benetton has streamlined and branched out worldwide in casual clothing market." Condensed version of Harvard Business School case 9-685-014. International Management, May, 79-82. Hoffman, K., and H. Rush. 1985. Microelectronics and Clothing: The Impact of Technical Change on a Global Industry. Brighton: SPRU, University of Sussex.

245 Textile and Clothing Industry International Textile, Garment and Leather Workers' Federation (ITGLWF). 1984. Multinational Companies in the Textile Garment and Leather Industries. Brussels: ITGLWF. Institut frangais de la mode (IFM). 1988. Rapport sur I'industrie textile. Paris. International Textiles and Clothing Bureau (ITCB). 1985. "Textiles and Clothing: Recent Developments in Trade, Technology, and Trade Policy." Agenda Item i, Seoul Meeting, 3—7 September. International Textile Manufacturers Federation (ITMF) International Textile Manufacture Delivery Shipment Statistics. Zurich: ITMF, various years. — International Production Cost Comparison Spinning/Weaving. Zurich: ITMF, *979' !983' !985' 1987Jang, Jung-Soo. 1984. "Korea's Textile Industry: A Battle on Two Fronts." Business Korea, October, 24—8. Kuramoto, Y. 1985. "Technological Development in Weaving Machinery of Japan." Digest of Japanese Industry and Technology, no. 207: 7—14. Lanone, P. 1985. "Textile et cuir: Le Maroc veut s'affranchir de 1'Europe." Usine Nouvelle 9 May, 46-7. Lardner, James. 1988. "Annals of Business: The Sweater Trade." The New Yorker Part i, January 11; Part 2, January 18. Lipietz, Alain. 1982. "Toward Global Fordism?" and "Marx of Rostow." New Left Review, no 132 (March-April): 35-58. — 1986. Mirages et miracles problemes de I'industrialisation dans les tiers monde. Paris: La Decouverte. Mody, Ashoka, and David Wheeler. 1987. "Towards a Vanishing Middle: Competition in the World Garment Industry." World Development 15. no. 10/11: 1269—84. Mytelka, Lynn K. 1981. "Direct Foreign Investment and Technologial Choice in the Ivorian Textile and Wood Industrie," Viertel jahresberichte der Entwichlungs — landerforschung (Friedrick-Ebert-Stitftung Institute. No. 83 (March): 61-79. - 1982. "In Search of a Partner: The State and the Textile Industry in France." In France in a Troubled World Economy, edited by S. Cohen and P. Gourevitch, 132—50. London: Butterworth. - 1985. "Stimulating Effective Technology Transfer: The Case of Textiles in Africa." In International Technology Transfer, edited by N. Rosenberg and C. Frischtak, 77-127. New York: Praeger. - 1986. "The Transfer of Technology: Myth or Reality?" In The European Community's Development Policy: The Strategies Ahead, edited by C. Cosgrove and J. Jamar, 243—81. — 1987. "Changements technologiques et nouvelles formes de la concurrence dans I'industrie textile et de l'habillement." Economie prospective internationale, no. 31 (3rd trimester): 5—28. Mytelka, Lynn K., and Rianne Mahon. 1983. "Industry, the State and the

246 Technological Development New Protectionism: Textiles in Canada and France." International Organization 37, no. 4 (Autumn): 551—82. Piore, Michael, and Charles Sabel. 1984. The Second Industrial Divide. New York: Basic Books. Toyne, B., et al. 1984. The Global Textile Industry. London: Allen and Unwin. Tuloup, A. 1986. "Quelles strategies pour I'habillement." L'industrie textile, no. 1162 (January): 63-4. United Nations Conference for Trade and Development (UNCTAD). 1984. Programme of Cooperation among Developing Countries, Exporters of Textiles and Clothing. "The Multi-fibre Arrangement in Theory and Practice." Back-up study on international trade in textiles and clothing. Karachi Workshop. United Nations Centre for Transnational Corporations (UNCTC). 1984. Transnational Conditions in the Synthetic Fiber, Textile &f Clothing Industries. N.Y.: United nations. United Nations Industrial Development Organization. (UNIDO). 1982. Handbook of Industrial Statistics. New York: United Nations. — 1986. Restructuring of the Mexican Textile Industry: Requirements and Policy Options. UNIDO Doc. No. UNioo/IS-595, 13 January. United Nations. Yearbook of National Accounts Statistics, various years. us International Trade Commission (USITC) 1985. Emerging Textile Exporting Countries. ITC Publication No. 1716, July. us Senate. Committee on Finance. 1985. State of the U.S. Textile Industry, Hearing before the Subcommittee on International Trade. Washington, DC: us Government Printing Office. Yoffie, David. 1983. Power and Protectionism: Strategies of the Newly Industrializing Countries. New York: Columbia University Press.

CHAPTER TEN

Engineering, Design Services, and Technology Transfers: The Case of the Republic of South Korea JACQUES PERRIN

The construction of a new industrial production unit, such as a steel mill, a petrochemical refinery, or a cement plant, involves the purchase of heavy industrial equipment from a widely diverse array of mechanical, electrical, electronic firms as well as calling upon the services of several companies for work in infrastructure — building construction, installation of metallic structures, piping, and electrical systems. The construction of a new plant equally necessitates a series of studies, technical (feasibility, process, project, execution) as well as organizational (planning, site co-ordination, equipment purchases). These organizational and technical activities, known as engineering services, can be wholly or partially carried out by newly formed design units (i.e. the owner), by the engineering departments of the company in charge of the project, or by specialized consulting engineering firms. Machinery manufacturing (motors, machine tools, paper presses) requires the integration and assembly of a large number of electrical, mechanical, or electronic components. Some of these components are manufactured by the firm, while others are subcontracted to specialized producers. Machinery building also requires the realization of a series of studies (financial, economic, marketing, assembly) as well as the supply of services in organizational planning and component purchasing. This type of activity - which we shall call here industrial machinery design — is generally supplied through design offices, marketing and procurement services, and engineering departments of industrial equipment manufacturing firms. In some cases, some of these services and studies can be subcontracted to specialized firms. Engineering services (or production-unit design) and industrial machinery design are considered a service-sector activity when they are supplied Translated by Kevin Fitzgibbons, CREDIT.

248 Technological Development by specialized firms and accounted for under the heading "business services." The objective of this paper is to draw attention to the economic importance of design services (engineering services and industrial machinery design) and particularly to their function in structuring the exchange of technical and economic information between the buyer (or project manager) and the component supplier both in plant construction and industrial equipment manufacturing. In the first part of this paper, we shall deal with organizational factors - or the state of social relations in the resolution of technical problems often attributed to the complexity of design activities. In the second section, we shall demonstrate that on an international level this sector is a strategic and lucrative market for trade, and industrialized countries, particularly the United States, are negotiating to improve their trade balance through the export of services. Finally, we shall show how the economic performance of the Republic South Korea can be partially explained by its desire to master design services and by its pursuit of a well-defined strategy of inward technology transfer of design capabilities.

THE ECONOMIC ROLE OF ENGINEERING AND MACHINERY DESIGN ACTIVITIES

The principal objective of design activities is to conceptualize machines and production units that would have the highest economic performance for the targeted production level and that would be in accord with previously established production factors (primary resources, intermediate products, labour, and capital). Such activities have up to now been the object of insufficient economic research. This is most likely due to the fact that machinery design is primarily done by machinery-manufacturing firms and that the movement to autonomous engineering activities as a separate industry is relatively recent and so far incomplete. It should be noted that the structure of this new industry tends to vary from one sector to the next and according to the country's industrial tradition (Perrin 1976). In order to increase plant and machinery performance, design units must maintain a close relationship with research and development to be in a position to improve on existing processes, introduce new materials, and so forth. Quite often, technology transfer from research organizations to industry must transit through design activities. Yet the role of design organizations in the transfer of research that results in the integration of new innovations should not hide the fact that technical change in machinery is primarily due

249 The Republic of South Korea

to an ongoing process of minor modifications. Many of these adjustments can be made by design departments themselves in response to needs for component standardization or in order to reduce delays or costs. Such changes can have an impact on component production and are generally initiated after consultation with suppliers concerning alternative solutions. Modifications can also be the result of requests from divisions in charge of building or assembling machinery or plants who seek to rationalize production, introduce new assembly and installation techniques, and increase standardization. The specific needs of machinery users or of project managers also contribute significantly to the steady flow of adjustments that design and engineering organizations must deal with. The user/producer dynamic plays an important role in the evolution of the machinery industry in several countries. Rosenberg (1976) in his study of American industrial history has demonstrated the role in the evolution of tooling machines of manufacturing industries, such as those that produce bicycles, sewing machines, and armaments. Modifications are also driven by component suppliers following changes in their production methods and materials. The introduction of such modifications from different sources, particularly within large design offices with several specialized departments, requires efficient organizational administration. Information technology (computers, CAD/CAM [computer-aided design/ computer-aided manufacturing]) provides new solutions for the management of change within these structures. It must be pointed out, however, that efficiency also depends on the state of internal social relations. This brief synopsis of the functions of design organizations demonstrates that linkages between machinery-design offices and engineering activities on the one hand and between R&D, suppliers, and users on the other are interactive. The schema proposed in Perrin 1976 that explains the function of engineering services and the structure of information exchange can be applied to design organizations (see Figure 10). The accumulation and integration of information that emanates from different sources in the industrial system (R&D, suppliers, and users), originally seen as undertaken by the engineering services, can equally be seen as undertaken by design structures as a whole. In this sense, a developing or industrialized country that buys a plant or a machine in another country participates in the accumulation of experience in the vendor country. Any operation of technology

250 Technological Development Figure 10 The Functions of Design Organizations

transfer induces a reverse transfer of experience benefiting the owner of the technology - a process we have called "reverse technology transfer" (Judet and Perri 1977). The function of information exchange between different poles in the industrial system and the accumulation/integration of this information undertaken by design services allow for the clearer identification of the economic role played by this type of service for enterprises. This functional analysis of engineering led us to see a link between the difficulties encountered by developing countries in technology transfer and learning, and turnkey contracts and the lack of local engineering capability (Judet and Perrin 1977). Identical conclusions have been drawn in regard to industrialized countries that integrate high-tech industries. This is one reason why in France in the late 19608 the Social and Economic Council, concerned by massive us imports of assembly-line automation equipment, called for the creation of consulting engineering firms specializing in automated production systems. The existence of such companies was considered a key instrument in promoting the French automation industry. The performance of a country's machinery-design activities has direct effects on the development of its industrial equipment sector. The industry's competitiveness depends very much on the technological and organizational capacity of its design services and on the state of its network relations (conflict, co-operation, or neutrality) with R&D firms, suppliers, and users. Several studies in Great Britain have pointed to the poor performance of design services as a factor in the decline of the mechanics industry (Rothwell, Gardiner, and

251 The Republic of South Korea

Pick 1983). For many industrial products like machinery, market competition is played out more in terms of quality (durability, production performance, easy maintenance) than in terms of price. A relatively recent study (Schott and Pick 1983) on quality and price factors in British trade performance indicates that 45 percent of exports and 30 percent of imports can be explained by quality factors. INTERNATIONAL TRADE AND DESIGN

SERVICES

Massive imports of design services have constrained in the development of their industrial system, the countries which have imported technology and many have made a concerted effort to control this type of transfer. One such country, Spain, was driven to decreeing control measures on technology transfer. These controls affected preliminary and feasibility studies, project-procedure plans, and services dealing with plant construction, operation, maintenance, and repair. The industry ministry's order accompanying the decree stated: "Organizations and enterprises contracting foreign engineering or consulting companies for studies or technical services must prove that they have made an attempt to procure these services from at least two national firms listed in the Consultants and Industrial Enterprises Register." In Brazil, the National Institute of Industrial Ownership (INPI) has also established controls on the import of design services (administration plans, feasibility and pre-project studies, management studies, engineering services). Importers must demonstrate that such services are unavailable locally. The company must provide a declaration from three Brazilian enterprises from the sector or from a recognized professional association so that the INPI can hear from the sector itself as to whether or not such services can be provided. The South Korean government, having recognized the limitations of massive importing of engineering services, issued a directive on 15 May 1969 that rejected turnkey projects in favour of agreements with foreign firms for the creation and development of local engineering capability. In 1973 the Korean government passed a law in support of local engineering-services development. Over and above fiscal and financial incentives, the law designated an increased role for national consulting engineering firms in the construction of new plants in the country. The state also directly participated in the creation of many of these firms.

252

Technological Development

It is important to point out that the massive import of design services is a handicap, not only for the construction of industrial machinery and plants, but also in regard to consumer and intermediate goods where the lack of design capability can have a negative impact on the company's control over product evolution (introduction of innovations, choice of modifications to be retained or introduced, etc.) and on user/vendor relations. While developing countries were primarily concerned with the transfer of production facilities during the ig6os, many are now more interested in setting up a design infrastructure capable of mastering process and product technology. This type of transaction is much more complex than that of production capacity and cannot be achieved through turnkey contracts. Successful transfer depends on strong interaction between the personnel supplying the technology and the receiver (Chanaron and Perrin 1987). International trade in services is difficult to estimate, since there is no direct statistical source of information beyond the engineering sector. Even in the case of engineering services, it is difficult to evaluate flows because this type of service is often integrated into turnkey contracts or into projects undertaken by large contractors. The only existing data of world trade in engineering services is an annual survey of 250 major engineering and construction firms and 200 design firms published in the American Engineering News Record. The international engineering and construction market (plants, buildings, infrastructure) was evaluated at $81.6 billion in 1986, which is considerably less than the 1981 level of $134 billion. Developing countries represent 80 percent of this market, while 75 percent of exports come from industrialized countries (40 percent from American firms). We can therefore estimate that contracting exports in 1986 involved between $8 and $10 billion in design services (engineering, machinery, building design). The world consulting engineering market is considerably smaller ($3.5 billion). Again 85 percent of the market is in the Third World and 95 percent of exports are controlled by industrialized countries involving American (30 percent) and European (50 percent) firms. It should be remembered that the world export market for the electrical and mechanical industry was $540 billion in 1986 and that industrial machinery alone was worth $320 billion. In industrialized countries, the development of high-tech industries has spurred an increase in the demand for design services for enterprises. It is difficult to estimate the evolution of this industry, since a large part of it is integrated into industrial production companies. In the United States, the rate of growth of business services

253 The Republic of South Korea

averaged 7.2 percent a year between 1980 and 1984. Some activities showed phenomenal annual growth between 1972 and 1982: computer services (20.9 percent), management consulting (17.6 percent), credit-leasing consulting (17.3 percent), and engineering (16.7 percent) (Ecalle 1986). In France, despite a marked decline in industrial production since 1974, sales volume in services to enterprises increased by 6 percent a year between 1977 and 1980, while industrial production rose by less than 3 percent. Services with the highest growth in sales volume to industry are computers (10 percent), advertising, and research. In French business services, 10,000 jobs a year were created between 1974 and 1982. It should be emphasized that international transactions have undergone major structural transformations since the early 19705. The international market share of service transactions has grown from 16 percent of world trade in 1970 to 25 percent in 1980. The export of professional services (banking, insurance, engineering, engineering consulting, legal, accounting, publicity, maintenance, tourism) was a $i2i-billion market in 1981 in which the United States (21 percent), France (9.5 percent), and Great Britain (8.5 percent) were major players. But it is above all the foreign investments of multinational firms that have experienced the most significant transformations. Between 1970 and 1985, multinational foreign investment in services rose from 25 to 45 percent of total multinational transactions. The United States exported five times more services through American multinationals than through direct exports, while other industrialized countries (UK, Canada, and Japan) exported twice the amount of services in this way. "Thus in the past, FDI (Foreign Direct Investment) was primarily in extractive industries, yesterday it was primarily in manufacturing and today it is primarily in service industries" (Sauvant 1986). These figures on international service transactions give us a better understanding of the major stakes in recent multilateral GATT (General Agreement on Tariffs and Trade) negotiations initiated by the United States and other industrialized countries on lifting trade barriers for services. In the early ig8os, a powerful lobby, the Coalition of Services Industries, was formed by the major American servicesector enterprises with the goal of promoting the total liberalization of international services trade. This group, led by American Express and American International Group (the largest insurance company in America), is also comprised of banks (Chase Manhattan), temporary employment services (Manpower), telecommunications companies (ATT), computers (IBM), and the two largest consulting engineering firms in the United States (Bechtel and Fluor). Accord-

254 Technological Development

ing to a recent survey taken of major American service companies, the countries most often mentioned for unfair import practices in services were Brazil, Mexico, Canada, Australia, Indonesia, Japan, and Venezuela. Developing countries in particular were singled out by the American firms. THE REPUBLIC OF SOUTH KOREA: A S T R A T E G Y FOR THE

THE

TRANSFER OF DESIGN TECHNOLOGY

With a gross national product (GNP) of over $100 billion in 1987, South Korea is today among the twenty most powerful economies in the world. Its per capita GNP is close to $2,500 a year, making it one of the wealthiest nations in the developing world. These figures point to a vigorous economy that has experienced an average real growth of around 10 percent over the period 1966-86 (12.2 percent in 1986). In contrast to other developing countries, South Korea did have an industrial base to build on when it began its intensive industrialization process in 1962. During the three decades of Japanese colonization (1910—45), industrial production on the Korean Peninsula (including North and South) expanded at an annual rate of 10 percent, and by 1940, industry employed some 300,000 Koreans. The industrial base was structured around heavy industries in the North (mines, fertilizers, mechanical engineering) and light manufacturing in the South (export agriculture, textiles) in order to respond to the needs of the Japanese colonial empire. When the country was partitioned into two separate and antagonistic states in the wake of the Korean War (1950—53), this previously complementary structure proved to be a serious handicap for both fledgling economies. During the 19505, industrial development in the South centred around the previously established light industries to meet the exclusive needs of the internal market (Chaponniere 1982). Beginning in 1962, state-initiated economic reforms and the opening of the country to external markets injected new dynamism into the Korean economy. The industrialization strategy was implemented through a series of four five-year plans (1962—81) based on setting up import substitution and export industries. The import substitution industries initially focused on consumer goods (first and second plans) and then on industrial machinery (third and fourth plans). This same gradual process can be found in the export sector, where labour-intensive products (textiles, garments, and electronic

255 The Republic of South Korea

components) were followed in later plans by more capital-intensive industries such as the shipbuilding, automobile, and machinery industries. Taking advantage of both the dynamic internal market and growing export trade, Korea was able to move back up the industrialization process from consumer manufacturing to intermediate goods. A good example of this type of process can be seen in the textiles industry, where Korea progressed from garment production to the production of synthetic fibres, petrochemicals, and even textile machinery. From Aid to Investment

In implementing its industrialization strategy, South Korea had the benefit of substantial financial aid. It was in fact one of the most heavily subsidized countries in the world. The United States was its principal backer, having donated $12.6 billion between 1946 and 1976, of which $5.7 billion was strictly economic aid and $6.9 billion designated for military assistance. Multilateral organizations (World Bank, Asian Development Bank, United Nations) provided $1.9 billion, while Japan paid $1 billion in war damages (Chaponniere 1982). American military assistance was a key factor in financing major infrastructure projects (roads, airports, etc.). It also stimulated the creation of Korean construction companies (in particular, Hyundai) under contract to the American armed forces that later became involved in projects in Vietnam during the war and in Saudi Arabia. These American military aid programs played a significant role in training highly qualified Korean construction companies, while economic aid provided financing for industrial investment. Multinational investment in South Korea began only after 1966 and was primarily Japanese (56.7 percent) and American (22.1 percent). The Korean law on foreign ownership underlined the state's bias toward joint (50-50) and minority ownership of branch plants; it also defined target sectors for foreign investment - textiles from 1966 and heavy industry after 1973. International Subcontracting Agreements

Although direct multinational investment was limited by state control, foreign firms nevertheless played a major role in the development of the Korean export industry through international subcontracting agreements (APO 1978). The earlier contracts were primarily in the garment and electronics sectors.

256 Technological Development

• In 1974 the garment industry represented 35 percent of exports in manufactured goods (427.1 million). This was realized by 300 Korean enterprises under contract to foreign firms. Korean companies signed an average of ten such subcontracting agreements. • Between 1970 and 1975, electronic component exports increased from $55 to $580 million, 52 percent of which was produced by 40 multinational branch plants, 22 percent under 108 different joint ventures, and 26 percent by 241 Korean enterprises. In the mid-1970s, this practice was extended to the transportequipment industry (shipbuilding, automobiles). Planned Technology Transfer One of the essential elements of the Korean industrial strategy was the linking of industrial planning to the import of increasingly sophisticated technology. In the electronics industry, progressive technological learning first involved the transfer of assembly technologies and then the design and production-management know-how. A study of the development of the Korean electronics industry identifies three major phases in this process, each phase being accompanied by specific government measure (Kim 1980). The first phase (adoption of production technology) involved the creation of television, radio, and communications-equipment assembly units. Technology transfer included assembly technology, product specification, machinery and component purchases, and technical assistance. Products were manufactured for the domestic market with the support of rigorous import controls imposed by the Korean government. The primary objective of the second phase (assimilation) was local component production. In order to broaden and protect the domestic market, the state extended its import substitution policy to include components and participated in the creation of specific export industries (calculators and electronic watches). Government aid entailed preferential credit rates for exporting companies, tax-free production zones, and support in establishing export services. This stage was characterized by a diversification of export products and by an upgrading of technology transfers in assembly as well as in product design. The third phase (mastery) was characterized by the creation and development of Korean R&D, engineering, production-management, and industrial organization services. This resulted in the manufac-

257 The Republic of South Korea

ture of specific Korean-designed products: mini-calculators, miniprocessors, telecommunications equipment, and the like. This phase could never have been realized without a concerted government effort in technical training over the years. The acquisition of engineering expertise in petrochemical industry also evolved over three periods. During the 19605, the construction of chemical plants (fertilizers, petrochemicals) was essentially done through turnkey projects undertaken by foreign enterprises. This stage had little impact on the development of the fledgling Korean engineering sector. By the end of this period, the government reevaluated the efficiency of this policy (lack of local machinery supply, high cost of foreign exchange, etc.) and, in a presidential policy statement on 15 May 1964, forbade the use of turnkey projects and obliged major contractors to sign co-operation agreements with foreign companies in order to create and develop local engineering capability. The 1973 Local Engineering Services Law marks the beginning of a phase in which state-supported Korean engineering firms have established themselves and are actively participating in the process of accumulating know-how. Certain firms formed under the auspices of the state, such as the Korea Engineering Company, have facilitated the acquisition of local expertise in detailed engineering, projectpurchasing management, and construction supervision. The second half of the 19708 can be characterized by the consolidation of the Korean engineering industry and its expansion into export activities. During this period, two strategies can be identified. The first involved new joint ventures with foreign firms for both domestic and external projects. In 1978 Kean Nan Enterprise Limited bought out Pritchard, an American firm. In November of that year, the Korean-owned Sinotech Engineering and Bechtel formed a subsidiary engineering and construction firm. The second entailed the acquisition and mastering of "process engineering" (detailed engineering and the adoption of production technology developed in research laboratories). Two different examples can be used to illustrate these goals: • In 1977 Korea Engineering was subcontracted by the Japanese firm Toyo Engineering for the construction of an urea facility for the Korean Nambahae-Chemical Company (a 75/25 percent joint venture between Korea General Chemical and the American firm Agrico Chemical). By 1979, for the building another Nambahae urea plant, Korea Engineering took responsibility for major basic and detailed engineering and subcontracted to Toyo (Perrin 1979).

258 Technological Development

• The first Korean ethylene plant (100,000 mt/year) was built for the Korean Oil Corporation in 1973 with totally imported technology. By 1982, for the construction of the third ethylene unit (300,000 mt/year), Korea Engineering had become the primary contractor for detailed engineering. Power Plant Design and Engineering Up until the mid-19705, the construction of electrical power plants for the Korea Electric Power Corporation (KEPCO) was undertaken by foreign multinationals on a turnkey basis. By the second half of the decade, promotion of national engineering capability had become a high priority for KEPCO and no further turnkey contracts were awarded. Foreign engineering companies came under the obligation to take on local counterparts. Again in this sector, the primary objective was national engineering autonomy. In the nuclear industry, KEPCO, together with the Korea Advanced Energy Research Institute (KAERI), created the nuclear engineering firm KOPEC (Korean Power Engineering Corporation). The first two nuclear plants were turnkey projects built by Westinghouse and engineered by Gilbert Commonwealth, an American firm. A third turnkey contract was awarded to Atomic Energy of Canada Limited and Canatom for the construction of a 600 MW CANDU reactor. The following nuclear units (5,6,7, and 8 — there was no unit 4) were done on a "de-packaged" contract basis, with Bechtel taking responsibility for construction supervision and material inspection and with KOPEC as a subcontractor. Units 9 and 10 were PWR (Pressure Water Reactor) reactors designed by the French firm Framatome, with conventional engineering being shared by Alsthom-Atlantique and KOPEC. Over the years, KOPEC'S participation in nuclear-engineering services progressed from 5 percent for the first unit to 46 percent for units 9 and 10 (see Table 63). The company, which now employs 200 engineers, is expected to take on 70 percent of engineering services for the next two planned units (11 and 12). KEPCO maintained this same policy for the construction of its thermal plants in refusing to offer turnkey contracts in 1976 for Pyunqtaeg i and 2. Engineering services was undertaken by the Hyundai Engineering Company with the assistance of the American firm Brown and Root. From that point on, all thermal projects were handled by Korean consulting engineering enterprises, with periodic subcontracts awarded to foreign firms for technical problem solving or for personnel training. This promotion of national engineering services gave the Korean government the opportunity to set in place a policy directed toward

259 The Republic of South Korea Table 63 KOPEC Participation in Nuclear Engineering

Units

Principal Engineering

Type of Contract

1,2 3 5, 6 7, 8 9, 10

Turnkey Turnkey De-packaged De-packaged Separated

Gilbert

11, 12

De-packaged

KOPEK

AECL

Bechtel Bechtel Framatone Alsthom

KOPEC

Participation (%)

5 15 26 31 46 70 +

Source: UNCTAD.

the integration of local electrical machinery for the energy sector. Such integration now stands at close to 55 percent. KEPCO played a key role in this policy by establishing a specialized mechanical construction subisidary, Korea Heavy Industries and Construction Company (KHIC), and through the Indigenization Division of its Quality Assurance Control Department. Industrial Machinery Design During the ig6os and early 19708, the Korean government implemented an import-support policy for industrial machinery (reduction of customs taxes, purchasing credits). As early as 1969, the state had already drafted the first elements of an industrial machinery development strategy with the Machinery Industry Promotion Act. But it was not until 1973 that the policy was put into action through major investments in heavy industry, such as in the chemical products industry. Import controls were imposed on machinery, and customs taxes were hiked up for all equipment that could be manufactured by Korean industry. In the case of machinery imports of over $1 million, a special organization, KOSAMI (Korean Society for the Advancement of Machinery Industry), was set up to study the possibilities of local production of an equivalent product. If it was considered to be within the capabilities of Korean industry, the import licence was refused. Government promotion of development in industrial machinery involved granting credits to and reducing taxes for firms operating in the sector. As a result, investment in the industry rose 54 percent between 1973 and 1981, while production increased by a factor of 15 (1970—79). The Korean heavy machinery industry is controlled

260 Technological Development

by a small number of large firms such as the Korea Heavy Industries and Construction Company (KHIC), which employs 12,000, and Hyundai Heavy Industries, which employs 10,000. KHIC was originally a private company bought out by the government after running into deep financial difficulty. The proportion of small and medium enterprises (in terms of employment) is considerably lower in Korea than in other countries (Amsden and Kim 1986). The technology necessary for the development of the Korean machinery industry was primarily acquired through licensing agreements. Between 1977 and 1980, 84 percent of all licensing agreements (close to 1,000) in Korea were signed in the mechanical (36 percent), electrical and electronic (17.1 percent), and petrochemical (16.1 percent) industries (Chudnovsky 1986). The government participated in R&D through the creation of major specialized research centres such as the Korea Advanced Institute for Science and Technology (KAIST) and the Korea Institute for Machinery and Metals (KIMM). The research effort in this sector, measured in terms of R&D expenditures as a percentage of sales, is higher than that in other industries (Amsden and Kim 1986). In the case of machinery-design technologies, it is interesting to note that the state-owned KHIC adopted a similar strategy to that of the Korean engineering companies. For the construction of electrical power plant boilers, for instance, it was fairly simple for KHIC to acquire the necessary manufacturing technology through licensing agreements with foreign firms such as General Electric and Combustion Engineering. But design technology was a much more complex process. In the first phase, KHIC placed sixty to seventy engineers and technicians in the design offices and workshops of the licensing company for training in standards and computer software operation. In the second phase, the objective was set of acquiring 10 percent of the remaining non-standardized or non-coded technology. This was achieved through experience and comparisons gained through the presence of the licensing country's engineers sent to work in the KHIC study offices. In the third phase, KHIC now does the major part of design work and the licensing company only intervenes at the end of the process through inspection (Amsden and Kim 1986). It should be noted that the government supported this strategy through the use of preferential interest-rate loans. The Export of South Korean Technology The Korean import control policy and the promotion of local engineering and machinery-design technology was originally conceived

261 The Republic of South Korea

as part of an import substitution strategy. By the mid-igyos, this policy evolved into a central element of the country's export strategy. Given the limited size of the domestic construction and machinery market, Korean industry depended on the export of part of its production for further development. A recent study (Westphal et al. 1984) examines South Korean technology exports between 1977 and 1983. The total value of these exports was $47 million, which can be broken down as follows: • 93-3 percent were building or infrastructure construction exports. In 1984 Korea was the second-largest exporter of construction services. The great majority of construction material is still foreign. Engineering services is undertaken by Korean companies only in the case of small projects. Korean labour salaries represent about 20 percent of the contract value. • 5.4 percent involves industrial infrastructure exports, including off-shore facilities (2.1 percent), plants (i percent), and desalination plants (0.6 percent). This result is below that targeted by the Korean government, particularly in the export of integrated machinery, which represents one-third of that sector's exports. The two largest contracts, the sale of a cement works to Saudi Arabia and a tire plant to Sudan, were the work of the major exporting company, the Chaebol. • i percent was the export of engineering services and technical assistance. Plant construction and industrial engineering exports were for the most part the result of subcontracting agreements, with foreign multinationals undertaking the basic engineering side of the contract. Given the learning potential, the Korean government actively encouraged this type of arrangement. We should point out that between 1970 and 1980 Korean mechanical and electrical equipment exports increased 45 percent and that in 1984 the country held 1.2 percent of the world-export market share compared with 0.6 percent in 1970. CONCLUSION

Korea's performance on the domestic and export markets is largely due to a long-term strategy of technology transfer and the accumulation of technical expertise. This was done through controls imposed on the import of design services for production facilities and machinery in conjunction with the promotion of local design

262

Technological Development

capability in consulting engineering firms and in the design departments of machinery-manufacturing companies. Korea's success in this field demonstrates the key role of design services in the effective implementation of technology policies. It is nevertheless certain that the liberalization of international services trade as proposed by the United States and other industrialized countries will make such strategies considerably more difficult to implement in the future. REFERENCES

Amsden A., and L. Kim. 1986. "Technological Perspective in the General Machinery industry in the Republic of Korea." In Machinery and Economic Development, edited by M. Fransman. London: Macmillan. Asian Productivity Organization (APO). 1978. International Sub-Contracting: A Tool of Technology Transfer. Tokyo: APO. Chanaron, J-J., and J. Perrin. 1987. "R&D. Facilities Transfer into Developing Countries: Analysis and Proposals." Fortune, October, 503-12. Chaponniere, S.R. 1982. La Republique de Coree: Un nouveau pays industriel. Paris: La Documentation franchise. Notes and documentary studies. Chudnovsky, D. 1986. "The entry into the Design and Production of Complex Capital Goods. The Experience of Brazil, India and South Korea." In Machinery and Economic Development, edited by M. Fransman. London: Macmillan. Ecalle, F. 1986. La revolution tertiaire aux Etats-Unis. Notes and documentary studies. Paris: La Documentation francaise. Judet, P., and J. Perrin. 1977. Transfert de technologic et developpement: Problematique economique. Paris: Librairies techniques. Kim, L. 1980. "Stages of Development of Industrial Technology in a Developing Country, a Model." Research Policy, no. 9: 254—77. Perrin, J. 1976. Engineering: Terminology and Economic Function. Paris: OECD Development Center. — 1979. Biens d'equipement pour la petrochimie et les engrais dans les pays en voie de developpement. Grenoble: IREP-D (Institut de Recherche Economique et de Planification de Developpement). March. - 1983. Les transferts de technologie. Paris: La Decouverte. Rosenberg, N. 1976. Perspectives on Technology. Cambridge, Mass.: Cambridge University Press. Rothwell, R., P. Gardiner, and K. Pick. 1983. Design on the Economy: The Role of Design and Innovation in the Prosperity of Industrial Companies. Design Council.

263 The Republic of South Korea Sauvant, K. 1986. International Transactions in Services: The Politics of Transborder Data Flows. Boulder, Col.: Westview Press. Schott, K., and K. Pick. 1983. "The Effect of Price and Non price in U.K. Export Performance and Import Penetration." Discussion paper, no. 35, University College, London. United Nations Conference on Trade and Development (UNCTAD). 1984. Technology Issues in the Energy Sector of Developing Countries: Technological Impact of the Public Procurement Policy. The Experience of the Power Plant Sector in the Republic of Korea. Id/B/C 6/105, 4 June. Westphal, L., Y.W. Rhej, L. Kim, and A.H. Amsden. 1984. "Exports of Technology by Newly Industrializing Countries: A Case Study of the Republic of Korea." World Development, May-June, pp. 505-34.

This page intentionally left blank

PART FOUR

Industrial Structure and Innovation

This page intentionally left blank

CHAPTER ELEVEN

Oligopoly, Innovation, and Firm Competitiveness BERNARD BONIN

What types of firms are the most innovative and under what market conditions are they most likely to be so? Ever since Schumpeter hypothesized that the market structure most conducive to innovation would be some form of imperfect competition (later dubbed a "competitive oligopoly" by Villard), this hypothesis has been statistically tested both by those attempting to confirm it and those trying to refute it. No firm conclusion emerged from this work, which involved testing the hypothesis directly or indirectly by positing a relationship between oligopoly, firm size, and innovation. Although most of the work reviewed in this chapter was done in the context of a national economy, a conclusion to the effect that international oligopoly is indeed the market structure most conducive to innovation, albeit understandable, would remain a premature generalization.

INTRODUCTION

What types of firms are the most innovative and under what market conditions are they most likely to be so? What kind of competitive edge can they gain from their capacity for innovation? Innovation occurs when a new product or method is successfully incorporated into a firm's production process. While invention refers to a new idea, innovation is the application or commercialization of the novel idea or the knowledge. This process involves taking the idea or invention, identifying a market niche (or desire), developing from the idea a useful product, technique, or service, and packaging the result in a form acceptable to consumers. The process of innovation

268 Industrial Structure and Innovation

also includes improving an existing product or procedure to make it commercially viable. In the economic literature, the propagation of new ideas is considered to be one of the essential components of economic progress, since its effect is to spread and propagate knowledge, invention, and innovation from industry to industry and through the national economy, as well as between countries. In this chapter I will trace economists' attempts, in the context of a national economy, to find answers to the two questions posed above.l

SCHUMPETER'S HYPOTHESIS'* Most of the credit for raising the question of the importance of market structures for innovation in a national economy belongs to J. Schumpeter. His intriguing hypothesis provided the impetus for a large body of research, some of it aimed at confirming his hypothesis, some at refuting it. The results of these efforts, however, have not provided firm answers to the two questions we are concerned with here. J. Schumpeter believed that regeneration and growth occur in an economic system through the replacement of existing products, processes, and modes of industrial organization with better ones. Thus, technological progress involves the destruction of capital investments in zones where competition from new products and new methods is evident. This can be considered a form of "creative destruction." In a perfectly competitive market, small businesses have to constantly adapt to changing market conditions if they hope to survive. However, when the market is not perfectly competitive and all sectors of the economy are controlled by a handful of large groups, the members of these groups may actively resist any technological progress that threatens the structure of their capital investment. Schumpeter also believed that because innovation requires a high volume of resources and carries an inherently higher risk for the party investing these resources, it must offer the prospect of higher returns to be worthwhile. Business managers introduce improvements when they feel that the new production methods will lead to production costs lower than those associated with the methods currently employed. But improvements will not be made as long as present facilities are not fully amortized or depreciated. Under perfect competition, an innovation will spread quickly, eliminating the prospect of high profits and reducing a firm's in-

269 Oligopoly, Innovation, Competitiveness

centive to innovate. But under other market structures, firms may find innovation a very profitable activity and will be willing to commit resources to it. Attempts to impede technical progress will be of no avail. Large firms, whether they comprise a whole sector or part of a larger industry, have to face up to competition (or rivalry) in the form of new products, techniques, or sources of financing or in the form of competitors' potential recourse to such measures. The conditions outlined above are not sufficient for industries and firms whose main objective is maintaining the value of their previous investments. In many cases, ambitious plans will not be carried out unless their authors can rely from the outset on strategies to discourage competition or to explicitly hinder their competitors. Thus, it is important to provide innovators with opportunities for action and with the time they need to develop their advantages (through patents, industrial secrecy, long-term contracts, etc.). In addition, the initiative is only possible, in most cases, when there is a guarantee of exceptionally favourable conditions so that enough profits will be generated to weather the exceptionally difficult period that will inevitably follow. Accordingly, only those industries in which there are only a few firms already holding tacit or explicit agreements on competition by price, quality, or quantity can hope to achieve strong and steady production growth, with profits high enough to warrant the introduction of new products or processes into the system. This would not be the case where there are many firms in the industry. The logical conclusion of Schumpeter's reasoning is that a "competitive oligopoly" (to use Villard's expression) is the market structure most conductive to innovation. C H A R A C T E R I S T I C S OF AN OLIGOPOLISTIC MARKET STRUCTURE

In market theory, a market structure is oligopolistic when demand is satisfied by a handful of producers; that is, production is concentrated in the hands of a small number of firms, as when, for example, the four largest producers account for 80 percent of production. Interdependence of decision making is one of the main characteristics of this kind of market structure. All producers realize that the decisions taken by at least some of their competitors will depend upon the producers' own behaviour, and thus they weigh their decisions accordingly. In trying to preserve the existing order in their corner of the market, members of the oligopoly engage in a certain

270 Industrial Structure and Innovation

degree of competition in the marketplace, and this may sometimes involve price competition, investment competition, product differentiation, or genuine innovation. However, because of the economic power held by these firms, they do not have to accept market conditions passively (as do firms in a situation of perfect competition); instead they are able to influence these conditions to varying degrees by acting co-operatively. Oligopolistic firms, therefore, do not necessarily have the Machiavellian goals sometimes attributed to them. It is interesting to note that it is precisely the interdependence of decision making that has made a comprehensive theory of oligopoly so difficult to develop. In fact, there is a trade-off in any such model between an accurate description of conditions relative to perfect competition and the accuracy of the conclusions that may be drawn from it. Thus, to say that an oligopoly is the most appropriate structure for innovation is to argue that innovation springs first and foremost from the large firms that account for most of the supply to a market. Ever since Schumpeter declared that innovation was the essence of entrepreneurship, big business, rightly or wrongly, has been considered virtually synonymous with innovation - with new production and marketing methods, new products, new services, and so on. A firm that regularly puts new products onto the market is better able to spread the potential effects of product life cycles over time. Consequently, developing new products - which requires that scientists and engineers come up with innovations through invention or adaptation — and marketing these products have become two of the major functions of big business. When a firm launches a new product, it cannot sit back and wait for its market to develop; it must actively persuade consumers of the virtues of its product and constrain them to purchase them (to use Schumpeter's expression). Schumpeter's hypothesis has been tested statistically. Let us now take a closer look at the results of these empirical studies. FACTORS OF TECHNICAL PROGRESS AND

MARKET

STRUCTURES

We will look at five studies that have attempted to confirm Schumpeter's hypothesis: Villard 1958; Phillips 1956; Kamien and Schwartz 1970; Loury 1979; and Dasgupta and Stiglitz igSo.3 This research sample will suffice to show that no firm conclusions can yet be drawn.

271 Oligopoly, Innovation, Competitiveness

Villard adopted a position similar to Schumpeter's, arguing that a market characterized by oligopolistic competition will progress more quickly than a perfectly competitive market. This will be the case when firms in the market are sufficiently large and sufficiently few that they can carry out research knowing they will benefit from its results. Progress will also be more rapid when firms have some incentive to innovate. Villard found that research is concentrated in large firms and in a small number of industries. In terms of the relationship between research and development and the number of firms in a market, Villard offered the hypothesis that because research is generally not a profitable activity (very few individuals live off their inventions), research will be undertaken mainly by and for those firms that have products to sell. Consequently, there will be no R&D in a highly competitive environment, but there will be R&D in oligopolistic markets, where improvements to a product or a production process may confer a temporary advantage over rival firms. While it is true that the reduced competition in oligopolistic market structures may hamper progress, inter-firm competition remains strong enough that innovation itself becomes a kind of competition. Under these circumstances, firms that do not innovate are courting disaster. Phillips attempted to test the hypothesis that there is a positive correlation between the degree of concentration of production in a small number of large firms and observed signs of technical progress. He found such a relationship between concentration, firm size, and the pace of technical change, but was unable to determine the direction of influence — that is, whether particular characteristics of concentration and firm size stimulate technical progress or whether technical change influences concentration and firm size. Kamien and Schwartz tried to identify the conditions under which a monopolistic industry structure would provide a stronger stimulus to invention than a competitive structure, as well as the conditions under which this stimulant would be weaker. They found the price elasticity of demand to be the critical factor. In another paper, they argued that a firm's choice regarding the optimum development rate of an innovation in an environment where the similar choices of its rival were uncertain would depend upon the expected structure of the returns accruing to the innovator and would likely fall short of a monopolistic position. Loury took into account the relative ease of firms' entry into the industry involved, and concluded that greater competition was not necessarily socially desirable. Given the steadily decreasing returns on R&D investment, optimum innovation effort arises from atomistic

272

Industrial Structure and Innovation

competition (since entry into the industry carries no cost). In a more realistic context that includes economies of scale, however, the optimum market structure would imply a finite number of firms. In all market structures, however, firms tend to invest more than the optimum amount in R&D because they fail to take into account that their respective efforts parallel one another. Thus, social welfare can be maximized by limiting firms' entry and investment through the imposition of royalties or limited-life patents. Lastly, Dasgupta and Stiglitz started from the assumption that competition itself can take different forms: price competition, product competition, and R&D competition aimed at invention and innovation. However, there may be a link between monopolistic competition in terms of products (differentiation) and competition in terms of research strategies. Among their various findings, perhaps the most important is that R&D rivalry cannot necessarily be considered a substitute for competition in the product market, nor does it necessarily lead to a competitive product market. In fact, R&D competition requires imperfect competition in the product market. What is striking about the above research, which was presumably intended to confirm Schumpeter's hypothesis, is the lack of robustness of the results. Even though the various studies are based on different models and use a variety of statistical techniques, not one of the authors has succeeded in clearly demonstrating the relationship between oligopolistic market structure and innovation. The proof that is offered tends to be approached by a circuitous route. Curiously, however (and this only serves to complicate our search for answers), some of the work ostensibly aimed at disproving Schumpeter's hypothesis ends up lending it some support. Scherer (1965), for example, concludes that high concentration in a market definitely has a beneficial effect on the production of inventions but the effect is felt only weakly. In his study, raising from 15 to 90 percent the market share of the four largest firms in an industry where those firms had sales of around $100 million (at that time) resulted in only a single additional patent. When the four largest firms had sales on the order of $1 billion, however, increasing the concentration rate from 15 to 90 percent produced five new patents. Comanor (1967) also showed that the average levels of research activity in industries with differentiated goods and in industries with non-differentiated goods tended to increase with greater concentration of production, although the correlation was stronger in industries with low product differentiation. In industries where products could be readily differentiated (and so where research was

273 Oligopoly, Innovation, Competitiveness

an important aspect of competitive capacity), concentration was not necessarily a crucial factor. With lower differentiation, on the other hand, concentration potentially played a role. In this case, research programs focused more on new production techniques than on new products. Thus, in industries where a positive link could be established between concentration and research, the neo-Schumpeterian hypothesis appeared to be valid. In industries where research was more closely linked to competitive conditions, however, Comanor found that higher concentration did not lead to more research. Lastly, Shrieves attempted to measure the relationship between innovation (measured by employment in R&D, i.e., a measure of input, not output) and levels of industrial concentration. Using a sample of 411 firms, he found that small firms active in R&D actually invested more in R&D as a proportion of their sales than large firms (a conclusion disputed by some authors). The relationship between concentration and innovation varied according to the type of product sold and the various kinds of markets served by the industry. The link was direct and significant for commodities (grain, wood, oil) and consumer goods, positive but weak for producer goods, and negative and not significant for capital equipment (machinery and other equipment; aviation). OLIGOPOLY, FIRM SIZE, AND INNOVATION

Let me briefly mention some other research that is relevant to the subject in that it indirectly touches upon particular aspects of the relationship between market structure and innovation: changes in industry structures over time; determinants of firms' R&D expenditures; research intensity, firm size, profitability, and growth; risk and innovation; and corporate strategy and innovation. To summarize the results of these efforts, one could say that the oligopoly may be judged superior as long as the following line of reasoning is valid. The larger the firm, the more active it is in R&D. Moreover, research is the primary producer of industrial innovation, which guarantees higher profitability and more rapid growth for firms. Since firms tend to be larger in research-intensive industries, and since innovation is inherently risky, larger firms definitely enjoy an advantage in this area. Larger firm size is characteristic of an oligopolistic market structure. Therefore, oligopolistic structure must indeed be conducive to innovation, and this conclusion would only be reinforced by observations that oligopolistic firms tend to take a leading role in innovation.

274 Industrial Structure and Innovation

In other words, despite the fact that research expressly aimed at establishing a relationship between market structure and innovation has failed to do so, can the link be inferred from theoretical and empirical research on specific aspects of the question? Unfortunately, up to now, this indirect route has not yielded any morepromising results than the direct approach. For example, the body of research on product life cycles,4 which has already served to show that corporate managers may play a larger role than previously thought, gives some indication of how market structures evolve over time. Only in the mature phase of the product life cycle — when technology is stable, there are few major innovations to be made, and capital intensity and the price elasticity of demand are high - will the number of firms decline and market structures approach an oligopolistic structure. Large firm size is not necessary for invention, development, and technological innovation when an industry is young (even though innovation is at its peak at this time). However, during the next stage, when there are a large number of firms in the industry, the high cost of building and maintaining R&D capacity becomes an entry barrier that represents an advantage for large industrial laboratories. Opportunities for young firms tend to be better during the early stages of the product life cycle (novelty), when scale economies are not yet a factor, market shares are highly changeable, and entry and bankruptcy rates are high. To sum up, larger firms generally enjoy more of an advantage for "older" products, and younger firms have better chances at the novelty stage. It is therefore difficult to find definite proof in research on product life cycles of the superiority of large firms (and so, oligopolies) in the area of innovation. Perhaps it is possible to establish a positive relationship between firm size and R&D spending.5 One might think that large firms spend proportionately more on R&D, relative to their sales volumes, than smaller firms, not only because they have more incentive to do so, but also because they have more resources at their disposal. Yet, statistical research has been unable to confirm such a relationship. Above a certain threshold, which varies by industry, the relative R&D effort seems to increase with the size of the firm up to a point, where it hits a plateau or even declines. The positive relationship is not valid for all firm sizes, therefore. This result is compatible with the finding that the vast majority of large firms have their own research programs, while the vast majority of small firms do not. In the form it was posited and tested, this relationship involves firms that carry out research, whether they are large or small, and not the concept of participation rates in the national research effort.

275 Oligopoly, Innovation, Competitiveness

Since a firm's degree of diversification bears some relationship to its size, if a positive relationship were found between research and the diversification of firms' operations, it might be concluded that large firms enjoy an advantage. They would be better placed to diversify their activities, attract the best and most innovative talent, and bring a greater percentage of their discoveries to market because their costs and failures could be spread among a wider range of activities. Such a relationship has indeed been confirmed by research, but the direction of the cause/effect link has not been established: greater diversification in firms' activities could be leading to higher R&D intensity, or it could simply be the result of higher R&D activity. Few would disagree that a successful innovation is profitable for a firm; this fact has been clearly established by research. But what is the nature of the relationship between research intensity, firm size, profitability, and growth (Dunning and Pearce 1981)? There is evidence of a statistical relationship between research intensity and firm size; in other words, firms in research-intensive sectors tend to be larger. Among the largest firms, there exists what might be termed a "U" relationship between size and profitability, as well as between size and growth. The nature of the relationships between research intensity, profitability, and growth is not clear. While most research-intensive industries exhibit high profitability ratios, there are also highly profitable firms among those with low research intensity. What seems most certain from the analysis is that industries with average research intensity fare the worst in terms of profitability ratios. Again, there is a U-shaped relationship between research intensity and profitability; high profitability may be associated with both high and low research intensity, but rarely with average levels. Moreover, the link between firms' research intensity and their subsequent growth is not strong, while the association between successful innovation and growth is more firmly established. Common sense would dictate (and empirical studies confirm) that success in technical innovation brings a firm rapid growth. On the other hand, innovation failure may cause a firm to slide into bankruptcy, regardless of its commitment to R&D. Thus, research by itself is not a sufficient condition for success. Most theories of innovation are based on three postulates: (i) the cost of innovating is sufficiently high that only large firms can attempt it; (2) enough product must be produced so that the successes and failures at least level each other out; and (3) in order for an innovation to be introduced, the firm must have adequate control over

276 Industrial Structure and Innovation

its market. More recently, however, in light of some marked differences by industry, these assumptions are being revised (Freeman 1982). When the process of innovation is broken down into phases, it appears that small firms have an advantage in the initial stages of invention, as well as an advantage for less expensive, but much more "radical," inventions. Large firms, on the other hand, enjoy an advantage in the later stages of the innovation process (since development costs are often much higher than the cost of the original idea); they also are in a better position to improve and extend technological breakthroughs in the later stages. In addition, the relative performance of small and large firms varies from industry to industry. Analysis reveals two distinct types of industry: • industries in which small firms have made little or apparently no contribution to innovation, both in absolute and relative terms (e.g., the aerospace industry, motor vehicles, dyeing, pharmaceuticals, cement, glass, steel, aluminum, resins, naval construction, coal and gas); and • industries in which small firms have made a significant contribution to innovation in their respective industries (e.g., scientific instruments, electronics, carpets, textiles, textile machinery, paper, leather and shoes, wood and furniture, and construction). RISK, UNCERTAINTY, AND INNOVATION

If innovation is inherently a very risky business, it would not be surprising to find that it is primarily conducted by large firms. The uncertainty involved may stem from the technique itself, the market, or economic conditions in general. The market risk frequently appears greater than the technical risk. Generally speaking, firms do not acccept projects unless they have a good chance of technical success (say, 70 percent). Since firms have much less control over markets, market-related risk is much more difficult to predict beforehand. In fact, a high degree of uncertainty is acceptable only to certain kinds of innovators (Freeman 1982, 156; Mansfield 1982). A number of empirical studies indicate that there is a tendency to exaggerate the technical risk involved in innovation. Mansfield et al. (1971) demonstrated this quite clearly. The majority of projects surveyed by their study were considered by the firms in question to be relatively safe from the technical point of view. Half of the projects

277 Oligopoly, Innovation, Competitiveness

included in the sample were judged by the individuals involved to have at least a 50 percent chance of technical success. When a project was technically complete, the probability of it being a commercial success was just over 50 percent. When it was already on the market, the probability that it would be a commercial success in the sense that it would generate profit (according to the economic meaning of the term) was about 40 percent. In fact, projects that were technically successful were expected to be worthwhile, since the expected median rate of return was generally about 30 percent, despite frequent cost overruns, longer-than-anticipated delays, and even errors in profit projections. More recently, Cooper (1983) showed, on the basis of the experiences of 103 firms with new products, that the rate of commercial success for new products was, at 59 percent, considerably higher than previously had been expected. In more than half of the remaining cases, moreover, the projects had already been terminated before reaching the commercialization stage. Another finding was that firms' ranking in terms of success rate was unrelated to the size of their R&D expenditures: those firms that spent the most on R&D had the smallest sales of new products by R&D dollar. This led Cooper to argue that if R&D did indeed promote the sale of new products, it did so most inefficiently. Firms' marketing resources appeared to play a more significant role in the commercial success of new products than their R&D expertise. This calls into question the contention that firms with large technical capabilities are better able than firms with weak technical capabilities to judge whether their new products have a better chance of success than those of their competitors. Lastly, the theory that each firm is a unique case in itself does not receive much support from Cooper's work, which indicates that the nature of the firm (i.e., the industry to which it belongs, its size, and its ownership) has no influence on actual performance indicators. The results show instead a remarkably consistent correlation between the success rates for new products and R&D efficiency, independent of firm size, industry, and ownership. These findings also indicate that firms' managers may adopt any of several different innovation strategies - offensive, defensive, imitative, or opportunistic. While space does not permit examining all of these approaches here, the first two have particular relevance for the discussion. A firm opting for an offensive strategy tries to stay one step ahead of its competitors in introducing new products in order to achieve a technological and commercial leadership role. A completely different approach is the defensive strategy, which by no means implies

278 Industrial Structure and Innovation

a lack of R&D activity; in fact, a defensive strategy may be just as research intensive as an offensive one. The label applies rather to the nature of the innovations and the timing of their introduction. "Defensive" innovators, while they do not strive to be the first in the world, have no intention of being shut out of technological change. In fact, most industrial research is essentially defensive or imitative in that it is aimed at minor improvements, variations in existing products and techniques, technical services, or other such short-term goals.6 In fact, defensive R&D is likely an inherent feature of oligopolistic markets and is closely linked with product differentiation. CONCLUSION

On the basis of this (hopefully) representative survey of the literature on economic analyses of industrial innovation, two conclusions may be drawn. First, J. Schumpeter's theory of innovation remains eminently plausible; however, the conjectured relationship between an oligopolistic market structure and innovation or technical progress continues to resist confirmation, despite the large body of theoretical and empirical research on the question. Second, Schumpeter's theory cannot be confirmed indirectly on the basis of work on the relationships between oligopoly, firm size, and innovation. Large firms do not appear to enjoy a clear advantage in terms of a particular phase of the product life cycle during which a greater number of innovations might be expected. Although the vast majority of large firms engage in research, it is not clear that they invest more in R&D as a proportion of their sales than smaller firms. And even if this were so, it would remain to be demonstrated that R&D is a necessary and sufficient condition for successful innovation. The relationships between firm size, profitability, and growth, on the one hand, and between research intensity, profitability, and growth, on the other, do not appear to be linear, even though successful innovation is clearly an important element in firms' growth. It may be that small firms have an advantage during certain phases of the innovation process and have much better opportunities in certain industries to exploit their particular talents along these lines. Lastly, it seems highly likely that most economic theories of innovation overestimate the extent of risk and uncertainty. It might be added in closing that if the analysis of the relationship between oligopoly and innovation had been extended to an international context, we would have found a great deal of research aimed at proving the existence of this relationship; indeed, we would have

279 Oligopoly, Innovation, Competitiveness

found many authors who are already firmly convinced of it. However, whether we are concerned with international relations or with national economies, it is only when we hypothesize the existence of such a relationship and logically deduce the conditions of an appropriate model that we can come to the conclusion that international oligopolies are the main sources of the creation and international propagation of innovation. This relationship remains to be statistically verified, however (Bonin 1984). Ever since the late S.H. Hymer pointed out the relationship between oligopolies and international investment in his doctoral thesis (which, although submitted in 1960, was only published some fifteen years later), much research has been devoted to exploring this relationship or certain aspects of it. This work has revealed a link, albeit a weak one, between foreign production and research intensity in production and between research intensity and firms' degree of multinationalism. Some theories of international investment refer explicitly to the behaviour of oligopolies (Raymond Vernon's localization theory, for instance), while others assume from the outset that markets are imperfect (e.g., the entire school of thought related to industrial organization or to the advantages of internalization). The possibility of a kind of "bandwagon" effect has already been raised by Yair Aharoni relative to foreign investment decisions, and F.T. Knickerbocker has verified a tendency for this effect to create "bursts" of entries in an oligopolistic context. When products' characteristics are finely distinguished, oligopolies may appear to be very common in international production. In a survey of world production mandates completed in 1984, 30 percent of respondents felt that they had no more than five competitors on world markets, and 80 percent no more than fifteen competitors (Bonin and Perron 1986). Moreover, since multinational firms appear to enjoy advantages from the point of view of the international diffusion of innovations (considering that a possible substitute, contractual agreements, has itself been a factor in oligopolistic rivalry in certain circumstances [Bonin 1987], the conclusion that international oligopoly is indeed the market structure most conducive to innovation is understandable. Nevertheless, this remains a premature generalization. NOTES

1 The first part of this text is based on Bonin and Desranleau 1988. 2 For a basic discussion of Schumpeter's thesis, see Schumpeter 1950. 3 For a critical overview of work on market structure and innovation, see also Kamien and Schwartz 1975.

280 Industrial Structure and Innovation 4 Besides the well-known work of R. Vernon, see also Hirsch 1967, Mueller and Tilton 1969, and Pavitt and Wald 1971. 5 A summary of the most important research in this area may be found in Kamien and Schwartz 1975, ECC 1983, and Pavitt and Walker 1976. 6 Among other sources, consult ECC 1983; Nelson et al. 1967; Schott 1975, 1981. REFERENCES

Bonin, B. 1984. L'entreprise multinationale et I'etat. Montreal: Les Editions etudes vivantes. - 1987. "Contractual Agreements and International Technology Transfers: The Empirical Studies." In Multinationals, Governments and International Technology Transfer, edited by A.E. Safarian and G.Y. Berlin. London: Groom Helm. See also the paper by Alan Rugman in the same volume. Bonin, B., and C. Desranleau. 1988. Innovation industrielle et analyse economique. Chicoutimi: Gaetan Morin. Bonin, B., and B. Perron. 1986. "World Product Mandates and Firms Operating in Quebec." In Managing the Multinational Subsidiary, edited by H. Etemad and L. Seguin-Dulude, 161-76. London: Groom Helm. Comanor, W.S. 1967. "Market Structure, Product Differentiation and Industrial Research. Quarterly Journal of Economics, November. Cooper, R.G. 1983. "Most New Products Do Succeed." Research Management, November—December. Dasgupta, P., and J. Stiglitz. 1980. "Uncertainty, Industrial Structure and the Speed of R and D." Bell Journal of Economics 11, no. i. Dunning, J.H., and R. Pearce. 1981. The World's Largest Industrial Enterprises. London: Gower Press. Freeman, C. 1982. The Economics of Industrial Innovation. 2nd ed. Cambridge, Mass.: MIT Press. Hirsch, S. 1967. Location of Industry and International Competitiveness. Oxford: Clarendon Press. Kamien, M.I., and N.L. Schwartz. 1970. "Market Structure, Elasticity of Demand and the Incentive to Invest." Journal of Law and Economics 13, no. i. — 1975. "Market Structure and Innovation: A Survey. Journal of Economic Literature 13, no. i. Loury, G.C. 1979. "Market Structure and Innovation." Quarterly Journal of Economics, August. Mansfield, E. 1982. "How Economists See R and D." Research Management,

July-

281 Oligopoly, Innovation, Competitiveness Mansfield, E., J. Rapoport, J. Schnee, S. Wagner, and M. Hamburger. 1971. Research and Innovation in the Modern Corporation. New York: W.W. Norton. Mueller, B.C., and J.E. Tilton. 1969. "Research and Development Costs as a Barrier to Entry." Canadian Journal of Economics, November. Nelson, R.R., J. Peck, and E. Kalacheck. 1967. Technology, Economic Growth and Public Policy. Washington, D.C.: Brookings Institute. Pavitt, K., and S. Wald. 1971. The Conditions for Success in Technological Innovation. Paris: OECD. Pavitt, K., and W. Walker. 1976. "Policies Towards Industrial Innovation: A Review." Research Policy, no. 5. Phillips, A. 1956. "Concentration, Scale and Technological Change in Selected Manufacturing Industries, 1899-1939." Journal of Industrial Economics, June. Scherer, P.M. 1965. "Firm Size, Market Structure, Opportunities and the Output of Patented Inventions." American Economic Review, December. Schott, K. 1975. "Industrial R and D Expenditures in the U.K.: An Econometric Analysis." Ph. D. thesis, Oxford University. — 1981. Industrial Innovation in the United Kingdom, Canada, and the United States. British-North American Committee. Schumpeter, J. 1950. Capitalism, Socialism and Democracy, 3rd ed. New York: Harper and Row. Shrieves, R.E. 1978. "Market Structure and Innovations: A New Perspective." Journal of Industrial Economics, June. Economic Council of Canada (ECC). 1983. The Bottom Line, chap. 4. Ottawa. Villard, H.H. 1958. "Competition, Oligopoly and Research. "Journal of Political Economy, December.