The Industrial Experience of Tanzania

The Industrial Experience of Tanzania Edited by Adam Szirmai and Paul Lapperre The Industrial Experience of Tanzania ...

Author: Adam Szirmai | Paul Lapperre

236 downloads 1386 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

The Industrial Experience of Tanzania Edited by

Adam Szirmai and Paul Lapperre

The Industrial Experience of Tanzania

Also by Adam Szirmai ECONOMIC AND SOCIAL DEVELOPMENT: Trends, Problems, Policies EXPLAINING ECONOMIC GROWTH: Essays in Honour of Angus Maddison (editor with B. Van Ark and D. Pilat) INEQUALITY OBSERVED: A Study of Attitudes towards Income Inequality ONTWIKKELINGSLANDEN: Dynamiek en Stagnatie

The Industrial Experience of Tanzania Edited by

Adam Szirmai Professor of Technology and Development Studies Eindhoven Centre for Innovation Studies (ECIS) Eindhoven University of Technology The Netherlands

and

Paul Lapperre Associate Professor of Technology and Development Studies Director of Education MSc Programme Technology and Society Eindhoven University of Technology The Netherlands

Editorial matter and selection © Adam Szirmai and Paul Lapperre 2001 Chapter 1 © Donné van Engelen, Adam Szirmai and Paul Lapperre 2001 Chapter 3 © Adam Szirmai, Menno Prins and Wessel Schulte 2001 Chapter 7 © Bartelt Bongenaar and Adam Szirmai 2001 Chapter 9 © Raymond Duijsens and Paul Lapperre 2001 Chapter 12 © Paul Lapperre 2001 Chapters 2, 4–6, 8, 10, 11, 13–17 © Palgrave Publishers Ltd 2001 Chapter 5 was previously published in Public Choice 89: 3/4, pp. 375–92, 1996. Reprinted with kind permission from Kluwer Academic Publishers. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P 0LP. Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2001 by PALGRAVE Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N. Y. 10010 Companies and representatives throughout the world PALGRAVE is the new global academic imprint of St. Martin’s Press LLC Scholarly and Reference Division and Palgrave Publishers Ltd (formerly Macmillan Press Ltd). ISBN 0–333–80019–2 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data The industrial experience of Tanzania / edited by Adam Szirmai and Paul Lapperre. p. cm. Includes bibliographical references and index. ISBN 0–333–80019–2 1. Industrialization—Tanzania. 2. Tanzania—Economic conditions—1964– 3. Technology—Tanzania. I. Szirmai, Adam, 1946– II. Lapperre, Paul, 1942– HC885 .I533 2001 338.09678—dc21 2001021199 10 9 8 7 6 5 4 3 2 1 10 09 08 07 06 05 04 03 02 01 Printed in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire

Contents Preface

vii

Conference Participants

viii

Notes on the Contributors Introduction Adam Szirmai and Paul Lapperre

1

Part I: Long-run Economic Performance

9

1

2 3

4

x

Public Policy and the Industrial Development of Tanzania, 1961–95 Donné van Engelen, Adam Szirmai, Paul Lapperre

11

Is African Manufacturing Skill Constrained? Howard Pack and Christina Paxson

50

Measuring Manufacturing Performance in Tanzania: GDP, Employment and Comparative Labour Productivity 1961–95 Adam Szirmai, Menno Prins and Wessel Schulte

73

The Role of Technological Factors in the Early Stages of Industrial Exports: A Note Charles Cooper

Part II: Innovation, Technological Capabilities and Choice of Techniques 5

6

7

8

114

133

Public Choice, Technology and Industrialization in Tanzania. Some Paradoxes Resolved Jeffrey James

135

The Form and Role of Industrial Innovativeness in Enhancing Firms’ Productivity: The Case of Selected Manufacturing Firms in Tanzania Haji H. Semboja and Josephat P. Kweka

153

Development and Diffusion of Technology: The Case of TIRDO Bartelt Bongenaar and Adam Szirmai

171

Technological Capabilities: A Core Element for National Development Opportunities? A Study of Technological v

vi Contents

9

Capabilities in the Dwelling Construction Sector for Lower-Income Households in Tanzania Emilia van Egmond-de Wilde de Ligny

194

Technical Education, Knowledge and Skills in the Metalworking Industry in Tanzania Raymond Duijsens and Paul Lapperre

218

Part III: Environmental and Energy Aspects of Industrialization

235

10 Industry and Environment: Methodologies for Environmental Assessment in Data-Poor Situations Lex Lemmens, Peter Scheren, Harro Zanting, Gregory Njau and David van Horen

237

11 Energy Conservation in the Industrial Sector in Tanzania Frank van der Vleuten, Lex Lemmens, Otto Bos, Caspar Samplonius, Dick Toussaint and Michel Yhdego

262

Part IV: Lessons from Past Experiences, Economic Reform, Prospects for the Future

281

12 Industrialization of Tanzania: Can Tanzania Learn from European History? Paul Lapperre

283

13 Macro-Economic Policy and Performance of the Manufacturing Sector in Tanzania: Has Liberalisation Helped? An Econometric Approach A.V.Y. Mbelle

300

14 The Urban Informal Manufacturing Sector in Tanzania: Neglected Opportunities for Socioeconomic Development Herman Gaillard and Amber Beernink

318

15 The Impact of Reforms in Tanzania: The Case of Privatized Manufacturing Industries Humphrey P.B. Moshi

341

16 Economic Reforms, Industrialization and Technological Capabilities in Tanzanian Industry Samuel M. Wangwe

349

17 Highlights of the Sustainable Industrial Development Policy in Tanzania, 1996–2020 A.K. Maziku

367

Bibliography

376

Index

405

Preface This volume presents the proceedings of a conference on ‘The Industrial Performance of Tanzania’ held from 25 to 27 June 1998 at Eindhoven University of Technology. The conference has been organized jointly by the Section of Technology and Development Studies (TDS) and the Centre for International Cooperation Activities (CICA) of Eindhoven University of Technology. Since 1973 CICA has been responsible for the cooperation between the Eindhoven University of Technology and organizations and universities in developing countries. The conference marked its 25th anniversary. The Technology and Development Studies section is responsible for research and teaching in the field of technology, innovation policy and development studies. The section is part of the Department of Technology and Policy, Faculty of Technology Management, at the Eindhoven University of Technology. It participates in the MSc programme Technology and Society. Its research activities are integrated into the research programme of the Eindhoven Centre for Innovation Studies (ECIS). Between 1992 and 1998 staff and MSc students of the TDS executed a wide range of research projects in firms, organizations and sectors in Tanzania, in the context of a research programme on industrialization and technological development in Tanzania. Over a hundred research assignments have been completed. These research projects have resulted in a wealth of new information about firm-level and sectoral developments in Tanzanian industry. One of the aims of the conference was to make part of this information available to the outside world. A second aim was to strengthen scientific interaction between African and European researchers and institutions. Nine researchers from Tanzania were invited to present papers at the conference. Many of their contributions have been included in this volume. We gratefully acknowledge financial support from the Faculty of Technology Management and the Department of Technology and Policy. We thank Jan van Cranenbroek of CICA for his work as co-organizer of the conference. Lutgart van Kollenburg acted as secretary of the conference and made contributions going far beyond the bounds of normal secretarial support. Thanks are due to Petra Heck, who provided invaluable support in the word processing of the manuscript. Adam Szirmai Paul Lapperre October 1999 vii

Conference Participants ‘The Industrial Performance of Tanzania’ Bartelt Bongenaar Eindhoven University of Technology IBM Deployment of New Technologies Amsterdam Charles Cooper The United Nations University/INTECH Maastricht, The Netherlands Jan van Cranenbroek Bureau for International Affairs, Eindhoven University of Technology The Netherlands Raymond Duijsens Eindhoven University of Technology Netherlands Red Cross The Netherlands Emilia van Egmond-de Wilde de Ligny Eindhoven University of Technology The Netherlands Donné van Engelen Eindhoven University of Technology The Netherlands

Herman Gaillard Eindhoven University of Technology The Netherlands Jeffrey James Tilburg University The Netherlands Josephat P. Kweka Economic and Social Research Foundation (ESRF), Dar es Salaam, Tanzania Paul Lapperre Eindhoven University of Technology The Netherlands Lex Lemmens Centre for Technology for Sustainable Development (TDO) Eindhoven University of Technology The Netherlands A.K. Maziku Ministry of Industries and Trade Tanzania A.V.Y. Mbelle University of Dar es Salaam Tanzania viii

Conference Participants ix

T.S.A. Mbwette University of Dar es Salaam Dar es Salaam, Tanzania Hubert Meena University of Dar es Salaam Dar es Salaam, Tanzania Humphrey P.B. Moshi Economic Research Bureau University of Dar es Salaam Tanzania Gregory Njau Environmental Association of Tanzania (ENATA), Dar es Salaam Tanzania

Wessel Schulte Eindhoven University of Technology The Netherlands UNIDO/UNDP Uganda Haji H. Semboja Economic and Social Research Foundation (ESRF), Dar es Salaam, Tanzania Adam Szirmai Eindhoven Centre for Innovation Studies Eindhoven University of Technology The Netherlands

Howard Pack World Bank/University of Pennsylvania USA

F. van der Vleuten Eindhoven University of Technology Free Energy Europe Eindhoven, The Netherlands

Menno Prins Eindhoven University of Technology ERICSSON, Sweden

Samuel M. Wangwe Economic and Research Foundation (ESRF), Dar es Salaam Tanzania

Notes on the Contributors Amber Beernink, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Bartelt Bongenaar, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology/IBM Deployment of New Technologies, Amsterdam, The Netherlands Otto Bos, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Charles Cooper, The United Nations University/INTECH Maastricht, The Netherlands Raymond Duijsens, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology/ Netherlands Red Cross, The Netherlands Emilia van Egmond-de Wilde de Ligny, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Donné van Engelen, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Herman Gaillard, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands David van Horen, Centre for Technology for Sustainable Development, Eindhoven University of Technology, The Netherlands Jeffrey James, Tilburg University, The Netherlands Josephat P. Kweka, Economic and Social Research Foundation (ESRF), Dar es Salaam, Tanzania x

Notes on the Contributors xi

Paul Lapperre, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Lex Lemmens, Centre for Technology for Sustainable Development (TDO), Eindhoven University of Technology, The Netherlands A.K. Maziku, Ministry of Industries and Trade, Tanzania A.V.Y. Mbelle, University of Dar es Salaam, Tanzania Humphrey P.B. Moshi, Economic Research Bureau, University of Dar es Salaam, Tanzania Gregory Njau, Environmental Association of Tanzania (ENATA), Dar es Salaam, Tanzania Howard Pack, World Bank/University of Pennsylvania, USA Christina Paxson, Princeton University, USA Menno Prins, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology/ERICSSON, Sweden Caspar Samplonius, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Peter Scheren, Centre for Technology for Sustainable Development, Eindhoven University of Technology, The Netherlands Wessel Schulte, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands, and UNIDO/UNDP, Uganda Haji H. Semboja, Economic and Social Research Foundation (ESRF), Dar es Salaam, Tanzania Adam Szirmai, Eindhoven Centre for Innovation Studies, Eindhoven University of Technology, The Netherlands

xii Notes on the Contributors

Dick Toussaint, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology, The Netherlands Frank van der Vleuten, Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology/ Free Energy Europe, Eindhoven, The Netherlands Samuel M. Wangwe, Economic and Social Research Foundation (ESRF), Dar es Salaam, Tanzania Michel Yhdego, ERC Consultants, Dar es Salaam, Tanzania Harro Zanting, Centre for Technology for Sustainable Development, Eindhoven University of Technology, The Netherlands

Introduction Adam Szirmai and Paul Lapperre

This collection of articles focuses on an analysis of the industrial experience of Tanzania since independence in 1961. Tanzania is taken as a case study of industrialization in sub-Saharan Africa because it represents many of the common features of industrialization processes in other African economies. What we need to understand is why some developing economies, which started from low levels of per capita income in the post-war period, achieved some measure of success in industrialization and economic development, while others did not. In this context the experiences of some Asian countries contrast sharply with the experiences of the majority of countries in sub-Saharan Africa. As Tanzania in many ways represents one of the models of African economic development, a careful analysis of its industrial experiences, and in particular the role of policy in these experiences, contributes to a better understanding of both Tanzanian and African economic development. Common features of African development include the lack of an industrial heritage, the overwhelming importance of the agricultural sector in the post-war period, dependence on primary exports and a very small share of manufacturing in GDP and employment. In policy and development thinking industrialization was, and often still is, seen as the key to economic development. The fiery enthusiasm for industrialization was coupled with distrust of free-market forces, which were negatively associated with colonial experiences. A developmental state soon came to be seen as the main catalyst of development. Thus – irrespective of the precise shadings of ideology– state interventionism, state planning and state ownership of industrial enterprises increased all over Africa. This is not to say that Tanzania has no distinctive features of its own. It has specific characteristics which distinguish it from many other countries. Apart from the intervention in Uganda in the 1980s, Tanzania succeeded in maintaining both peaceful relations with its neighbours and internal peace. Although the process of rural resettlement during the Ujamaa period, in the late 1960s and early 1970s, resulted in some measure 1

2 The Industrial Experience of Tanzania

of forced relocation and disruption of agricultural populations, it avoided the extremes of Stalinist socialism, exemplified for instance by Ethiopia. Also there were undoubted successes in fields such as education and health care. Julius Nyerere was a political leader of exceptional integrity, who played a major role in maintaining ethnic harmony and building a sense of national identity. Tanzania was also characterized by peaceful changes of leadership and a smooth transition to a multiparty democracy in 1996, though marred by rising corruption. In the late 1950s and early 1960s, after most countries gained their independence, hopes for rapid economic development were high across the African continent. Initially these hopes seemed to be well founded. Until 1973 income per capita and manufacturing GDP per capita tended to increase. From 1973 onwards, with country to country variations, a long period of stagnation set it. Apart from economic factors, political instability, corruption, military conflict and ethnic strife were major factors contributing to economic stagnation in most African countries. As time went by, the perception of the role of the state started to change. Instead of being seen as a driving force in development, the state even started to be interpreted by some observers as a predatory and corrupt institution, diverting resources and human talent and effort from productive investment in economic development. In the course of the 1980s, the international and national economic policy climate changed. Almost all countries, whether voluntarily or involuntarily, embarked on a process of structural adjustment, liberalization, deregulation, privatization and return to the market. Although there is a consensus concerning the shortcomings of state led industrialization, structural adjustment and economic reform have had mixed results in Africa. Despite some intimations of change there has, so far, seldom been a resumption of rapid sustained growth. Some sub-Saharan countries are even experiencing de-industrialization, and many are now faced with economic stagnation. In the case of Tanzania, political commitment to reforms seems to be quite strong and authentic. But, as illustrated by the articles in this volume, there is still a lively debate on the successes and failures of reform and on the future role of the state in the context of reform, liberalization and structural adjustment. This book is structured into four parts. Part I focuses on long-run economic performance. It consists of two chapters on Tanzania and two chapters which place Tanzanian industrialization in a comparative international perspective. Part II brings together five articles dealing with issues of technology and innovation. Part III presents two articles on environmental and energy aspects of industrialization. Section IV brings together six articles focusing on economic reform and its prospects. The volume combines contributions from economics, sociology and engineering. It complements articles focusing on macro trends, with

Introduction 3

articles based on a wealth of information derived from firm-level surveys in various sectors of Tanzanian manufacturing. These provide new insights into the mechanics of industrialization, technological change and the responses to reform. Part I opens with a chapter by van Engelen, Szirmai and Lapperre. The chapter provides an overview of the evolution of industrial policy in Tanzania since 1961 and the impact of policy on industrial performance. It describes the familiar process of increasing government intervention and the emergence of inward-looking import substitution, followed by painful attempts at adjustment, deregulation, opening up and reform. It argues that policy has had very significant, and after 1973 negative, effects on industrial performance. Instead of looking at policy rhetoric, the article focuses on indicators of actual policy implementation. This leads to different periodizations of policy regimes. The importance of such an approach is underscored by James’s observations in Chapter 5. He shows that at the onset reform was more rhetoric than reality. During the first reform period after 1986, the role of the state in industrial investment initially even continued to increase. Any discussion of economic policy requires adequate monitoring of economic performance by statistical institutions. The National Bureau of Statistics of Tanzania is involved in ongoing attempts to improve and revise national accounts and industrial statistics. In Chapter 3, Szirmai, Prins and Schulte try to make a contribution to these efforts by presenting newly revised estimates of manufacturing levels and trends. The authors find that the level of industrial value added has been seriously underestimated in existing statistics. The pattern of growth followed by decline, however, is even more pronounced than in earlier estimates. The article also presents new indices of real labour productivity, which show dramatic secular declines in productivity since 1973. An international comparison of labour productivity indicates that Tanzania has been falling behind in a period when labour productivity in Asia was catching up with the lead economies. In 1989 value added per person engaged in mediumand large-scale manufacturing had declined to 3.9 per cent of that of the world productivity leader, the USA. In a comparative study of the role of education in three African economies, Pack and Paxson present a critical discussion of the orthodox notion that African growth is skill constrained, and that expansion of education will in itself provide an impetus to growth of productivity. They argue convincingly that unless a complementary inflow of new technology takes place, educational investments will not lead to industrial development. In a non-competitive, inward-looking and stagnant setting, investment in human capital will have disappointing results. They are careful to point out, however, that once the economy starts becoming more dynamic, human capital and increased technological capabilities will also


become more important. In this respect it is important to remember that improved educational levels preceded economic growth in settings as diverse as Scandinavia, South Korea and Japan. In a thought-provoking article, based on an analysis of the experiences of the Asian newly industrializing economies, Cooper attacks the preoccupation with technological upgrading. He argues that it is much more important for sub-Saharan Africa to find out how and why Asian NICs transformed themselves into labour-intensive exporters, and why nonNICS got stuck in import substitution which created vicious cycles of inefficiency. Though final answers are not forthcoming, the orientation of government policy was certainly one of the relevant factors in the transformation. A second issue raised in the article is that of technological change in so-called early, labour-intensive, industries. Even in labour- intensive sectors such as textiles or leather, the pace of global technological change is accelerating. This creates special challenges for late late industrializers who now want to enter world markets. Even countries with a potential comparative advantage in labour-intensive production have to make substantial efforts to keep abreast of world technological developments. In the early stages of industrial development, technological innovation is primarily a question of taking over and adapting technologies developed in the advanced economies. However, the process of technology transfer is not effortless. It is itself a type of innovation which requires considerable effort and capabilities. The success with which developing countries take over technologies, depends to an important extent on their technological capabilities: their capacities to select, acquire, adapt and further develop internationally available technologies and to integrate them into the domestic economy. The five articles in Part II reflect different aspects of this complex of problems. In Chapter 5, James presents a political economic analysis of the role of the Tanzanian state in technology selection, and provides interesting examples of the lack of economic rationality in the investment process. The main driving force in technology selection consists of bureaucratic managers maximizing project size and budgets in negotiations with international donors. James uses this perspective to clarify two paradoxes in technology selection: different technology choices in similar industries, and a preference for capital-intensive projects in a country with a labour surplus. In Chapter 6, Semboja and Kweka discuss innovative strategies of firms. They note that, with the exception of a few firms with international partners, innovation and technological changes have been very limited in scope. They go on to discuss the factors influencing innovativeness and the obstacles to innovative behaviour. These obstacles include a shortage of technical skills in the labour force, a low volume and ineffective coordination of firm-level R&D, insufficient orientation to export markets – the

Introduction 5

literature emphasizes that exporters are forced to innovate and to meet international quality standards – , insufficient financial resources, a weak institutional framework for science and technology support, and weak links between innovation research centres and users. The issue of technological capabilities is also central to the article by van Egmond in Chapter 8. She defines technological capabilities in terms of four components: stock of technology, stock of human resources, stock of natural resources, and technology infrastructure. She argues that technological capabilities are very important for successful technology transfer. Operationalizing the four components, she goes on to map the capabilities of the dwelling-construction industry, identifying weaknesses and bottlenecks. The literature on national systems of innovation emphasizes the crucial importance of fruitful interaction and strong networks and linkages between enterprises and science and technology institutes. The importance of the science and technology infrastructure is stressed by various authors in this volume. Chapter 7, by Bongenaar and Szirmai, focuses specifically on a case study of the functioning of the Tanzanian Industrial Research and Development Organisation (TIRDO). TIRDO is an R&D organization, which explicitly aims at developing and adapting technology for domestic industry. A detailed analysis of TIRDO projects indicates a considerable degree of technical project success. In terms of transferring technologies to enterprises, however, the record is very disappointing. The organization tends to operate on the assumption that good technologies should sell themselves. Therefore, insufficient attention is paid to networking, marketing and transfer activities. The case study provides a clear illustration of the weakly developed linkages between publicly funded R&D institutes and enterprises. In Chapter 9, Duijsens and Lapperre focus on one important component of technological capabilities: technical education. In a survey study of 28 enterprises, they investigate the supply and demand for technical education in the metalworking sector. At all educational levels, major shortcomings are identified with regard to knowledge of modern metalworking techniques, awareness of safety-related aspects, awareness of preventive maintenance, striving for excellence, and mastery of English. In an analysis of the institutional sources of shortcomings, interestingly enough, the use of English as the language of instruction is identified as one of the major problems, as both teachers and pupils have insufficient mastery of the language. Part III opens with Chapter 10, by Lemmens and associates, on the assessment of environmental impacts of productive and household activities. This article develops an ingenious methodology to assess environmental impacts in data-poor situations, which is of special relevance for developing economies such as Tanzania. It uses standard pollution factors for productive outputs and human activities.


The method is initially applied to Lake Victoria where, as it turns out, industry is far less important as a source of different types of pollution than households, agriculture and wet deposition. Though industry may not yet be important at an aggregate level, point pollution turns out to be more serious, especially in major urban centres, such as Dar es Salaam, and in small-scale gold mining. A link between industrial productivity and environmental considerations is provided by energy efficiency. An increase in energy efficiency increases firm productivity and profitability, while at the same time reducing the burden on the environment associated with increasing industrial production. Though some might argue that in a poor country such as Tanzania economic development and industrialization take precedence over environmental considerations, van der Vleuten, Lemmens and associates argue persuasively in Chapter 11 that there are ample opportunities for leapfrogging to cleaner production. Their conclusions are based on meticulous studies in beer brewing and cement production. Tanzania is now trying to move towards export orientation and privatization. As various authors indicate, Tanzania is swallowing the IMF and World Bank medicine of deregulation, export orientation and privatization more wholeheartedly than many other developing countries. The chapters in Part IV focus on the reform process. On the one hand, they reflect on the lessons to be learned from past experiences; on the other they analyse the impact of reforms and make suggestions about future paths to be followed. In the opening chapter in Part IV, Lapperre puts Tanzanian industrialization in a historical comparative perspective, and asks what lessons can be derived for Tanzania from the industrial history of Western Europe. He makes systematic comparisons between the historical prerequisites for industrialization in the past, and present conditions in Tanzania. Such historical comparisons are valuable since every ‘advanced’ country was once a ‘developing’ one. However, one has to be careful, because countries do not necessarily follow identical paths of development, and the international context facing late developers may be very different from those of the first movers. This point is well illustrated in the chapter by Cooper, who points to accelerated global technological change in labour-intensive sectors such as textiles. One of the important prerequisites cited by Lapperre is that industrialization requires prior increases in agricultural productivity. Other important preconditions for successful industrialization have to do with the crucial importance of political stability and efficient, predictable and non-corrupt government. Lapperre also presents an analysis of attitudes and social institutions, such as for example extended family ties, which may be unfavourable for industrialization. With regard to education, the message is mixed. Lapperre indicates that increased formal education has not always

Introduction 7

been essential for early industrial development in Europe. Like Pack and Paxson, he doubts that simply increasing formal educational levels will provide solutions to Tanzania’s economic problems. On the other hand, many of the articles in this volume (van Egmond, Duijsens and Lapperre, Semboja and Kweka) emphasize the importance of improving education and in particular technical education. An article on the informal urban manufacturing sector by Gaillard and Beernink has a dual purpose. On the one hand, the authors criticize the neglect of the informal sector in past policy thinking and practice. On the other hand, they present a state of the art of informal-sector studies, both in Tanzania and worldwide. They are very critical of the scientific quality of informal-sector studies, concluding that this branch of research is still in a pre-scientific stage. Therefore, it is of less use for the formulation of well-specified informal-sector policies than is to be desired. Nevertheless, the authors do conclude on the basis of the available information that the relevance of this sector for economic development is very substantial in terms of employment, output, income generation and learning potential. In 1991 no less than 56 per cent of the urban active population in Tanzania was working in the informal sector, generating 32 per cent of total urban income. Gaillard and Beernink try to identify the main constraints for the future growth of this sector, which include: limited access to credit, lack of equipment and spare parts, and limited access to more permanent sites. Training is mentioned least frequently by respondents, as the informal sector is largely self-supporting with regard to training. However, the authors conclude that from a long-run perspective, training for upgrading labour in the informal sector should be part of the policy package. The articles by Mbelle, Moshi and Wangwe all focus on the process of reform and structural adjustment. It is important to note that there is consensus concerning the shortcomings of the model of state-led importsubstitution industrialization dominant in the past. None of the authors advocates a return to this model. However, their assessments of ongoing reform policies and their outcomes are mixed. In Chapter 15, Moshi provides two interesting case studies of major firms that have experienced a marked turnaround since their privatization, with increases in sales, productivity, competitiveness and profitability. In an econometric article on the impact of liberalization in chapter 13, Mbelle starts by noting that important macroeconomic indicators, such as manufacturing growth and employment, have started improving since 1994. For 1998 he even records a GDP growth in manufacturing of 8.1 per cent. GDP growth in the total economy has also picked up in recent years (Table 1.9). Like Wangwe, Semboja and Kweka, and Moshi, he also points to increased technological dynamism and ‘offensive strategies’ in some of the manufacturing firms.


On the basis of econometric analyses Mbelle concludes that manufacturing exports responded positively to liberalization after 1993. However, the share of manufacturing in investment tended to decline. Mbelle concludes that macroeconomic policies in themselves are not enough. Sector-specific policies are needed to ameliorate some negative sectoral impacts of macropolicies. Much the same conclusion is drawn by Wangwe in chapter 16. On the basis of a survey of recent literature he concludes that restructuring entails more than macroeconomic reform. He is particularly concerned with the fact that, while liberalization does expose formerly protected firms to competitive pressures, there are still no strong inducement mechanisms for technological change and improving technological capabilities. In view of Cooper’s conclusion that developing countries that want to enter world markets in labour-intensive exports are faced with more technological change than before, this is an important finding. Wangwe goes on to identify the main factors which need to be tackled by policy. They include: improving institutional capacity to provide supportive technological services to industry, improving linkages between economic activities and between establishments and R&D institutions, improving the provision of infrastructure, investing in human resource development and making labour processes more conducive to learning. He identifies three major priorities: policies aimed at technological learning and building competitiveness, priority of agro-based small and micro-enterprises, and enhancing the capability of government to complement the opening-up process of the economy with supportive policies. In the final chapter Maziku summarizes Tanzanian policy intentions contained in the recently drafted Sustainable Industrial Development Policy (SIDP) document for the period up to 2020. The new policy formulation addresses many of the issues discussed in this book. Maziku rightly emphasizes that one of the important new elements in policy making is a consultative mechanism for policy implementation, in which a revived private sector and the government are expected to cooperate and interact. Though this volume stresses that policy implementation is far more important than policy intentions, the SIDP certainly provides an indication of the extent of change in policy thinking. Adam Szirmai and Paul Lapperre Eindhoven, October 1999

Part I Long-run Economic Performance

This page intentionally left blank

1 Public Policy and the Industrial Development of Tanzania, 1961–95 Donné van Engelen, Adam Szirmai and Paul Lapperre*

1

Introduction

Throughout the past decades the economic and industrial performance of less developed countries in sub-Saharan Africa can be captured in three words: boom, crisis and adjustment (Little et al., 1993). The sequence of rapid growth, followed by crisis and an ongoing process of structural adjustment, has received a great deal of attention, both from neo-liberal and structuralist points of view. Neo-liberal doctrine emphasizes the primacy of markets and the negative effects of excessive state intervention in the process of economic development. Structuralists, on the other hand, seek explanations of economic underdevelopment in the combination of negative external influences, unfavourable initial conditions and a variety of structural and institutional constraints that developing countries have had to face (Lensink, 1996; Szirmai, 1997a) Overcoming these structural constraints requires active intervention on the part of governments. For most of the period since 1961, Tanzanian industrial policy has been overwhelmingly structuralist in nature. This article provides an analysis of 34 years of industrial development in Tanzania (1961–95), in order to gain insight into the reasons why manufacturing has so far failed to become an engine of economic growth. The central aim of this article is to analyse the combined effects of initial conditions, external influences and policy responses on the performance of the manufacturing sector in different periods. After a brief discussion of initial conditions in Section 2, Section 3 presents a discussion of industrial policies. Periods of coherent policy are identified in an empirical manner: by monitoring the actual use of instruments of industrial policy. In section 4, the emphasis is on the external influences and external shocks, that have provided challenges to industrial policy makers. Incorporation of these external influences into the analytical framework results in a revised periodization. In Section 5, we examine the impact of industrial policies and external influences on manufacturing performance. 11


2

Initial conditions

During the colonial interlude lasting from 1885 to 1961, Tanzania is subjected to German rule (until 1918) and later to British rule. During most of the colonial period, industrialization receives little attention. There is some domestic manufacturing in the nineteenth century (handicraft production of iron products, cotton cultivation, spinning and weaving), but under German rule (1890–1918) this is largely replaced by imports. Between the two world wars British colonial policy discourages investment in manufacture for the local market (Coulson, 1982b). In the twilight of colonialism after World War II, the previously almost non-existent Tanzanian industrial sector begins to take shape. Initially all industrial efforts are directed at the processing of primary agricultural commodities such as sisal, cotton, coffee and tobacco. These commodities are processed for the export market. In the 1950s production of consumer and intermediate goods for the domestic market gains in importance. But the share of these products (for example, woodwork, paint, soap, steel) in total manufacturing remains small (Silver, 1984). The demand for intermediate and capital goods (for example, iron, vehicles and machinery) is largely satisfied by imports (World Bank, 1987a). To encourage the establishment of industrial enterprises, the colonial government provides fiscal advantages (tax relief on capital expenditures, refunds of custom duties) and some measure of protection (import duties) to investors. Furthermore, the government is willing to participate in investment projects, provide specialist research and advisory services, create industrial sites and improve infrastructural facilities. Since Tanzania lacks a well-developed capitalist entrepreneurial class (Rweyemamu, 1973), the incentive package in this period aims at stimulating private foreign investment. Unfortunately for Tanzania, the other member states of the East African Community, Kenya and Uganda, are providing similar incentive packages. Moreover, Kenya possesses the largest industrial base, the best financial and physical infrastructure, the most concentrated domestic market and the largest settler community. Thus it is a more attractive country for foreign investors than Tanzania. Silver (1984) concludes that the policy incentive package of the 1950s cannot have been very effective in stimulating foreign investment in Tanzanian manufacturing. As a result of the unstimulating policy environment and the continued focus on agricultural commodity processing, Tanzania has not experienced a great deal of industrial development at the eve of independence. In 1961 the manufacturing sector contributes 3.6 per cent to GDP at factor costs (Rweyemamu, 1973). Of the total number of 891 industrial establishments, including cottage establishments (Silver, 1984), 245 are engaged in primary products processing (sisal, tea, tobacco, cotton), 248 in basic food processing and manufacturing (sugar, salt, dairy products, grain milling, bakeries,

Public Policy and Industrial Development 13

beverages), and 225 in consumer goods manufacturing (textiles, footwear, tailoring, woodworking, printing and publishing, rubber products, oil milling, soap). The remaining establishments (173) are engaged in motor vehicles manufacturing, engineering and electricity repair, and metal products manufacturing. There is no doubt that in the early 1960s, Tanzania is still overwhelmingly a producer of primary commodities (Chenery, 1979). In this period, industrial establishments vary considerably in size. In most industries there are large numbers of small-scale enterprises producing for the local (rural) market, alongside a limited number of large-scale enterprises producing either for the urban markets or for export. In 1961, an industrial census year, 62 per cent of a sample of 76 of the large-scale enterprises report to be working below full capacity. Reasons for underutilization of capacity include insufficient demand, weather conditions, inadequate transport and communication facilities, inefficient use of raw materials, breakdowns and a shortage of skilled labour (Silver, 1984). These problems are typical for the initial conditions Tanzania faces when embarking on the road to import substituting industrialization in the decades to come. More generally speaking, initial conditions are rather unfavourable for rapid industrial development in Tanzania (see Chapter 12 by Lapperre, see also Szirmai, 1997a, b; Collier and Gunning, 1999). These unfavourable conditions include a large share of agriculture in employment and GDP, low productivity in the agricultural sector, poor infrastructure, low levels of general education, and lack of technical skills and human capital in the workforce. In the political sphere there is a newly established central nation state, lacking a long history of political centralization, with low levels of fiscal and managerial capabilities in the civil service. Successful industrialization in the so-called NICs is usually associated with a longer period in which industrial experience is acquired, covering some 30–40 years. Tanzania lacks an industrial heritage and a tradition of industrial entrepreneurship. As mentioned above, an indigenous entrepreneurial class is weakly developed in the Tanzanian economy. What entrepreneurship there is, is disproportionately concentrated in the Asian community of Tanzania. Infrastructure in communications, transport and energy supply is underdeveloped, hampering the logistics of newly established industrial activities. Unfavourable initial conditions for Tanzanian industrial development also include a low population density and the small size of the domestic market for industrial goods.

3

Policy periods

Identifying the periods of coherent policy which have affected the Tanzanian manufacturing sector is the next step towards the development of an analytical framework for the examination of Tanzanian industrial development between 1961 and 1995. Such periods of coherent policy can


be established in two ways: by looking at objectives and plans, or by monitoring the use of policy instruments. The first option is to look at the changes in industrial objectives and strategies as laid down in Tanzania’s development plans. A serious shortcoming of this approach is that the objectives and strategies set down in the plans may not have actually been implemented during the planning periods. There are numerous reasons why gaps between policy plans and policy implementation exist in developing countries. For instance, development plans are at times used for obtaining local and international support, leading to the formulation of propaganda documents rather than implementable plans (Todaro, 1997). Furthermore, the unstable economic and political situation of LDCs often forces governments to implement ad hoc policies, rather than carefully spelt out plans (Killick, 1976). Finally, as Gulhati (1990) points out for the African continent, policies mentioned in plans are often subverted in the process of implementation. The pitfall of distinguishing policy periods on the basis of development plan classification can be avoided by using the second option for distinguishing periods of coherent policy: by empirically monitoring the use of policy instruments over time. As there are a large number of policy instruments at a government’s disposal (for general economic classifications see Chenery, 1958; Killick, 1981; Haggblade et al., 1990), it is necessary to cluster these policy instruments in such a manner that they allow for the identification of periods of coherent policy affecting Tanzania’s manufacturing sector. Following Weiss (1988), this is done by taking four general aspects of industrial strategies in LDCs as a point of departure. These general aspects are the industrial trade strategy, the degree of direct regulatory control, the relative roles attributed to the public and the private sector, and the nature of dependence on foreign finance. 1 Policy instruments affecting the manufacturing sector (Donges, 1976; Cody et al., 1980, 1990; Kirkpatrick et al., 1984; Killick 1990; Greenaway and Milner, 1993) can be subsumed under these four general aspects of industrial strategy, resulting in the classification presented in Table 1.1. Each aspect of industrial strategy in Table 1.1 represents a dimension of which the poles represent the use of a different combinations of policy instruments. Industrial trade strategy is scaled from import substitution to export promotion; regulatory control varies from low to high degrees of control in five policy domains; the relative importance attributed to the private and the public sectors is scaled from private to public; and nature of dependence on foreign finance varies from dependence on direct foreign investment to dependence on foreign aid. Using secondary literature and statistical indicators, the implementation of policy instruments over time is monitored, in order to identify shifts along the four dimensions discussed above. These shifts can be both gradual in nature and radical. In this

Table 1.1

Classification of policy instruments according to general aspects of industrial strategies

Aspect of Industrial strategy

Scaling

Policy instruments

I.

a. Import substitution

Tariffs, import quotas, import licences, (real) exchange rate appreciation, export taxes, export licensing, export duties

b. Export promotion

Tax concessions, export credits, foreign exchange retention, export subsidies, export processing zones, (real) exchange rate depreciation

II. Degree of direct regulatory control in the areas of: a. International trade

Low/moderate/high

Import and export licensing, import quotas and prohibitions, foreign exchange allocation, exchange rate controls

b. Monetary sector

Low/moderate/high

Interest rate control, credit allocation, money supply

c. Labour

Low/moderate/high

Minimum wage legislation, immigration and emigration quotas, legislation regarding working conditions and fringe benefits

d. Ownership

Low/moderate/high

Industrial licensing, legislation regarding foreign investment

e. Prices and internal trade

Low/moderate/high

Producer and consumer price controls, confinement

Industrial trade strategy

15

16

Table 1.1

(continued)

Aspect of industrial strategy

Scaling

Policy instruments

III. Relative roles attributed to the private and the public sector

a. Private sector

Privatisation, investment incentives (tax reductions, monopoly privileges, subsidies), openness to DFI, land allocation and tenure

b. Public sector

Public enterprise investment, nationalisation, regulation of joint ventures, preferential allocation of resources, prohibitions of investment

a. direct foreign investment (DFI)

Subsidies and tax incentives for foreign investors, monopoly privileges

b. aid

Requests for aid, restrictions for private foreign investment, requirement for domestic majority shares, constraints on profit remittances and capital repatriation, exclusion from key industries, regulation of TNCs

IV. Inflow of foreign finance: relative roles attributed to

Source: van Engelen (1996). Notes: policy instruments which have been omitted are the provision of infrastructure and vocational training. Nevertheless, these unclassifiable policy instruments are taken into account in Section 5 of the paper.


article, the beginning of a new policy period is defined by the occurrence of radical shifts in policy implementation along at least two of the four main policy dimensions. Applying this approach to Tanzanian industry, one can distinguish five separate periods of coherent policy. From 1961 to 1967 the Tanzanian policy climate is characterized by import substitution through foreign private investment, along with a low level of direct regulatory control. In 1967 a jump to a moderate degree of control is made, along with an abrupt shift from reliance on the foreign private sector to reliance on the public sector as the motor for industrial development. From 1973 onwards a high degree of direct regulatory control is imposed on the manufacturing sector. At the same time, a high level of dependency on foreign aid is attained. In 1984 the first shift in trade strategy is observable, from import substitution towards export promotion. Simultaneously, a shift away from a high level of direct regulatory control can be discerned. The final policy period distinguished in this paper starts in 1990. It is then that the Tanzanian government further decontrols the sector, shifts towards a higher participation in investment of private investors, whether foreign or local, and opts for an export-oriented trade strategy. Reliance on foreign aid for industrial development is considerably reduced during this policy period. Before elaborating further on the policy periods, it is important to point out that the policy instrument approach has yielded results differing substantially from a policy periodization based on development plans. On the basis of development plans, one can only distinguish four policy periods: 1961–69 (three-year development plan, TYDP; first five-year plan, FFYP); 1969–81 (second and third five-year plan SFYP, TFYP; basic industrial strategy BIS); 1981–86 (national economic survival programme, NESP; structural adjustment plan, SAP) and 1986–94 (economic recovery programme, ERP; economic and social action programme, ESAP; rolling plan forward, RPF). Furthermore, during the first four years of the third policy period (NESP, SAP) no significant changes in implemented policies affecting the manufacturing sector occur (Stein, 1991; Bagachwa and Mbelle, 1993). It would have been misleading to use these plans as indications for changes in implemented policy. It should also be noted that the development plans do not always contain policy information specific to the manufacturing sector. For instance, for the last three development plan periods it is not possible to identify the nature of foreign financial inflows (van Engelen, 1996). In the following paragraphs, the five policy periods relevant for the Tanzanian manufacturing sector are discussed in detail. 3.1

Import substitution through foreign investment (1961–67)

The first policy period starts at independence (1961) and comes to an end when the Arusha declaration is signed (1967). During this period policy


instruments are used to encourage foreign investors to engage in first-stage import substituting industrialization. Similar policy instruments had been used in the late colonial period. However, before independence policy efforts aimed at industrialization had been constrained by the agreements and the economic status quo within the East African Community. From 1961 onwards this situation changes when the Tanzanian government starts using policy instruments which offer extremely favourable investment conditions for (foreign) entrepreneurs. The granting of tariff protection is the most striking example of the use of such policy instruments. According to Rweyemamu (1973) firms could obtain tariff protection by bargaining with the government, resulting at times in effective rates of protection amounting to 500 per cent or more. On average, the production of consumer goods received the most protection (see Table 1.5), suggesting that overall protection of the manufacturing sector is in accordance with the strategy of first-stage import substitution. Other policy instruments used to attract foreign investors are guarantees against nationalization (Foreign Investment Act of 1963), land provision for the creation of industrial estates, accelerated depreciation allowances and guarantees for the repatriation of capital (Wangwe and Bagachwa, 1990). A final policy instrument used to stimulate private investment is the granting of monopoly or near monopoly power. According to Coulson (1982b) this is ‘a condition without which the multinationals in particular would not invest’. Besides tariffs, import licences and quotas are also introduced as policy instruments during the 1961–67 period (Ndulu and Semboja, 1994), giving an extra incentive to manufacturers to produce for the local market. This form of direct regulation in the area of international trade is accompanied by legislation regarding working conditions and fringe benefits. Nevertheless, the degree of direct regulatory control is still low during this policy period. The government refrains from using monetary policy instruments or exchange rate controls, and few attempts are made to curtail the freedom of action for foreign private entrepreneurs. 3.2 Nationalization, self-reliance and increased regulatory control (1967–73) The signing of the Arusha declaration in 1967 puts an end to the low level of direct regulatory control and the reliance on foreign private investors. A new policy period commences, in which Tanzania aims to become selfreliant in industrial production. Amongst others, this results in the nationalization of large (foreign owned) industrial enterprises. The emerging public sector is developed by the National Development Corporation (NDC), which has to expand very fast in order to keep pace with the nationalization process (Skarstein and Wangwe, 1986). The state’s control


of the manufacturing sector is facilitated by the introduction of an industrial licensing procedure under the National Industries (Licensing and Registration) Act of 1967 (Musonda, 1992). After the Arusha declaration foreign investors can only participate in joint ventures in which the Tanzanian government is the major partner. According to Barker et al. (1986) considerable numbers of these joint ventures come into existence. Along with the nationalization and licensing procedures, other signs of direct regulatory control can be observed from 1967 onwards. In 1967 a socalled confinement strategy is introduced in which imports and exports are channelled through the State Trading Corporation (STC). Nationalization of the internal wholesale trade by the STC follows in 1971, forcing manufacturers to sell specified goods to and purchase inputs from this parastatal organisation. At the same time a modest price control system is set up under the National Price Control Advisory Board, initially controlling the prices of a very limited number of manufactured products (Maliyamkano and Bagachwa, 1990). Finally, it should be noted that the Tanzanian government controls the exchange rate. In 1967 this has few consequences. However, when Tanzanian inflation rates start rising vis-à- vis those of trading partners in the late sixties, exchange rate rigidity leads to real exchange rate appreciation (Lipumba, 1991). A real exchange rate index is presented in Table 1.2. The real exchange rate index is set at 100 in 1966, a year for which it is assumed that the nominal exchange rate is not significantly overvalued. The index is defined in such a manner that decreases indicate real exchange rate appreciation, whilst increases indicate depreciation. As the data in the table show, the real exchange rate appreciates considerably between 1967 and 1971. It recovers slightly in 1972 and 1973, but the nominal exchange rate remains overvalued. This can also be deduced from the ratio of the parallel and nominal exchange rate. Whilst the nominal exchange rate does not change between 1967 and 1973, the parallel market rate rises steadily, roughly doubling the nominal rate between 1971 and 1973. This implies exchange rate shortages, caused by the overvaluation of the nominal exchange rate. Together with continued protection through tariff barriers, import quota and import licensing, the real appreciation of the exchange rate is an incentive for import substituting industrialization. In this respect no change in observable industrial strategy occurs. The major differences between this policy period and the previous policy period are the increased degree of direct regulatory control, the change in relative roles attributed to the private and the public sector, and the decreased reliance on private foreign investment to fuel industrial development. When discussing the following policy period it will become clear that the inflow of direct private foreign investment is substituted by inflows of foreign aid.

20 The Industrial Experience of Tanzania Table 1.2 Year 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

Exchange rates: nominal, parallel, real, 1967–94

Nominal exchange rate (TSh/US$) 7.14 7.14 7.14 7.14 7.14 7.14 6.90 7.14 8.26 8.32 7.96 7.41 8.22 7.96 8.18 9.52 12.46 18.40 16.50 51.72 83.72 125.00 192.30 196.60 233.90 335.00 479.87 523.45

Parallel market rate (TSh/US$)a

Ratio parallel/ Real exchange rate nominal (%) (1966 = 100)b

8.68 8.25 9.10 10.45 15.00 15.40 13.45 14.00 25.00 20.40 15.15 11.75 13.50 23.80 26.00 30.90 39.80 80.00 125.00 160.00 180.00 210.00 245.08 309.08 384.58 422.92 506.23 560.07

1.22 1.16 1.27 1.46 2.10 2.16 1.95 1.96 3.03 2.45 1.90 1.59 1.64 2.90 3.12 3.24 3.19 4.42 7.58 3.09 2.15 1.68 1.27 1.57 1.64 1.26 1.05 1.07

90.24 76.77 68.23 69.58 70.02 71.14 74.68 78.21 70.86 74.78 72.27 68.01 70.33 59.60 51.54 47.57 46.14 42.16 41.88 60.01 87.97 103.82 131.00 148.70 n.a. n.a. n.a. n.a.

Sources: Lofchie (1988); Maliyamkono and Bagachwa (1990); Musonda (1992); Tanzania Economic Trends (1994); Bank of Tanzania (1995); Economic Research Bureau of the University of Dar es Salaam, calculations made by A.V.Y. Mbelle and J.J. Semboja (1996). Notes: a Estimates of parallel rates have been made by van Engelen (1996) using sources mentioned above. b The real exchange rate index is defined as the Tanzanian inflation rate relative to that of the USA, times 100. n.a. = not available.

3.3 Dependence on foreign aid and a high degree of direct regulatory control (1973–84) A shift towards increased dependence upon foreign aid is initiated in the early 1970s, when Tanzania is able to obtain funding at favourable terms from countries such as China, Denmark, the United States, Canada, the Federal Republic of Germany and the Netherlands. During these years approximately 4 per cent of capital expenditures in manufacturing is financed by foreign aid flows. From 1975 onwards this share increases to


one third of total capital expenditures. When the volume of aid inflows starts declining in the early 1980s, manufacturing can still count on a substantial share of foreign aid. Most of the aid provided between 1973 and 1984 is in the form of donor-tied capital intensive projects (World Bank, 1987a). There is an increased reliance on foreign know-how. Apart from the shift towards a higher degree of dependency on foreign aid, this policy period is characterized by the emergence of a high degree of direct regulatory control. From 1973 onwards a new import licensing system is introduced. Along with a system for the administrative allocation of foreign exchange (Finance Act of 1973), this system allows for full control of (legally) imported goods (Musonda, 1992). The allocation of the import licences to manufacturers is considered on a firm-by-firm basis, and is driven by ‘the objective of supporting fiscal revenue earners … and keeping most existing enterprises alive, on the one hand, and by particular pressures, representations and ad-hoc decisions on the other’ (World Bank, 1987a). The resulting allocation pattern is presented in Table 1.3, which shows that from 1979 onwards the manufacturing sector can on average count on being allocated less than one fifth of the requested amounts of foreign exchange for the purchase of raw materials, machinery and spares from abroad. In 1981 no foreign exchange at all is made available for the importation of machinery and spares for manufacturing purposes. The same is true for raw material imports for manufacturing during the first six months of 1984. Another way in which the government augments the degree of direct regulatory control of the manufacturing sector is by introducing a full-scale Table 1.3 Allocation of foreign exchange, 1977–1984 (as percentage of amounts requested) Foreign exchange requested for: Industrial raw materials Raw materials (manufacturing) Petroleum Machinery Spares Machinery & spares (manufacturing) Transport % of total requests allocated

1977 1978 1979 1980 1981 1982 1983 1984a 43 60

52 51

24 22

14 41

17 14

16 11

n.a. 15

n.a. 0

100 26 44 29

100 23 43 29

93 1 14 12

71 0 17 16

67 7 20 0

89 3 1 5

n.a. n.a. n.a. 11

n.a. n.a. n.a. 16

37

35

13

10

8

2

n.a.

n.a.

50

52

38

32

31

33

n.a.

n.a.

Sources: manufacturing data from Mbelle (1988); all other data from World Bank (1984). Notes: a The 1984 data concern the January–June period. n.a. = not available.


price control system (Price Control Act of 1973). Price controls are introduced with the dual purpose of limiting the monopoly pricing power of domestic producers, and at the same time ensuring that these producers are guaranteed satisfactory financial profitability (Mongi, 1980). For the domestic producers the financial profitability is taken care of by the use of a costplus pricing method, allowing for a 30 per cent pre-tax rate of return on assets. The price of imported goods is set using a fixed percentage mark-up (Maliyamkono and Bagachwa, 1990). Trends in the numbers of price-controlled products can be derived from Table 1.4. As a result of a government directive, the number of price-controlled items is reduced from 1980 onwards (United Republic of Tanzania, 1983). Nevertheless, until 1985 almost all consumer goods, a significant number of agricultural intermediate inputs and various construction materials, such as iron sheets, rolled steel and cement, remain price controlled. Apart from import licensing and price control, the government uses confinement as yet another policy instrument to ensure a high degree of direct regulatory control over the manufacturing sector. As described in the discussion of the previous policy period, trade is confined to the State Trading Corporation STC. In order to perform more effectively the STC is decentralized and reorganized in 1973. The STC’s tasks are taken over by 6 parastatal importing companies and 18 regional trading companies (Musonda, 1992). Up to 1984 wholesale (and a part of retail) trade is confined to these trading companies. According to the World Bank (1987a) this ‘has been associated with poor service, lack of payments and high marketing costs’. The three policy instruments mentioned above not only allow for increased control of the manufacturing sector. They also allow the government to provide protection to the locally based industries. Import licensing restricts competitive imports from entering the country, whilst the price controls and the confinement system raise the prices of imported products Table 1.4 ucts)

Price controlled items, 1973–91 (product groups and separate prod-

Controls

1973

1979

1982

Total number of items locally produced imported Total number of products locally produced imported

n.a. 329 72 n.a. 154 54 n.a. 175 18 1 000 1 031 333 n.a. 456 241 n.a. 575 92

1983

1986

1987

1988

1991

56 50 6 225 216 9

21 21 0 107 107 0

22 22 0 83 83 0

12 n.a. n.a. n.a. n.a. n.a.

2 n.a. n.a. n.a. n.a. n.a.

Sources: 1973 data from World Bank (1984); 1979–1988 data from Kiondo (1991); 1991 data from World Bank (1991). Note: Items are groups of separate products; n.a. = not available.


relative to domestic products, and limit their distribution opportunities. Along with an increasingly overvalued exchange rate (see Table 1.2), which discourages exports and lowers the price of imported inputs, this results in the continuation of the import substitution trade strategy which prevailed during the first two policy periods. By 1984, the overvalued exchange rate and the price controls have resulted in extremely high effective rates of protection (ERPs). In Table 1.5 ERP calculations are presented for 1966, 1984 and 1986. In 1984, the effective rate of protection for the total manufacturing sector is no less than 470 per cent. A comparison of selected industries in 1966 and 1984, indicates the dramatic increases in levels of protection (from 134 per cent to 526 per cent for the total sample of industries). It also shows that a shift in protection has occurred. In 1966 protection is clearly in favour of consumer good manufacturers. By 1984 producers of intermediate and capital goods can count on degrees of protection which are at least as high as the ERPs for consumer good manufacturers. From 1981 onwards the first attempts are made to counterbalance the strong incentives towards import substitution. In that year an export rebate scheme (ERS) is introduced, which is to serve as an export subsidy for the producers of manufactured and horticultural commodities. The scheme compensates exporters for losses incurred by selling at world market prices, and also compensates them for duty and sales taxes on imported inputs. Along with the ERS a presidential award is introduced for the best performing exporter. The winner of this award is given preferential treatment with regard to the acquisition of scarce resources. The most important export incentive introduced during this policy period is the general retention scheme (GRS) of 1983. This scheme allows exporters to retain between 10 and 100 per cent of the foreign exchange they earn, for the purpose of importing inputs (Board of External Trade, 1989). On average, however, the export incentives are not very effective, since they lack the desired degree of comprehensiveness to stimulate the export sector adequately (World Bank, 1987a). In short, there is no reason to conclude that a shift away from import substitution is initiated in 1981. Neither is there any indication that substantial shifts occur in the relative roles attributed to the private and the public sector. The emphasis remains on the development of the public sector, at the expense of the private sector. In the words of Henley and Assaf (1993), ‘until … 1986 the private sector is crowded out of official thinking, access to loan capital and foreign exchange allocations’. The systems of credit allocation and foreign exchange allocation are interdependent. Both serve to ensure the survival of existing large-scale publicly owned firms. Apart from credit priority, the public sector can also count on lower interest rates than the private sector. High inflation rates and a policy of interest rate rigidity mean that real interest rates for the public and the private sector are negative between

24

Table 1.5

Effective rates of protection in manufacturing, 1966, 1984 and 1986 (%)

Industry

1966

1984

Canned fruits and vegetables Soft drinks Beer Tobacco Textiles Sisal and jute bags Tanning and leather Footwear Pharmaceutical products Soap

184 –24 187 528 269 1 130 123 0 151

335 5 1 300 317 240 inf inf inf 2 952 5 258

Tyres and tubes Glass products Paper products Cement Metal products

270 31 26 12 95

59 424 6 682 101 inf

Sample average

134

526

Branch

1984

1986

Food products Beverages and tobacco

53 172

65 84

Textiles

811

55

inf

41

1 762

45

inf 309

2 28

3 780 1 347 470

28 25 n.a.

Tanneries and leather Plastics and pharmaceuticals Chemicals and fertilizers Rubber, glass, wood, paper and cement

Iron, steel and metal products Machinery and transport equipment Sectoral average

Sources: 1966 data from Rweyemamu (1973); 1984 data from World Bank (1987a); 1986 data from Ndulu et al. (1987). Notes: inf = infinite; n.a. = not available.


1973 and 1984. Nominal interest rate rigidity, preferential credit allocation and interest rate discrimination once again emphasize the high degree of direct regulatory control which prevails throughout this policy period. 3.4

Decontrol and the shift away from import substitution (1984–90)

The introduction of the own funds import scheme in June 1984 marks a radical change in government policy. It marks the end of 11 years of full scale direct regulatory control of the manufacturing sector. The scheme encourages Tanzanian citizens to obtain foreign exchange on the parallel foreign exchange market, from friends, family, accounts abroad and from foreign investors, so that imports can be financed without making use of official foreign exchange (Bank of Tanzania, 1988). In Table 1.6 the various sources of foreign exchange are presented. The table shows that within two years the own funds import scheme has become the largest foreign exchange window. Administration of an own funds import licence is not subject to bureaucratic procedures, and the imports under the scheme are not subjected to any price controls or confinement rules. Due to the magnitude of the own funds window, these conditions imply that considerable trade liberalization takes place. This trade liberalization is only partial, since not all goods are allowed to be imported. Nevertheless, the decontrol of trade marks a jump in policy. Partial liberalization of imports funded with official foreign exchange follows in February 1988, when the open general licence (OGL) is introduced. The foreign exchange available through this window is provided by the World Bank and other donors, and serves to finance high priority imports. It is allocated on a non-discretionary and automatic basis, Table 1.6 Magnitudes of foreign exchange windows, 1984–92 (% of total foreign exchange available) Year

Free resourcesa

Own fundsb

Export retention

Loans/ grants

Import support

OGLc

Barter

1984 1986 1987 1988 1989 1990 1991 1992

35.3 26.1 32.5 18.6 29.1 22.3 13.7 *

19.4 37.2 34.6 35.0 26.3 26.1 28.7 27.3

– 0.1 0.2 0.4 1.1 n.a. 4.6 n.a.

33.4 27.6 23.6 33.5 23.0 24.3 20.8 22.2

6.7 8.2 8.2 9.2 11.1 9.1 4.8 2.6

– – – 2.7 9.1 18.0 27.0 47.4

5.2 0.8 0.9 0.6 0.3 0.2 0.4 0.5

Source: Lipumba and Mbelle (1993). Notes: a The 1992 figure is included under the open general license (OGL). b The 1984 own funds figure covers the period July–December. c The 1988 figure covers the months February–December. The 1992 figure includes free resources; n.a. = not available.


provided that an importer adheres to the conditions of utilization (legal business premises, minimum limit, red-list of forbidden imports is observed, 100 per cent cash cover up front). Major advantages of the OGL are less bureaucratic procedures and guaranteed utilization once the licence has been approved (de Valk, 1992). From Table 1.6 it can be seen that the use of the OGL rapidly increases. However, during this policy period, which ends in 1990, the own funds import scheme remains the most used (partially) decontrolled foreign exchange window. Decontrol does not remain limited to foreign exchange allocation. As Table 1.4 shows, between 1983 and 1986 the number of domestic products which are price controlled is halved, and the price control of imported products is abolished. By 1991 only two items (fertilizer and petroleum products) are still price controlled, compared to 56 items in 1983. These two items are also the only two items which remain confined after 1991. Thus, the trend towards decontrol can also be observed with regard to the system of confinement. Between 1984 and 1989 a large number of exemptions from the confinement system are granted. In 1989 the system of confinement is almost entirely abolished. The level of regulatory control is further reduced by the government’s decision to abandon the policy of nominal exchange rate rigidity. Table 1.2 shows that from 1986 onwards the nominal exchange rate is continuously devalued, resulting in real exchange rate depreciation. In 1985 the real index has reached its lowest point, 41.9, after which it steadily increases. In 1990 the real exchange rate index equals 148.7. This implies that price competitiveness of Tanzanian manufactured exports has improved, relative to the base year 1966. The continuous devaluation of the exchange rate provides increasing incentives for Tanzanian manufacturers to produce for the export market. Further incentives for export production are provided by decreases in effective rates of protection. Devaluation of the exchange rate and price decontrol are among the reasons why production for the domestic market becomes less attractive. According to the World Bank (1991) the own funds import scheme is also of importance in this respect: ‘the own funds import policy exposes the industrial sector virtually overnight to a trade regime in which levels of protection have fallen dramatically. Average levels of effective protection for industry decline from about 500% in early 1984 to about 150 per cent in 1985, and some firms become effectively disprotected’. The decline in protection through price controls, import licensing, confinement and cheap imported inputs because of overvalued exchange rates means tariffs once again become the main instrument by which manufacturing is still protected. Table 1.5 shows what the effective rates of protection would be in 1986, if tariffs were the only means of protection. As can be deduced from the table, (hypothetical) ERPs have dropped


considerably since 1984, ranging from 2 to 84 per cent at the branch level in 1986. Furthermore, a shift in relative protection has taken place in favour of the manufacturers of consumer goods. These conclusions should be drawn with some caution, though. In 1986 the exchange rate is still overvalued, and price controls, import licensing and confinement are still used as policy instruments. It is more likely that the 1986 ERPs in Table 1.5 are good approximations for ERPs in 1988. By then the 1973–84 system of protection is almost fully dismantled. The remaining protection from this system is cancelled out by tariff reductions (Musonda, 1992), tariff evasions and exemptions (see also the 1990–95 policy period on this issue). A final incentive for production for export is given by the introduction of several export promotion schemes. In 1986 a new retention scheme replaces the general retention scheme described in the previous section. Under the new scheme, producers of non-traditional exports products are granted a 50 per cent retention rate. In 1985 the seed capital revolving scheme (SCRS) is introduced. Under this scheme a producer is provided with foreign exchange when starting production for the export market. Between 1985 and 1989 the number of manufacturers which benefits from the scheme rises from 18 to 51 (Ndulu and Semboja, 1996). Yet another incentive for export-oriented production is provided by commodity exchange programmes (CEPs), which allow manufacturers of export products to exchange their goods for raw materials and spare parts from abroad. Finally, the export duty drawback scheme (DDS) is established for manufacturers of exports, compensating them for the import duties they have to pay when acquiring foreign inputs (Mbatia, 1993). Apart from the decrease in direct regulatory control and the shift in the industrial trade strategy, no other significant changes occur in the other dimensions of policy during this policy period. With regard to foreign financial inflows, there is actually an increase in foreign aid flows to Tanzania and to Tanzanian industry. From 1987 to 1989 the share of official development assistance (ODA) in GDP rises from 25.5 to no less than 32.3 per cent (Szirmai, 1997a, table 12.6). UNDP data show that the share of total industry in external assistance rises from 7.8 per cent in 1986 to 24.8 per cent in 1989 (Table 1.7). From table 1.7 it becomes evident that external assistance is increasingly provided in the form of financial transfers, rather than in the form of technical assistance (transfer of skills and technology). This change in emphasis should be understood in the context of attempts to ameliorate the financial position of the often heavily indebted public enterprises (Agrawal et al., 1993). Apart from supporting public enterprises through aid inflows, the predominance of the public sector is also enhanced by continued lending (at negative real interest rates) by the government owned banking sector: ‘the financial sector has continued to lend to ailing firms, particularly

28 The Industrial Experience of Tanzania Table 1.7 External assistance to Tanzania and the Tanzanian manufacturing sector 1986–94 (US$ m) Year

1986 1987 1988 1989 1990 1991 1992 1993 1994

Total external assistance (1)

Assistance to industrya (2)

Ratio (%)

670.1 814.9 905.5 905.0 956.2 1 059.9 1 112.7 905.4 895.0

52.6 81.8 112.1 224.6 163.5 171.7 109.0 48.9 10.7

7.8 10.0 12.4 24.8 17.1 16.2 9.8 5.4 1.2

(2/1)

Technical assistance to industryb (3) 25.9 25.0 24.8 26.0 14.8 29.8 32.2 8.1 3.2

Ratio (%) (3/2) 49.3 30.6 22.0 11.6 9.0 17.4 29.5 16.5 30.0

Source: United Nations Development Programme (various issues). Notes: a For the years 1989–94 industry includes manufacturing, construction and mining; for the years 1986–1988 industry also includes the energy sector. b Technical assistance is defined as activities undertaken to promote economic and social development and well-being by enhancing human and institutional capacities through the transfer, adoption, mobilization and utilization of skills and technology.

parastatals, many of which are in serious arrears. The public manufacturing sector … draws about 70% of the total loans to the industrial sector’ (World Bank, 1991) – Furthermore, public enterprises are also favoured regarding exemptions from duty and sales taxes, emphasizing once again the importance of the development of the public sector in the minds of policy makers. Incentives for the stimulation of private entrepreneurship cannot be discerned between 1984 and 1990. In sum, in spite of marked changes in the trade regime and regulation in this period, the dependence on foreign finance increases and the state sector remains predominant. 3.5 Continued liberalization, export promotion, private (foreign) entrepreneurship and decreased inflows of foreign aid (1990–95) During the final policy period the private sector is rediscovered by policy makers. In 1990 the government passes the National Investment (Promotion and Protection) Act (NIA), an act which acknowledges the importance of the private sector. By providing tax holidays, customs duty exemptions, foreign exchange retention possibilities, constitutional safeguards against expropriation and guarantees against nationalization without compensation, the government attempts to encourage foreign and local private investment in industry. Additionally an institutional investment framework is established: the Investment Promotion Centre (Tanzania Economic Trends, 1990). Priority areas for investment are mostly


consumer goods industries, steel and metal engineering, cement and ceramics, bottles and glassware, paints and automotive engineering industries. Iron and steel production, machine tool manufacturing and chemical and pesticides production are designated to remain controlled by public enterprises (United Republic of Tanzania, 1990b). In 1992 and 1994 the NIA is amended to allow for slight expansions of the incentive package (United Republic of Tanzania, 1992a, b, 1994). Along with the stimulation of the private sector, the government sets out to reform the parastatal sector. The legal basis is provided by the Public Corporations Act of 1992, which is amended in the same year to allow for the establishment of the Presidential Parastatal Sector Reform Commission. This commission is empowered to decide which public enterprises should be retained, privatized or wound up. Up to 1 January 1995, 40 of the 170 manufacturing parastatals are divested, of which six are closed down (United Republic of Tanzania, 1995a). Furthermore, as a consequence of the liberalisation of the financial sector in 1991 (Bank of Tanzania, 1992), the parastatal sector can no longer count on preferential treatment regarding credit allocation. Since the banking system is forced to operate on commercial principles, negative real interest rates also belong to the past, again implying that from 1990 onwards the public manufacturing sector faces a changing policy environment. Apart from the liberalisation of the monetary sector and liberalisation of private investment, there is also further liberalisation of the trade regime. As Table 1.4 shows, price control is no longer operative as a means of direct regulatory control from 1991 onwards. Table 1.2 indicates that the nominal exchange rate is depreciated to such an extent that by 1994 nominal and parallel market rates no longer differ significantly. This is caused by the liberalisation of the foreign exchange market under the Foreign Exchange Act of 1992. This act enables the setting up of ‘bureaux de change’, which are allowed to sell and buy foreign exchange at freely determined market rates. As a result of the liberalisation of trade, Tanzanian manufacturers are to rely on the tax system for protection against foreign competition. However, as a document prepared by the Textile Manufacturing Association of Tanzania (TEXTMAT, 1994a) shows, the tax system performs poorly in this respect. Import duty collection rates are far below official tariff rates, implying large scale exemptions and evasions of taxes on imported goods (TEXTMAT, 1994a). The report concludes that importers manage to evade both import and sales taxes on a large scale. A confidential 1995 study by the World Bank and the Tanzanian government would, if published, shed more light on the magnitude of exemptions and evasions. The fact alone that such a study has been carried out shows that the threat to Tanzanian firms manufacturing for the domestic market from untaxed imports is being taken seriously.


From 1990 to 1995 further decreases in protection due to tax exemptions and evasions do not stimulate production for the local market. However, production for the export market is stimulated, resulting in a situation in which export promotion has become the dominant industrial trade strategy. Production for the export market is stimulated during this policy period by further devaluation of the nominal exchange rate and by several export promotion schemes (see National Investment Act and previous policy period). Additionally the Bank of Tanzania introduces the export credit guarantee scheme (ECGS) in 1990. This scheme is designed to encourage pre- and post-shipment credit provision to exporters of non-traditional exports. The final criterion distinguishing this policy period from the previous one is the decreasing flow of foreign aid to the manufacturing sector. Total aid flows to Tanzania start declining, after peaking in 1992. The share of industry in total external assistance decreases rapidly from 1990 onwards (see Table 1.7). In 1990 this share amounts to 17.1 per cent. It drops to 9.8 per cent in 1992, and declines further to 1.2 per cent or US$10.7 m in 1994. These data show that foreign aid no longer plays a significant role in industrial policy. In terms of reliance on foreign financial inflows, the emphasis is once again on private foreign investment. 3.6

Overview of the policy periods (1961–95)

The characteristics of the five policy periods which have been distinguished are summarised in Table 1.8.

4

Synthesis of external influences and policy periods

An overview of the external influences affecting the performance of the Tanzanian manufacturing sector is the final step towards the completion of the analytical framework used in this paper. Two types of external influences are distinguished: changes in the macroeconomic environment in Tanzania and influences external to the Tanzanian economy. Changes in the external environment can take the form of shocks or of more gradual changes. Both kinds of changes in the external environment provide challenges to which industrial policy may or may not respond. If there are changes in external influences which are not accompanied by policy responses, the periodization presented in Table 1.8 needs to be refined. This is the case for the period 1980–84, when the policy regime shows no change, while the macroeconomic situation deteriorates dramatically. When changes in the external environment of manufacturing are followed at short notice by changes in policy implementation, no changes in the periodization are required. An example of the latter situation is the Arusha declaration of 1967, which was followed by nationalization of large

Table 1.8

Overview of policy periods

Aspect of industrial strategy Import substitution (IS)/export promotion (EP) Degree of direct control (low/moderate/high) Private ownership (priv.)/public ownership (publ.) Relative inflows of foreign investment (DFI) and aid

1961–67

1967–73

1973–84

1984–90

1990–95

IS low priv. DFI

IS moderate priv. → publ. DFI → aid

IS high publ. aid

IS → EP moderate publ. aid

EP low publ. → priv. aid → DFI

31


scale enterprises and other measures signalling a turn towards a centrally planned economy. Table 1.9 presents macro-indicators for the total Tanzanian economy. This table indicates that up to 1973 the macroeconomic situation in Tanzania is relatively stable. According to Ndulu (1993) this is facilitated by the ‘relatively benign’ external environment the country faces. The negotiations regarding industrial investment between Tanzania and the other partner states of the East African Community are an example of this relatively benign environment. Thus, in 1964 it is agreed that Tanzania has the sole right to the manufacturing of radios, motor tyres and tubes, and aluminium foil, circles and plain sheets for the East African market (Gulhati and Sekhar, 1982). Furthermore, during this period the world economy and world trade are growing at historically unprecedented rates. The first worldwide oil crisis takes place in 1973. This causes the Tanzanian current account deficit as a percentage of GDP to rise from 6.5 per cent in 1973 to 14.3 per cent in 1974. Simultaneously, inflation rates rise from 7.6 per cent in 1972 to 27.0 per cent in 1975. Thanks to policy measures and the coffee boom of 1977, Tanzania initially manages to stabilize the external and internal imbalances (Green et al., 1980). Inflation decreases to 6.9 per cent in 1976, and the current account deficit shrinks to 9.5 per cent of GDP in 1975 and 1.3 per cent of GDP in 1976. However, in 1978 the current account deficit increases substantially again to 12.7 per cent of GDP. This time the external imbalances are less manageable. Reasons for the second balance of payments crisis include the dissolution of the East African Community in 1977, the second oil crisis of 1979, the war with Uganda in 1979, severe droughts and deteriorating terms of trade. Inflation rates are on the rise again from 1979 onwards. In 1980 a 30.3 per cent inflation rate is recorded, setting the standard for inflation over the next fifteen years. Most important, it should be noted that macroeconomic growth comes to a halt in 1981. In this year a negative real growth rate of –0.5 per cent is experienced, followed by close to zero growth in 1982 and another year of negative real growth (–2.4 per cent) in 1983. Negative real growth rates, high inflation rates and current account deficits clearly show that the Tanzanian economy is in crisis from 1980 onwards. The crisis years are highlighted in Table 1.9. The crisis persists up to 1984, the year in which the own funds import scheme is introduced (see Section 3.4). From then onwards positive real growth rates of GDP are recorded. Nevertheless, inflation rates remain high and current account deficits start rising again in 1986. Between 1986 and 1993, deficits of 10 to 12 per cent of GDP are not uncommon. Obviously, external and internal imbalances persist after the crisis. Within this context it is important to point out the worsening of Tanzania’s debt position. Total debt rises from 74 per cent of GDP in 1985 to 241 per cent of GDP in 1993. The relatively low levels of indebtedness during the crisis years

Indicators of changes in the external environment of the Tanzanian manufacturing sector, 1970–94

Year

GDP growth rate (%)a

Inflation (%)b

Curr. acc./GDP (%)c

Terms of Trade (1987 = 100)c

Debt/GDP (%)c

GDP growth 9 African Economiesd

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

4.0 3.5 6.3 2.8 1.3 5.1 6.1 0.4 2.1 2.4 3.0 –0.5 0.6 –2.4 3.4 4.6 6.6 5.9 4.8 2.2 7.1 2.8 1.6 –0.6 1.2 3.2 4.6 3.2

3.1 5.0 7.6 10.6 18.4 27.0 6.9 11.4 6.6 12.9 30.3 25.7 28.9 27.1 36.1 33.3 32.4 30.0 31.2 30.4 35.9 28.8 21.9 25.2 34.1

–3.1 –8.0 –4.6 –6.5 –14.3 –9.5 –1.3 –2.3 –12.7 –8.8 –11.4 –7.7 –8.7 –5.5 –7.4 –6.6 –8.1 –16.1 –11.6 –10.5 –12.5 –11.9 –12.2 –12.8 n.a.

n.a. n.a. n.a. 145.7 173.6 142.0 151.7 182.0 151.9 138.7 142.0 129.0 127.1 128.0 131.0 126.1 140.5 100.0 107.2 103.2 92.6 93.8 84.7 n.a. n.a.

n.a. n.a. n.a. 34.2 43.1 48.3 56.2 60.5 62.2 64.3 65.0 58.3 55.4 64.1 77.2 74.0 123.6 211.4 188.0 171.1 202.1 188.5 210.9 240.9 n.a.

9.1 5.3 2.5 5.0 6.5 1.4 5.8 4.3 1.4 4.6 5.1 2.0 2.1 –1.3 3.7 4.1 3.9 1.9 5.3 3.2 2.2 1.4 –0.6 n.a n.a

Notes: a Source: Bureau of Statistics (1995a, 1995b, 1995e, 1999). b Source: Bank of Tanzania (various issues). c Source: World Bank (various issues). d Source: Maddison (1995). n.a. = not available.

33

Table 1.9


(1980–84) reflect the drying up of external financial flows due to disagreements with the IMF concerning economic reforms. On the basis of the preceding analysis of the changes in the external environment of the manufacturing sector one may conclude that industrial entrepreneurs operate in a relatively stable macroeconomic climate prior to the early 1970s. During the mid- and late 1970s the first signs of external and internal imbalance can be distinguished, and from 1980 to 1984 Tanzania suffers a severe economic crisis. This crisis is characterized by negative real growth rates of GDP and external and internal imbalances. Growth rates pick up from the mid-1980s onwards, but external and internal imbalances persist well into the 1990s. Calling to mind the periods distinguished in Table 1.8, it becomes apparent that an additional period of analysis is to be introduced if external influences on the sector are to be taken into account. This is the crisis period 1980–84. Although no changes in policy implementation occur during these years, the macroeconomic situation changes to such an extent that it is useful to make this distinction. Thus, the analytical framework drawn up to explain Tanzanian manufacturing development is characterized by six periods with differing combinations of implemented policies and external circumstances. These are summarized in Table 1.10.

5

Tanzanian industrial development: trends and explanations

In this section the framework drawn up in the previous sections is used to analyse trends in manufacturing development. It will be shown that turning points and breaks in trends in manufacturing growth, productivity or structure coincide to a considerable extent with the six periods of analysis presented in Table 1.10. As may be expected, there is not always a oneto-one relationship between the beginning of a new period and a trend change. There may be lags caused by the time it takes for implemented policies to take effect or external influences to exert their impact. Nevertheless, close relationships can be observed between the periods defined and the changes in manufacturing trends. On the basis of the analysis of the periods, it can be argued that public policy has been one of the important factors determining the development and performance of the Tanzanian manufacturing sector. The following four figures provide a picture of long-run trends in manufacturing performance in Tanzania. Figure 1.1 presents the changing share of total manufacturing in national income. Figure 1.2 reproduces a newly revised index of manufacturing value added in constant prices for the period 1961–95 (see Chapter 3). The index refers to the formal sector and excludes establishments with less than ten persons engaged. In the long run, one can distinguish a period of rapid growth from a very low level

Table 1.10

Periods of analysis

Characteristics Industrial strategy IS/EP Control Priv./public DFI/aid Macroeconomic situation

1961–67

1967–73

1973–80

1980–84

1984–90

1990–95

IS low priv DFI

IS moderate priv. → publ. DFI → aid

IS high publ. aid

IS high publ. aid

IS → EP moderate publ. aid

EP low publ. → priv. aid → DFI

relatively stable

relatively stable

first signs of imbalance

crisis

internal and external imbalance

internal and external imbalance

35

36 The Industrial Experience of Tanzania Figure 1.1

The share of manufacturing value added in GDP (%)

14 12

MVA/GDP (%)

10 8 6 4 2

0 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 Year Sources: Bureau of Statistics (1995a, 1995b, 1995e).

Figure 1.2 Index of manufacturing gross value added, 1961–95 (establishments with 10 or more persons engaged, 1976 = 100) 120

Index

100

80

60

40

20

0 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995

Year Source: Chapter 3.

Public Policy and Industrial Development 37 Figure 1.3 Index of manufacturing labour productivity, 1965–90 (establishments with 10 or more persons engaged, 1976 = 100) 120 100

Index

80 60

40

20

0 1965

1967 1969 1971 1973 1975 1977 1979 1982 1983 1985 1987

1989

Year Source: Chapter 3.

from 1961 to 1978, a period of collapse from 1978 to 1985 and a period of uncertain stabilization from 1985 to 1995. Figure 1.3 presents an index of labour productivity in constant Tanzanian shillings using the output series of Figure 1.2 and newly revised employment series (see Chapter 3). This figure illustrates the secular decline in productivity after 1973. Figure 1.4 puts Tanzanian manufacturing productivity in comparative perspective. This figure derives from a level comparison between Tanzania and the world productivity leader, the USA, for 1989 (see Chapter 3). The binary comparison is based on the calculation of industry of origin unit value ratios, which are used to convert Tanzanian gross value added for purposes of international comparison. In this paper, the benchmark comparison is extrapolated using national time series of GDP per person. For Tanzania, we used the series described in the source notes to Figure 1.2; for the USA we used national accounts series (see Timmer and Szirmai, 1999). One of the interesting results documented in Figure 1.4 is that relative labour productivity in Tanzania shows secular decline from a rather high initial level. Having started at around 9 per cent of US levels in 1965, comparative labour productivity increases until 1973, starts to decline after that year and falls to 3.7 per cent in 1987. In the following paragraphs, each period of analysis is discussed in terms of the interrelatedness of policy, external influences and manufacturing

38 The Industrial Experience of Tanzania Figure 1.4 Comparative labour productivity in Tanzanian manufacturing, 1965–90 (USA = 100) 12

Tanzania/USA (%)

10 8 6 4 2 0 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 Year Sources: 1989 benchmark from Szirmai et al. (Chapter 3); index of GDP per person in Tanzania from Figure 1.2; GDP per person in USA from national accounts sources (see Chapter 3).

performance. Within this context the effects of infrastructural constraints and low levels of industrial skills are also mentioned. 5.1

1961–67

Between 1961 and 1967 the growth performance of the manufacturing sector is impressive. Double-digit growth rates of real value added are recorded for the whole sector, causing the share of manufacturing value added in GDP to rise to 10.2 per cent in 1967 (Figure 1.1). During this period the index of real value added for enterprises with ten or more persons engaged rises 68 per cent (from 22 to 37; see Figure 1.2), and there are also increases in labour productivity (see Figure 1.3). Between 1965 and 1967 one can also discern some catch-up in labour productivity levels vis-àvis the world productivity leader, the USA (Figure 1.4). The investment thrust necessary to raise output and labour productivity levels is mainly provided by foreign owned firms. As Perkins (1983) emphazises, many of the new firms established after 1961 are owned by transnational corporations (TNCs), or their Kenyan subsidiaries. In 1965 no more than 32 per cent of all industrial firms are exclusively owned by Tanzanians. The predominance of private foreign owned firms can be explained by the negotiations within the East African Community


(see Section 4), and even more by the policy climate prevailing at the time. As described in Section 3.1, between 1961 and 1967 policy instruments are used to encourage private foreign investment. The most notable incentives are the granting of monopoly or near monopoly power and tariff protection. As pointed out in Section 3.1, tariff protection results in high effective rates of protection, especially for the manufacturers of consumer and intermediate goods. Changes in degrees of import substitution and changes in industrial structure can be explained in this context. Table 1.11 presents data on trends in the total supply of manufactured goods in Tanzania. The total manufactured supply (TMS) is defined as the sum of locally produced goods (for the domestic market, DM and exports, EXP) and imported goods, IMP. A declining IMP/TMS ratio, combined with a rising DM/TMS ratio, implies that imports are being substituted for locally manufactured goods.2 Table 1.11 confirms that import substitution is taking place between 1961 and 1965. From the last three columns of the table it can be seen that import substitution is primarily taking place in the consumer goods sector, and to a lesser extent in the intermediate goods sector. This coincides with the structural changes taking place in the share of branches in gross value added (10+). As Table 1.12 shows, the shares of consumer goods (ISIC 31 and 32) and intermediate goods (ISIC 33, 34, 35 and 36) in manufacturing value added rise between 1965 and 1967, whilst the share of capital goods declines. These changes can be traced back to the pattern of effective protection (Table 1.5), which stimulates (foreign) entrepreneurs to invest in consumer and intermediate good import-substitution activities. 5.2

1967–73

Between 1967 and 1973 the role private foreign investors play in the development of the manufacturing sector is considerably reduced. From 1967 onwards the public sector provides the main development thrust, as can be deduced from the rising shares of the public sector in total manufacturing value added and total manufacturing employment. In 1966 the public sector accounts for only 5 per cent of total value added. This percentage has increased to 32 per cent by 1973. The share of public sector employment in total manufacturing employment rises from 15.5 per cent in 1967 to 46.7 per cent in 1973 (Skarstein and Wangwe, 1986). These major changes are realized through the nationalization of private enterprises and by public investment in additional capacity. Public investment in manufacturing peaks in 1970, when the share of the public sector in total manufacturing investment reaches 63.7 per cent. 3 Thereafter, the rate of public investment in manufacturing slows down, since, as the World Bank (1988) puts it, ‘little is left to nationalize [and] fewer joint ventures are established’.

40

Table 1.11

Import substitution of manufactured goods, 1961–90

Year

1961 1965 1973 1980 1984 1990

Total manufactured supply

Domestic market/total manufactured supply

Domestic market (%)

Exports (%)

Imports (%)

Consumer (%)

Intermediate (%)

Capital (%)

29.6 36.0 42.5 50.9 58.6 46.6

8.2 7.8 6.4 6.6 5.7 8.5

62.2 56.2 51.1 42.5 36.3 44.9

50.0 59.5 72.2 81.2 87.4 89.4

25.7 32.1 39.8 55.2 58.6 47.1

16.4 16.6 21.4 24.3 23.7 27.7

Sources: 1961–73 data from World Bank (1987a); 1980–90 data calculated from United Republic of Tanzania (1995c).

Public Policy and Industrial Development 41 Table 1.12 Structure of medium- and large-scale manufacturing gross value added, 1965–95 (%) Year

31a

32b

33/34c

35d

36e

37/38/39 f

1965 1967 1973 1980 1984 1990 1995

32.4 26.4 26.8 21.4 31.9 44.2 51.4

23.0 33.2 36.1 33.9 25.4 14.8 12.2

10.4 9.9 7.2 9.9 9.5 9.9 12.8

6.2 11.5 10.6 15.4 14.4 13.1 11.9

1.5 4.6 3.3 3.3 2.1 3.1 3.7

26.5 14.3 16.0 16.1 16.7 14.7 7.8

Source: Prins and Szirmai (1998). Notes: a Food, beverages and tobacco. b Textiles and leather. c Wood products, furnishing and fixtures, paper products, printing and publishing. d Chemicals, petroleum, rubber and plastic products. e Non-metallic mineral products. f Metals, machinery and equipment and others.

The shift in the relative importance attached to the public and the private sector finds its origin in the Arusha declaration. The subsequent change in the use of policy instruments (industrial licensing, nationalization, regulation of joint ventures) explains the growth of the public sector. As documented in Figures 1.1 and 1.2, the increasing importance of the public sector has been accompanied by accelerating growth of value added and increased shares of manufacturing in GDP. The period 1967–73 records the most rapid growth of 10+ manufacturing value added in the history of Tanzania. This growth in value added is accompanied by absolute and relative labour productivity growth (Figures 1.3 and 1.4). In 1973 labour productivity is 44 per cent higher than in 1967. The process of import substitution continues throughout this period. By 1973, 42.5 per cent of the total manufactured supply is produced for the domestic market, whilst 51.1 per cent is imported (Table 1.11). In 1965 these percentages were respectively 36.0 per cent and 56.2 per cent. Again, most import substitution is taking place in the consumer and intermediate goods sector, as is to be expected, given the pattern of tariff protection. Along with other policy instruments (overvaluation of the exchange rate, import licensing and import quota), the tariff barriers explain the continued import substitution taking place. 5.3

1973–80

In terms of industrial performance, 1973 marks an important trend break. From 1973 onwards, labour productivity in 10+ manufacturing enterprises starts declining steadily (see Figure 1.3). The labour productivity index drops from 110 in 1973 to 75 in 1980, a decline of 5.3 per cent per year.


Comparative labour productivity declines from its peak value of 11 per cent to 7.2 per cent of the US level. There have been declines in value added between 1973–75 and 1978–80, but the index of manufacturing value added in 1980 is higher than that in 1973, indicating a net expansion of output over the whole period (see Figure 1.2). Given the fact that between 1973 and 1980 investments are largely taking place in capital intensive projects (Perkins, 1983), the combination of declines in labour productivity and expansion of output suggests that throughout this period manufacturing enterprises are producing less and less efficiently. The decline in efficiency, which also manifests itself in the drop in capacity utilization (World Bank, 1987a; Maliyamkono and Bagachwa 1990), is closely related to all four characteristics of the policy period prevailing at the time. First, it should be noted that the increased reliance on foreign aid results in investments in inefficient, large-scale and capital-intensive publicly owned enterprises. On this matter Perkins (1983) writes that ‘in many cases it appears that the government approves the establishment of new parastatals … merely because a foreign equipment supplier or TNC is willing to implement the project, and to provide equity capital, commercial credit or a soft loan through the aid programme of its home country’. These newly established publicly owned enterprises are not in accordance with resources available within the country, since they prove to be extremely import intensive, and require highly developed managerial skills and highly trained workers (see Chapter 5). On the latter issue, Gulhati and Sekhar (1982) point out that in subSaharan Africa inefficient production in public manufacturing results to a considerable extent from an ‘acute scarcity of skilled managers, administrators and technicians’. By expanding the Tanzanian public sector through aid programmes, the strain on the scarce availability of managerial skills and skilled labour is increased. On top of this, by allowing for the establishment of import-intensive enterprises, the already deteriorating balance of payments (see Section 4) is put under even more strain. Because of the balance of payments crisis and the continued establishment of new importintensive firms, available foreign exchange has to be spread out more thinly over claimants. This results in shortages of machinery, spare parts and raw materials (see Table 1.3). It is not surprising that the paradox of capital intensive industrial expansion in a situation of scarcity of material, financial and human resources results in declines in labour productivity. Another cause of the decline in efficiency should be sought in the combination of the emphasis on the public sector and the high degree of direct regulatory control prevailing from 1973 onwards. This combination, which is not uncommon in developing countries (White, 1984; Szirmai, 1997a), is known to result in inefficiency, since the imposition of direct regulatory controls (for example, foreign exchange allocation, price control, confine-


ment, industrial licensing) in a government-led sector leads to corruption and increased lobbying activities. Parastatal managers ‘spend their time escorting their papers through government offices’, rather than attempting to find solutions to the constraints hampering production (Krueger, 1992). In Tanzania this phenomenon is indeed observed (World Bank, 1988): ‘[The use of direct regulatory controls] has encouraged [parastatal] managers to focus their efforts on seeking additional import allocations and permissions to charge higher prices rather than on efforts to control costs, search for new markets and improved technologies.’ The imposition of direct regulatory controls also makes for extremely high degrees of protection. From 1973 onwards, trade and price controls, together with an overvalued exchange rate, protect both the private and the public sector from foreign competition. Such sustained importsubstituting policies have generally been associated with a loss of efficiency in the longer run (Page, 1990; Krueger, 1992). Protection allows manufacturers to produce above world market prices, ‘entailing production that is increasingly high cost and less economic’ (Meier, 1990). This mechanism exposes yet another manner in which policy instruments used between 1973 and 1980 are related to declines in efficiency. Table 1.11 shows that import substitution is indeed taking place throughout the period 1973–80. The domestic market–total manufactured supply ratio rises from 42.5 to 50.9 per cent, whilst the share of imports in the total manufactured supply declines from 51.1 to 42.5 per cent. Most import substitution takes place in the intermediate goods sector, a phenomenon which can be explained by the shift in relative protection in favour of the manufacturers of intermediate and capital goods (see Table 1.5: the 1984 effective rates of protection presented can be taken as indicative for levels of protection during the 1973–80 period). The shift in relative protection also contributes to the structural change taking place in this period. Table 1.12 shows that in 10+ manufacturing the share of intermediate goods in total manufacturing value added rises at the expense of the consumer goods sector. 5.4

1980–84

A second turning point in Tanzanian manufacturing performance occurs in 1979. In this year the trend of manufacturing (10+) growth is reversed. As Figure 1.2 reveals, the index of medium- and large-scale manufacturing GDP declines rapidly after 1979. In 1984 it is already beneath its 1973 level. The four years of negative growth cause the share of manufacturing value added in GDP to shrink from 12.4 per cent in 1979 to 9.1 per cent in 1984 (Figure 1.1). Between 1979 and 1983 the decline in labour productivity accelerates. Comparative labour productivity declines dramatically from 8.1 per cent in 1979 to 4.8 per cent in 1983. The divergence is the result of the combination


of labour productivity growth in the USA, and steeply declining levels of labour productivity in the Tanzanian manufacturing sector (see Figure 1.3). To a certain extent these steep declines in Tanzanian labour productivity levels can be explained by the policy environment prevailing in this period. The line of reasoning is similar to that used to explain the decline in labour productivity between 1973 and 1980. High levels of protection do not stimulate competitive behaviour, and contribute to inefficient production. Furthermore, the combination of high degrees of direct regulatory control and an emphasis on public ownership stimulates corruption and lobbying activities, rather than stimulating improvements in productive efficiency. Finally, by relying on foreign aid as a source of investment finance, capacity has been installed which cannot be used to its full potential, since the human and material resources to attain such a goal are in short supply. It should be noted, however, that the same mix of policy instruments used between 1973 and 1978 does not cause such a steep decline in labour productivity. Neither does the policy environment result in persisting negative growth rates of real output throughout that period. It is clear that factors other than policy alone are needed to understand the performance trends between 1979 and 1984. As stressed in section 4, the analytical difference between the period 1973–80 and the period 1980–84 can be found in the change in the external circumstances. From 1980 onwards the total Tanzanian economy faces an economic crisis, and it is precisely the combination of this crisis with the prevailing policy environment that brings about the radical deterioration in manufacturing performance. The aspect of the general crisis which has the largest impact on the performance of the manufacturing sector is the shortage of foreign exchange (de Valk, 1992). As Table 1.3 shows, during the early 1980s on average only 30 per cent of all requests for foreign exchange within the Tanzanian economy are awarded. The percentage of manufacturing requests awarded is even lower (on average 10 to 15 per cent). In fact, in some years no foreign exchange at all is made available for imports of raw materials, machinery and spare parts. Furthermore, as a World Bank (1987a) study shows, the bulk of the foreign exchange available is allocated to the least efficient enterprises. This is caused by the high degree of direct regulatory control the government has assumed over the manufacturing sector (see Section 3). Between 1982 and 1985 inefficient (mostly public) enterprises receive 74 per cent of the available foreign exchange. Given that the manufacturing sector is import intensive (see analysis of previous period), it is very likely that the shortage and inefficient allocation of foreign exchange have had negative effects on output and productivity levels. The 1980–84 crisis is also characterized by frequent interruptions in water and electricity supply. Along with the foreign exchange shortage, these infrastructural problems are reckoned amongst the principal production


constraints prevailing at the time. Other production constraints are caused by firms supplying inputs, by a lack of supervisory and technical manpower, and by a lack of labour discipline (de Valk, 1992). In the publicly owned enterprises the latter production constraint is the result of the socialist approach to management adopted after the drawing up of the Arusha declaration (Weaver and Kronemer, 1981). Infrastructural problems (including inadequate transport and communication facilities) and the lack of supervisory and technical manpower are closely linked to the unfavourable initial conditions discussed in Section 2. The investments in large scale manufacturing enterprises have not been accompanied by sufficient investments in infrastructure and training of the industrial workforce by the Tanzanian government. In conclusion, the weaknesses of the oversized and already inefficiently operating manufacturing sector are revealed by the 1980–84 macroeconomic crisis. The roots of the crisis in the manufacturing sector are to be found in government policy, which has stimulated the establishment of an import-intensive, publicly owned manufacturing sector, which is not compatible with material and human resources available at the time. Hence, aspects of the overall crisis, such as a lack of foreign exchange and insufficient infrastructural provisions, result in scarcity of inputs and frequent work stoppages. Labour productivity levels drop further, and declines in real output levels are recorded. The continued import substitution taking place between 1980 and 1984 (Table 1.11), is more a result of a shortage of foreign exchange for the financing of imports, rather than the result of deliberate import substitution policy (Wangwe and Bagachwa, 1990). The shares of intermediate and capital goods in gross value added do not change throughout the period (Table 1.12). This lack of structural change in an environment of high levels of protection for the manufacturers of intermediate and capital goods also illustrates the seriousness of the 1980–84 crisis in the manufacturing sector. 5.5

1984–90

In 1984 the decline in output and labour productivity levels comes to an end. As Figures 1.2 and 1.3 show, output and labour productivity levels stabilize between 1984 and 1989. So does the trend in the share of manufacturing value added in GDP: 1985 and 1989 shares both equal 8.4 per cent. The declining trend in comparative labour productivity levels, however, is not reversed. In 1984 the ratio of Tanzanian and US manufacturing labour productivity levels equals 5.8 per cent. This ratio has dropped to 3.9 per cent in 1990. Comparative productivity in medium- and large-scale manufacturing is well below that of developing economies in Asia, such as China, India and Indonesia (between 5.7 and 10 per cent of the US level in 1987; see Timmer and Szirmai, 1999). The further decline in comparative


manufacturing labour productivity levels is the result of the combination of labour productivity increases in the lead country and the already noted stabilization of labour productivity levels in Tanzania. From an international perspective, this is indicative of a process of falling behind. The introduction of foreign exchange windows (e.g. own funds, open general licence) other than loans and grants and the free resources allocated by the Bank of Tanzania, forms the main reason why output and productivity levels do not decline further from 1984 onwards. Through the newly introduced foreign exchange windows, import-starved manufacturers can obtain raw materials, machinery and spare parts necessary to improve capacity utilization. As a consequence of these new import possibilities, the trend in import substitution is reversed. Table 1.11 shows that the share of imports in the total manufactured supply rises from 36.3 per cent in 1984 to 44.9 per cent in 1990, whilst the DM/TMS ratio drops from 58.6 per cent to 46.6 per cent over the same period. Between 1984 and 1990 there is also a significant rise in the share of exports in the total manufactured supply for the first time since independence. This coincides with the trade policy shift taking place throughout this period. As explained in Section 3.4, because of trade liberalization, nominal exchange rate devaluations and tax evading importers, production for the home market has become less attractive than it was in previous policy periods. Consumer goods manufacturing, and manufacturing of food products in particular, once again becomes the most attractive sector in terms of protection (Table 1.5), a change which is reflected by the structural changes occurring during the period (Table 1.12). Another disincentive for production for the home market is provided by export promotion schemes introduced from 1984 onwards. It is often argued that trade liberalisation and incentives for export oriented production ‘result in the contraction of inefficient sectors and the expansion of new, efficient ones’, because trade liberalisation and export promotion force domestic producers to comply to international price and quality standards (Michalopoulus, 1987). However, Kirkpatrick and Maharaja (1992) point out that in the case of less developed countries there is little empirical evidence supporting this view. Trends in Tanzanian manufacturing performance between 1984 and 1990 do not provide any support for this view either (Figures 1.2 and 1.3). Moreover, a World Bank (1991) study states that the extremely inefficient firms which survived during the heyday of protectionism still manage to survive from 1984 onwards, despite the liberalisation of internal and external trade. This remarkable observation is closely linked to the preferential treatment of the public sector, which is still cherished by the government between 1984 and 1990. In 1989, 56 per cent of the publicly owned manufacturing enterprises are operating at a loss (World Bank, 1991). But despite this troublesome figure, the public manufacturing sector is exempted from


duties and sales taxes, is allocated foreign exchange on favourable terms (through aid inflows), and can count on the bulk of available (subsidized) credit. As pointed out in Section 3.4, public enterprises receive about 70 per cent of total loans made available to manufacturing firms. Preferential treatment of public enterprises regarding the allocation of loans is enabled by the control the government still has over the monetary sector. 5.6

1990–95

With the 1990–95 decontrol and shift away from the predominance of the public sector in the industrial strategy, it would seem that the many policy barriers to efficient production are removed. But this is not (or not yet) reflected in the indicators of manufacturing performance. The share of manufacturing value added hardly changes between 1990 (7.8 per cent) and 1994 (7.6 per cent) (Figure 1.1). Output levels in medium- and largescale manufacturing do not rise substantially either: the 1995 output level is only just above the 1989 output level (Figure 1.2). Furthermore, a sample of 20 manufacturing activities shows only modest improvements in the use of capacity throughout the period. In 1990, 32 per cent of the installed capacity in the sample industries is used. In 1994 capacity utilization in these industries has increased to 37 per cent (United Republic of Tanzania, 1995b). The absence of significant improvements in manufacturing performance can be traced back to the macroeconomic situation, the unfavourable initial conditions Tanzania has had to face and the policy changes in the period studied. Regarding parastatal reform, for instance, it cannot be expected that privatization of 34 of the 170 manufacturing parastatals immediately results in productivity gains. In the words of Adam (1994), public sector reform is accompanied by ‘changes … in terms of managerial attitudes, and a greater perception of operating autonomy, which will gradually be translated into efficiency gains and increased profitability’. It is unlikely that these desirable effects take place at short notice though, a point which can also be made regarding the inflow of direct foreign investment after the opening up of the sector to private entrepreneurs. Another reason why policy changes do not result in improvements in manufacturing performance can be found in the combination of the liberalisation of the banking system and the high inflation rates prevailing between 1990 and 1994 (Table 1.9). The high inflation rates force the banks to charge extremely high nominal interest rates in order to maintain positive real interest rates (see Section 3.5). These high nominal interest rates are a threat to the liquidity of manufacturing enterprises. Furthermore, the high inflation rates erode the purchasing power of the bank overdrafts. Since banks are reluctant to raise bank overdrafts, manufacturers have less and less access to short-term credit to finance procurement of raw materials and intermediate inputs. On top of that, non-creditworthy enterprises are


no longer allocated medium- and long-term credit, since this would contradict the commercial principles banks operate on from 1991 onwards. As such, credit has replaced foreign exchange as the major production constraint the manufacturing sector faces. Other production constraints are related to the conditions Tanzania has had to face since independence. As discussed in Section 2 and in this section (1980–84 period), the manufacturing sector has been operating in an environment characterized by poor infrastructure and a shortage of supervisory and technical manpower. In a more general sense, Helleiner (1994) points out that ‘industrial development hinges, in any case, upon local capabilities and appropriate supporting institutions no less than on incentives, and, in sub-Saharan Africa these have been typically lacking’. In Tanzania, this lack of local capabilities and supporting institutions has continued to hamper manufacturing development between 1990 and 1995 (United Republic of Tanzania, 1996b). Along with the unfair competition by tax-evading importers (see Section 3.5), the production constraints mentioned above result in an environment which is not very conducive to rapid industrial growth. Neither is the environment conducive to structural change. As Table 1.12 shows, in 1995 more than 50 per cent of manufacturing value added (10+) is produced in the food, beverages and tobacco branch. In 1995 the total consumer goods sector accounts for 64 per cent of manufacturing value added, higher than the share recorded 30 years earlier. Thus, in terms of structural change little has changed throughout these 30 years, a conclusion which can also be drawn regarding the importance of the manufacturing sector for the economy as a whole (see Figure 1.1).

6

Concluding remarks

Although a promising start is made in the period 1961–67, both in terms of policy and performance, changes in the policy climate from 1967 onwards affect manufacturing development adversely. This is due to the fact that the policy regimes prevailing between 1967 and 1980 stimulate the establishment of an oversized and increasingly inefficient sector, which is not well adapted to Tanzania’s resources in terms of factor proportions, skills, human capital, infrastructure, availability of foreign exchange and other relevant factors. The macroeconomic crisis of the 1980s reveals this weakness, and the manufacturing sector plunges into a crisis from which, to this day, it is still attempting to recover. The policy responses to the crisis in the period 1984–90 seem conducive to industrial stabilization, but further changes in policy between 1990 and 1995 have so far not been followed by industrial recovery. A recent review of African economic performance by Collier and Gunning (1999) indicates that the industrial experience of Tanzania is representative of that of many economies in Sub-Saharan


Africa. The detailed analysis of the Tanzanian experience provided in this volume should be seen as a case study in African industrialization. Looking at Tanzanian industrialization from a long-term perspective, one may conclude that post-war industrial policy has been informed by a mainly structuralist interpretation of the process of economic development. This interpretation called for heavy government involvement in order to force the pace of industrialization. The problems discussed in this paper suggest that this approach is flawed and that many of the problems of the manufacturing sector are policy related. However, the fact that the return to neo-liberally inspired policies since 1984 has so far not led to a resumption of manufacturing growth suggests that a return to the market alone will not automatically solve all problems either.

Notes *

1

2

3

Section of Technology and Development Studies and Eindhoven Centre for Innovation Studies, Faculty of Technology Management, Eindhoven University of Technology Some classifications focus on the level of dependence on foreign finance, but in the case of Tanzania this dimension does not help distinguish policy periods. The dependence on foreign finance continues to be high throughout the whole period studied. For total manufacturing, one can also take production for the domestic market (DM) as a proportion of production for the domestic market plus imported goods (IM). However, these ratios are not available for subcategories of manufacturing. Calculated from United Republic of Tanzania, Analysis of Parastatal Accounts, and Bureau of Statistics, National Accounts (various issues).

2 Is African Manufacturing Skill Constrained? Howard Pack and Christina Paxson*

1

Introduction

In most of the sub-Saharan African economies neither the levels of total factor productivity (TFP) nor the growth rates of TFP in manufacturing have been high. All studies of cross-country performance find that the subSaharan African economies are the largest bloc of nations that have not converged on the US. Most of the inter-country explanations have focused on easily measured aggregate variables such as the ratio of investment to GDP, education levels, and, in some models, proxies for political stability (Barro and Lee, 1993; Easterly, 1993). These models have as their underlying theoretical framework a view that the national economy can be modelled with a set of multiplicative inputs – an increase in the right hand side variables such as the investment rate or education level will produce an increase in growth rates. However, as is increasingly recognised, by, among others, the authors of the many papers on convergence, the particular specification of the implied production function is open to question, and the right-hand-side variables may themselves be endogenous. Moreover, in the case of the African nations, close observers question whether a simple increase in investment rates will generate the impact implied by the crosscountry regressions – many countries have experienced growing marginal capital–output ratios over the last two decades (Husain, 1993). There are two divergent though complementary views of the lack of productivity growth in sub-Saharan Africa.1 The first holds that pervasive government intervention imposes large costs on individual firms and reduces incentives to become efficient (Collier and Gunning, 1999). It is not only protection from foreign competition that has such effects but also problems stemming from bad macroeconomic policies, such as highly variable real exchange rates, which make long-term planning difficult for firms. A second interpretation focuses on the paucity of technological competence as measured by the small supply of trained managers and engineers.2 This paper examines the determinants of manufacturing productivity and the 50

Is African Manufacturing Skill Constrained? 51

role of education and technology variables at the level of individual firms. A major implication of our results is that, ceteris paribus, it is unlikely that an increase in education levels will have much impact on manufacturing TFP levels. We suggest a more complex view of the role of education, emphasizing that, in the absence of technology inflows, higher skills may have very low productivity. Using the results of surveys of industrial firms in Ghana, Kenya and Zimbabwe, we consider the connection between technological abilities and productivity in the three nations. Section 2 sets out the determinants of productivity. Section 3 provides some aggregate measures of factors that may affect productivity in the three economies in question. Section 4 presents empirical results on the analysis of the determinants of productivity in a large group of firms in the three countries. Section 5 offers an interpretation of the results, and Section 6 contains conclusions.

2

The determinants of productivity levels

While economy-wide and sectoral shortages of human capital are widely assumed to limit efficient industrial development, there is little empirical evidence on whether this is indeed the case in individual African firms. The presence of trained and experienced managers and workers does not guarantee high firm productivity for several reasons. First, individuals with formal credentials may not possess the cognitive skills necessary to improve TFP. Second, if there is limited rivalry, firms may view investment in costreducing efforts as not being necessary to remain competitive. Third, an industry’s organization, such as vertical integration between manufacturers and retailers, may result in low productivity if, for example, each firm produces a large range of products, the profitability of firms being guaranteed by protection from imported goods. Hence, the frequently drawn policy conclusion of the need for higher levels of education and training based on the assumed shortages of skilled manpower may be unwarranted. In theory technological abilities are important for cost reduction in industrial firms, facilitating learning by doing. The simplest interpretation of pure learning by doing envisions a manual worker tightening a bolt in 40 per cent less time the hundredth time she does it as compared to the first time. But a complete specification of ‘pure’ learning by doing would recognise that there is usually an ongoing reorganization of the flow of work, the design or distribution of improved machinery, and additional training. Thus, even in the simplest case of learning by doing, one would expect that the presence of more educated or trained supervisors and production workers will facilitate realization of greater TFP growth. However, there is little empirical support, even in industrialised countries, for a close relation between technological abilities and cost reduction for industrial firms.3


By itself, the purchase of modern equipment does not guarantee that high TFP will be realised with such machinery – many complementary firm-level actions are required. Production engineering must be mastered, the flow of intermediate inputs to the plant and spare parts must be arranged, workers must be trained, and marketing must be enhanced so that it is not necessary to produce short production runs to order, and incur excessive set-up costs. Education, vocational training, foreign training, firm-level research and development, expenditures on technology licences, and the nationality of owners may all affect the ability of firms to master the software of technology, and are thus important determinants of firm-level productivity. The economic environment established by macro- and micro-economic policies will also impinge on the firm and discourage or partly offset positive efforts at the firm level. Difficulty in obtaining raw and intermediate inputs of a given quality due to exchange controls, erratic supply of public services such as electricity, the incentive provided by import restrictions to expand the range of products, and the general absence of incentives to improve productivity all affect firm-level performance. For any given set of firms the impact of the economic environment and firm-level factors jointly determine productivity. Assuming that all firms within a given industry face the same economic environment, it is possible to test whether firm technological abilities affect TFP or its mirror image, unit costs. There are two alternate views of the role of education and technology in the production function. The first views them as multiplicative inputs in a standard production function. Thus, if a production function is estimated, elasticities of output with respect to technology inputs can be calculated. A second view that we believe is more useful (Nelson and Phelps, 1966) argues that education will have its greatest impact when there is rapid technological change. If the basic technology (a loom used in weaving) is largely unchanged over time, the production process becomes routine and the ability to deal with change is not germane – high education results in only limited productivity gains. In contrast, where technology is rapidly evolving, learning about the existence of new processes, learning to use them when they are deployed, and staying abreast of new developments require the adaptability provided by formal education. The preceding can be formalised following a model of Nelson and Phelps (1966). Firms in LDCs operate with a technology level equal to A(t) in period t. Their peers in industrial countries operate technology T(t). The rate at which the DC technology is introduced into the LDC depends on the level of human capital, h, and is  T (t ) − A(t )  A(t ) / A(t ) = ( h)   A(t )  

(2.1)


The extent to which local LDC technology improves is a positive function, ’(h) > 0, of the level of human capital and proportional to the magnitude of the differential between current and ‘best practice’ technology. As the technology T(t) does not have to be invented, the potential productivity gain from the transfer of this technology is the benefit of relative backwardness (Gerschenkron, 1962). Assume that the DC technology improves each year by per cent, so that T(t) = T0et

(2.2)

Given (2.1) and (2.2), the underlying differential equation implies that the potential equilibrium path of technology of an LDC firm is A(t) = [(h)/((h) + )]T0et

(2.3)

The potential level of technology realised by an LDC firm at a moment in time will thus be higher: (1) the greater its ability to deal with new technologies as a result of the presence of qualified individuals on its staff; (2) the larger the inflow of technology to the firm in the form of new equipment, new material inputs, new knowledge obtained from consultants, licensers and foreign owners. The potential level of technology characterizing a firm will evolve along Equation 2.3 and depends solely on its own level of h and the rate of technical progress in the DCs that becomes available to the LDC firm. If foreign exchange shortages or arbitrary rules prohibiting some forms of technology imports reduce the inflow of new knowledge, the benefit conferred by h is reduced. This might stem, for example, from regulations that limit expenditures on technology licensing payments so that a firm cannot purchase a potentially useful technology, implying that F < where F represents the lower rate of technology inflow. Thus Equation 2.3 can be rewritten as A(t) = [(h)/((h) + F)]T0eF t

(2.4)

Equation 2.4 underlines the fact that, at a given point in time, the level of productivity A(t) for a given firm may be weakly associated with h unless firms’ ability to obtain new technology varies systematically with h. Unlike formulations which treat education or purchases of knowledge as conventional inputs in the production function, Equation 2.4 implies that human capital will have no effect on the level of output obtained with conventional inputs unless F > 0, which can only occur if new productive inputs are introduced. Schultz (1975) argued that in addition to the ability to absorb new technology, education would be important if there were a changing sectoral structure or internal product mix of firms, more highly educated workers


being better at reallocating factors, exploring new profitable opportunities and securing markets. Changing sectoral structure and internal product mix in turn reflect growth in per capita income in the internal market combined with different Engel elasticities and changes in the net trade balance in individual commodities. If sectors are not changing, it is likely that product mix within a sector is also stable, putting little premium on the ability to cope with change.

3.

Aggregate evidence on technology

The payoff to education depends in the above view on two measurable magnitudes: (1) the inflow of imported machinery and industrial intermediate inputs that may embody new technology and (2) the sectoral structure of production. To measure the first Table 2.1 shows the rate of growth of all imports of goods and services, in constant prices, for the three African countries as well as Korea and Malaysia, two Asian countries that are widely viewed as having benefited from the inflow of technology. The second is measured by the change in the split between light and heavy industry and the standard deviation of the percentage change in sectoral shares of value added (Table 2.2). In the period 1980–93 the three African economies exhibited a negative or very slow growth of all imports, reflecting their inability to earn foreign exchange through exporting, and the limited inflow of concessional aid. Kenya, for example, experienced a decline in constant price imports of 0.8 per cent per year over the period. As the share of machinery and transport equipment and other manufactured inputs in total imports was roughly constant over the 1985–93 period, the absolute level of imports that could embody new technology declined.4 In Ghana and Zimbabwe there were significant declines in the share of imported equipment with only slight growth of exports; thus a significant decline in the level of imported capital goods occurred. The figures also suggest that, at most, a modest increase occurred in manufactured intermediate inputs. In contrast, in Korea and Malaysia, two NICs in which the level of education is widely perceived to have been an important contributor to industrial growth, the level of imports was rising very rapidly, reflecting the growth in export earnings, and the share of the two technology-embodying inputs was constant or increasing, implying potential for significant returns from the ability to effectively utilise the new technology component of these imports. The African nations also show a striking lack of change in sectoral structure, compared to dramatic structural changes in the Asian NICs. Table 2.2 shows the distribution of manufacturing value added among light and heavy manufacturing for the five countries in question, and an index of structural change given by the standard deviation of the percentage change

Table 2.1 Country

Kenya Ghana Zimbabwe Korea Malaysia

Imports of goods embodying new technology Growth rate of imports, 1980–93

–0.8 2.7 0.2 11.4 9.7

Share of machinery and transport equipment in total imports

Share of other manufactured imports in total imports

1985

1993

1985

1993

23 40 65 34 46

25 26 36 34 54

28 28 26 23 28

29 31 31 29 30

Source: World Development Report (various issues).

55

56 The Industrial Experience of Tanzania Table 2.2

Sectoral distribution of value added in manufacturing Ghana Kenya Zimbabwe Korea Malaysia 1980 1990 1980 1990 1980 1990 1980 1990 1980 1990

Light industry 57.8 59.9 52.7 55.2 49.3 53.3 41.5 29.6 47.8 30.8 Heavy industry 42.2 40.1 47.2 45.1 50.7 46.7 58.5 70.4 52.2 69.2 SD of the % change in sectoral value shares in 28 ISIC sectors 1.58 1.25 2.05 2.18 2.98 Source: UNIDO (1993).

between 1980 and 1990 in value added, calculated over 28 three-digit industrial sectors. As can be seen, both the sectoral distribution between heavy and light and the summary standard deviation measure show that the three African economies have experienced relatively little change in the structure of sectoral production, a phenomenon also replicated at the firm level.5 The firms in our sample manufacture relatively simple products, and it is unlikely that the internal product mix of firms has changed much. Indeed, many of the items produced by the larger firms are not that different from those fabricated in the informal sector. These results suggest that unless some firms were able to deviate from the overall pattern of stagnation of new inputs, it would be surprising if firms with higher technical skills realised greater levels of productivity.

4. 4.1

Empirical results at the firm level The data

We analyse the relationship between skills and production costs using firmlevel data from three African countries, Kenya, Zimbabwe, and Ghana. The survey data were collected in 1993 (for Kenya and Zimbabwe) and 1992 (for Ghana) by the Regional Program on Enterprise Development of the World Bank. Although the data are by now panel data, with up to three years of information for each firm, only the first year of data was available for this study. The surveys for each country used similar questionnaires and sampling procedures. All collected information on the value of output, production costs, the skill level of the owners and managers in each firm, and each firm’s involvement in foreign markets, such as raw material imports, exports and foreign ownership.


The Kenyan data are drawn from a survey of 223 manufacturing firms. The original sample consisted of 162 formal sector firms drawn from the Central Bureau of Statistics register of firms, and 61 informal (that is nonregistered) firms. Four sectors are represented: food processing, textiles and garments, woodworking, and metalworking. The sample is stratified by sector and (for formal firms) by firm size. Detailed information on the survey design and the procedures used to compute weights for the strata are in World Bank (1993b). Although the descriptive statistics presented below make use of data on all 223 firms, we had to exclude 51 firms when estimating cost functions, because of missing or incomplete data on costs, output, capital values and other variables used in the analysis. In Zimbabwe 201 firms were surveyed in the summer of 1993. The firms were selected from two sources. Large firms, with 50 or more employees, were drawn from a firm registration list provided by the Central Statistical Office. Smaller firms were drawn from a list compiled by the Gemini Survey of small-scale firms, conducted in 1991 (see World Bank, 1994b, for further details on the sample design and the sample weighting procedure). We use data on 200 firms (one firm had no sector information and was excluded), and lose an additional 30 firms for the estimates of cost functions because of missing data. The sample from Ghana consists of 186 firms surveyed in the summer of 1992. The sample design is described in World Bank (1993c). The design of the survey appears to have been complicated by the fact that two censuses, an industrial census and a population census, yield inconsistent information on the numbers of small-scale enterprises (World Bank, 1993c, p. 28.) Because of this ambiguity, we do not use population weights, the number of firms in the population in each stratum, in calculating the summary statistics that follow, but the means of all firms in the sample. It is thus possible that our sample statistics may not accurately reflect the characteristics of the population of firms in Ghana. The firms in the Ghanaian sample tend to be smaller than the firms in Zimbabwe and Kenya. Furthermore, there is no simple definition of “formal” versus “informal” firms, as there is for the other two countries. In what follows we sometimes split the Ghanaian sample between ‘large’ firms, defined as those with 10 or more employees (as opposed to more than 50 employees for Zimbabwean formal-sector firms), and small firms. The Ghanaian data also have a larger fraction of missing values than the other two surveys. Of the original 186 firms, only 132 can be used to estimate cost functions. Valuation of the capital stock of firms appears to have been a particular problem in Ghana, and so we do not include capital stock in the cost functions: doing so would have resulted in losing 19 more firms. The results for Ghana should be treated with extra caution, because of the small sample and the selection problems that may result from missing dependent variables.


4.2

Descriptive information on the firms

The surveys indicate that the majority of firms are engaged in producing fairly simple products. In all three countries the modal product of firms in the food sector is bread-making. Garments and wooden household furniture are the modal products for the textile and woodworking sectors. The metal sectors in each of the three countries are somewhat more diverse, although the majority of sampled firms report metal doors, windows, gates and burglar bars as their primary products. The picture does not change much in Kenya and Zimbabwe when sample weights are used to calculate the percentage of firms in the population (and labour shares) in each activity.6 Most of the firms in each of the countries are sole-proprietorships, partnerships or limited liability enterprises, and for almost all of these firms a person identified as ‘the owner’ took part in a survey about his education, experience and personal background. Owners in each country were asked about the highest educational level obtained, and were also asked if they had additional vocational, technical, or (for Kenya and Ghana) professional training. In each country the majority of surveyed owners had at least some secondary education, and a third to a half of these had extra training as well (see Table 2.3). Without reliable data on the distribution of skills in the populations of each of these countries it is impossible to tell if owners of manufacturing firms have better than average skills. Data from the Table 2.3a

Kenya: education of firm’s owners (193 firms)

Education level None Primary Secondary University

Total 6 54 87 46

5.4% 51.0% 40.0% 3.7%

Vocational 0 5 7 0

0.0% 7.6% 2.0% 0.0%

Technical 0 3 13 3

0.0% 1.7% 4.2% 0.0%

Professional 0 0 10 8

0.0% 0.0% 4.4% 0.5%

Note: Numbers indicate numbers of sampled firms. Percentages indicate estimated percentages of firms in the population, and are calculated using population weights provided by the survey study.

Table 2.3b

Zimbabwe: education of firm’s owners (134 firms)


Total 3 32 64 35

Vocational or technical training 2.4% 45.3% 46.5% 5.8%

0 14 30 6

0.0% 8.8% 20.2% 1.0%

Note: Numbers indicate numbers of sampled firms. Percentages indicate estimated percentages of firms (with owners) in the population, and are calculated using population weights provided by the survey study.

Is African Manufacturing Skill Constrained? 59 Table 2.3c

Ghana: education of firm’s owners (153 firms)


Total 17 12 107 17

Vocational

11.1% 7.8% 69.9% 11.1%

0 1 13 0

0.0% 0.7% 8.5% 0.0%

Technical 0 3 18 0

0.0% 2.0% 11.8% 0.0%

Professional 0 0 4 3

0.0% 0.0% 2.6% 2.0%

Note: Numbers indicate numbers of sampled firms, and percentages indicate percentages of sampled firms. Population weights are not used.

Ghanaian Living Standards Measurement Study, conducted in 1988 and 1989, indicate that average years of schooling among people aged 15 to 55 was six years (Jolliffe, 1995), with a lower average for people in older age groups, so it is likely that owners of manufacturing firms have better than average skills than the working population. Owners may hire managers with skills that exceed their own, in which case the appropriate measure of skill for the firm may be that of the manager rather than the owner. Tables 2.4a, 2.4b and 2.4c tabulate the education levels of ‘general managers’ and ‘production managers’ for firms that indicate that they have a manager of either type. The distribution of education among managers is not too dissimilar from the distribution among owners, although a higher fraction of managers have attended a university. However, this does not imply that owners hire managers who are more educated than themselves, and may reflect the fact that it is larger firms with more educated owners that hire managers. Among firms that have both owners and managers present, there is a high correlation between the skill levels of the two.7 The survey also asked about the average education levels of ‘managerial employees’. Tabulations of these for Kenya and Ghana are shown in Table 2.5 (the results for Zimbabwe were miscoded Table 2.4a Kenya: distribution of managers’ education General manager (N = 110) None Primary Secondary University – nontechnical University – technical Postgraduate (Kenya) Postgraduate (abroad)

2 8 47 25 21 3 4

0.1% 33.4% 40.2% 3.4% 17.8% 0.3% 4.9%

Production manager (N = 58) 1 3 25 5 21 1 2

15.2% 1.2% 57.7% 1.4% 23.8% 0.3% 0.5%


60 The Industrial Experience of Tanzania Table 2.4b

Zimbabwe: distribution of managers’ education General manager (N = 190)

None Primary Secondary University – nontechnical University – technical Postgraduate (local) Postgraduate (abroad)

3 27 86 31 31 2 10

Production manager (N = 108)

2.4% 42.4% 47.1% 3.7% 2.7% 0.1% 1.7%

1 16 49 10 27 3 2

0.9% 14.8% 45.4% 9.3% 25.0% 2.8% 1.9%


Table 2.4c

Ghana: distribution of managers’ education General manager (N = 47)

None Primary Secondary University – nontechnical University – technical Postgraduate (local) Postgraduate (abroad)

4 4 10 7 18 1 3

8.5% 8.5% 21.2% 14.9% 38.3% 2.1% 6.4%

Production manager (N = 51) 0 3 11 2 28 1 6

0.0% 5.9% 21.6% 3.9% 54.9% 2.0% 11.8%

Note: Percentages are calculated from firms in the sample.

and cannot be used), and these indicate that the majority of managerial employees have attended secondary school, and roughly a quarter of the sampled firms with managerial employees list the average education level of their managers at the university level. In the cost functions presented below we use a variety of skill measures that reflect the education levels of both owners and managers. We have no direct measures of the rate at which different firms acquire new technology (that is, the term F in Equation 2.4). However, it is likely that the acquisition of new technology from abroad is positively correlated with activities in foreign markets, and Table 2.6 presents information on these activities. The predominant picture is of firms that have little involvement in international markets. The vast majority of all firms import none of their raw materials, export none of their output, and do not have any foreign ownership. The firms were also asked about the use of foreign technology licences and foreign technical assistance. In Kenya only 9 out of 220 firms that answered these questions had foreign licences, and 17 out of 220 received

Is African Manufacturing Skill Constrained? 61 Table 2.5

Average education of managerial employees

Education level Kenya (126 firms) None Primary Secondary University Ghana (64 firms) None Primary Secondary University

Total

Vocational

Technical

Professional

1 8 82 35

0.2% 14.2% 74.2% 11.4%

0 0 7 1

0.0% 0.0% 6.4% 0.0%

0 2 8 0

0.0% 8.0% 14.3% 0.0%

0 0 11 9

0.0% 0.0% 1.5% 1.6%

2 1 45 16

3.1% 1.6% 70.3% 25.0%

0 0 4 0

0.0% 0.0% 6.2% 0.0%

0 0 8 0

0.0% 0.0% 12.5% 0.0%

0 0 14 2

0.0% 0.0% 21.9% 3.1%

Note: Numbers indicate numbers of sampled firms. For Kenya, the percentages are calculated using population weights provided by the survey. The data from Ghana are unweighted.

foreign technical assistance. In Zimbabwe licences and technical assistance were somewhat more common: 28 out of 200 firms had foreign licences, and 26 out of 200 received technical assistance. The questions asked of Ghanaian firms were phrased differently, but also indicate little use of licences or assistance. Of the 186 firms in the sample, only 6 firms indicated that they had held at least one foreign licence during the past 10 years, 6 said they had received foreign technical assistance in the last year, and 2 said they had used foreign consultants in the past two years. 4.3

Estimates of cost functions

The data described above have been used to estimate cost functions for firms in each of the three countries. Our primary interest is to see whether firms with more skills (as measured by the education levels of owners and/or managers) and more involvement in foreign markets have lower productions costs, given output. We start with the short-run cost function for firm i: ci = C(pi, ki, qi/Ai)

(2.5)

where c is costs defined as the cost of labour, raw materials and indirect costs such as utilities and rent; p is a vector of input prices, q is the value of output, k is the replacement value of the firm’s capital equipment, and A is a measure of the technology level of the firm. This form of the cost function follows from the assumption that the short-run production function has the form qi = AiQ(Mi, ki), where Mi is a vector of variable inputs and Ai depends on the level of ‘technological’ inputs such as education. The Nelson-Phelps view discussed in Section 2 implies that Ai is a function of both the skill level of the firm (measured by hi) and also by factors that

62

Table 2.6

Involvement in foreign markets

Kenya Percentage of raw materials imported Percentage of output exported Percentage of foreign ownership of the firm Zimbabwe Percentage of raw materials imported Percentage of output exported Percentage of foreign ownership of the firm Ghana Percentage of raw materials imported Percentage of output exported Percentage of foreign ownership of the firm

None

0–25%

25–50%

50–75%

75–100%

121 (76.1%)

32 (11.4%)

15 (3.3%)

25 (3.3%)

25 (5.9%)

170 (78.0%) 184 (96.5%)

29 4

(6.6%) (0.2%)

7 (0.1%) 6 (1.7%)

4 (0.2%) 10 (0.3%)

8 (0.3%) 14 (1.3%)

92 (93.0%)

57

(4.4%)

20 (0.9%)

13 (0.4%)

12 (1.3%)

104 (92.2%) 155 (96.7%)

58 12

(4.4%) (0.5%)

23 (1.1%) 7 (0.6%)

9 (1.1%) 8 (0.9%)

6 (1.3%) 15 (1.3%)

127 (72.6%)

12

(6.9%)

6 (3.4%)

11 (6.3%)

19 (10.9%)

163 (90.1%) 154 (82.8%)

11 2

(6.1%) (1.1%)

2 (1.1%) 12 (6.5%)

1 (0.6%) 13 (7.0%)

4 (2.2%) 5 (2.7%)

Note: The numbers in parentheses indicate percentage of firms. The percentages for Kenya and Zimbabwe are calculated using population weights. The percentages for Ghana are unweighted.


affect the firm-specific rate of flow of new technologies from abroad (measured by F), the two entering the cost function interactively. We start by estimating basic cost functions that do not include any variables that may affect Ai. Because we have small samples of cross-sectional data, and no data on input prices p, we employ simple specifications for the short-run cost functions. We work with the following: ln(ci) = s + q ln(qi) + qq[ln(qi)]2 + k ln(ki) + kk[ln(ki)]2 + qk ln(qi)ln(ki) + i

(2.6)

where s is a sector-specific intercept meant to capture the effects of differences in input prices across sectors. As discussed above, the data from Ghana contain many missing values for the value of capital, and so for it the terms in Equation 2.6 that involve the logarithm of capital are excluded. Estimates of these basic cost functions are shown in Table 2.7. The top panel shows estimates for formal firms (or, in the case of Ghana, firms with ten or more employees), and the bottom panel shows results for the full set of firms. The estimates for the three countries are quite similar, and are also similar for the formal and full samples. In all specifications the sectoral dummy variables are jointly insignificant.8 The results for the simplest specification, in which the logarithm of costs is regressed on the logarithm of output and a set of sector dummies, indicate that the underlying production technology displays constant or slightly increasing returns to scale. Variation in output explains a high fraction of variation in costs, with R squares that exceed .92 for all countries. Including the logarithm of output squared, the measures of capital, and the interaction of capital and output does not greatly improve the fit of the cost functions. The cost function shown in Equation 2.6 was modified to include variables that are likely to affect A. As discussed above, A is meant to capture the effects of skills and determinants of technology inflows on costs. We begin by including measures of the skill level of the firm’s owner or, if no owner was present, the skill level of the firm’s manager. For Kenya and Ghana the education levels of owners and the average education of managerial employees were coded in the same way, and we construct dummies that indicate whether the owner (or managerial employees) had no schooling, primary schooling, secondary schooling or university training. We also experimented with using more finely detailed educational categories that made use of the information on whether owners/employees had received occupational or vocational training. These results are not reported, but were essentially no different from the results that are discussed below. For Zimbabwe the variable pertaining to the average education of managerial employees was miscoded, and so for firms without owners we substitute the education level of the general manager or, if there was no general manager, the production manager. There are only three firms in Zimbabwe

64 The Industrial Experience of Tanzania Table 2.7 Basic cost functions for formal sector firms or (for Ghana) large firms (absolute values of t-statistics in parentheses) Kenya Wood Textiles Metal ln(q)

–0.11 (0.92) –0.03 (0.23) 0.06 (0.48) 0.915 (44.94)

ln(q)2 ln(k) ln(k)2 ln(q)ln(k) Obs 125 R2 .95 Full sample of firms Formal 0.35 (3.56) Wood –0.18 (1.59) Textiles –0.10 (0.85) Metal –0.08 (0.70) ln(q) 0.90 (47.93) ln(q)2 ln(k) ln(k)2 ln(q)ln(k) Obs R2

172 .96

Zimbabwe –0.09 (0.70) –0.02 (0.20) 0.04 (0.32) 0.75 (5.68) 0.02 (1.57) 0.24 (1.65) –0.01 (0.70) –0.01 (.43) 125 .95

–0.23 (1.55) 0.01 (0.08) –0.02 (0.12) 0.88 (43.20)

0.28 (2.15) –0.17 (1.50) –0.09 (0.79) –0.09 (0.79) 0.78 (9.94) 0.022 (2.03) 0.10 (1.54) 0.003 (0.33) –0.02 (1.11) 172 .96

0.32 (2.40) –0.11 (0.83) 0.04 (0.42) 0.07 (0.56) 0.89 (46.29)

142 .93

170 .96

Ghana –0.22 (1.60) –0.04 (0.38) –0.01 (0.09) 0.42 (3.87) 0.05 (1.90) 0.37 (3.65) –0.01 (.73) –0.02 (0.36) 142 .94

–0.32 (1.58) –0.18 (0.96) –0.18 (0.95) 1.00 (27.10)

0.31 (2.26) –0.06 (0.51) 0.03 (0.27) 0.04 (0.30) 0.69 (9.46) 0.06 (3.06) 0.13 (2.69) 0.016 (1.42) –0.06 (2.09) 170 .96

0.12 (0.72) –0.14 (0.73) –0.14 (0.77) 0.02 (0.09) 1.109 (25.78)

80 .93

132 .92

–0.32 (1.56) –0.15 (0.79) –0.19 (0.98) 1.33 (6.45) –0.022 (1.64)

80 .93 0.10 (0.65) –0.11 (.55) –0.11 (0.63) 0.02 (0.12) 1.19 (8.21) –0.01 (1.17)

132 .92


with owner/managers with no education, and so we combine the ‘no schooling’ and ‘primary education’ categories. For all countries we experimented with altering the definition of ‘skills’ in a variety of ways, for example by using the education level of managerial employees and only using the owner’s education if no managerial employees were present. Doing so had no effect on the results. The first columns of Tables 2.8a, 2.8b and 2.8c show estimates of cost functions when the ‘skill’ measures are included. As the coefficients on the variables shown in Table 2.7 are largely unchanged, only coefficients for the skill variables are reported. The results are easily summarised. In no country are higher managerial skills associated with lower costs, given output. There is one anomalous result for formal sector firms in Ghana (Table 2.8c): firms with educated owners appear to have higher costs than those with no education. However, the three educational dummy variables are not jointly significant. In the second column of the three tables we include variables that measure involvement in foreign markets: the percentage of raw materials imported, the percentage of output exported and the percentage of foreign ownership. If firms with more involvement in foreign markets have more access to lower-cost technology, then these variables should be negatively correlated with costs. However, interpreting the parameter estimates requires caution: it could be that firms whose costs are initially lower export their output. This could also produce a negative correlation between exports and costs, and it would be incorrect to interpret the results as reflecting the effects of the rate of technology transfer on costs (Clerides et al., 1999).9 However, the results indicate that there is no correlation between involvement in foreign markets and costs in any of the countries, and the hypothesis that the variables are individually and jointly insignificant cannot be rejected. The lack of a statistically significant relationship between these variables and costs may not be surprising given that very few of the firms have any involvement in foreign markets. If only high-skill firms are able to take advantage of new technologies from abroad, the Nelson-Phelps view, there should be a negative correlation between costs and an interaction term between skills and the variables reflecting access to foreign markets. The third column of Tables 2.8a, 2.8b and 2.8c show estimates of cost functions when these interactions are included. Overall, the results provide very little support for the hypothesis that high skills combined with access to foreign markets result in reductions of unit costs. With only one exception, the coefficients on these interactions are insignificant. The exception is for Kenya, for which firms with university-trained owner/managers that export higher fractions of their output have lower costs, all else being equal. One should not, however, make too much of this finding. First, there are only 17 formal

Cost functions with skill measures included, Kenya

66

Table 2.8a

Sample means

Kenya: formal firms (obs = 125) Owner/manager’s education: Primary school Secondary school University

–0.038 (0.32) –0.062 (0.47) –0.056 (0.47)

% raw materials imported % output exported % foreign ownership

–0.017 (0.15) –0.065 (0.49) –0.041 (0.33)

–0.053 (0.45) –0.009 (0.07) 0.032 (0.23)

0.0006 (0.46) –0.0034 (1.78) –0.0019 (1.16)

0.0000 (0.04) 0.0014 (0.59) 0.0001 (0.04)

26.8 8.9 10.2

0.0010 (0.43) –0.0115 (3.20) –0.0023 (0.72)

16.5 4.3 6.5

Owner/manager education is university, times: % raw material imported % output exported % foreign ownership R2

.95

.95

0.32 0.18 0.31

.96

Kenya: all firms (obs = 172) Owner/manager’s education: Primary school Secondary school University

–0.054 (0.56) –0.005 (0.05) 0.016 (0.15)

% raw materials imported % output exported % foreign ownership Owner/manager education is university, times: % raw material imported % output exported % foreign ownership 2 R

.96

–0.034 (0.35) 0.004 (0.03) 0.039 (0.34)

–0.042 (0.42) 0.033 (0.26) 0.076 (0.60)

0.0000 (0.02) –0.0026 (1.26) –0.0020 (1.27)

–0.0008 (0.47) 0.0027 (1.09) –0.0021 (0.84)

22.3 6.5 8.2

0.0017 (0.74) –0.0123 (3.33) 0.0011 (0.32) .97

12.4 3.1 5.3

.96

0.31 0.15 0.24

Notes: Absolute values of t-statistics in parentheses. Also included in cost functions were the logarithms of output and capital, each of these terms squared, and an interaction term between output and capital (as in Table 2.7).

Table 2.8b

Cost functions with skill measures included, Zimbabwe Sample means

Zimbabwe: formal firms (obs = 142) Owner/manager’s education: Secondary school University

0.0216 (0.15) –0.0373 (0.25)


.0107 (0.08) –0.0321 (0.21)

–0.041 (0.29) 0.102 (0.63)

0.0004 (0.24) 0.0030 (1.46) 0.0011 (0.80)

0.0024 (1.19) 0.0033 (1.33) 0.0032 (1.62)

19.2 13.0 14.0

–0.0044 (1.50) –0.0025 (0.61) –0.0041 (1.58)

8.7 5.7 8.0

Owner/manager education is university, times: % raw material imported % output exported % foreign ownership .95

.95

.95

0.0500 (0.44) –0.0256 (0.20)

0.0376 (0.33) –0.0204 (0.16)

0.0008 (0.01) 0.1104 (0.77)

0.0005 (0.33) 0.0031 (1.48) 0.0012 (0.86)

0.0025 (1.21) 0.0034 (1.37) 0.0030 (1.50)

16.0 11.5 11.8

–0.0043 (1.45) –0.0023 (0.56) –0.0037 (1.36)

7.2 4.8 6.7

Zimbabwe: all firms (obs = 170) Owner/manager’s education: Secondary school University % raw materials imported % output exported % foreign ownership Owner/manager education is university, times: % raw material imported % output exported % foreign ownership R2

.96

.96

.97

0.46 0.37


R2

0.48 0.42

Cost functions with skill measures included, Ghana

68

Table 2.8c

Sample means

Ghana: large firms (10 or more employees) (obs = 80) Owner/manager’s education: Primary school Secondary school University

0.043 (0.22) –0.087 (0.42) 0.427 (1.71)


0.038 (0.19) –0.091 (0.44) 0.426 (1.70)

0.044 (0.22) –0.135 (0.61) 0.340 (1.22)

–0.0006 (0.29) –0.0023 (0.78) 0.0044 (1.36)

–0.0041 (1.29) –0.0034 (0.77) 0.0099 (1.71)

20.0 6.7 10.6

0.0059 (1.40) 0.0022 (0.38) –0.0070 (1.14)

11.5 3.7 6.6

Owner/manager education is university, times: % raw material imported % output exported % foreign ownership R2

.93

0.39 0.33 0.16

.93

.94

0.386 (2.24) 0.215 (1.15) 0.519 (2.07)

0.389 (2.23) 0.190 (0.96) 0.465 (1.72)

–0.0002 (0.09) –0.0027 (0.75) 0.0027 (0.80)

–0.0019 (0.57) –0.0038 (0.72) 0.0045 (0.92)

12.8 4.0 7.2

0.0035 (0.75) 0.0024 (0.35) –0.0037 (0.54)

7.1 2.3 4.0

Ghana: all firms (obs = 132) Owner/manager’s education: Primary school Secondary school University

0.389 (2.29) 0.216 (1.16) 0.526 (2.12)

% raw materials imported % output exported % foreign ownership Owner/manager education is university, times: % raw material imported % output exported % foreign ownership R2

.92

.92

0.45 0.27 0.11

.92

Notes: Absolute values of t-statistics in parentheses. Also included in cost functions were the logarithm of output and the logarithm of output squared (as in Table 2.7).


sector and one informal sector firms with university-trained owner/managers who export any of their output. Second, as discussed above, without better information on exogenous determinants of access to export markets, it is not possible to know whether low-cost firms with educated owners are better equipped to sell their output abroad, or whether involvement in foreign markets makes it easier for more educated owners to produce using lower-cost methods.

5.

Interpretation of the results

There are a number of potential explanations for the absence of significance of the components of technological ability (A) in countries in which it is generally believed that a constraining factor is human capital and technological infrastructure. Consider the specification that assumes the components of A are multiplicative elements in the cost function. First, the relatively non-competitive environment may lead to limited efforts to utilise those skills that are present. Second, even where skilled managers are present or R&D takes place, these may simply be devoted to efforts to redress the productivity-depressing effects imposed by the protectionist regimes. For example, engineers may spend their time adjusting equipment to cope with low-quality intermediate inputs that are locally available. While this may lead to higher productivity than in their absence, they are mainly redressing the productivity-depressing effects of policy, and the net effect of their effort is likely to be small. Surveys of firms in Africa repeatedly find that firms frequently encounter difficulties with arbitrary foreign exchange allocations, zoning, building and licensing codes, and uncertain quality of telecommunications and electricity. For example, the surveys discussed in this paper indicate that the availability of foreign exchange is a special problem among formal sector firms. In Kenya, 35.4 per cent of firms indicate they have ‘very severe problems’ regarding delays in obtaining foreign exchange, and 41.6 per cent report very severe problems in foreign exchange availability. Of formal firms in Zimbabwe, 27 per cent firms report very severe problems with delays in obtaining foreign exchange, and 52.1 per cent report very severe problems with foreign exchange availability. Reports of problems with infrastructure are also common. With such a large set of governmentimposed handicaps affecting them, it would be surprising if their technological abilities, however measured, would reveal themselves in a systematic fashion across firms. Some companies may use technologists to address regulations, some to correct infrastructure deficiencies, as in Nigeria (Lee et al., 1990), and others to improve productivity. Only after liberalization are the benefits from improved technological capacity likely to be measurable and lead to a significant reduction in costs.


The firms in our sample have not been challenged by newer technologies where the payoffs to domestic skills are likely to be largest. Simultaneously the failure of agricultural productivity to grow has limited the economywide growth in per capita income. Thus the change in sectoral structure engendered by different income elasticities among industrial goods has not had an effect. The absence of a need to alter production methods or sectoral mix has reduced the payoff to education. Even if there were an inflow of technology, and high education made its absorption potentially productive, the absence of a competitive environment limits these gains. Much of the empirical testing of the role of education in adjusting to technical change has been in the agricultural sector, which is usually much more competitive than the industrial sector, particularly in developing countries (see the references in footnote 3). In contrast, in Ghana, Kenya and Zimbabwe the survival environment is lax. Tariffs and quotas protect industrial firms from external competition, while internal rivalry is reduced by the relatively small number of firms producing a specified product. The shift in short-run supply curves induced by abnormal profits in competitive markets is not forthcoming, as there are a limited number of firms capable of entering most industries. Firms possessing technical abilities may choose not to utilise them to increase productivity or reduce costs. The abilities of skilled personnel may be devoted to insuring the existence and exploitation of profit possibilities made possible by tariffs, quotas and cheap credit. Firm-level efforts to improve productivity are the result of an income and a substitution effect. High levels of protection, ceteris paribus, increase the firm’s profits per unit of effort. If the foregone ‘leisure’ or easy life is a normal good, protection will tend to reduce cost saving efforts. However, higher levels of effective protection increase the opportunity cost of forgoing additional output from productivity-augmenting activities. Thus, the impact of firm-level abilities on productivity in the presence of protection is ambiguous. An absence of an association between firm skills, however measured, and productivity levels among firms is not an indication that such skills do not provide the potential to raise productivity. Rather, such a capacity may not be deployed in pursuit of greater cost reduction but in the search for greater rents or an easier life.

6.

Conclusions

The absence of a significant private payoff to technological capability may be surprising given the frequently heard statement that the binding constraint on African industrial development is an insufficient supply of technologically capable manpower. This is also the implication of the many cross-country studies that find that an important source of low growth in per capita income in Africa is the low level of human capital, however


measured. Our results are not necessarily in conflict with this view. They should be interpreted as indicating that in the non-competitive industrial sector with very little inflow of new technology the contribution of technological abilities, however measured, is limited. If there were a liberalisation of the economy that generated greater competition or if there were an acceleration of export growth, permitting the import of new technology embodying inputs, the contribution of local skills would become more significant in raising output. The experience of other countries also suggests that as the economy becomes open to flows of international knowledge, whether through informal transfers from purchasers of exports or through technology licences, the technological capacity of local industry becomes important.10 The policy implications of the empirical analysis are clear. Absent greater prospective competition, continued efforts to develop high level industrial skills may be wasteful. The conundrum is that the absence of these skills may limit the benefits to the industrial sector from future liberalisation, and lead to a weak supply response to improved incentives.

Notes *

1 2

3

4

5

University of Pennsylvania and Princeton University. We are grateful to Tyler Biggs of the Africa Technical Department of the World Bank for making available the data upon which this paper is based. We received helpful comments from participants at a symposium at the Eindhoven Technical University. A. Szirmai provided many helpful suggestions. Part of the research was financed by the World Bank’s Regional Program on Enterprise Development. Pack gratefully acknowledges partial support from the University of Pennsylvania Research Foundation and the World Bank Development Research Group. Meier and Steel (1989) contains a good selection of representative views. For a general discussion of the technological requirements for industrial development see Lall (1990). Pack (1993) provides an overview of industrial development in Africa and extensive references to the literature. For studies demonstrating the interaction of education and technological change in agriculture see Welch (1970), Foster et al. (1995a, 1995b), Rosenzweig (1995). The measures described in the text and shown in Table 2.1 could be improved. For example, standard international trade data contained in the UN Yearbook of International Trade Statistics allow a disaggregated breakdown of imports that would allow more precision with respect to both the equipment and intermediate imports. However, the latest date for which these are available for the African economies is 1990, two years before the firm-level surveys were carried out. Thus we have cited the aggregate data generated by the World Bank. In almost all cases the sectoral shares of the 28 sectors, not reproduced in Table 2.2, show little change, with the exception of the Zimbabwean food and beverage sectors. The greater summary measure in Zimbabwe compared to Ghana and Kenya is largely due to the expansion in the share of the beverage sector, a sector in which the technology employed has not changed.

72 The Industrial Experience of Tanzania 6 It should be noted that the use of sample weights together with small sample sizes can yield a misleading picture of the activities of these firms. For example, the Kenyan numbers indicate that 58.7 per cent of wood-sector workers are in the ‘miscellaneous’ category that includes ‘carvings, camping equipment, and automobile trim.’ This is due to one 70-person firm that makes camping equipment, and which was given a sample weight four times the sector average. Likewise, the result that 41 per cent of food and beverage workers make animal feed and hay is due to one poultry-feed firm with 420 workers and a sample weight three times the sector average. 7 The survey does not ask whether the owner and manager are the same person, and this might account for the similarity between the education of owners and managers. 8 We also experimented with including dummy variables for the city in which firms are located, and these were also insignificant. The results are consistent with there being neither regional nor sectoral variation in input prices. 9 One of the benefits of panel data would be the ability to address the question of whether efficient firms are more successful exporters or whether exporting generates learning. 10 See the East Asian Economic Miracle (World Bank, 1993d), chapter 6.

3 Measuring Manufacturing Performance in Tanzania: GDP, Employment and Comparative Labour Productivity, 1961–95 Adam Szirmai, Menno Prins and Wessel Schulte*

1

Introduction

Tanzania is a late late-comer to the process of industrialization: the first steps towards industrialization were taken after World War II. These took the form of processing for export markets. The expansion of manufacturing activities for the local market started in the mid-1950s. This late start is illustrated by the modest numbers of manufacturing establishments in Table 3.1. Of the establishments with ten or more persons engaged in operation in 1961, only 101 predated 1945. Including small scale establishments, the number of establishments in 1933 was 321 (Silver, 1984, p. 42). According to official figures, the manufacturing sector contributed 3.5 per cent to GDP at factor cost in 1961 (Central Statistical Bureau, 1964b). 1 Since independence, the number of establishments increased sharply to 569 by 1965, peaking at 1282 in 1978 and subsequently declining to 886 establishments in 1989. In 1961, the large- and medium-scale manufacturing sector (10+) employed 22,000 persons. Manufacturing employment increased to 110,000 in 1978, in which year the sector registered a peak share of 12 per cent of GDP. In 1989 employment in medium- and largescale manufacturing had increased to 124,000 persons, while its share in GDP dropped to 8 per cent. One may conclude that Tanzanian industrialization started almost from scratch in the mid-1950s. The aim of this article is to chart the course of industrialisation since independence, by contributing to the statistical measurement of manufacturing performance. To this end, we will scrutinize the available statistics on medium- and large-scale establishments (with ten or more persons engaged), for which the most reliable data are available. This article summarizes and integrates the results of two bodies of research on 73


Indicators of Tanzanian manufacturing performance, 1946–89a

Year

Number of establishments In production (number)

1946 1957 Independence (1961) 1965 1978 1989

101 b 293 b 380 b 569 1 282 886

Increase c (annual average) 17 22 38 51 –33

Persons engaged (thousands)

22 28 110 124

Value added (% of GDP)

3.5 9.2 12.3 8.4

Source: Number of establishments 1946–65 from Survey of Industries 1965 (Central Statistical Bureau, 1966), table 5; persons engaged and value added 1961 from Skarstein and Wangwe (1986), p. 79; persons engaged 1978 and 1989 from Census of Industrial Production 1978 (Bureau of Statistics, 1995b) and Census of Industrial Production 1979 (Bureau of Statistics, 1994a); value added 1978 and 1989 from Bureau of Statistics (1995c), table 7.2. Notes: a All data refer to establishments with 10 or more persons engaged, except for value added data which refer to the entire manufacturing sector. Data until 1965 include activities in sisal decortication and motor vehicle repair, which are excluded in 1978 and 1989. b Number of establishments which were still in production in 1965. c Average annual increase in the number of establishments.

Tanzanian manufacturing performance: a reconstruction of nominal and real output and employment in Tanzanian manufacturing (Prins and Szirmai, 1998) and a benchmark comparison of levels of real output and labour productivity in manufacturing between Tanzania and the world productivity leader, the USA, for 1989 (Szirmai and Schulte, 1998).2 The article has the following objectives: • to present consistent newly revised time series of 10+ manufacturing value added at current prices for the medium- and large-scale manufacturing sector as a whole and for major branches of manufacturing (Sections 2 and 3) • to present consistent indexes of industrial output and labour productivity for major branches of manufacturing and total manufacturing, for the period 1961–95 (Section 4) • to present a benchmark comparison of real output and labour productivity in Tanzania vis-à-vis the USA in 1989, using the methodology developed in the International Comparisons of Output and Productivity project (ICOP) (Section 5)

Measuring Manufacturing Performance 75

• to combine US and Tanzanian time series and the 1989 benchmark, in order to measure comparative trends of Tanzanian labour productivity (Section 6) • to discuss the implications of the newly constructed series for our understanding of growth, structural change and productivity in Tanzanian manufacturing (Section 6)

2

Data sources for time series on 10+ manufacturing

The main data sources for the manufacturing sector in Tanzania are the annual survey of industrial production (Bureau of Statistics, ASIP, various issues), the 1978 and 1989 censuses of industrial production (Bureau of Statistics, 1985b, 1993a, 1994a) and the quarterly surveys of industrial production covering establishments with 50 or more persons engaged (Bureau of Statistics, QSIP, various issues). Additional sources are input-output tables (Bureau of Statistics, 1986), informal sector surveys (Planning commission, 1991, 1995), and price indices and economic surveys (Bureau of Statistics, 1972, 1977, 1982, 1985a). Our main source for nominal value added in 10+ manufacturing are the surveys and censuses. Figure 3.1 summarizes these sources and compares their coverage.

Figure 3.1

Comparison of coverage of statistical sources

Coverage Census

a

I/O table

b

Census

IS-

IS-

surveyd

surveyd

1+

Census

5+ ASIP

10+ 1961

1965–74

Ec. surveyc 1976 1978

ASIP 1979–88

ASIP 1989 1990 1991 1994/95

Year

Notes: ASIP = Annual Survey of Industrial Production. IS-survey = Informal Sector Survey. I/O table = Input–Output table. Ec. Surv. = Economic Survey. a 1961 Census results are not comparable to later industrial surveys/censuses. z The coverage of the I/O table is not based on a canvassing procedure of 1+ establishments, but on estimates based on data such as the ‘final demand’ for certain manufacturing products. Estimates thus attempt to reflect 1+ coverage. c Although no ASIP was published in 1975–77, data were collected through the ASIP questionnaires and the aggregates were published in the Economic Survey. For 1978, 10+ estimates have been published in the Economic Survey, since census results did not became available till 1985. d The Informal Sector Surveys covered establishments with 5 or fewer paid employees not using high technology. This does not correspond to our categorization of 1–4 persons engaged.


3

Reconstruction of nominal 10+ GDP, 1961–95

Careful analysis of the primary sources identifies three main sources of error (Prins and Szirmai, 1998): (a) undercoverage of 10+ establishments in the directory of establishments (b) treatment of non-response (c) conceptual errors. Adjustments for these sources of error are made for two periods: 1978–90 and 1965–78. Additional sources of error, which will not be discussed in this article, are errors in the raw data of the 1989 census, the changing categorization of activities in economic sectors and inadequate treatment of the small-scale and informal sector in the national accounts, using rules of thumb (see Prins and Szirmai, 1998, sections 2.2.2, 2.4 and 2.5). The issue of changes in categorization has been dealt with by aggregating manufacturing sectors into six major branches. The informal sector merits treatment in a separate article (see also Chapter 14). This paper limits itself to medium- and largescale enterprises with ten or more persons employed. 3.1

Coverage

The framework for data collection is the directory of establishments maintained at the industrial section of the Bureau of Statistics. The coverage of medium- and large-scale establishments in the directory varies over time and does not adequately reflect changes in the real volume of activities. There are particularly large differences in coverage between survey years and census years (see Prins and Szirmai, 1998, table 2.2). For instance, the directory for the 1974 survey lists 499 enterprises. This jumps to 1282 in the census year 1978. In the survey year 1988 the directory includes 711 establishments; this jumps to 866 in the census year 1989. In some cases establishments, such as furniture making and tailoring, have deliberately been omitted from the annual surveys (omitted establishments). In most cases, the differences are unintentional (undercoverage). 3.1.1

Adjustments for undercoverage

For the period 1978–90 coverage adjustments have been based on analysis of a sample of 102 50+ establishments drawn from the 1989 census. For these establishments files were available with the original survey and census returns for all years. These files also indicated which years an establishment was in operation, but was not covered. Coverage rates are calculated as percentages of the establishments in operation in a given year. These coverage rates were used to make upward adjustments of value added.


For the period 1965-78 we made a rough adjustment. We compared gross output from the 1978 economic survey, which is consistent in coverage with the surveys prior to 1978 with gross output from the 1978 census. We used this ratio to make an upward adjustment for value added in the pre-1978 years. 3.1.2

Adjustment for omitted establishments

Value added in omitted establishments (1965-78) was calculated as follows (see Prins and Szirmai, 1998, appendix D). For 1966, we calculated omitted value added as a residual, by subtracting the following from total 5+ value added in the national accounts: unadjusted 10+ value added, estimated value added for non-response (see below) and estimates for 5–9 value added. The 1966 proportion of omitted to total value added was used to adjust value added for the whole period. 3.2

Treatment of non-response in official statistics

Between 1961–71 the numbers in the ASIP reflect responding establishments only. No adjustments have been made for non-response. Between 1972 and 1974 the ASIP data have been adjusted for non-response. In 1976, the data in the input-output table have been adjusted for non-response. For the period 1978–90 no published information is available on treatment of non-response. Interviews within the bureau of statistics revealed that the methodology for dealing with non-response was that of simple repetition. If an establishment does not respond in a given year, one takes the previous year’s figures. If an establishment does not respond for several years, one takes the figures from the last year in which it responded. 3.2.1

Adjustments for non-response, 1978–90

For the calculation of the effects of non-response on value added, we have examined records of a sample of 102 50+ establishments from the 1989 census. The aim of this exercise is to make an estimate of the understatement of value added of non-responding establishments, resulting from the method of simple repetition of value added. Our basic approach is to inflate the repeated value added figures with the consumer price index, to account for rapid price changes.3 For the sample it is known for each year between 1978 and 1990 whether an establishment is a non-respondent in that year, and, if so, for how many successive years it has been a non-respondent. 4 The establishments in the sample are subdivided into three size classes (50–99, 100–499, 500+). Within each of the three size classes distinguished we can express the degree of underestimation of value added in year t (UEt in percentages) as a sum of underestimations (ue in percentages) caused by subsets (or cohorts) of establishments of which nominal value added data are repeated for a given number of successive years, because of non-response. For one


cohort the degree of underestimation equals the proportion of all establishments which are not responding for a given number of years (for example, two successive years), weighted by the price indexes over the period of successive non-response.5 In formula, the degree underestimation of value added in a size class is calculated as follows: UEt =

St

∑ ue

t ,i

and uet ,= i

i= 1

 tnrt nrt (t − i )  CPI t − 1  Nt tnrt  CPI t − i 

(3.1)

UEt is the underestimation of value added due to non-response, expressed as a percentage of unadjusted value added Nt is the number of establishments in the sample which are in production in year t tnrt is the total number of non-respondents in the sample in year t St is the maximum number of successive years value added is repeated for a given establishment in year t, nrt (t–i) is the number of non-respondents of which value added has been repeated since year t–i CPIt is the consumer price index in year t The estimate for the overall underestimation of value added was calculated as a weighted average of the degrees of underestimation per size class, weighted with value added weights per size class from the census.6 3.2.2

Adjustments for non-response, 1965–74

The records of the annual surveys enabled us to make estimates for value added of non-responding establishments for the years 1965–71. For the years 1965–71 we have calculated non-response rates and calculated non-response value added at branch level (see Prins and Szirmai, 1998, tables E-2 and E-3). 3.3

Conceptual errors

A change in the questionnaire design for the ASIP, gradually introduced from 1980 onwards, gave rise to errors in the calculations of manufacturing value added. These flaws were discovered during the in-depth analysis of the results of the 1989 census, which was based on the same flawed questionnaire design. Owing to an ambiguous definition of an intermediate input category labelled all other costs, responding establishments were inclined to allocate huge amounts of interest payments to intermediate inputs. Since interest costs are a component of value added rather than intermediate inputs, value added is wrongly defined and, as a consequence, substantially underestimated. Examination of the cost structure of 10+ establishments in the census revealed that some establishments have


enormous amounts of costs allocated to the residual all other costs category. In particular they have allocated large amounts of interest payments to this category, where they do not belong.7 This conclusion is based on the following reasoning. The questionnaire distinguishes 16 categories for production costs. One of these categories is bank charges and insurance paid, which explicitly excludes interest costs. However, the instructions for the category all other costs do not indicate that interest costs should be excluded. They explicitly exclude labour costs, sales taxes, corporate taxes, excise duties and depreciation, but there is no mention of interest at all. In Figure 3.2 we can see how important all other costs are compared to most other intermediate input cost categories. From an analysis of firm level-data for 175 large firms accounting for 96 per cent of ‘all other costs’, we conclude that the all other costs category indeed includes interest payments which should have been allocated to value added. Some other components of value added have also been erroneously included in all other costs, because they were not identified elsewhere in the questionnaire. These cost categories are bad debts, directors’ fees and donations. A less important conceptual error is that profits from sale of fixed assets should not have been included in the calculation of gross output. As gross output is too high, gross value added at factor cost is overstated by 0.3 billion TSh. Figure 3.2

Cost structure, 1989 census

50 40 30 20 10

All other costs

Miscellaneous

Professional & office exp.

Bank charges & insurance

Cost of resales

Transport

Industrial services

Fuel, Lubr., Electr. & Water

Chem. & Pack. materials

0 Raw materials

Intermediate inputs (billion TSh)

60


3.3.1

Adjustments to value added, 1978–90

The most important conceptual adjustment to the census data has been achieved by reclassifying a portion of the cost category all other costs from intermediate inputs to value added. Our estimate of the total amount of non-intermediate inputs incorrectly allocated to all other costs was 11 billion out of a total of 17 billion TSh. in this category (Prins and Szirmai, 1998, tables A-1 and A-4). Reallocation of these cost categories to value added resulted in a downward adjustment of intermediate inputs from 98 billion TSh to 87 billion TSh. Correspondingly, gross value added was adjusted upwards by 11 billion TSh. Thus, huge adjustments have been made in 1989 value added for all branches (except branch 39). Total value added has been adjusted upward by 51 per cent. The most notable adjustment is the 241 per cent increase in value added for ISIC category 32 (textiles and leather), which increases its share in total value added from 7.9 per cent to no less than 17.8 per cent! The questionnaire used for the 1989 census had gradually come into use since 1980. We have made an estimate of the rate of adoption of the new questionnaire and have backcast and forecast the 1989 adjustments to value added and intermediate inputs at branch level. The questionnaire adoption rate was estimated from the sample of 50+ establishments taken from the 1989 database (see Prins and Szirmai, 1998, appendix B). For each establishment we were able to identify the year the ‘new’ questionnaire design was first used. The adjustments made in the 1989 census have been extrapolated to 1980–88 and to 1990, multiplying 1989 adjustments to gross output and intermediate inputs with the adoption rate of the new questionnaire for each year.

3.3.2

Estimates of nominal value added 1961–65 and 1991–95

Although no annual surveys were held between 1961 and 1964, we do have data for nominal value added in 1+ manufacturing from 1960 to 1966 (OECD, 1971, table 2-4). Assuming that the growth rate of nominal value added for 10+ manufacturing equals that of 1+ manufacturing , we have applied the growth rates of the 1+ series to backcast 1965 10+ value added to 1961. No annual survey has been published since the 1990 survey. Therefore, we use a real output index and a price index to extrapolate 1990 value in order to arrive at value added estimates in current values for the years 1991–95. The price index has been derived from two sources: the consumer price index (CPI) (1990–92) and the producer price index (PPI) (1992–95).8 The quantity index used is the improved index of industrial production, which will be discussed in Section 4.

Table 3.2

Level adjustments to nominal MVA, 1960–95

Unadjusted MVA

OC-adj.b

NR-adj.c

OE-adj.d UC-adj.e (Millions TSh)

Change of MVAa due to adjustment for: Extrapolated

OC-adj.b

NR-adj.c

OE-adjd (%)

UC-adj.e

Total

32 35 34 30 33 34 33 35 36 33 50 45 29 34 3 8 10 14 21 88 12

52 49 45 38 39 37 36 35 36 33 50 45 29 34 3 11 25 47 83

243 272 276 350 267 295 319 378 475 561 643 806 914 1 157 1 246 1 480 2 075 2 186 2 842 2 927 2 891 3 108 3 204 3 620 4 417

282 297 323 390 497 571 657

2 842 2 927 2 900 3 148 3 333 3 994 5 269

2 842 2 985 3 279 3 983 4 862 5 706 7 241

307 327 344 402

405 441 460 522 659 766 874 1 087 1 244 1 537 1 863 2 143 2 678 2 926 2 926 3 238 3 622 4 555 5 873 6 812 8 094

6 1 1 3 5 2 2

0 0 0 1 4 10 19

0 2 13 27 46 43 37

9 10 6 3

19

83

81

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1978* 1979 1980 1981 1982 1983 1984

MVAa after adjustments for:

82

Table 3.2

(continued)

Unadjusted MVA

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

5 112 6 412 11 062 11 358 21 474 23 956

MVAa after adjustments for: OC-adj.b 6 373 8 525 14 545 15 866 32 443 37 478

NR-adj.c 8 915 11 714 20 790 25 849 32 653 38 681

OE-adj.d UC-adj.e (Millions TSh)

Change of MVAa due to adjustment for: Extrapolated

9 812 12 360 21 679 26 642 32 653 38 681

OC-adj.b

NR-adj.c

25 33 31 40 51 56

40 37 43 63 1 3

OE-adjd (%)

UC-adj.e 10 6 4 3 0 0

Total 92 93 96 135 52 61

48 128 53 473 63 535 82 159 98 818

Sources: Unadjusted data: 1965–74: ASIP; 1975–78: Economic Survey; 1978: Census; 1979–1988, 1990: ASIP; 1989: 1989 Census; adjusted data: Prins and Szirmai (1998) (appendixes A–G). a Change calculated as the percentage change of adjusted value compared to the pre-adjustment value. b Other costs based adjustment. c Non-response based adjustment. d Omitted establishments based adjustment. e Undercoverage based adjustment.

Table 3.3

Adjusted nominal value added, 1961–1995 (thousands TSh) Unadjusted

Adjusted ISIC Branch

131 318 135 661 12 627 180 020 202 029 235 746 286 292 310 905 334 000 390 272 403 249 477 801 627 467 694 642 800 919 774 318 1 134 567

32 Textiles & leather

93 178 144 031 152 849 143 557 193 533 278 753 300 707 368 850 448 992 481 125 572 779 675 596 900 910 892 085 970 942 1 227 974 1 345 426

33/34 Wood, furnitures & fixtures, paper, printing & publishing

42 032 42 794 45 803 47 739 55 176 70 593 64 198 61 144 89 133 122 775 150 762 177 831 213 552 287 409 323 294 359 672 494 176

35 Chemicals, petroleum, rubber & plastics

25 154 26 787 52 882 53 640 62 789 65 486 81 270 121 236 131 646 249 184 286 029 337 360 339 244 347 488 471 689 558 216 676 847

36 Non-metallic minerals

5 916 11 749 21 394 23 361 28 923 30 090 33 817 49 636 41 022 48 491 96 173 60 140 62 566 97 185 99 560 120 741 189 175

37/38/39 (Basic) metals, machinery & equipment and other manufacturing

107 189 79 507 65 912 73 202 116 294 84 841 107 580 174 880 199 329 245 445 354 489 414 570 534 159 606 824 571 271 581 366 714 357

3 Total manufacturing

243 439 271 745 275 520 350 061 404 787 440 530 460 466 521 519 658 743 765 510 873 864 1 086 650 1 244 123 1 537 293 1 863 480 2 143 298 2 677 899 2 925 632 3 237 676 3 622 286 4 554 547


266 701 295 162 318 625 378 324 475 411 560 616 642 871 806 328 914 327 1 156 652 1 245 622 1 480 345 2 074 758 2 842 316 2 927 333 2 890 897 3 108 206

83

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981

31 Food, beverages & tobacco

84

Table 3.3

(continued) Unadjusted

Adjusted ISIC Branch









1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

2 026 862 2 067 669 2 582 369 3 148 163 3 099 117 5 157 851 7 666 631 12 441 829 17 149 464 21 232 538 25 522 856 30 743 364 43 202 958 53 931 263

1 198 534 1 812 035 2 022 511 2 238 412 2 577 458 5 331 612 6 865 440 5 826 626 5 597 233 6 459 244 7 033 218 7 567 179 8 666 806 12 257 957

499 515 549 270 760 200 870 414 1 107 314 1 546 226 2 044 644 3 017 332 3 755 861 4 328 090 4 138 842 7 639 361 6 716 699 8 543 564

982 171 889 370 1 228 673 1 398 033 2 189 489 4 903 191 4 960 480 4 901 342 5 432 448 6 944 192 7 650 103 8 002 358 10 917 273 12 284 511

246 378 326 831 170 774 384 734 917 068 1 237 203 1 359 874 1 422 835 1 176 377 2 059 175 1 809 982 2 427 974 2 781 754 3 770 135

919 253 1 167 309 1 329 517 1 772 483 2 469 293 3 503 219 3 744 791 5 042 905 5 570 034 7 105 139 7 317 953 7 154 866 9 873 702 8 030 217

5 872 714 6 812 485 8 094 045 9 812 238 12 359 739 21 679 302 26 641 860 32 652 870 38 681 417 48 128 378 53 472 955 63 535 101 82 159 192 98 817 647

3 203 832 3 619 760 4 417 219 5 111 606 6 412 236 11 062 008 11 357 863 21 474 018 23 955 853

Source: Prins and Szirmai (1998).


3.4

Summary of adjustments

Table 3.2 provides a summary of all the adjustments to aggregate 10+ manufacturing value added. The adjusted figures for nominal value added by branch of manufacturing are reproduced in Table 3.3. In the right-hand columns of Table 3.2 adjustments are expressed as percentage changes relative to value added after previous adjustments. The overall level adjustments are substantial. The census year 1978 is the only year in which there have been but modest adjustments. The greatest upward adjustment to value added was made for 1988 (135 per cent). In 1990 the adjustment was 61 per cent. Over the whole period the adjusted series tend to be smoothed out, as jumps in coverage have been eliminated and the whole series have become more consistent.

4

New indices of industrial production

For most years official real value added series have only been constructed for total manufacturing. An exception is the index of industrial production (IIP), published in the quarterly survey of industrial production (QSIP) from 1985 onwards. Real value added series have been constructed at 1966 base prices (1960–80), 1976 prices (1964–94) and 1985 prices (1986–94). For 1976–84 an indirect approach has been followed, deflating current value added (1+) from the national accounts with a cost of living index of clothing and footwear (Bureau of Statistics, 1985a). For all other years an direct approach has been followed, in which quantity relatives have been weighted with base year value added shares. For the 1966 series no sources and methods are available. The 1976 series are based on quantity relatives for 14 commodities. The 1985 series are based on 120 commodities derived from the quarterly survey among 50+ enterprises. We constructed a new index 1965–95 based on the following principles: use of as much commodity information as possible, consistent application of the direct approach of weighting quantity relatives to calculate a Laspeyres fixed base weighted index, and construction of indexes for six branches of manufacturing as well as for total manufacturing. 4.1

IIP, 1965–85

For the construction of an improved index of industrial production as many quantity data as possible were collected for the period 1965–85 (see Prins and Szirmai, 1998, appendix H). To construct an IIP for a large time span it is preferable to rebase the index regularly. The choice of base years was mainly determined by the availability of quantity information and the availability of value added weights for 10+ establishments from the surveys and censuses. The following base years have been chosen: 1966 (coverage 16 commodities), 1970 (coverage 19 commodities), 1975 (coverage 25 commodities), and 1980 (coverage 33 commodities).


The method of constructing the IIP is to select commodities representing given industries and to weight the commodity quantity relatives with the value added of industries which the commodities represent. If no quantity relatives can be calculated for a given industry, this industry is combined with a related industry for which a quantity relative is available. That quantity relative is then weighted with the value added of the combined sectors. Since we construct a fixed base index, we need value added weights for base years. The following years have been selected as base years: 1966, 1970, 1975 and 1980. The weights are constructed as follows. In a few cases there is more than one commodity representing a four digit manufacturing industry. In such cases, so-called intra-industry weights are needed. For the years 1965–85 we have used the gross output values of commodities within an industry from the 1989 census as intra-industry weights.9 The four-digit industry indexes within a three-digit ISIC branch are weighted with their industry value added to get a three-digit branch index. Industry weights are taken from the ASIP, censuses and the 1976 input-output table. (see Prins and Szirmai, 1998, tables H-5 and H-6 for weights). Each three-digit branch index is weighted with its branch value added, to arrive at an index for six major branches of manufacturing and an index for total manufacturing. In the calculation of the index for total manufacturing, we use the adjusted nominal value added series presented in Table 3.3 as weights. 4.2

Silver’s chain index, 1965–72

In The Growth of Manufacturing Industry in Tanzania, Silver (1984) provides a sophisticated index of industrial production for the years 1965–72. Utilizing an impressive amount of data derived from the QSIP questionnaires, Silver constructs indexes for 38 subindustries conform the 1958 ISIC classification. The Silver index differs from our Laspeyres fixed weighted index. Silver uses a Laspeyres chain index which will yield higher growth rates, since industries growing more rapidly will get higher weights in a chain index. Although we are aware of the methodological differences, we have nevertheless chosen to integrate Silver’s index for 1965-72 into our index, because of the far better commodity coverage of Silver’s chain index compared to our IIP 1965–85. We have recalculated Silver’s chain index for our six branches for the period 1965–72, using our adjusted nominal value added weights for our base years (1966 and 1970). 4.3

IIP, 1985–95

For 1985–95 there is an index of industrial production published in the QSIP. We found that there were significant discrepancies between several of the 1985 base weights applied in the QSIP and the value added


data from the 1985 survey. Moreover, the analysis of the surveys made clear that the reliability of the value added data in the mid-eighties is questionable. In our opinion the most reliable source for reweighting the IIP (1985-95) is the data from the 1989 census of industrial production, as adjusted in this report (see Section 3.4). We have reweighted the index of industrial production utilizing (adjusted) 1989 census value added as base year weights. The weighting procedures are the same as explained above for the 1965–85 index (see Prins and Szirmai, 1998, appendix I and Table I-3). 4.4

The new index of industrial production, 1961–95

For the period 1961–65 no commodity information was available for the construction of an index. To get an estimate for total manufacturing, we have retropolated the index figure for total manufacturing in 1966, using a real value added index derived from national accounts real manufacturing value added data for the early period (OECD, 1971, Table 2-5). For the period 1965–72 we have used Silver’s chain index. For 1972–85 we have used our index; for 1985–95 we have used the reweighted QSIP index. The resulting indexes for the period 1961–95 are reproduced in Table 3.4. 4.5

Employment and labour productivity

As is the case for nominal and real manufacturing value added, no consistent long-run series of manufacturing employment is available at branch level from published statistical sources. We have reconstructed an employment series, consistent in time and consistent with the adjusted nominal value added estimates for 10+ manufacturing. For the period 1965–78 we have incorporated the level adjustments for non-response and undercoverage in the employment series, under the assumption that the labour productivity of the covered (or responding) establishments is equal to the labour productivity of the non-covered (or non-responding) establishments. Thus we can apply the ratio of value adjusted for coverage and nonresponse to non-adjusted value added to the employment figures. Similar adjustments to the employment figures have been made for the years 1978–89, for all branches except the textiles/leather branch. This branch was characterized by major discrepancies in employment figures between census and non-census years. For this branch we have applied an ad hoc adjustment, interpolating between the census years (see for details Prins and Szirmai, 1998, appendix M). The resulting employment figures for 1965–90 are consistent with our adjusted value added figures. Table 3.5 presents the employment figures for six manufacturing branches and for total manufacturing. Table 3.6 presents the corresponding labour productivity indices.

Indexes of industrial production 1965–95

ISIC Branch




88

Table 3.4





(1976=100) 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981

37 37 41 43 51 51 59 84 86 81 78 100 92 103 90 77 67

24 30 31 45 47 63 62 80 90 94 92 100 105 97 104 99 90

100 114 101 116 124 123 145 141 159 142 123 100 151 124 100 108 97

26 55 57 62 62 65 78 91 88 90 95 100 94 98 80 77 75

28 57 59 113 118 124 114 130 150 106 115 100 115 105 111 84 107

24 27 24 33 29 29 48 61 81 91 92 100 112 133 138 130 114

22 24 26 28 31 37 38 48 52 57 65 84 93 93 91 100 105 107 103 96 88

Table 3.4

(continued)

ISIC Branch



64 67 62 57 54 48 57 55 62 64 62 65 66 69

83 62 60 57 58 75 82 76 77 73 65 68 66 58

1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

33/34 Wood, furnitures & fixtures, paper, printing & publishing 91 68 58 62 83 114 111 109 101 90 71 101 72 66

35 Chemicals, petroleum, rubber & plastics (1976=100) 71 70 120 77 71 70 64 76 84 84 78 72 79 75


92 68 102 103 121 126 133 136 157 214 154 164 139 153


132 167 194 137 152 115 102 110 141 133 100 88 125 69


83 77 88 72 73 71 75 75 84 84 74 75 79 72

Source: Prins and Szirmai (1998), table K-2.

89

Persons engaged in 10+ manufacturing, by branch (1965–90)

ISIC Branch



10 038 9 767 12 666 14 930 13 478 16 340 17 652 21 334 18 880 20 035 22 026 22 333 24 793 28 111 28 455 30 291 32 437 38 041 34 885 30 458 29 138 38 641 39 858 43 200 45 282 47 397

12 102 14 504 14 265 19 033 21 770 22 791 24 245 24 795 28 609 33 200 34 140 34 661 38 168 43 926 43 072 45 407 48 670 45 308 49 642 44 545 42 191 43 499 43 426 41 069 38 128 37 674

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

Source: Prins and Szirmai (1998), table M-10.

33/34 Wood, furnitures & fixtures, paper, printing & publishing 9 203 9 597 8 235 9 596 9 737 8 885 9 635 10 896 11 117 12 219 13 125 13 302 15 308 14 781 14 909 14 637 14 513 14 852 15 793 14 262 13 722 14 452 14 211 15 229 16 930 24 360

35 Chemicals, petroleum, rubber & plastics 955 2 330 1 767 2 507 1 997 2 208 2 782 4 876 4 769 5 309 5 235 5 304 5 604 7 900 7 480 8 890 9 388 9 772 9 439 8 085 6 447 6 930 7 042 6 741 6 677 6 534

90

Table 3.5


652 1 744 1 422 2 380 2 622 2 572 3 511 2 866 3 343 3 061 2 942 2 982 3 476 3 773 3 333 3 218 4 247 3 300 3 687 3 614 3 821 5 344 4 517 4 484 5 069 4 237

37/38/39 (Basic) metals, machinery & equipment and other manufacturing 6 204 7 116 5 977 5 402 5 538 5 719 7 582 8 277 9 304 10 316 11 253 11 402 13 947 14 194 12 836 13 566 14 090 14 242 14 546 13 156 12 402 11 303 11 041 10 919 13 793 14 211


39 154 45 058 44 332 53 847 55 142 58 515 65 407 73 044 76 022 84 139 88 721 89 984 101 295 112 685 110 084 116 009 123 345 125 515 127 991 114 120 107 721 120 168 120 096 121 642 125 879 134 413

Table 3.6

Labour productivity index for six branches of manufacturing, 1965–90 (1976 = 100)

ISIC Branch


1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

82 85 73 65 84 70 75 88 102 90 79 100 83 82 70 57 46 38 43 45 44 31 27 30 27 29


68 71 75 83 75 95 89 112 109 98 93 100 96 76 84 75 64 64 43 47 46 47 59 70 69 71





145 158 162 160 169 184 200 173 191 155 124 100 131 112 89 99 89 81 57 54 60 77 107 97 86 55

146 126 171 131 165 155 148 99 98 90 96 100 89 66 57 46 42 38 39 79 63 54 52 50 60 68

128 98 123 141 135 144 97 135 134 103 117 100 99 83 99 78 75 83 55 84 81 68 83 88 80 111

44 43 47 71 60 59 72 84 99 101 93 100 91 106 123 109 92 106 131 168 126 153 119 107 91 113


71 73 76 81 84 88 90 104 110 100 93 100 93 86 85 74 64 60 54 70 60 55 53 55 53 56 91

Source: Tables 3.4 and 3.5.


5

The 1989 benchmark

In this section we present a level comparison of real labour productivity in manufacturing between Tanzania and the world productivity leader, the USA, for 1989 (see Szirmai and Schulte, 1988). Level comparisons are a useful complement to time series analysis, because they provide information on the absolute size of the gaps in real output and productivity between economies at given points in time. The productivity gaps provide indications of the size of the technology gap between economies. The benchmark comparisons can also serve as an anchor for trend comparisons. International comparisons require adequate converters. It is well known that comparisons based on exchange rates can substantially underestimate levels of national income and product in developing countries (see Kravis et al., 1982; Ren, 1997; Szirmai and Ren, 1998). From the perspective of welfare comparisons, expenditure-based purchasing power parities are more realistic converters than exchange rates. However, expenditure-based PPPs are less suitable for sectoral comparisons of real output and productivity, because they are not based on producer prices and because final products include productive contributions of many different sectors of the economy. For sectoral output and productivity comparison, industry of origin converters are required. In this paper, industry-of-origin unit value ratios (UVRs) are calculated for branches of manufacturing, using the industry-oforigin methodology developed in the International Comparisons of Output and Productivity project (ICOP). The UVRs are used to convert 1989 Tanzanian manufacturing GDP into US dollars. So far data availability has been a major constraint for the application of the ICOP methodology to an African economy. The methodology requires reliable data on product quantities and ex-factory output values, which are seldom available in published form. In the case of Tanzania, the Bureau of Statistics has made the basic files underlying the industrial census of 1989 available to us. This allowed us to reconstruct the necessary data on production quantities and values, from a very basic level. It should be stressed that the benchmark results presented in this paper are preliminary. 5.1

Methodology for the 1989 level comparison

The ICOP methodology for level comparisons has been described in detail in several publications (see Maddison and van Ark, 1988, 1994; Szirmai and Pilat, 1990; van Ark, 1993; Timmer, 1996). In this paper, we apply the methodology as refined in Timmer (1996). Here, we provide only a brief outline of the methods used. The primary sources used in this study are the US 1987 Census of Manufactures (US Department of Commerce, 1990b), the 1989 Annual Survey of Manufacturing (US Department of Commerce 1990a), the Tanzanian Census of Industrial Production (vols. IV–V, Bureau of Statistics


1993a, 1994a) and the data files underlying the Tanzanian published census. These sources provide information on product quantities and corresponding gross output values, making it possible to derive unit values for products or groups of products for both economies. The basic approach is to make matches of comparable products or product groups from the two censuses and to calculate unit value ratios (UVRs) for each of the matches. Subsequently these are aggregated into average unit value ratios for industries, branches and total manufacturing. These unit value ratios are used as conversion factors. Matches were made in 16 sample industries. On the US side these consist of one or more four-digit industries. For Tanzania, the commodity and output information collected in the 1989 industrial census has not been published. The information from the basic census data files on products and their output value was rearranged into four-digit ISIC industries (1968 version of ISIC; see United Nations, 1968b). These were combined into sample industries comparable to those on the US side. The conversion factors are calculated in a number of steps. 1. For each product match, UVRs are calculated (TSh/$). The initial unit value ratios derive from 1989 Tanzanian unit values and 1987 US unit values. 2. UVRs are put on a 1989–89 basis, using US 1987–89 price movements for each product group from the Bureau of Labour Statistics (1998). The 1989/1989 UVRs are used in all the subsequent calculations. 3. All the UVRs are aggregated into average UVRs at sample industry level using output quantities of either countries as weights: S

UVR

XU ( X ) j

=

∑Q

S

X ij

∗ PijX

i= 1 S

∑

UVR

XU (U ) j

Q ijX ∗ PijU

i= 1

where UVRjXU(X)

=

∑ (Q

U ij

∗ PijX )

i= 1 S

∑

(3.2)

(Q Uij ∗ PijU )

i= 1

is the unit value ratio of the Tanzanian shilling against the US dollar in sample industry j, at quantity weights of Tanzania

UVRjXU(U)

is the unit value ratio of the Tanzanian shilling against the US dollar in sample industry j, at quantity weights of the USA

i = 1…s

is the sample of matched items.

4. Sample industry UVRs are aggregated into branch UVRs. Manufacturing branches in this study consist of one or more ISIC three digit major


sectors. Where there is more than one sample industry in a branch (as in food products), the 1989 sample industry UVRs are aggregated at manufacturing branch level by taking the weighted average of sample industry UVRs using 1989 sample industry gross output as weights: O

UVR

XU (U ) k

=

∑ [GO

U (U ) j

]

(U ) ∗ UVR XU j

j= 1

(3.3)

O

∑ GO

U (U ) j

j= 1 O

UVR

XU ( X ) k

=

∑ GO

X (X ) j

j= 1

∑ [GO O

j= 1

X (X ) j

(X ) / UVR XU j ]

where GOjU(U) GOjX(X)

is gross output in US sample industry j in dollars is gross output in Tanzanian sample industry j in Tanzanian shillings k is the branch of industry j=1…o is sample industries belonging to a branch k However, following Timmer (1996), if the reliability of the UVR for a given sample industry is too low, we weight the original UVRs of the matched items with their output in Equation 3.2.10

5. The branch UVRs are aggregated into UVRs for total manufacturing, using branch gross output weights according to Equation 3.2. However, if the reliability of branch UVRs is too low, we use the sample industry gross output weights of step 4 to weight sample industry UVRs. 11 The general rationale behind these stepwise weighting procedures is to ensure that unit value ratios in large sample industries and large branches receive heavier weights than in small ones (see van Ark, 1993). However, where the unit value ratios are insufficiently reliable, we use gross output of matched items in a sample industry as weights at branch level and sample industry gross output weights at manufacturing level (see Timmer, 1996, for more detail). Thus, the more reliable a UVR the heavier the weight it receives in the aggregation procedure. 6. At each level of aggregation – sample industry, branch and total manufacturing – the UVRs can be used to convert 1989 value added into the currency of the other country for purposes of real value added comparisons. In binary comparisons one gets two UVRs at every level of aggregation, one at quantity weights of country X, the other at quantity


weights of country U. We use the Fisher geometric average of the two UVRs as a summary measure. We have made 76 product matches in 16 sample industries, representing 11 of the 15 branches of manufacturing. The important Tanzanian food manufacturing branch is represented by 6 sample industries. Another 10 branches were represented by a single sample industry. No matches were realized in metal products, machinery, electrical machinery and other manufacturing. For these branches we used the calculated UVRs for total manufacturing based on 11 branches. Appendix Table A.3 shows the coverage ratios at branch and sample industry levels. The matched value of output represents 31.6 per cent of the total gross value of output in Tanzania and 7.1 per cent in the USA. The low coverage on the US side is due to the fact that we are comparing a tiny industrial sector with a very large one. 5.2 5.2.1

Data sources for the 1989 benchmark comparison Tanzanian product listing

For Tanzania our basic source for the calculation of unit values consisted of the unpublished data files of the 1989 Census of Industries (Bureau of Statistics 1993a/1994a). The questionnaire for establishments with ten or more persons employed (form 1C-89-1) contained a question D: ‘Principal products manufactured during the year’, requesting information on the names of products, the units, installed capacity, production quantities and production values in 1000 shillings. The commodity data thus collected have not been published. However, Bureau of Statistics made the raw data files underlying the census available to us.12 These files contain information on quantities of different products produced by establishments and their corresponding output values. We rearranged this list of products into ISIC four-digit categories and created product listings for each industry. First, firms were allocated to an ISIC industry on the basis of their most important products. However, firms tend to produce a wide range of primary and secondary products. Therefore, the next step was to reallocate products to the ISIC industry producing those products as primary products. In the course of the data screening it was found that product names were frequently misspelt and units, quantities or values were frequently incorrectly reproduced. For many products information on either unit, quantity or value was lacking. A process of checking and cross-checking resulted in a valuable listing of commodities and their quantities and values. The provisional nature of this listing needs to be stressed. 5.2.2

Tanzanian output, value added and employment

For Tanzania we use our revised estimates of gross output, value added and employment for 10+ manufacturing in 1989 (see Tables 3.3 and 3.5; for


gross output, see Prins and Szirmai, 1998, table C-2). These estimates, especially for value added, are substantially higher than the official ones. Gross value added in the US census is measured without deducting the cost of services purchased from outside the manufacturing sector. Thus the US census concept of value added involves a degree of double counting. To ensure comparability with the USA, the Tanzanian value added concept has been adjusted to the US census concept of value added, which is gross of service inputs from outside the manufacturing sector. This was done by adding bank charges, insurance costs, transport costs, communication services, accountancy and professional services (see Bureau of Statistics, 1993a, codes 408–411). The basic data for 1989 are reproduced in Appendix Table A.1. 5.2.3

US product listing

For the USA our basic source was the 1987 Census of Manufactures (US Department of Commerce, 1990b), which lists approximately 11 000 products. For most, though not all, products the US census provides both quantity and value information for 1987. First, unit value ratios were computed on the basis of the two censuses, using Tanzanian 1989 unit values and US 1987 unit values. Subsequently the 1989/1987 unit value ratios were adjusted to a 1989–89 basis using US price indices for each product category for the period 1987–89 from the Bureau of Labour Statistics (1998). 5.2.4

US gross output, value added and employment

The data for 1989 on gross value of output (here, value of shipments), gross value added and employment by industry derive from the 1990 Annual Survey of Manufactures (US Department of Commerce, 1990a). The Tanzanian census refers to establishments with ten or more persons engaged. To ensure comparability, the US data have to be adjusted to a 10+ basis. As the ASM data for 1989 only provide information on output, value added and employment for total manufacturing, we used proportions of 10+/total value added and employment from the general summary of the 1987 census to adjust the US data to a 10+ basis. The basic data for the USA are reproduced in Appendix Table A.2. 5.3

Unit value ratios

The UVRs for different branches of manufacturing are reproduced in Table 3.7. The aggregate UVR (geometric average) for total manufacturing is 120 shillings to the US dollar, lower than the 1989 exchange rate of 143. 13 The price level, defined as the UVR divided by the exchange rate, is 0.8. On first sight, this result is surprising given the general complaint that Tanzanian exchange rates tend to be overvalued. However, the discussion concerning overvalued exchange rates refers primarily to internationally

Measuring Manufacturing Performance 97 Table 3.7 Unit value ratios and price levels by major manufacturing branch Tanzania/USA (TSh/$), 1989 UVR (TSh/US$)

Relative price level Tanzania At US At Tanzanian Geometric (USA = 100) quantity quantity average weights weights Food manufacturing Beverages Tobacco products Textile mill products Wearing apparel Leather products & footwear Wood products, furniture & fixtures Paper products, printing & publishing Chemical products (incl. oil) Rubber & plastic products Non-metallic mineral products Basic & fabricated metal productsa Machinery & transport equipmenta Electrical machinery & equipmenta Other manufacturing industriesa

102 69 51 105 42 36 90

45 81 51 110 28 20 86

68 75 51 107 35 27 88

0.5 0.5 0.4 0.7 0.2 0.2 0.6

160

139

149

1.0

470 333 90 177 177 177 177

110 325 143 78 78 78 78

227 329 114 117 117 117 117

1.6 2.3 0.8 0.8 0.8 0.8 0.8

Total manufacturing, census weightsb Total manufacturing, implicit UVRsc Exchange rate

177

78

117

0.8

198

73

120

0.8

143

143

143

Notes: a No sample industries for this branch. We used the UVR for the total of branches. b The UVR for total manufacturing is the gross output weighted average of branch or sample industry UVRs (see Timmer, 1996). c Implicit UVRs calculated from the summed branch value added totals in Table 3.8. These are the preferred UVRs.

tradable goods. A considerable portion of Tanzanian industrial output is directed to the domestic market and does not enter into international trade. The finding that UVRs for developing countries are well below the exchange rate has been found in several studies (for example, on China, Indonesia and India). However, low UVRs in part also reflect unrecognized quality differences for identical products and a predominance of lowquality items in the product mix. Lowest UVRs are found for food, beverages, wearing apparel and leather, and highest UVRs for rubber, chemicals, paper products and textiles. UVRs at US quantity weights are much higher than at Tanzanian quantity weights. This is standard in comparisons between high income and lowincome economies. Products which are relatively cheap and common in


the USA will tend to be expensive and rare in a low-income country such as Tanzania. Products which are cheap and common in Tanzania will tend to be rare in the USA. Therefore matches with high unit values will have high quantity weights in the USA and low quantity weights in Tanzania. Matches with low unit values will have high quantity weights in Tanzania and lower weights in the USA. 5.4

Productivity comparisons

Applying the branch unit value ratios from Table 3.7 to the gross value added figures from Appendix Tables A.1 and A.2 results in real comparisons of gross value added. Dividing these figures by the employment figures from Tables A.1 and A.2 results in real labour productivity comparisons. These comparisons are reproduced in Table 3.8. Aggregate real labour productivity in Tanzanian manufacturing in 1989 is 3.7 per cent of that in US manufacturing. There is substantial variance in branch productivity performance, varying from 2.6 per cent in chemical products and 3 per cent in paper products to 12.4 per cent in leather products and footwear and 10.5 per cent in electrical machinery and equipment. These productivity differentials are an indication of the vast technology gap between a developing economy such as Tanzania and economies operating at the technological frontier such as the USA. Productivity in the tiny Tanzanian manufacturing sector is lower than that found for large and dynamic Asian developing economies such as China and Indonesia. Two qualifying remarks are in order. In the first place, the product listings are not sufficiently detailed to allow for quality adjustments. It is likely that Tanzanian products are of lower quality than the corresponding US products. More detailed analysis of each of the matches, using information from outside the census, should be performed. In the second place, the comparison excludes the important small-scale and informal sector in Tanzania, characterized by highly labour-intensive activities. In most developing countries, productivity in the informal sector is much lower than in the formal sector, so that productivity comparisons for total manufacturing need to be adjusted downward even further. At this stage, we can safely say that our unit values and UVRs are a lower bound and our productivity comparisons are an upper bound. In spite of the low levels of productivity found, real Tanzanian productivity will be even lower than our estimates, after adjustments for quality differences and inclusion of small-scale labour-intensive enterprises.

6

New insights

In this section, we discuss the new insights in Tanzanian manufacturing performance arising from revised estimates of value added, employment and comparative productivity discussed above.

Table 3.8

Gross value added (census concept) per person, Tanzania and the USA (1989) At Tanzanian prices (in TSh) Tanzania

Food manufacturing Beverages Tobacco products Textile mill products Wearing apparel Leather products & footwear Wood products, furniture & fixtures Paper products, printing & publishing Chemical products Rubber & plastic products Non-metallic mineral products Basic & fabricated metal products Machinery & transport equipment Electrical machinery & equipment Other manufacturing industries Total manufacturing

USA

At US prices (in US$) Tanzania/ USA (%)

Tanzania

USA

Tanzania/ USA (%)

Geometric average, Tanzania/ USA (%)

203 370 816 907 712 352 205 282 48 669 120 675 150 251

8 387 897 10 413 211 15 691 221 4 229 796 1 306 552 1 305 577 3 828 677

2.4 7.8 4.5 4.9 3.7 9.2 3.9

4 511 10 027 13 871 1 871 1 734 6 111 1 754

82 572 150 548 305 552 40 281 30 798 36 648 42 331

5.5 6.7 4.5 4.6 5.6 16.7 4.1

3.6 7.2 4.5 4.7 4.6 12.4 4.0

321 831

11 517 612

2.8

2 318

72 141

3.2

3.0

845 705 726 943 390 347 470 994 419 871 778 159 267 095

66 647 074 17 537 943 5 703 193 10 680 022 12 819 273 11 234 739 12 307 000

1.3 4.1 6.8 4.4 3.3 6.9 2.2

7 711 2 234 2 737 6 072 5 413 10 032 3 443

141 811 52 669 63 071 60 344 72 431 63 478 69 537

5.4 4.2 4.3 10.1 7.5 15.8 5.0

2.6 4.2 5.4 6.7 4.9 10.5 3.3

312 562

13 844 486

2.3

4 278

69 787

6.1

3.7

Source: Gross value added and employment from Appendix Tables A.1 and A.2; UVRs from Table 3.7.

99


6.1

Level adjustments in nominal value added

It is beyond doubt that the level of medium- and large-scale manufacturing performance has been underestimated in the published statistics of Tanzania. In Figure 3.3 our level adjustments are graphically presented for the series 1965–78 and 1978–90. Manufacturing value added increased for the entire period, but most pronounced adjustments appeared between 1982 and 1990. Value added increased by 83 per cent in 1982, 135 per cent in 1988 and 61 per cent in 1990. Between 1965 and 1978 value added increased on average by 40 per cent. Though the adjusted levels are much higher, it can be seen that for 1965–78 the adjusted nominal trend closely follows the unadjusted trend, while this is not true for 1978–90. Especially between 1986 and 1990 the level adjustments are more pronounced than is the case for earlier years. 6.2

Changes in the structure of manufacturing

Table 3.9 compares the adjusted and not-adjusted value added shares of six branches of manufacturing for the years 1966, 1978 and 1989. For 1978, there is no difference between the unadjusted and adjusted structure of production. In 1966 the adjusted shares for the branches ISIC 37/38/39 (basic metal, machinery and other manufacturing) and ISIC 32 (textiles and leather) are substantially higher, while they are much lower for ISIC 31 (food, beverages and tobacco). In 1989 the most marked difference involves ISIC 32 (textiles and leather), where the adjusted value added share increased is 18 per cent, compared to an unadjusted share of 8 per cent. Overall, there is much less structural change in the revised data compared to the unrevised data. Changes are less marked, and the textile sector remains a major contributor to manufacturing value added in the late Table 3.9

Structural changes in Tanzanian 10+ manufacturing (%) 1966

ISIC

Branch

31 Food, beverages & tobacco 32 Textiles & leather 33–4 Wood, furniture & fixtures, paper, printing & publ. 35 Chemicals, petroleum, rubber & plastic products 36 Non-metallic mineral products 37–9 (Basic) metal products, mach. & equipm. & Other man. 3

Total manufacturing

Source: Prins and Szirmai (1998), table G-2.

1978

1989

Unadj. Adj. Unadj. Adj. Unadj. Adj. 42 25 12

31 33 10

24 30 10

24 30 10

42 8 11

38 18 9

8

6

12

12

17

15

3 10

3 18

3 21

3 21

5 17

4 15

100

100

100

100

100

100

Measuring Manufacturing Performance 101 Figure 3.3 1965–90

Unadjusted and adjusted nominal value added for 10+ manufacturing,

Unadjusted and adjusted nominal value added,1965–1978

3 500 000

3 000 000

Unadjusted Adjusted

2 500 000 2 000 000 1 500 000 1 000 000 500 000

1978

1977

1976

1975

1974

1973

1972

1971

1970

1969

1968

1967

1966

1965

0

Unadjusted and adjusted nominal value added,1978–1990

Nominal value added in 1000 TSh

45 000 000 40 000 000

Unadjusted Adjusted

35 000 000 30 000 000 25 000 000 20 000 000 15 000 000 10 000 000

Source: Table 3.3.

1990

1989

1988

1987

1986

1985

1984

1983

1982

1989

1979

1978

0

1981

5 000 000


eighties. The light industries food, beverages, tobacco, textiles and leather together account for 56 per cent of value added, as against 54 per cent in 1978 and 63 per cent in 1966. 6.3

Trends in real growth

In Figure 3.4 the new index of industrial production for 1961–95 is compared with a national accounts based real manufacturing index for total manufacturing (Bureau of Statistics, 1995b) and the 50+ QSIP index (see Prins and Szirmai, 1998, table L-3). Inspection of Figure 3.4 reveals that real value added in 1961 is lower for the new index and that the BoS 1995 index shows less growth between 1961 and 1978. A steep decline in the early eighties is identified by the BoS 1995 index as well as by our new index. From 1985 onwards the various indexes show very divergent trends. Where the Bureau of Statistics (1995b) registers relatively rapid growth, the QSIP index indicates growth between 1985 and 1990 and decline after 1990. Our index shows the same pattern as the QSIP index: recovery from 1985–91, followed by renewed stagnation in the nineties. In general, the periods of growth and stagnation of the Tanzanian manufacturing sector are more clearly distinguished by our index. The main turning points in the industrialisation pattern of Tanzania are the years 1978 and 1985. The new index shows more rapid growth before 1978, and a more dramatic collapse between 1978 and 1985. After 1985 performance is uneven, though there is some slight improvement compared to 1985. Figure 3.5 combines our level and trend adjustments. The top line represents adjusted value added for 10+ manufacturing in 1989, extrapolated Figure 3.4

National accounts, QSIP and the new index of real value added, 1965–94

140

Index 1985 = 100

120 100 80 RMVA series BoS 1995b RMVA series QSIP New IIP

60 40

Sources: see text.

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

1973

1971

1969

1967

1965

1963

1961

20

Measuring Manufacturing Performance 103 Figure 3.5 Manufacturing value added at 1989 prices, 1961–1995: comparison of the new index of industrial production with index based on the national accounts and QSIP 50 000

Million constant 1989 TSh

45 000

NA&QSIP New IIP

40 000 35 000 30 000 25 000 20 000 15 000 10 000

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

1973

1971

1969

1967

1965

1961

0

1963

5 000

Sources: Adjusted and unadjusted nominal value added for 1989 from Table 3.2. Adjusted value added extrapolated with new index of industrial production from Figure 3.4, unadjusted value added extrapolated from 1985 to 1995 with QSIP index from Figure 3.4 and from 1961 till 1985 with the national accounts index (Bureau of Statistics, 1995b) from Figure 3.4.

with the new index of production. The bottom line represents unadjusted 1989 value added extrapolated with an unadjusted index of industrial production. This unadjusted index consists of the national accounts index up to 1985 (Bureau of Statistics, 1995; see Prins and Szirmai, 1998, table L-3), linked with the QSIP index available since 1985. Figure 3.5 has a double message. On the one hand, the rise and collapse of manufacturing are more dramatic in the new estimates than in the old ones. Real value added in 1995 is about the same as in 1972. But, on the other hand, the adjusted level of manufacturing value added in 1995 turns out to be substantially higher than in the old estimates. An important outcome of this research consists of indices of real output for different manufacturing branches for the period 1965-85. No such series were previously available. In Table 3.4 above we have presented the new indices for six manufacturing branches for the period 1965–95. The growth trends for the branches food, beverages and tobacco, and textiles and leather more or less correspond to the pattern for total manufacturing, represented in Figure 3.4. The trends for the branch wood are irregular and show long-term decline. Chemicals, petroleum, rubber and plastics show modest growth until 1977 and stability thereafter.


Non-metallic mineral products showed growth until 1973, stagnation between 1973 and 1983 and recovery in the post-1983 period. Real GDP in metals and machinery (ISIC 37, 38 and 39) increased until 1984, followed by a period of decline between 1984 and 1995. 6.4

Trends in labour productivity

Figure 3.6 presents the new index of labour productivity for total manufacturing. Initially, labour productivity increased rapidly after 1965, reaching a peak in 1973. Well before the turning point in real output trends in 1978, labour productivity started declining after 1973. The decline continued throughout the seventies and eighties, probably as a result of the inability of Tanzanian firms to shed labour as their output contracted. By 1990 labour productivity in total manufacturing was about one half the level in 1973, and 21 per cent below the level of 1965. Labour productivity trends by branch of manufacturing have already been presented (Table 3.6). Figure 3.6

Labour productivity index for total manufacturing, 1965–90 (1976 = 100)

120 110

Index, 1976 = 100

100 90 80 70 60 50 40 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 Source: Table 3.6.

6.5

Comparative labour productivity

Figure 3.7 combines the benchmark productivity comparison for 1989, with Tanzanian and US time series on labour productivity. Comparative labour productivity trends show an interesting pattern. Starting from a fairly high level of almost 9 per cent of the US level, there is productivity catch-up until 1973. Increase in the amount of capital per worker is one of the causes of this. After 1973 one sees an extended period of comparative productivity decline. Initially, this is due to absolute declines in labour productivity in Tanzanian manufacturing, but the absolute decline evens out


Tanzania/USA (%)

Figures 3.7 Comparative labour productivity in Tanzanian manufacturing, 1965–90 (USA = 100) 12 11 10 9 8 7 6 5 4 3 2 1 0 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 Year Source: 1989 Tanzania/USA benchmark from Table 3.8; index of GDP per person in Tanzania from Figure 3.6; GDP per person in USA from national accounts sources. 1965–82 GDP at constant prices from US Department of Commerce (1986); 1977–90 from Survey of Current Business, various issues; persons employed, 1965–90, from NIPA; 1959–88, US Department of Commerce (1992), and Survey of Current Business, various issues.

after 1983. After that year comparative productivity decline is caused by stagnation in Tanzania and increasing productivity in the lead country. The trend is also worth noting. In most productivity studies for Asian low-income economies, such as China, India and Indonesia (Timmer and Szirmai, 1999), we see productivity starting at lower levels than in Tanzania, little change in productivity performance over time and some catch-up in the nineties. Tanzania starts at much higher levels in the sixties and ends up doing much worse in the eighties, reflecting the inefficient pattern of industrialisation discussed in Chapter 1.

Table A.1

Basic data on output and employment for Tanzania, 1989 (establishments with 10 or more persons engaged) Number of Gross value establishments of output at factor cost (mill. TSh)

1 Food manufacturing (311/12) 2 Beverages (313) 3 Tobacco products (314) 4 Textile mill products (321) 5 Wearing apparel (322) 6 Leather products & footwear (323/324) 7 Wood products, furniture & fixtures (331/2) 8 Paper products, printing & publishing (341/2) 9 Chemicals products (351–3) 10 Rubber and plastic products (355/6) 11 Non-metallic mineral products (361–9) 12 Basic & fabricated metal products (371–81) 13 Machinery & transport equipment (382/4)

Gross value added at factor cost (mill. TSh)

Gross value added at factor cost, US census concepta (mill. TSh)

Gross value added in branch as % of total

Employment (person)

Gross value added per person employed TSh

146 19 3 83 54 21

24 531.6 8 959.6 6 237.2 17 701.0 519.5 1 883.2

5 630.9 3 003.1 3 359.7 5 256.4 105.8 401.1

7 177 4 116 3 528 6 367 136 520

18.24 10.46 8.97 16.18 0.35 1.32

35291 5039 4952 31014 2803 4311

203 370 816 907 712 352 205 282 48 669 120 675

214

4 399.7

1 115.4

1 454

3.69

9674

150 251

62

7 411.8

1 865.7

2 335

5.94

7256

321 831

51 15

11 680.3 4 900.7

3 515.8 954.0

4 201 1 243

10.68 3.16

4967 1710

845 705 726 943

22

5 608.1

1 413.7

1 979

5.03

5069

390 347

94

13 583.2

2 231.6

2 882

7.32

6118

470 994

72

8 062.1

1 768.3

2 234

5.68

5320

419 871


Appendix

Table A.1

(continued) Number of Gross value establishments of output at factor cost (mill. TSh)

Total manufacturing

Gross value added at factor cost, US census concepta (mill. TSh)

Gross value added in branch as % of total

Employment (person) person

Gross value added per employed TSh

6

2 303.4

710.0

830

2.11

1067

778 159

24

1 126.4

278.9

344

0.87

1288

267 095

886

118 907.7

39 345

100.00

125 879

312 562

31 610

Sources: Prins and Szirmai (1998): gross output and value added table A-4, employment table A-5. Original source: data files of 1989 census of production. Note: a US census value added defined as: gross value of output at factor cost minus intermediate inputs, except intermediate service inputs from outside the industrial sector.


14 Electrical machinery & equipment (383) 15 Other manufacturing Industries (385–90)

Gross value added at factor cost (mill. TSh)

Gross value of output (mill. US$) Food manufacturing Beverages Tobacco products Textile mill products Wearing apparel Leather products & footwear Wood products, furniture & fixtures Paper products, printing & publishing Chemicals, incl. petrol. refining Rubber & plastic products Non-metallic mineral products Basic & fabricated metal products Machinery & transport equipment

Annual Survey of Manufactures Gross value Gross value added added in (mill. US$) branch as % of total

Employmentb (1 000s)

GVA/person (US$)

310 109.7 49 695.9 25 789.9 67 072.6 61 451.2 9 757.4

106 053.8 24 103.4 18 916.2 27 123.1 31 361.7 4 543.6

8.31 1.89 1.48 2.12 2.46 0.36

1 284.4 160.1 61.9 673.4 1 018.3 124.0

82 572.0 150 547.9 305 552.2 40 280.7 30 798.2 36 648.4

108 550.5

47 791.0

3.74

1 129.0

42 331.0

271 117.7

152 708.2

11.96

2 116.8

72 140.8

415 580.0

168 705.0

13.21

1 189.6

141 811.3

96 725.5

46 850.0

3.67

889.5

52 668.9

60 938.3

32 895.4

2.58

521.6

63 071.2

308 697.4

132 557.2

10.38

2 196.7

60 344.0

618 891.6

282 524.5

22.13

3 900.6

72 431.1


Table A.2 Basic data on output and employment in manufacturing, USA, 1989 (establishments with 10 or more persons engaged)a

Table A.2 Basic data on output and employment in manufacturing, USA, 1989 (establishments with 10 or more persons engaged)a (continued)

Gross value of output (mill. US$)

Total manufacturing

Employmentb (1 000s)

GVA/person (US$)

190 906.6

105 044.9

8.23

1 654.8

63 478.2

150 519.0

95 611.5

7.49

1 375.0

69 536.7

2 745 803.2

1 276 789.4

100.00

18 295.6

69 786.6

Source: US Department of Commerce (1990a). Notes: a Ratio of 10+ to total manufacturing from 1987 Census of Manufactures, general summary (US Department of Commerce, 1990b). b Including head office and auxiliary employment. Totals distributed across branches using 1987 proportions from Census of manufactures.


Electrical machinery & equipment Other manufacturing industries

Annual Survey of Manufactures Gross value Gross value added added in (mill. US$) branch as % of total

Number of UVRs, coverage rates and reliability Number of UVRs

Food manufacturing Dairy products Preserved fruits & vegetables Fats and oils Grain mill products Bakery products Sugar & confectionary, food not elsewhere specified Beverages Malt and malt beverages Tobacco products Tobacco stemming and redrying Textile mill products Textile mill products Wearing apparel Wearing apparel Leather products and footwear Leather footwear Wood products, furniture & fixtures Wood products & furniture Paper products, printing & publishing Paper, printing & publishing

Coverage, USA

Coverage, Tanzania

Reliability PPP at US quantity weights

Reliability PPP at Tanzanian quantity weights

22 4 4 4 4 1 5

20.3 38.6 31.1 51.5 59.2 34.4 41.0

46.6 39.0 3.2 53.4 80.3 16.3 34.1

0.10 0.1 0.5 0.1 0.1 0.0 0.2

0.27 0.0 1.4 0.2 0.1 0.0 0.7

2 2 1 1 6 6 7 7 2 2 13

28.0 89.5 8.6 86.5 15.4 48.4 20.3 61.0 39.9 90.5 27.6

15.7 37.1 21.4 42.1 39.5 45.9 17.3 19.1 2.1 7.8 41.8

0.18 0.1 0.00 – 0.09 0.1 0.26 0.2 0.43 0.0 0.13

0.79 0.7 0.00 – 0.08 0.1 1.20 1.2 0.11 0.1 0.13

13 9

57.7 11.6

41.9 40.0

0.1 0.16

0.1 0.15

9

24.1

40.2

0.1

0.2


Table A.3

Table A.3

(continued) Number of UVRs

Chemicals, incl. petrol. refining Chemical products Rubber & plastic products Rubber tyres & tubes Non-metallic mineral products Cement & bricks Basic & fabricated metal products Machinery & transport equipment Electrical machinery & equipment Other manufacturing industries Total manufacturing

Coverage, USA

Coverage, Tanzania

Reliability PPP at US quantity weights

Reliability PPP at Tanzanian quantity weights

9 9 2 2 3 3

3.4 44.1 6.1 50.0 6.5 67.1

19.4 38.1 24.1 69.2 38.3 109.0

0.67 0.5 0.00 0.0 0.41 0.2

0.86 0.8 0.02 0.0 0.01 0.0

76

7.1

31.6

0.14

0.12

Note: Coverage refers to matched output as percentage of total gross value of output. The measure for reliability is calculated as the variation of unit value ratios divided by the UVR for the sample industry or branch. The 90 per cent confidence interval equals the UVR plus or minus a percentage equal to twice the reliability measure.

111


Notes *

1 2

3

4 5

6

7

8

9

10

Eindhoven Centre for Innovation Studies and Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology. Menno Prins is currently working with Ericsson/Sweden, Wessel Schulte with UNIDO/Uganda. Thanks are due to John Komba and Russell Freeman for supervision and support in Tanzania, and to Marcel Timmer and Cees Withagen for valuable comments and advice. Unless indicated otherwise, GDP always refers to the factor cost concept. The work on Tanzanian time series is based on fieldwork carried out by M. Prins at the Tanzanian Bureau of Statistics, from April to October 1996 (M. Prins, ‘Manufacturing Statistics: Reconstructing Tanzanian Manufacturing Value Added 1965-1995’, MSc thesis, Technology and Development Studies, August 1997). Lex Lemmens collected the original data files for the benchmark study. We thank him for making these data available to us. We thank Marcel Timmer for advice and assistance with the calculation of unit value ratios and their reliability. This means that we can correct for price changes, but not for real output changes of the non-responding establishments. Inspection of the real index of industrial production discussed in Section 4 of this article shows that aggregate real output was more or less stable between 1983 and 1989. Therefore, our adjustment for price changes does not seriously underestimate aggregate value added. Note that this information is not available for the period 1978–90 for the full population of establishments in the census. The assumption here is that value added is roughly proportionate to the number of establishments. We have to make this assumption as no complete value added data are available for the sample for all the intervening years between 1978 and 1990. The important assumption here is that the degree of underestimation calculated from our sample of 50+ establishments is considered to be representative for 10+ manufacturing. Intermediate inputs or intermediate consumption consists of the value of the goods and services consumed as inputs by a process of production. (United Nations, 1993, 143). Cost categories such as interests costs, directors’ fees and donations are not intermediate inputs, but part of value added. The construction of a price index in this way derives from Mr R. Freeman, who has prepared a similar estimate in draft data files which he has kindly made available to us. The CPI is used for food manufacturing (food index), beverages and tobacco (beverages and tobacco index), textiles and apparel (clothing and footwear index), and furniture (furniture and utensils index) and for all other branches the total CPI is used. From 1992 onwards the PPI is available at industry level. The CPI and the PPI are spliced in 1992 to construct a series for 1990–95. Commodity gross output values were not available for the period 1965–85. We have used 1985 unit value data to cross check and incidentally adjust our 1989based intra-industry weights. The reliability of a UVR depends primarily on the variation in unit values within the given category. The smaller the variation of UVRS around the weighted average, the more reliable the calculated UVR is as a representation of the underlying UVR for the category. Reliability is measured as variation of unit value

Measuring Manufacturing Performance 113 ratios divided by the average unit value ratio; see Timmer (1996). We used .10 as the cut-off value for reliability. 11 If matched output value was taken as the sample industry weight in step 4 of the aggregation procedure, it is used as the sample industry weight in step 5 as well. 12 This file contained quantity information for approximately 720 enterprises. Though this is not explicitly indicated, we may safely conclude that this file refers to 10+ establishments. 13 There is a typical index-number type of discrepancy between the directly calculated UVRs for total manufacturing and the implicit UVRs calculated from summed branch values added at US and Tanzanian prices. We choose for the lowest degree of aggregation and therefore use the implicit UVR for total manufacturing.

4 The Role of Technological Factors in the Early Stages of Industrial Exports: A Note Charles Cooper*

1

Introduction

Over the past decade or so, the economic performance of the Southeast Asian NICs, particularly their performance in the export of manufactured goods, has exercised a considerable influence on policy thinking in developing countries. In many countries there is the hope that the remarkable demand growth generated in export markets might be emulated and lead to similar achievements to those of the NICs in terms of full employment accompanied by rising real wages and labour productivity. Admittedly, the desirability of an NIC pattern of development may be more questionable now, in the wake of the Southeast Asian financial and economic crisis, but the attraction of some, if not all, elements of what is seen as NIC exportpromoting policy still has strong influence elsewhere in the world. This paper is not about the economic and financial crisis. It proceeds from the assumption that, by and large, the crisis may be seen as a latterday phenomenon, afflicting the NICs in the past tumultuous two years, but not necessarily discrediting all past NIC industrialisation policies. After all, as many writers have pointed out, different countries have followed different economic policies, so that it is misleading to think that some generally defined group of common NIC policies have been put in question. Moreover, whatever may have been put in question, the reality is that there were historically unique successes in economic and especially industrial growth over a period of 30 years, and it remains relevant to ask how they were achieved, and especially (for present purposes), what role technological change and policies to induce it, played in the three decades of income growth which have so recently and abruptly come to an end. There are many new questions to be asked about the NICs at the moment, but the crisis does not gainsay the continuing relevance of the question which is asked in this paper. Particularly not, perhaps, because of the way it is intended to approach the question here. The first contention of the paper is that the technological 114

Technological Factors in Industrial Exports 115

and industrial policies followed in some Southeast Asian economies in the relatively recent past – particularly in South Korea, and particularly a group of interventionist policies which are described in various ways, but are often called ‘technological upgrading’ – have received far too much attention. They have exercised an influence out of proportion to their relevance in the rest of the developing world, including sub-Saharan Africa. The second and main contention is that earlier developments in various NICs of the ‘first generation’ may be much more relevant and important to developing countries of nowadays, than the more recent NIC experiences. And it can be argued that the earlier and by and large overlooked successes of policy in various NICs in relation to labour-intensive exports remain highly relevant to all developing countries, especially those in the subSaharan region that have faced difficulties in dealing with the rapid opening up of the world economy in recent years. There are two rather obvious and essentially empirical reasons why it makes sense to examine the immediate relevance of the labour-intensive manufactured export phase in the present sub-Saharan countries. The first is a matter of economic history. All countries that are currently active in world export trade in manufactures started out in the labour-intensive sectors, and it is to be expected that those that enter such trade in the future will go through a similar economic experience.1 The second point is about the current situation of the sub-Saharan economies as exporters of manufactures. In a 1995 UNU/INTECH discussion paper (Cooper, 1995), we established that out of 118 developing countries for which consistent data on manufactured exports are available, about 40 – or one third – showed evidence of sustained growth in these exports over the period 1970–90. 2 Out of the 118 countries, 34 were sub-Saharan. Of these only 4 or about 12 per cent, showed sustained growth in manufacturing exports. So there are proportionately far fewer sub-Saharan economies showing sustained growth of manufactured exports than in the rest of the world. If these countries are to achieve sustained growth of manufactured exports in the future they will have to pass through a phase of labour-intensive exports. It makes sense therefore to redress the balance of much recent writing, and to explore in more detail some of the main features of the labour-intensive export phase – as opposed to the process of ‘technological upgrading’ which has, perhaps for obvious reasons, attracted far more attention in the literature. In Section 2 there is an account of some important attempts to understand the role of manufactured exports in economic growth of developing countries. These grew out of the debates on the dual economy, whose point of departure was the seminal paper by Arthur Lewis (1954), which was developed by Fei and Ranis (1964, 1976). The paper examines the conclusions of the Fei and Ranis analysis. Section 3 outlines some of the problems that may be associated with entry into world markets for labour-intensive


manufactures. It takes the analysis beyond the limits of the Fei–Ranis model and considers the role of technological change in world market entry – even in labour-intensive manufactures. It also examines the longrun path dependencies this initial entry may generate. Section 4 gives the main conclusions.

2

The dual economy model and manufactured exports

The Lewis dual economy model and its development in an open economy form by Fei and Ranis had a profound influence on thinking about economic development. This part of the paper examines what these models of economic development had to say about technology. In particular it will deal with the implications which can be drawn from the Fei and Ranis form of the model for technology in the early stages of entry into world markets for manufactures. The Lewis ‘unlimited supplies’ model deals with the processes of capital accumulation in a labour surplus economy, leading to the emergence of a modern sector in the context of a large subsistence-oriented rural sector. There is a labour surplus in the rural sector in the sense that the migration of workers to the modern sector will not cause a fall in output. It is assumed that arrangements in the subsistence sector are such that all persons working there enjoy access to the average product of labour in the sector – and this average product of labour is what determines the minimum real wage in the modern industrial sector. This is one of the more debatable and debated assumptions of the model, but we will not enter into that. The level of output in industry is determined by the prevailing modern sector technology and the minimum real wage. Production is expanded to the point where the marginal product of labour is equal to the real wage. At this point the surplus value added in production above the wage bill accrues as profit to the owners of capital. It is this surplus, properly reinvested, which provides for reinvestment and expansion, and which therefore drives the economy. Reinvestment of surplus and the accumulation of capital stock will expand the modern sector so that eventually surplus labour will be fully absorbed. Lewis addresses the problem of international trade in his original paper, but it is probably fair to say that his main concerns are with the internal operation of the dual economy, and the process of accumulation of capital in the modern sector. Thus, a central matter for Lewis is the effect of movements in the internal terms of trade – between the urban and rural sectors – on profits and the capital accumulation process in the modern sector. In this particular sense, and despite the discussion of the international economy in the second half of the paper, it is probably fair to say that the Lewis formulation deals essentially with a closed economy. Twenty years


later Fei and Ranis consider the implications of the Lewis type of accumulation in an open economy and apply their framework to the (early) development of Korea and Taiwan (Fei and Ranis, 1976). More recently Ranis (1988) has given a useful reformulation of the original ideas of the earlier paper. The centrepiece in both analyses is the onset of a phase of ‘export substitution’, starting at the point where traditional exports are replaced by exports of labour-intensive manufactured goods. 3 This is a key turning point, because thereafter the absorption of surplus labour is greatly accelerated. So much so, claim Ranis and Fei, that debates on trade-off between growth and employment, which were characteristic of the seventies, became largely irrelevant. Once the economy had got into the export substitution phase it was expected to move rapidly to the next turning point, called by Ranis and Fei the ‘commercialization point’. At the commercialization point, surplus labour is fully absorbed, the real wage is no longer ‘institutionally’ determined, but becomes equated to the marginal product of labour in the rural sector. What implications does this model of entry into the international market for manufactured exports have for technological change and factor productivities? The question is simply answered (in the model) and the answer follows from labour market conditions. After the process of export substitution has started, and up until the commercialisation point, the idea is that the institutionally determined low real wage will rule. Once labour is fully absorbed, that is, once the commercialization point is past, the real wage will naturally rise. In the first, ‘precommercialisation’ phase, ‘the existence of relatively constant (and low) … real wages … should induce labour-intensive technology choices and, more importantly, labour-using technology change … in the dual economy’ (Ranis, 1988, p. 82). Then, after the commercialization point and full absorption of surplus labour, ‘increase in real wages … is expected to be accompanied by a shift towards more capital and skill intensive technology and output mix’ (Fei and Ranis, 1976). In short, labour productivity and real wages will remain low and stagnant after the initial shift to what Fei and Ranis call ‘export substitution’, whilst manufactured exports will rise rapidly. Thereafter, when surplus labour is absorbed, wages and labour productivity will rise more rapidly. As Fei and Ranis tell the story, the key role of the labour-intensive export phase is to accelerate the rate of absorption of excess labour in the economy above the closed economy rate of the Lewis model in its early form. The implication for technology is simple: labour intensity will prevail until the commercialization point is reached and the real wage begins to rise. In broad outlines Fei and Ranis give a fairly convincing framework for understanding the actual pattern of development in Korea and Taiwan in the early stages of their post-war development. There was, of course, a shift away from primary product exports as these economies successfully


entered trade in manufactured exports. Also, just as Fei and Ranis predicted, basing themselves on the experience of Korea and Taiwan, the rate of growth of exports and employment was accelerated strongly by the shift. And finally, it is clear that in all cases the initial shifts in the pattern of trade and output in manufacturing were towards the simpler types of manufactures in the first two or three ISIC two-digit groups – in other words towards higher labour intensities as expected in terms of the Fei and Ranis analysis. In addition, some other countries bear out the Fei–Ranis expectations in a more detailed way. Countries such as Mauritius and Sri Lanka, for example, have had precisely the low and more or less constant real wages, and the low productivity growth, which were predicted for the period of continued surplus labour. Furthermore, the recent historical pattern followed by Malaysia also seems similar to the conventional anticipation. Over the whole period 1970–90 Malaysia had a low growth of productivity (1.7 per cent), and a slightly higher growth of real earnings (2.8 per cent). Manufacturing employment grew rapidly (at 7.5 per cent), and by the mid-eighties labour shortages were beginning to be felt, and an import of unskilled labour started from neighbouring countries. At the same time as the labour surplus phase came to an end, a technological shift took place. Labour productivity growth accelerated to more than 4 per cent per annum in the second period (1980–90). (Evidence on movements of the real wage in manufacturing in this period is not available.) This Malaysian pattern is very close to the expectation that technology will be predominantly labour intensive in the first period of manufactured exports, whilst there is labour surplus, and will then shift to higher capital intensity and higher labour productivities as full employment levels are reached. 4 It is also notable that countries which show no significant growth trend in the pattern of manufactured exports, have not – in the main – shown any significant trend in value added per worker. To this extent they conform to the Fei–Ranis prediction5 – which carries the corollary that, in the absence of manufactured exports, labour absorption will be slow, and, in the presence of surplus labour, there will not be much of an incentive for capitalists to invest in higher productivity technologies. All in all Fei and Ranis give a pretty clear picture of the role of labourintensive exports in the initial stages of entry into world markets, and in the process of absorption of excess labour. It is a simple picture, in which the technological behaviour of firms in the manufacturing sector is determined in a direct way by the labour market. Whilst there is excess labour, the real wage stays constant and labour-intensive technology prevails. It is only after the commercialization point that new technologies become necessary and labour productivities will rise.


The model gives a helpful insight into the importance of labour-intensive exports at an early stage of development. There are, however, two major issues, which are to some degree empirical, which are not clearly addressed. The first is that Fei and Ranis give no clue as to why the crucially important ‘export substitution’ phase should have started. Other turning points – the commercialization point for example – have a clear economic rationale, but not the inception of the export substitution phase. That remains unexplained, and it is a crucial theoretical lacuna. Why do a group of firms that enjoy a fair degree of market protection (Fei and Ranis assume that there is a preceding phase of import substitution) suddenly launch themselves into the international market for labour-intensive manufactured goods? Indeed, the question is even more puzzling. It is well established that in most countries which are not NICs, in South Asia, Latin America or in present day sub-Saharan Africa, no such shifts to ‘export substitution’ took place. The import substitution process did not ‘incubate’ an export substitution phase. It created a vicious circle of economic inefficiency in which international competitiveness became virtually impossible for the individual firm to achieve. In a sense, therefore, Fei and Ranis beg a key question. One of the more important tasks for historians of the NICs is to elucidate what were the factors that led to an export substitution phase. Economists and economic historians of various schools of thought plainly join issue on this question. There are those who argue that the role of the state, in particular in Korea, was critical in the initiation of an export substitution phase. There are others who argue, often using Taiwan and Hong Kong as exemplars, that the state role was not necessary and indeed distorted the process. However, the latter kind of argument about the role of the state putatively slowing down a process that would have been even more successful without it is strictly counterfactual, and to that extent must be regarded as scientifically dubious. The fact is that we simply cannot know, in any normal scientific way, whether the processes of development and industrialisation in these economies would have been faster or more efficient without the intervention of the state. We can only observe that there were important state interventions, which certainly appear to have been positively influential in industrialization; and we can also observe that – though the comparisons are complex – some at least of the interventionist states seem to have had striking success. These points have a sounder historical and scientific foundation than the counterfactual arguments which at one point were very popular in Bretton Woods institutions, especially the World Bank. For the present, therefore, it seems sensible to rest on the commonsense view that the export substitution phase – where it happened – was importantly associated with state action and intervention. The second issue is that, on closer examination, the technological ‘predictions’ of the Fei and Ranis analysis are not borne out in all countries and are plainly wrong in some very important ones. The analysis predicts a


simple correspondence between labour market conditions and technology, whereby with an initial excess supply of labour, productivities do not change and export development is labour intensive. However, a closer look suggests that this correspondence is not always present. A number of countries have plainly experienced considerable technological advance and rising labour productivity, whilst still having large amounts of surplus labour. This is certainly true of China, India, Indonesia and Pakistan, and probably also Thailand. In addition to this, historical evidence on Korea also suggests rather strongly that there was a vigorous growth of labour productivity well before the point of full absorption of labour was reached. So we are left with the problem of explaining the apparently anomalous behaviour of the economies that, whilst they demonstrate the general importance of the export phase, have followed a growth path more characterized by technological change than the Fei and Ranis model would have supposed. The main question is: why did these economies follow dynamic productivity growth paths while they were still in the labour surplus phase of economic development? A number of reasons can be suggested. Firstly, the Fei–Ranis expectation that a commitment to labour-intensive technology would be sustained until surplus labour is fully absorbed is linked to strong assumptions about the working of the labour market, and in particular to the idea that, during this period, the real wage will be more or less constant – or ‘slow growing’ (Ranis, 1988). In practice, this assumption has not been borne out in many countries. In most countries there has been a strong upward shift in industrial real wages. A cursory examination bears this out. Amongst the high value-added growth countries there are a number which, throughout the period 1970–90, had excess supplies of labour in the Lewis sense. China, India, Indonesia, Pakistan and Thailand were certainly in this category. Despite this, the average rate of growth of real earnings per worker for these countries was 3.5 per cent per annum over the period. In addition, although surplus labour has been absorbed in Korea, the evidence shows that, in the early part of the period, before this had been accomplished, Korean real wages were already rising. So it could be argued that the reason why productivity increases in these economies took place so early (in the sense that there was still a labour surplus when they occurred) may be found in the ‘untimely’ increase in real wages. It might be argued that the only way to maintain competitiveness in the face of rising real wages was through a higher rate of technological change. There are, however, some problems with this argument. In the first place, it assumes that real wage rises took place independently of changes in technology. In fact, real wage increases could just as easily have been a result of the incorporation of technology that raised factor productivities. 6 On the other hand, this argument shifts the burden of explanation from one area to another. Differences in real wage growth between economies may have


been the cause of differences in the rate of growth of labour productivity, but then what causes the differences in real wage growth between economies in the first place? A second possibility is that the acceleration of technological change during the labour surplus phase may have resulted from pressures generated by technological change in the international economy. In order to remain competitive, firms in the domestic economy must reduce costs or modify products, either through technological change, or through some other means of cost reduction. So some countries deal with the competitive threat by holding down real wages, or even reducing them, whilst others respond by technological advances. This may be a more plausible explanation than the first, but it still leaves unanswered questions. In particular it is not clear why the acceleration of technological change internationally should affect economies in the highly protected import-substituting phase. And for this explanation to work, we must suppose it did have effects in the case of countries such as China and India. Indeed, on consideration, there is every reason to expect that the combined influences of increasing capital accumulation and increasing capital per worker would lead to labour productivity increases, perhaps associated with changes in the pattern of outputs, without necessarily being associated with increasing real wages. Labour productivity changes do not need to be associated with changes in the real wage. They may be stimulated by the search for higher profitability without any change in wage levels. Thirdly, it may be that the early onset of higher productivity in some countries is due to important supply-side differences – in particular the fact that some countries might have a better endowment of factors of production that make it possible to adopt new technologies. So, if the technologies becoming available internationally require proportionately large demands for particular factors – such as skilled labour – they may become profitable in countries where there is a supply of relatively low-waged skilled labour, even though there is a large excess supply of unskilled labour. It is possible that the countries mentioned above, some of which have long and substantial traditions in scientific and technical education, and substantial science and engineering capabilities, are differentiated from the technologically weaker economies in this way. Evidence on the supply of scientific and technically trained people would support this idea in the case of countries such as China, India, Thailand and Singapore – perhaps also for other countries. But puzzles remain, since the large Latin American countries also have long traditions of technical education and a comparatively highly educated workforce, and have nevertheless shown very limited increases in labour productivity. In short, there is no single explanation which can easily encompass the comparisons between all the countries in the analysis. This is not necessarily a major problem, since contingent conditions may vary widely between


countries, and there may therefore be more than one explanation for the various differences. It is not surprising that such a complex set of phenomena cannot easily be reduced to a single simple pattern. Plainly, despite the basically sensible framework which the Fei–Ranis model provides, there are some complexities of the real world that it does not address in an adequate way, and some of the points made above point in their direction. These complexities indicate that the framework of the model, though convincing and in part empirically supported, leaves something out. Essentially the Fei–Ranis framework says very little about the implications of technological change for the process of entry into world markets in manufactures. In this respect it is in line with most development economics of its time. In fact, for the sub-Saharan countries seeking to enter international trade in manufactures there are conditions in the international market which strongly influence the nature and requirements of the Fei–Ranis export substitution phase. The next section touches on implications of generic technological change7 and innovative competition, as well as on the question of path dependencies resulting from development of labour-intensive export development. In the modern world economy these considerations are an essential complement to the preceding analysis.

3 Some problems for new entrants: generic technological change and path dependencies The preceding section suggests that, in reality, technological changes involving increases in labour productivity have been important at a much earlier stage than suggested by the Fei and Ranis framework. That framework gives the prediction that advances in labour productivity will become important only after the commercialization point is reached; experience suggests they are important at an earlier stage than that. The probable reason for this is that although the Fei–Ranis model is a valiant attempt to ‘open’ the original Lewis model, it remains, in certain respects, essentially focused on closed economy ideas. It does not take into account the importance of changes in the international environment, which begin to influence the economy as soon as it is opened. In particular, it does not consider the significant influence which technological change in the international manufacturing economy is likely to have on new entrants, even at the earliest stages of ‘export substitution’. So the first point to be made in the following discussion is about the impact of generic technological change and innovative competition in the international economy on the ‘export substitution’ phase. Although we have chosen to express the argument in these terms, that is, in terms of ‘generic technological change’ – a terminology appropriate to the modern discussions on technology – it is


in fact a perfectly general point which can be made in a much more reductionist way; it depends only on the fact that Fei and Ranis leave out the fact that factor productivities in the world economy are changing all the time, mainly increasing, and that this has important implications for the new entrant developing country – that is for the country entering on the export substitution phase. Essentially it implies that improvements in factor productivities – labour productivity in the simplest case – have to be achieved from the very beginning of the export substitution phase and not just from the later ‘commercialization point’. So countries need to organise for technological change even when there is excess labour available. The problem of ‘generic technologies’ is tackled in Section 3.1 This need for technological advance and improvements in factor productivities at an early stage in the process of ‘opening up’ does not gainsay the argument that these early stages will be specialized in labour-intensive manufactured outputs. The important point to be made in the following is that precisely because technological changes are increasingly generic, the sectors producing labour-intensive outputs will not be exempt from technological change. These sectors – textiles, garments and the like – are for that reason no longer technologically stagnant, as they were generally considered to be in the period in which the Fei–Ranis paper was written. But the fact that the early stage of export substitution must still be labour intensive remains: labour intensivity characterizes the immediate source of static comparative advantage from which the economy must start. This, of course, is quite consistent with the fact that there is a need for improving labour productivity at the same time: there is no reason in principle why the factor productivities of labour-intensive technologies should not increase. The question is whether this initial commitment to technological change and technological learning processes associated with labour intensive outputs in the initial phases of export substitution may not set the economy on a trajectory of technological skill development, which creates a continuing commitment to a more labour-intensive, less skill-intensive and less capital-intensive path in the longer run. Does an initial entry into world markets for manufactures in labour-intensive lines of production create a continuing dependency on such lines of production, and restrict the skills available for technological upgrading at a later stage? In other words, to use the language of Paul David, does it not create a ‘path dependence’? Although this may seem to lie outside the line of sight of countries which are just entering the world market, it is sooner or later a question which they will be obliged to face. If there is likely to be such a path dependency, it is important to know about it from the beginning – in order to do something about it later on, when a shift away from labourintensive production may become desirable and necessary. So Section 3.2 of this paper deals with the question of ‘path dependencies’ which an


initial commitment to labour-intensive production may create, largely because of the pattern of technological skills it generates through ‘learning by doing’. 3.1

Generic technological change

It is no doubt risky to generalize about the rate and direction of technological change in the international economy. However, some of the trends which seem to be emerging are especially important – sufficiently so to justify a few generalizations, however risky they may be. A common generalization is that the international economy is increasingly influenced by generic technological changes. From the point of view of the developing countries, the appearance of generic technologies is potentially very important. This term is usually used to describe the fact that many of the new technologies have fields of application across a multiplicity of sectors. There is not much doubt that such a pattern is developing. Some examples: Alcorta (1995) has shown the way in which industrial automation technologies spread across sectors; a UNIDO paper (UNIDO, 1995) demonstrates with great clarity the generic nature of new materials technology; a paper by Steinmueller and Bastos (1995) shows the situation for information and communication technologies. Generic, or multisectoral, technological changes have many implications for the production system. One which is especially significant for the present discussion is that they have introduced innovative competition into sectors which were previously seen – especially in the Fei–Ranis scheme of things – as essentially stagnant from a technological point of view, and are therefore characterized by straightforward price competition of the Marshallian kind. Competition through innovation is distinctive and different to Marshallian competition as described in standard economics textbooks. Price competition – based on minimizing the costs of production on a given type of technology – is a mechanism for re-establishing an equilibrium in the economy. Innovative competition, as it was first described by Schumpeter (1912), is a means whereby firms create unique advantages for themselves through temporary sole possession of a piece of technological knowledge, and so profit from a temporary disequilibrium. The effect of a high rate of generic technological change in the economy is that this type of competition will prevail in many sectors of the economy. This has important implications for firms in developing countries seeking to enter international trade in manufactures, since it importantly affects the terms of entry. Christopher Freeman (1982) has classified the competitive responses of firms in industries which are characterized by innovative competition. At the leading edge of these industries (from the technological point of view) are the innovative firms, seeking to capture a lead over the rest of the industry by establishing a unique process or product. Follower firms may


pursue different strategies in response. Some will seek to innovate themselves. Others will try to exploit the advantages of being a follower, by imitating the original innovator – if necessary by licensing its technology. Still others will seek alternative, more defensive strategies. For example, if the new product arising from the innovation is an imperfect substitute for the old, firms may continue to produce the older product. Or they may continue to use the old methods of production – if there is a process innovation. As Freeman points out (1982, pp. 169ff.), follower firms of this kind require some compensating advantages in order to maintain themselves in competitive production. Follower firms in developing countries usually attempt to exploit low labour costs, or advantageous access to materials in response to innovative competition, though in the more industrially advanced developing countries many firms will follow imitative strategies based on the international transfer of technology. An important aspect from the point of view of developing countries is that the generic nature of technological change has meant that patterns of innovative competition are appearing in many of the sectors which before were considered to be technologically stagnant. Amongst these are the sectors which have long been regarded as the ‘traditional’ sectors for early industrialisation, such as textiles and garments production, for example. It does not follow that firms in developing countries have to become innovators in order to compete; nor do they necessarily have to adopt new technologies at a high rate. It does mean, however, that even in the older traditional industries, which are so important in early industrialisation, the pressure of innovative competition will be felt. The terms of entry will be more severe than in the past, and the requirements for maintaining competitiveness will be more severe as innovative competition develops. Steinmueller and Bastos include some interesting reflections on what this means in practice (1995, p. 9). They point out that whilst the working out of comparative advantage means that countries will always have some sectors in which they are competitive, the terms of trade are determined by relative productivity of trading partners. So ‘if developed nations’ productivity advances substantially outstrip those in developing nations, the consequence is slow growth or even a decline in real wages offered in developing countries’. This has important implications for the terms on which industrialisation may take place in the technologically less advanced developing countries. Fortunately, the impact of innovative competition, even though generic in form, is uneven across sectors. It is probably a fair generalization that all sectors in the manufacturing system have experienced accelerated generic technological change, but it remains the case that it has been more pronounced in some than in others. The traditional sectors of developing country industrialisation have been less exposed to innovative competition than others. The route to industrialisation through initial production of technologically simple products in a comparatively labour-intensive way is


still open, though it is narrower than before. In these products, there is still a high degree of conventional price competition, which developing countries are in a better position to meet. 3.2

Patterns of path dependency

A particularly striking instance of path dependency and its implications for new entrants into international markets comes from some recent developments in trade theory. These developments are mainly derived from ‘new growth theory’, in which technological change is treated as an intrinsic part of economic activity. Their implications for trade theory rest on differences in factor productivities arising from differential learning, or from differences in the production functions facing different economies. A recent review is Grossman and Helpman (1995). Barros (1993) is an interesting attempt to draw conclusions from the new growth/new trade theory approaches for developing countries. For present purposes, the main point of interest is from Krugman (1987). The first point is that where learning effects are important in determining the relative productivities as between trading countries, there will be a tendency for the existing trading pattern to get ‘locked in’. Essentially countries become relatively more productive in those branches in which they are specialized, and the short-run pattern of comparative advantage is reinforced by this; to quote Krugman, ‘once a pattern of specialisation is established, it remains unchanged, with changes in relative productivity acting to further lock the pattern in’ (Krugman, 1987, p. 46). In some Latin American countries, for example, the relative efficiency of production of resource-based industries is probably reinforced by the exporting from them. In principle this presents advantages, of course, but it also means that it is increasingly difficult as time goes by to make changes in the trading pattern. Furthermore, if the learning elasticities in such sectors are lower than in the sectors where advanced country trading partners have a comparative advantage, the Latin American economies could be committed in a long-term sense to a low productivity growth trajectory. 8 The same points would apply, of course, to economies such as Mauritius and Sri Lanka, with their heavy commitment to ‘low’ technology, labour-intensive lines of production.9 This line of analysis could have special importance to the ‘low productivity exporters’ as they reach the point where labour is fully employed. If at that time they are locked in to a pattern of relative productivities inherited from the past, they could experience serious difficulties in shifting to new lines of production with higher labour productivity (in the way the Fei–Ranis approach suggests is necessary). 10 We will discuss this problem briefly in the next section.11 The idea that trade patterns may get ‘locked in’, as described above, is derived from the learning process. Once firms are committed to a particular


line of production, the learning processes this sets in train – whether ‘automatic’ in the Arrow tradition (Arrow, 1962) or the result of conscious managerial decision and resource allocation12 – reinforce the inter-industry pattern of comparative advantages, and, since the same thing is happening in trading partner countries, it becomes increasingly difficult to change the pattern. This is an example of ‘path dependence’ – which might briefly be described as a recognition that ‘history matters’. Learning processes13 will, obviously, produce many situations of path dependence. From the present point of view, path dependence is important because it will influence the possibilities of shifting between the types of growth path (that is, high versus low productivity growth paths) which differentiate developing countries. There are two levels at which relevant kinds of path dependency may be set up. First, the technological learning processes within firms are path dependent. David recognizes this: ‘Because technological learning depends on the accumulation of actual production experience, short sighted choices about what to produce, and especially about how to produce it using presently known methods, also in effect govern what subsequently comes to be learnt’ ( David, 1975, p. 4). Dosi (1988) describes the cumulative learning processes which underlie the accumulation of technological capability in enterprises. 14 There are three distinctive features of these learning processes. First, they tend to have important firm-specific features. Although there may be spillovers of technological know-how between firms, a good deal of the learning process in a firm differentiates it from its competitors. Secondly, learning processes create a good deal of ‘tacit’ knowledge – that is knowledge specific to the application of particular processes inside the firm, and which is neither codified nor easy to codify. This is the type of technological capability that can only be acquired by ‘doing’. Thirdly, whilst some knowledge may accumulate ‘spontaneously’ through the experience of production, for the most part the accumulation of technological capabilities depends on the allocation of time and effort by the personnel of the firm, and depends on explicit management decisions. But though accumulation of technological capabilities takes place in the first instance within production units (and increasingly in service enterprises too), the broader institutional environment within which firms operate is also important. In recent times this environment has become called the ‘national system of innovation’, and important attempts have been made to describe it systematically (Nelson, 1993). The national system of innovation is the second level at which there are important path dependencies. It has a number of components other than enterprises. These differ in form from country to country, but are present in most. In the first place there is the education system – especially those parts concerned with scientific and technical education. The early creation of a highly skilled and educated workforce is generally agreed to have been a


key element in the success of the first generation of NICs. On its importance in Korea see Pack and Westphal (1986). Second, there are the various institutions engaged in scientific and technological research (outside of enterprises). These normally include the universities, as well as various national laboratory organisations. Sometimes, especially in developing countries, a large part of the scientific and technical capability of a country is ‘tied up’ in these institutions, and a major policy problem is how to relate this capability to national development objectives. Sometimes also, as in the United States for example, these institutions grew out of major national programmes – such as the space programme or defence programmes. Third, there is a set of important ancillary institutions – survey systems, technical information systems, standards systems, technology transfer organisations and so on. In most countries the institutions making up the national system of innovation play an important part in technological development within enterprises, whether through creating a supply of skilled persons, or through facilitating the acquisition of technology from abroad, or through provision of technological information, or through the support of university or other research activities on which enterprises can draw. It is important not only that the institutional structure of the national system of innovation should be present, but that it should be functionally related to the requirements of the enterprises which are at the sharp end of the process of acquisition of technological capability. The longterm development of these institutions and their organic relations to the enterprise sector has played a large part in the process of technological development in many of the high productivity growth developing countries. How then are high and low productivity growth economies distinguished from one another as far as the technological capabilities are concerned? We can give an impressionistic but probably reasonably accurate response along the following lines. First, we expect that in the high productivity growth economies we will find production and service enterprises, especially in the export sectors, in which there are considerable concentrations of technically skilled persons, and where – more importantly – there is a vigorous process of technical learning taking place within firms. Second, we would expect that there will be close links between production and the rest of the national system of innovation. In the low productivity growth economies we would expect to find firms which are solely concerned with repetitive production tasks, in which there is no concern with learning or change. Very little research has been done on these differences, but there is a good deal of impressionistic evidence to support the picture we have drawn. This hypothetical description will also make clear that the shift from low productivity to high productivity paths is not as easy as may appear. It will depend on generating learning processes within firms on the


one hand, and on linking the key elements of the national system of innovation to services and production on the other.

4

Points in conclusion

This note explores the problem, especially important in the sub-Saharan economies, of entry into the world market for labour-intensive manufactured goods. It argues that this mode of entry into industrial markets is unavoidable for sub-Saharan countries. There has, however, been relatively little theoretical attention to how countries acquire initial entry into world export markets for manufactures – compared, that is, to the amount of attention given to the processes of technological upgrading which some countries have successfully followed after their initial entry. Indeed, the Fei–Ranis model, which puts strong emphasis on the importance of labour-intensive manufactured exports in generating employment in the industrial sector, gives no indication of the policies which countries may follow to initiate this process.15 There is much that is unknown on this topic. Focus on the latter-day technological accomplishments of the NICs in such advanced areas as large memory chips and the like has obscured the question that is much more important from the point of view of economies such as those of the sub-Sahara: namely, how did the whole process start? What policies did today’s NICs follow to initiate what Fei and Ranis call ‘export substitution’? Unfortunately Fei and Ranis do not address the question, despite its central relevance to their model. And we cannot, without a good deal more historical research, solve it ourselves. The discussion in Section 2 of this paper, however, does indicate that government intervention seems to have been one of the factors involved. The Fei–Ranis model is also weak in the treatment of technological change. There is a presumption that as the period of excess labour comes to an end (that is, as the commercialization point is reached), there will be a switch to more capital-intensive technology in the manufacturing sector, in order to deal with the need for increases in labour productivity. In fact, it is very likely that both labour productivity increases and new products will become important long before this. The circumstances of the world economy – in particular the incidence of generic patterns of technological change – will require new entrants to world markets for labour-intensive manufactured goods to follow more active policies of technological change than is usually assumed. And this requirement will assert itself at an early stage of industrial export development, well before the Fei and Ranis ‘commercialization point’ at which labour is fully absorbed in the industrial economy. Essentially, Fei and Ranis failed to encompass the relevance (indeed, the potential threat) of world technological change in labour-intensive outputs for new entrants to world


markets in these products. They do not address this point, perhaps because there was not much technical change going on at the time, in labour-intensive outputs. But there is now, mainly because of the generic nature of much technological change, and this is a fact that undermines some of the model’s predictions. Technological change and advances in labour productivity become important from the beginning of the Fei–Ranis ‘export substitution’ phase, not just from the point where excess labour has been fully employed. International patterns of technological change acting on the open economies of the modern developing countries make that inevitable. A question that lies outside the immediate range of the Fei–Ranis discussion is whether an initial entry into international markets through labourintensive exports, which is an inevitable point of entry for most sub-Saharan countries, implies a long-run commitment – through the path dependencies it creates – to a low productivity growth trajectory, and hence to slower rates of growth of factor productivity and real wages in the industrial sector. The kinds of technological capability which producers develop tend to determine the technological options open to them at their next round of investment. It must remain unclear for the moment how serious a problem this is for developing countries, and whether policy intervention or the market is needed to solve it, supposing it is important. The question, though, is highly relevant, because in the longer run a shift towards production and export of higher value added goods at higher real wages seems to be an essential part of economic development. Countries will not be satisfied with a permanent commitment to labour-intensive lines of production, however essential they may be to the initial entry into world markets.

Notes * The author is Director of the United Nations University Institute for New Technologies (UNU/INTECH). This paper draws heavily on two discussion papers in the UNU/INTECH series, which are referred to in the text. The author benefited from the comments of his colleagues at UNU/INTECH as well as from others including Prof. Ed Steinmueller of MERIT, Prof. Jorge Katz of ECLAC in Santiago, Prof. Larry Westphal at Swarthmore College, USA, as well as participants at the INTECH/ECLAC conference in Marbella, Chile, in 1995, and at the UNIDO Global Forum on Industrialisation in Delhi. Prof. Eddy Szirmai, of TU Eindhoven, as Editor of the conference volume in which this paper appears, gave invaluable comments which are acknowledged at various points in the text. 1. It seems to me likely that some people will see this observation as an example of a kind of determinism. If so, they have an ahistorical case to make, and it is their obligation to do so. 2. Sustained growth of exports is defined as follows. The growth rate of exports is determined by fitting the logarithmic form of the standard expression for com-


3.

4.

5.

6. 7.

8.

9.

10.

11.

pound growth to the time series of manufactured exports (1970–90). The growth rate is determined from the coefficient of the logarithmic term. If this coefficient is significant for the whole time series, at the 1 per cent level, growth is defined as sustained in our term. Some more recent research will result in slight modifications to these statistical observations, but not sufficient to require modifications to the points made in the ensuing text. As Szirmai pointed out in a private note, the implicit symmetry between the terms ‘import substitution’ and ‘export substitution’ is potentially misleading. Import substitution refers to the competition between imports from abroad and domestic production for the domestic market. Export substitution refers to the increasing share of manufacturing in exports, but it is not clear whether there is a process of substitution. Nevertheless, in this paper, we will continue to use the accepted terminology. The data on which this conclusion is based are discussed in Charles Cooper, ‘Technology, Manufactured Exports and Competitiveness’, UNU/INTECH Discussion Paper no. 9513, December 1995. Although for these countries the matter is somewhat different since – as their export data indicate – they have not really entered the ‘export substitution’ phase. The data available are too weak to support tests of causality. The terminology ‘generic technological change’ is used very frequently in a certain part of the ‘technology’ literature, often in an unclear way. The dictionary definition of ‘generic’ is ‘pertaining to a class, not specific or special’. So generic technologies are not sector specific or unique to particular lines of production, but are widely relevant to many. They are technologically and not industrially defined. This is what Barros seems to have in mind when he writes of specialization having a ‘negative effect on productivity increase’ (Barros, 1993, p. 545), but his discussion is vague and unconvincing. The analysis is based on the assumption that whilst there may be international spillovers of technological capability within industries, there are no spillovers between industries. The lock-in effect would be much less severe if there were inter-industry spillovers. It might be argued that one of the implications of generic technological change is that such inter-industry spillovers will be important. Krugman’s analysis has not been extended to the case of labour surplus economies operating with a constant institutionally determined wage rate, but that does not change the validity of the present line of argument. The second point to emerge from Krugman’s analytics is that there is clearly a way out of the ‘lock-in’ – along lines which he identifies with Japanese industrial policy, and which is nowadays more commonly associated with the policies of selective protection followed by Korea. The idea is that governments may use ‘temporary protection to permanently shift comparative advantage’. The protection will be directed to goods which are just outside the present pattern of national comparative advantage, and applied for just so long as is necessary to raise relative productivities in their production to the point where a new area of comparative advantage is established. Krugman refers to this as the policy of a ‘narrow moving band’ of protection (Krugman, 1987, pp. 48–9). It is an interesting reflection of the notion of ‘technological upgrading’, and has considerable empirical foundation in the history of industrial policies in some countries.

132 The Industrial Experience of Tanzania 12. The recognition that learning processes involve important resources has a substantial history. As far as work on developing countries is concerned, Katz’s work in Latin America provided the essential empirical basis (Katz, 1987) and was the point of departure for a substantial literature. Much later the point became embodied in theories of endogenous technological change. 13. On the kinds of technological learning which are important in firms in developing countries see Dahlmann et al. (1987). 14. This and other basic material on accumulation of technological capability is surveyed in Cooper (1993). 15. There is an implication that the transition will be from import substitution to ‘export substitution’, but it is not clear what mechanisms will make this happen.

Part II Innovation, Technological Capabilities and Choice of Techniques

This page intentionally left blank

5 Public Choice, Technology and Industrialization in Tanzania: Some Paradoxes Resolved Jeffrey James*

1

Introduction

The public choice approach has already been used in the African context to explore the political rationality of policies that seem difficult, if not entirely impossible, to justify on purely economic grounds. A well-known study by Robert Bates (1981), for example, sought to explain why governments in Africa tend to adopt agricultural policies that are blatantly harmful to the interests of most farmers in the region. More recently, rent-seeking behaviour was used by Gallagher (1991) to explain variations in growth rates across a wide range of African countries. What has not been applied to any of those countries, however, is the area of public choice theory that deals specifically with the preferences and behaviour of government bureaucrats: the so-called political economy of bureaucracy. Yet, as we shall argue below, this important strand of the public choice literature helps to explain some of the most paradoxical aspects of technology and industrialization in the public sector of one particular African country, Tanzania. More specifically, our argument will be that these paradoxes can be explained by the following propositions. The first is that, contrary to what is almost always assumed in the literature on development economics, bureaucrats do not in fact have preferences defined over technologies (that is, they do not choose technologies in the usual sense of the word). Rather, their preferences are thought to be defined over projects and particularly those projects that enable the institutions to which they belong to grow as rapidly as possible.

2 Technological behaviour in the public sector – two paradoxes To a much greater extent than elsewhere in the Third World, the industrial sector in Tanzania (as in sub-Saharan Africa more generally) has been dominated by enterprises owned by the state and to a correspondingly greater 135


extent than in the other regions, therefore, the technological aspects of industrialization in Tanzania need to be understood in relation to the behaviour of those enterprises.1 This understanding is made more difficult, however, by several paradoxical features of public sector behaviour in the industrial sector of that country. One such paradox is the marked discrepancy between actual technological behaviour in the public sector and the behaviour that would have been consistent with the particular type of socialism pursued by Tanzania since 1967. For, whereas that type of so-called ‘African socialism’ laid considerable emphasis on small-scale, labour- and local-input-intensive technologies, the public sector has tended instead to use technologies with precisely the opposite features (that is, large-scale, capital- and import-intensive technologies). What is just as paradoxical, however, is that this apparent preference for certain types of technologies has not been uniformly applied across the public sector. Indeed, one can also find examples in the public sector of precisely the opposite forms of technological behaviour from those that have just been described. In some cases, these pronounced technological variations occur even within the same sector at around the same point of time.

2.1 The first paradox: planned versus actual technological behaviour in the public sector In order to make this first paradox as clear as possible, it is necessary to recognize that socialism in Africa after independence took a number of different forms. The first form, known as African socialism, is associated mainly with the governments of Nyerere in Tanzania and Nkrumah in Ghana during the 1960s. This form of socialism has been described as ‘Afrocentric’ and ‘non-aligned’, in that it purported to be adapted specifically to African conditions (Chazan et al., 1988, p. 150). African socialism needs to be contrasted with the ‘Afro-Marxist’ type of model that emerged in countries such as Mozambique, Angola and Ethiopia during the 1970s. In these and other countries that formed part of the ‘second wave’ of socialism in sub-Saharan Africa (Rosberg and Callaghy, 1979), the distinctiveness of the African situation tended to be rejected in favour of the established principles of scientific socialism (that is, of Marxism-Leninism). This distinction is important because it bears so heavily on the type of technology that the developing country is predisposed to select. Whereas socialism of the Marxist-Leninist variety demands the most modern, advanced technologies, this is not at all true of the African socialism that was practised in Tanzania after 1967. Immediately after the Arusha declaration, for example, President Nyerere made his position clear when he argued that:

Public Choice, Technology and Industrialization 137

even when we are building factories which serve the whole nation, we have to consider whether it is necessary for us to use the most modern machinery which exists in the world. We have to consider whether some older equipment which demands more labour, but labour which is less highly skilled, is not better suited to our needs, as well as being more within our capacity to build and use. (Nyerere, 1968b, pp. 98–9) There were indeed many cases, he believed, where the needs of society could better be met by labour-intensive, small-scale technologies than by large-scale mass production. These cases, furthermore, were closely in accord with his view that industry should be decentralized to the maximum possible extent. ‘In so far as there is a choice, we in Tanzania would infinitely prefer to see many small factories started in different towns in our country rather than one big factory started in any one of them’ (Nyerere, 1968a, p. 107). These early ideas were consistent with and indeed embodied in the key planning documents that were to form the basis of Tanzania’s industrialization strategy in the post-Arusha period (see Chapter 1). The second fiveyear plan, covering the years 1969 to 1974, for example, not only provided explicit encouragement of labour-intensive techniques, but also emphasized the importance of more decentralization and more effective linkages between large- and small-scale industries (ILO, 1982). For the period after 1974, the so-called ‘basic industry strategy’ strongly reaffirmed the principle of self-reliance, which played so central a role in the Arusha declaration and in subsequent policy pronouncements by the top political leadership. This goal was to be achieved, among other ways, by the encouragement of a local capital goods industry which would lessen Tanzania’s dependence on imports of foreign technology. Moreover, the major objectives laid down by the ‘basic industry strategy’, such as employment creation, equality of income distribution and dispersion of industry, suggested implicitly or explicitly that where a choice existed, technologies should generally be relatively low cost, labour intensive, simple and smallscale (Williams, 1976). Beginning in the 1970s, however, a long list of scholars in Tanzania and elsewhere pointed out that the technologies actually being used in the public sector typically had just the opposite features. That is to say, they were usually large-scale and inefficient, as well as being capital and import intensive.2 As such, therefore, these technologies also tended to have the effect of centralizing, rather than decentralizing, the location of industry in Tanzania. Far, therefore, from promoting the most important goals of the state, public enterprises in the industrial sector seemed to be doing just the reverse. It was not that there were no choices available to the managers of state-owned enterprises. On the contrary, over a wide range of


manufacturing industries a number of very different technological alternatives usually presented themselves. Table 5.1, for example, shows that the available techniques for ricemilling range from a small-scale huller (producing four to eight tons per day) to a large-scale roller (producing 120 tons per day), with the investment requirements per worker of the latter exceeding those of the former by a factor of around seven. Table 5.1 also shows that the estimated benefit–cost ratios vary inversely with the scale of the different techniques: the rice huller has the highest ratio of benefits to costs, while the rice roller has the lowest ratio. If the choice of the former could thus have been justified on the grounds of employment creation and efficiency, it was also preferable from the standpoint of industrial decentralization and self-reliance. Although one might have expected these considerations to prevail in the light of our earlier discussion, the state-owned milling corporation chose instead to expand its capacity on the basis of the largescale (120 tons per day) roller technology. Given the overriding importance that was apparently attached to the need for self-reliance in the period after 1967, one would also have expected state-owned enterprises, such as the milling corporation, to pay explicit attention to the acquisition of technological capabilities of various kinds. For there was certainly no lack of awareness on the part of national planners that a strategy of self-reliance had necessarily to be based on the substitution of domestic for foreign technological capabilities (an awareness that was perhaps most clearly articulated in relation to the ‘basic industry strategy’ in the 1970s), and the public enterprise, as an extension of the state, ought, one would think, to have been a key instrument in effecting that transformation. In practice, however, not only has there tended to be an increase rather than a decrease in the public sector’s

Table 5.1

Alternative rice-milling technologies Initial fixed cost (TSh 000)

Rice huller 4–8 tons per day Rice roller 24 tons per day 60 tons per day 120 tons per day

Total capital TSh per worker hour

Benefit–cost ratio (shadow prices, full capacity)

93.0

2.4

7.94a

1 752.5 4 410.0 7 927.0

14.3 11.9 16.7

2.52 1.69 1.52

Source: Bagachwa (1992), tables 1 and 5. Note: a 6 tons per day.


reliance on imported technology, but also a tendency for that technology to supplant rather than to stimulate the acquisition of indigenous technological capabilities (Wangwe, 1986, 1992a). 2.2 A second paradox: pronounced technological variations in the public sector Though these various departures from what one would have expected well describe the technological behaviour of the vast majority of firms in the public sector, they do not describe that of all of them. Indeed, one can also find examples in the public sector of precisely the opposite forms of technological behaviour from those that have just been described. And, especially when these pronounced technological variations occur within the same sector at around the same point in time, the possibility of a second paradox emerges, namely, that the state holds inconsistent technological preferences – preferences, that is to say, which, under similar conditions, lead to the choice of one particular technology at one point in time and to a completely different technology at another point in time. More formally, the paradox arises in that the state’s preferences for such sharply diverging methods of production need to be represented by intersecting rather than non-intersecting indifference curves, and the former, unlike the latter, violate the transitivity assumption of traditional micro-economic theory and welfare economics. Two examples describe this paradoxical type of behaviour especially clearly. Research on Tanzania’s textile industry, for example, has revealed a very wide range of factor intensities among state-owned enterprises, and in some of those enterprises, the choice of technology was made at approximately the same time and under similar conditions (see below). The other example comes from the brick-manufacturing industry, where, ‘The [Tanzanian] state has been involved in technological development, as well as in innovation. But what is especially interesting about this is the diversity of state actions. Two very different types of technological choice have been made by the ‘same’ state, one of an almost unbelievably inappropriate nature, the other much more relevant’ (Kaplinsky, 1990, p. 93). Mainly because it is the better documented of the two examples, we shall use data from the textile industry to illustrate the apparently inconsistent – and hence paradoxical – nature of the state’s technological behaviour. In particular, we shall confine ourselves to a comparison between two large-scale integrated textile mills, that, in spite of having been established more or less simultaneously at the end of the 1960s by the same public institution, the National Development Corporation, nevertheless exhibit a number of rather remarkable technological (and other) variations. One plant, for example, was highly labour intensive, whereas the other was highly capital intensive. One textile firm was one


of the most efficient in the entire industry, whereas the other performed poorly on virtually all the usual indicators. In one enterprise, indigenous technological capabilities were successfully acquired, while in the other they were not. Still another case that is difficult to reconcile with the general pattern of technological behaviour in the public sector is to be found in the farm implements subsector, for there is one state-owned enterprise in that sector that has used relatively simple labour-intensive technology (in conjunction with other factors) to attain a highly competitive position (as measured by domestic resource costs) (World Bank, 1987a). This firm is also unusual in the rapidity with which it was able to dispense with foreign technical expertise – that is, in the rapidity with which it was able to acquire indigenous technological capabilities of various kinds (Barker et al., 1986).

3

Existing explanations and their limitations

The earliest and most detailed attempt to explain the foregoing paradoxes was made by David Williams (1976) with particular reference to the textile industry. He showed, among other things, that the pronounced disparity between the technologies chosen in the industry could not be explained on the basis of the existing literature on the choice of technology in developing countries. It is often argued in this literature, for example, that technological differences between firms reflect differences in the types of products they manufacture (where differences refer to variations in the characteristics that the various products embody). In particular, it is commonly argued that labour-intensive technologies tend to produce goods with a higher proportion of functional or ‘low-income’ characteristics than capital-intensive techniques. Williams, however, found ‘no grounds for assuming that any particular type of technology in the observed range was dictated by product characteristics’ (Willams, 1975, p. 3). Nor was he able to find much evidence in support of another well-known category of explanations in the choice of technology, namely, those that impute various types of non-economic preferences to decision makers in developing countries. Perhaps the best known of these explanations is the ‘engineeringman’ hypothesis advanced by Louis Wells (1975). His contention, in brief, is that under conditions of imperfect competition the preference of ‘engineering-man’ for sophisticated automated technology and modern products dominates the concern of ‘economic-man’ to minimize costs of production. Yet, while there were certainly enough departures from perfect competition in Tanzania’s textile industry at the end of the 1960s to enable ‘engineering-man’ to hold sway, one would then have expected a uniform bias in favour of capital-intensive techniques, rather than the coexistence of techniques with markedly different factor intensities, as is most clearly


illustrated by the comparison between the two textile mills (see Table 5.2 below). As Williams puts it, ‘The engineering-man hypothesis would not predict that a single investor … would set up both capital-intensive and labour-intensive plants at the same time’. (Williams, 1975, p. 7). The same problem, one should note, would apply just as much to the various other technology-related objective functions that have been proposed in the literature, such as, for example, the sense of ‘national pride’ which is sometimes said to be evoked by the use of the most modern technologies in developing countries (Winston, 1979). 3.1

Project versus technological preferences

For this reason, Williams (1976) suggests that one should investigate instead the nature of non-technology-related objective functions in the public sector. He argues that managers are concerned essentially with projects rather than technologies, and that they are especially concerned with maximizing the number of projects that can be initiated and implemented. In seeking to meet this goal, managers tended to favour highly packaged projects – a tendency that was already present among parastatals by virtue of skill constraints and by lack of information about technology and other markets. The main reason was that highly packaged projects generally also offered distinct advantages, not just in terms of finance but also in terms of the relative ease and brevity with which they were able to be implemented (as is perhaps most obviously the case with turnkey projects). In fact, ‘a project package which included financing and which was “ready to go” would often be accepted with little question’ (Williams, 1976, p. 163). What evidence is available does tend to support the view that public enterprises favour packaged projects, at least in comparison with similar privately owned firms. A study of public and private sector industrial projects by Wangwe (1986), for example, found that it is mainly in the public sector that turnkey projects have been adopted in Tanzania. Under these circumstances, so the argument goes, the ‘choice of technique’ is just the ‘fall-out’ or residual from the particular package of foreign finance and other related project inputs that happen to be chosen. What seems to be decisive in this process, which, one should emphasise, actually excludes technological issues from the purview of the typical project, is not the value of one project in relation to others (as measured by the major development goals), but rather the demands made by each of them on scarce equity finance. For their part, The parastatals responded to the bureaucratic realities by presenting projects in a manner best calculated to ensure acceptance. Thus, a lowcost labour-intensive textile mill, completely financed outside the budget, would be preferred to any mill for which more financing from


the budget were required. But if circumstances were such that the more acceptable package available consisted of a high-cost capital-intensive mill, then that would be chosen. (Williams, 1976, p. 165) Let us now consider how this observation can be applied to the concrete case of the two textile mills that were established, as noted earlier, at more or less the same time by the National Development Corporation. 3.2

A comparison of the two textile plants

Table 5.2 sets out the main technological differences between these two factories. It shows that the one plant, Friendship, is considerably more labour intensive than the other plant, Mwanza, using as it does two and a half times as many workers per unit of output, together with an appreciably lower amount of capital. In circumstances where labour is comparatively cheap, as in Tanzania, it is not surprising that the labour-intensive alternative should be the more profitable of the two textile technologies, as shown in Table 5.2. What then were the particular circumstances that surrounded the more or less simultaneous selection by the same public sector institution of these two factories? Friendship, it seems, was originally conceived after high-level political contacts between the People’s Republic of China and Tanzania, and the project was financed entirely by a long-term interest-free loan from the Chinese government. Because the proposed textile factory was both politically sanctioned and externally financed it was approved without much ado. And the technology that resulted from this externally financed project in no way reflected any technological concerns on the part of the Tanzanian bureaucracy, whose objectives tended to be defined purely in terms of projects. Rather, the technology that was ‘chosen’ for the Friendship project appeared to reflect instead the conditions in the supplying country, and in particular the fact that the Chinese were familiar with older and relatively labour-intensive vintages of textile technology (indeed, the particular vintage used at Friendship happened to be the most labourTable 5.2

A comparison of two textile plants Friendship

Capital cost up to 1969 (million shillings) Production of woven fabrics in 1975 (million linear metres) Number of employees in 1975 Profit in 1975 (million shillings) Cost of carded cotton in 1973 (shillings per tonne) Labour hours per tonne of carded cotton in 1973 Source: Coulson (1982b).

61.5 24.0 5057 2.8 2512 247

Mwanza 106.5 22.5 2486 2.3 2910 98


intensive of the technologies that were being produced at the time in China). A year later, by contrast, a very different – albeit perhaps equally attractive – package of project characteristics presented itself to the Tanzanians in the form of the Mwanza textile mill. One reason why this project differed from the Friendship case was that is was financed by a supplier’s credit rather than by bilateral foreign aid. The most important difference, though, was that the financial source of the project was located mainly in France rather than in China, and ‘the latest and most automated equipment’ used at Mwanza tended accordingly to reflect conditions in the former rather than the latter country. It is thus by defining bureaucratic objectives over projects that Williams is able to resolve the second paradox. For the seemingly inconsistent technological preferences are replaced in his framework by a consistent set of project-related preferences, from which technological outcomes are derived rather than chosen. What this framework is unable to clarify, however, is the first of the two paradoxes described above. For whereas that paradox has to do with both an observed bias towards large-scale, capital-intensive and inefficient techniques and a systematic tendency towards sophisticated, ‘high-income’ products, Williams (1976) argues instead that the factor intensity of technologies associated with packaged, externally financed projects would be randomly distributed: that is, that there would be no systematic bias in any one direction. This particular weakness of the approach, however, can be overcome by a more realistic analysis of the relationship between the sources of external finance and the types of industrial technology that have been transferred to the public sector in Tanzania. For one thing, even after the Arusha declaration in that country, most of the foreign finance supplied to the public sector originated in the developed market economies (and took the form mainly of foreign aid, as shown in Table 5.5 below). And the types of products and processes associated with this type of finance are not randomly distributed, but are instead closely reflective of the socioeconomic conditions prevailing in the supplier countries (especially a relative scarcity of labour, high average incomes and large markets).3 On the other hand, though, when foreign finance originates instead in developing countries, the same historical line of argument leads one to expect the transfer of relatively appropriate forms of technology, an expectation that was clearly met in the case of the Friendship textile mill and also in the case of the labour-intensive farm implements enterprise referred to earlier. For in the latter case as well, what mattered was not ‘that the Tanzanian side took any greater degree of interest in the choice of technology than in any other project’ (Barker et al., 1986, p. 123), but rather that the finance was provided by the Chinese who (as in Friendship) favoured relatively simple technology. By thus combining the basic insight provided by Williams – that managers seek to maximize external project finance – with a historically


oriented view of how this behaviour leads to particular technological outcomes, one can account for much of what seems paradoxical about the behaviour of state-owned enterprises in Tanzania. Yet, for all its centrality to this composite argument, the foreign-exchange maximizing behaviour by the bureaucracy remains far from well understood. Most importantly, it is not at all clear how this behaviour actually enters the utility function of the manager of a public enterprise. What is lacking, in other words, is a detailed analysis of bureaucratic objectives and an assessment of the manner in which those objectives are furthered by foreign-exchange maximizing behaviour in the public sector. Though it was not designed to address this particular question, we shall now argue that the public choice approach to bureaucracy nonetheless throws considerable light on it.

4

The public choice approach to bureaucracy

The public choice approach to bureaucracy emphasizes the multiplicity of ways in which an expansion of the size of an institution promotes the particular interests of its members. In so doing, we shall argue, this approach provides precisely what was missing from the previous section, namely, an analysis of how the maximization of foreign exchange (and more generally budget maximization) promotes the objectives that were really pursued by managers of state-owned enterprises. Although an early contribution by Downs (1967) had emphasized how ‘The expansion of any organization normally provides its leaders with increased power, income and prestige’ (Downs, 1967, p. 17), and how those persons would therefore tend to favour organizational growth, the specific notion of a ‘budget-maximising bureaucrat’ is most closely associated with William Niskanen (1973). The latter makes essentially two claims, the first of them being that the manager of a public bureau has the following among his major goals: ‘salary, perquisites of the office, public reputation, power, patronage, output’ (ibid., p. 22). Niskanen’s second major claim is that since these goals are all assumed to vary directly with the size of the budget, the aim of the bureaucrat can be reduced to one of maximizing this financial variable during his time in office. This time-element bears emphasizing in part because most of the gains that accrue to the budget maximizing bureaucrat ‘are nearly unrelated to the ‘net worth’ of his organisation after his departure’ (ibid., p. 33), and in part because his tenure in office itself may be relatively short in many developing countries.

5 Foreign-exchange maximising bureaucracy and the political economy of Tanzania It is easy enough to show that these goals find frequent expression in the writings of political scientists who work on sub-Saharan Africa in general


and on Tanzania in particular. Indeed, one of those goals, bureaucratic power, is one of the most pervasive themes of that part of the political science literature which deals with the post-independence period in Africa. After noting how the phenomenal growth in the size of the civil service in Tanzania (and elsewhere) created ‘a privileged group’ with ‘corporate interests of its own’ and with considerable opportunities for ‘personal aggrandisement’, Chazan et al. (1988, p. 53), for example, are not alone in suggesting that the bureaucracy ‘emerged as the core of a new dominant class in the postcolonial period’. Moreover, in its analysis of how the economic and political power thus acquired by the bureaucracy is exercised, the political science literature almost uniformly assigns a paramount role to patronage, another of the goals emphasized by Niskanen (1973). There is, however, only one study that deals in detail with the political economy aspects of the growth in the public sector in Tanzania and which, at the same time, analyses the role played by foreign-exchange maximisation in that growth. This study (Mukandala, 1988) takes as its point of departure the notion that since 1969 the ‘political stratum’ of the Tanzanian state has been locked in an intense and prolonged conflict with the ‘managerial stratum’ of the state (where the former category refers essentially to those who make or design policy, and the latter refers largely to those who actually carry it out). It is not that there were no parastatals in Tanzania prior to the Arusha declaration, but it was only in the years thereafter, where the parastatal sector had emerged as a ‘strategic and complex branch of the state’ (ibid., p. 29) that the political and managerial strata came into direct and open conflict. For by then, following an extraordinarily rapid increase in their numbers and their net assets (see Table 5.3 below), parastatals had become ‘too big, diverse and strategic to be left to the managerial stratum yet proved too difficult to control through conventional government methods’ (ibid.) It was already all too plain, for example, that the parastatals were behaving less like vehicles in the transition to socialism and more like ‘bastions of capitalism’, with only a very tenuous connection to the political institutions of the state. What particularly concerned the ‘political stratum’ was the National Development Corporation (NDC), which accounted for no less than half the total investment by parastatals in new firms, and which was in control of the majority of the state’s assets in manufacturing and other sectors of the economy. (So rapid in fact was the growth of this institution that its total net assets almost doubled over the period from 1967 to 1969.) The ‘political stratum’s’ response to this situation was the introduction of a series of measures whose objective was to limit the degree of freedom then being enjoyed by the parastatal sector as a whole and the NDC in particular. These measures marked the beginning of what has been described as the ‘rationalisation’ phase in the parastatal sector (Mukandala, 1988, p. 29). According to one such measure, for example, particular groups

146

Table 5.3

The growth of the public sector in manufacturing Public sector fixed assets in manufacturing (million shillings)

1966 1967 1971 1979 1980 1981 1982

21a 971 2 851 3 278.4

Public sector employment in manufacturing (absolute numbers) 2 330 5 300 20 113 53 000

Share of public sector in manufacturing employment (per cent) 15.5 46.4 50.0 50.0 47.7 52.7

Share of public sector in manufacturing value added (per cent) 5 14 29 31 37.1 48.2 56.8

Sources: (Col. 1) World Bank (1988, p. 4); (Col. 2) Perkins (1980), Clark (1978) and World Bank (1987a); (Cols 3 and 4) Skarstein and Wangwe (1986). Note: a Refers to 1964.


of parastatals were to be placed under the control of newly created parent ministries. Other measures involved a heavier degree of reliance on some parastatals (such as the National Price Commission and the Bank of Tanzania) in the regulation of other parastatals, and a strengthened role for central ministries (such as the Treasury and the Ministry of Development Planning). By far the most extreme of all the measures employed by the ‘political stratum’ after 1969, however, was the attempt at ‘rationalization’ of the parastatal sector. What was intended by this was basically the fragmentation of the then existing corporations into increasingly smaller units, in the hope that they would thereby become less powerful and more manageable. The policy of rationalization did produce some results: between 1971 and 1974, for example, the NDC was stripped of 17 of its operating companies and 19 of its projects. On the whole, though, these and other attempts to limit the rapidly growing influence of the ‘managerial stratum’ met with only a very limited degree of success. To some extent this was a reflection of a somewhat predictable set of administrative and logistical problems (predictable because the relatively scarcity of administrative resources must have been apparent right at the outset of Tanzania’s attempted transition to socialism). To a large extent, however, the failure of the political stratum to achieve its objective was due to the vigorous countermeasures that the parastatal institutions themselves undertook. It is in these measures and their outcomes that one finds rather striking support for a public choice interpretation of bureaucratic behaviour in Tanzania, for what occurred was a particularly clear demonstration of the political advantages that accrue from an increase in the size of a public institution. Consider from this point of view Mukandala’s telling description of how the parastatal sector as a whole responded to what, in the guise of rationalization, was nothing less than a familiar ‘divide and rule’ tactic by the ‘political stratum’. Thus, On the one hand the managerial stratum established new subsidiary companies to replace those hived-off by the state … . This was deemed necessary because the more subsidiaries a corporation had, the bigger it was and consequently a) more resources were allocated to it in the national budget for developmental purposes; b) the more bargaining power it had when negotiating for higher salaries (especially for its management). Size showed ‘umuhimu’ or importance relevance; c) strengthened its hand in mobilising external finance on its own. … External finance was also useful because once obtained, it assured approval for the projects from the Treasury, the Bank of Tanzania and the Ministry of Development Planning. Foreign capital’s need to invest in new projects rather than expanding or rehabilitating old ones reinforced this trend. (Mukandala, 1988, p. 32).


The growth of the public sector in general Established posts in the civil service

1966 1967 1971 1980

65 708 80 239 99 564 295 342

Total number of parastatals 43 73 380

Sources: (col. 1) Mukandala (1985), (col. 2) de Valk (1992).

From the perspective of a sector engaged in a struggle for its very survival, therefore, what seemed to matter much more than its efficiency or profitability were its size and power. And it is from these political points of view, we suggest, that the central role of new projects and foreign capital needs to be understood. To what extent then did the public sector actually grow in Tanzania? Table 5.3 provides evidence on this question for the manufacturing sector, while Table 5.4 refers to the public sector more generally (in both cases the data cover the period from the 1960s to the early 1980s). As one would expect, the figures in the tables show a marked growth in the public sector after the Arusha declaration. For there was then a greatly enhanced need for civil servants and other officials to run the institutions that had just been nationalized (Mukandala, 1985). What is remarkable, though, is that the public sector continued to grow very rapidly after the early post-Arusha period, when no such obvious need for personnel was apparent and at a time when in fact the ‘political stratum’ of the state was bent on reducing the sector’s size and power. Between 1971 and 1981/2, for example, public sector fixed assets in manufacturing grew more than threefold, while the share of public enterprises in manufacturing value added almost doubled (see Table 5.3). These pronounced increases in the absolute and relative size of the public sector were closely related to the external sector, for in the ‘frantic spate of subsidiary creation’ referred to above ‘there was a renewed aggressiveness toward recruiting foreign project partners; the call for self-reliance notwithstanding. … This in turn led to a faster increase in the importation of capital rather than intermediate goods and over-installation of new industrial capacity’ (Mukandala, 1988, pp. 128–9). Tanzania’s increasing reliance on imported capital goods over this period was to a large extent made possible by growing amounts of foreign aid, which, as Table 5.5 shows, financed an ever-greater proportion of gross investment and government expenditure. The effect of all of this activity on the technological dimensions of the industrialization process was hardly surprising. On the one hand, the

Public Choice, Technology and Industrialization 149 Table 5.5

The growth of external sources of finance Total aid (million US dollars)

1960s 1970 1975 1980 1982

29.9 114.9 493.6 514.9

Total aid as percentage of monetary gross investment 11.4 25.3 46.6 54.3

Share of foreign financing in government expenditure (%) 30–35 40 45 50

Sources: (col. 1 and 2) Skarstein and Wangwe (1986); (col. 3) Wangwe (1992a).

powerful political pressure to grow as a means of institutional survival, with its attendant dependence on external sources of finance, meant that little time was available to public sector managers for consideration of factors such as labour intensity, self-reliance and industrial decentralisation that were described earlier as being important to the ‘political stratum’. Nor was there time for undertaking the various types of effort (such as search activities) on which the acquisition of local technological capabilities is known to depend. A management or technical collaboration agreement with a foreign partner, for example, was much more likely to have been seen by the parastatal as a way of overcoming short-run skills constraints on its expansion (via projects), than as an issue which carries important long-run implications for the acquisition of technological capabilities in the public sector. On the other hand, in its neglect of the managerial function the parastatal was further contributing to the inattention paid to capabilities issues, because it was effectively ignoring the range of issues that bear on technological learning and mastery at the operational level (that have to do, for example, with project implementation, repair and maintenance). 5.1

The size of the public sector under structural adjustment

If the public sector in Tanzania continued to grow rapidly into the early 1980s, in spite of attempts to curb its size and power, it also managed to withstand the privatisation schemes and other measures that were introduced a few years later by the World Bank, as part of the wider process of structural adjustment. On the one hand, some notable increases in the absolute size of the sector were recorded in the period after the reforms were introduced (that is, after 1986). The total number of state-owned enterprises, for example, grew from 380 in 1980 to 425 in 1988, 4 while the civil service as a whole increased in size by more than 5 per cent over the period 1985–92.5 On the other hand (and relatedly), some World Bank data show that with respect to most macroeconomic variables, the public sector share

150

Table 5.6

The public sector share in Tanzania before and after structural adjustment Share of state-owned enterprises in GDP

Share of state-owned enterprises in non-agricultural economic activity

Share of state-owned enterprise investment in gross domestic investment

Share of state-owned enterprises in employment

1978–85 1986–91

10.8 13.7

18.1 23.8

24.6 30.0

22.3 22.3

Change

+2.9

+5.7

+5.4

0

Source: World Bank (1995a).


increased rather than decreased in the reform period, as compared with the seven previous years. In particular, one can see from Table 5.6 that the state-owned share of the GDP, non-agricultural economic activity and gross domestic investment actually increased in the later period, while its share in employment remained unchanged between the two periods. Since the period covered by Table 5.6, however, the pace of liberalisation and privatisation has quickened somewhat. Between 1990 and 1995, for example, 34 of the 170 parastatals in manufacturing were privatised. This, in turn, will tend to have increased the share of the private sector in investment and other macroeconomic variables.

6

Conclusions

In this paper we have sought to explain several of the most paradoxical aspects of technology and industrialization in Tanzania since the Arusha declaration of 1967. Our explanation has turned mainly on two assumptions about bureaucratic behaviour in that country: the first is that bureaucrats have preferences defined over projects rather than technologies, and the second is that in their capacity as managers of state-owned enterprises these agents of the state have sought to initiate as many new projects as possible, mainly on the basis of foreign aid. Such behaviour, we suggested, was rooted in a political context where the drive to institutional expansion was viewed as a key element in the survival of the bureaucracy. It meant that the various branches of the public sector continued to grow very rapidly well beyond the early post-Arusha years, when there had been an initial need for large numbers of extra civil servants to manage and run the newly nationalised industries. In fact, as we showed in a number of tables, the public sector grew very rapidly throughout the 1970s and early 1980s, at a time when the ‘political stratum’ of the state was bent on reducing its size and power. It also meant that even from 1986 to 1992, during the period of structural adjustment reforms which began in 1986, the public sector share of most macroeconomic variables increased rather than decreased.

Notes * Department of Economics, Tilburg University, PO Box 90153, Tilburg, The Netherlands. This article was originally published in Public Choice, 89, 3/4 (1996), pp. 375–392, and is reprinted with kind permission of Kluwer Academic Publishers. 1. For details see James (1995). This book was the first attempt to apply public choice theory to African industrialization in general. 2. This was already clear from Clark’s study (1978) of the period from 1964 to 1973. He pointed to a ‘failure of the post-Arusha companies to distinguish themselves significantly and favourably from the pre-Arusha companies’

152 The Industrial Experience of Tanzania (Clark, 1978, p. 140), and concluded that ‘There has been as yet no developmental innovation on the part of parastatals to make themselves more consistent with the Tanzanian ideology’ (ibid.). Some years later a very detailed study by Perkins (1980, 1983) of more than 300 firms in ten industrial sectors arrived at a similar conclusion, namely, that despite the existence of efficient labour-intensive technologies and ‘Despite the rhetoric, Tanzania’s industrialization programme has in general promoted the establishment of [publicly owned] enterprises using large-scale capital-intensive, often technically and invariably economically inefficient technologies’ (Perkins, 1983, p. 231). 3. This argument is most clearly elaborated in Stewart (1977). 4. These data are taken from de Valk (1992), who does not cite his primary source. 5. See World Bank (1994a). This source does not however reveal the amount over 5 per cent by which the civil service grew over this period.

6 The Form and Role of Innovativeness in Enhancing Firms’ Productivity: The Case of Selected Manufacturing Firms in Tanzania Haji H. Semboja and Josephat P. Kweka*

1

Introduction

Development experience has shown that the countries that have undergone fast economic growth and structural change in the second half of the twentieth century are those that have based their productivity growth on rapid absorption of foreign technology and on increasing efficiency in the use of technology over time, and that have successfully developed and applied innovation through learning. Generally, the scale of operation, firm size, organisational structure, characteristics of the market, industrial policies, infrastructure, resource availability and the supporting institutions are crucial factors determining the nature and level of innovative activities. At the broader level, human resource development and access to finance and the legal and regulatory framework contribute to innovative achievement. At the firm level, technological development has occurred through innovation which has become not only an important determinant of the level of productivity but also of firms’ international competitiveness. However, during the pre-reform era innovation was never a major determinant of the level of productivity in Tanzanian manufacturing. Technological development has been limited. What development there was, has occurred through absorption and adaptation of foreign technology rather than indigenous technological change. It is upon this understanding that the process of technological development, the nature of innovations and the factors influencing innovative activities need to be understood as a basis of formulating policies for stimulating productivity and competitiveness of the manufacturing industry. This article intends to make a contribution in this direction.1 We set out to describe the various forms of innovative activities at firm level and to determine the factors influencing these activities,


assessing their role in enhancing the productivity of manufacturing firms in Tanzania.2 Section 2 discusses the form and role of innovativeness from a general and conceptual perspective. Section 3 maps the various forms of innovative efforts and activities in the selected manufacturing firms in Tanzania, while Section 4 deals with factors determining firm innovativeness. Finally, Section 5 provides conclusions and policy recommendations.

2 Forms and role of innovation from a conceptual and general perspective In this article we shall consider innovation in its broader sense, that is, as all that is new to the organisation. At firm level, innovation refers to both the technology of an enterprise and its products (Rothwell, 1977; Souder and Chakrabarti, 1980).3 Technical and innovative capability is an asset for adaptation to all sorts of societal changes. Innovation is caused through interactions between individuals, business units and social-economic sectors. In the developing world, public socioeconomic supporting institutions can make important contributions to importing, modifying and diffusing new technologies. Innovation contributes to the creation of wealth and induces growth, structural change and development. The innovative capacity of a country constitutes its ability to adapt to shocks, be they internal or external (DeBresson, 1994). This capacity may break the constraints of the internal and external environment, and enable a country to forge ahead, thus improving the market position of firms and challenging competitors and the external environment, making them adapt. According to DeBresson (1994), innovative activity is an aspect of competition, growth and development. Competition by technology is an important force for a firm and an industry in the context of both industrially developed and developing economies. With the opening of markets, for an organisation to survive, it is imperative to undertake innovative activities. Import substitution strategies or industrial protection introduced to enable an industry to learn and catch up often have the perverse effect of inducing inertia. Some forms of learning must exist in an organisation if it is to survive. Even the so-called traditional manufacturing industries are no longer exempt from the impact of new techniques. Innovation has also been identified as an important determinant of success in small-scale enterprises. It is through product rather than technological innovation that small-scale enterprises are able to develop their competitive edge. Romano (1990) found that, for high-growth firms, the determinants of product innovation that influenced the success of small business included technology, R&D, the product life cycle, market change, product/market mix and customer base. For low-growth firms, customer base was a major determinant of product innovation.

Innovativeness and Productivity 155

Innovations can be characterised on the basis of whether they are process or product innovations, and whether innovation is original to a specific country (say Tanzania) or to the world, or whether it is imitative. We also need to distinguish revolutionary/radical innovation from adaptive/ incremental innovation. Radical innovation is supposed to induce a new growth cycle in its sector and perhaps in the whole economy. It increases utility and dramatically reduces costs by setting up a totally new production function. Schumpeter postulated that important innovations always required the setting up of new productive units and new fixed investments. Adaptive innovation or adaptive technical change are those activities directed at modifying the technical basis of the production process prior to full-scale use, or changes to a product before it is introduced to the market. Adaptive innovations draw mostly on known techniques and the existing knowledge base. Incremental innovations are those evolutionary technical changes aimed at correcting technical imbalances, and modifying minor (yet significant) core or peripheral systems. In this analysis both adaptive innovation and incremental innovation are opposed to radical innovation. Technological knowledge is cumulative. There are no clear borderlines between imitation and creation. Both radical innovations and adaptations are essential elements of technological change. On the origins of technological innovation, two oversimplified theories of innovation have dominated the debate, ‘technology-push’ and ‘demandpull’. According to the demand-pull theory, innovation arises as a result of perceived and often clearly articulated market needs, which leads to focused R&D activities, creating a host of products for the market (Rothwell, 1985). The rationale behind the theory is that production units within the markets recognise customer needs and direct their efforts to fulfil those needs through technological activities (Dosi, 1984). The actors in the market, that is, the adopters, can be private firms, government or domestic consumers. Under the technology-push theory, discoveries in basic science and R&D eventually lead to technological development which results in the flow of new products and processes to the market place (Rothwell and Zegveld, 1985, cited in Diyamet et al., 1998). Innovations await technological progress. Thus, according to technology push proponents, the market initially plays a minimal role in the innovation process, acting only as a repository for R&D results. The entrepreneur occupies a central position. However, Fransman (1986) pointed out that it is a statement of the obvious to say that the market will influence a firm’s technological activity. Firms must market their products in order to survive. In the pursuit of innovation, firms need to lay down consistent strategies for achieving or acquiring the desired forms and types of innovation, or adopting certain technical advances. Some strategies develop in the drive to sharpen the competitive edge, others develop as guidelines for


implementing predetermined technological progress. We can identify two kinds of innovation strategies: offensive and defensive. An offensive innovation strategy is aimed at achieving technical and market leadership by being ahead of competitors in the introduction of new products and investing in technological advances. A defensive innovation strategy, on the other hand, may take the form described as ‘defensive’, ‘imitative’, ‘dependent’ or ‘opportunist’. A defensive strategy does not necessarily translate into a lack of research-intensiveness. Though the defensive strategist does not aspire to technical and economic leadership, he or she equally does not want to suffer the penalties of laggardness by being saddled with obsolete technology. Risk-taking and incurring financial losses are avoided. The defensive strategist aspires to benefit from the mistakes of the market leaders, and may lack the capacity to undertake the fundamental research necessary to assume and maintain technical leadership. Innovation capabilities consist of search capabilities and the capacity to integrate the results of such searches (Wangwe, 1994). Search capabilities are those required to find new ways of carrying out the firm’s investment, production, marketing and organisation, activities. The search for new routines is likely to be reflected in efforts to create new patterns of human resource development (through in-house or outside training), in R&D (formal and informal),4 in searches for technological information from local and foreign sources and in technological adaptations and market research (study of market trends and potential markets and of the possibilities of introducing new products). Given the circumstances existing in developing countries such as Tanzania, adaptive innovation is likely to be more prevalent than radical innovation. Sources of innovation are most likely to be foreign. Only the uses and applications may be new. The experience of Asian countries (or even industrial countries like Italy) has demonstrated that much growth, development and technical accumulation can be obtained without much invention or basic or applied research. Innovative activities need not always be related to intensive scientific and technological research. In this sense, innovative activity, as a social and economic factor of change, is not the monopoly of advanced industrial countries (as some of the recent trade and new growth models assume).

3 Innovative efforts and achievements in the manufacturing sector in Tanzania Tanzania aims to strengthen her technological capability for processing basic commodities and upgrading her export industries, in order to produce high quality goods that can withstand domestic, regional and global market competition. This has been done through endogenous capacity building in science and technology and efforts to create a conducive


environment for unleashing the creative and innovative potential of the people of Tanzania. The use of imported technologies provides learning opportunities and thus help to develop an indigenous technological base when these technologies are actively selected and acquired by users. A study by Wangwe (1986) on technology imports, technological learning and self-reliance in Tanzania found that technological changes and innovations have been limited by the localization of major technological decisions and a high degree of heterogeneity of technological imports. Technological change via adaptations resulting in the improvement of performance was limited. Evidence from various ESRF surveys of industrial performance 5 has shown that innovation-relevant types of linkages, such as inter-firm linkages, financial institutions linkages and business support linkages, do exist within the Tanzanian manufacturing sector. However, these have mainly been utilized on an ad hoc basis. Divestiture programmes in Tanzania have made it possible for some local firms to access and link with foreign firms in the procurement of technical expertise, machinery, equipment, product technology as well as new markets. These are mainly companies that have bought and invested in existing or former public/parastatal firms. For instance, in the case of the electrical equipment company ABB tanelec (by then tanelec) the design and production of transformers on one hand is based on the technology adopted from the ABB National Transformers, which has been producing electrotechnical equipment for more than 75 years.6 This has thus contributed to significant improvements in the firm’s technological capability and competitiveness through innovative efforts. In the case of Tanzania Breweries Ltd, a successful joint venture and divestiture process in 1993 was followed by a dramatic improvement in the firm’s profitability and performance (see Chapter 15). A more recent study by the World Bank, the Regional Program on Enterprise Development (RPED, 1994) covered four sectors, namely textiles, wood, food and metal fabrication. Given the quality of technical personnel in these firms, innovations were of a limited nature. Quality or refined technology design and innovation were generally beyond their capabilities and resources. Sources of innovation included both the ideas and information about innovation as well the technology and the support for carrying out innovation. Major sources are established business channels, personal contacts, inhouse R&D efforts, customers and research institutes (Chungu et al., 1994). There is a need for more emphasis on the absorption aspects of innovation, coupled with technical extension services from the producers and the management and maintenance of the technology. Studies by Semboja et al. (1997b) and ESRF-PSRC (1996) suggest that the sectors with the highest levels of innovation and strongest innovative linkages in the Tanzanian manufacturing sector are predicable by previous levels of domestic forward


and backward linkages. Also, the relative frequency of domestically initiated innovations rises with the level of firm performance. According to the RPED (1995) study, firms show marked differences as to the renewing or upgrading of their equipment. Between 1988 and 1992, 57 per cent of the firms surveyed undertook major additions to or changes in plant and equipment. The main reasons given for new direct investment were to increase capacity in existing product lines (50 per cent) and to improve the production process (35 per cent). Only 12 per cent of firms invested in a new equipment for producing a different product or a different type of a similar product. The metals branch is the leading innovative manufacturing sector. According to Semboja et al. (1997b) 7 larger firms and private firms in Tanzania have more innovative capability than small firms and public firms The more innovative firms concentrated on increasing capacity and to a certain extent on improving the production process. These additional capacities led to a reduction in the unit production costs and to improved product quality. Thus the larger firms managed to reduce costs much more than the smaller ones (an aspect of a defensive innovation strategy). This shows that large firms have a greater capability to adjust production processes in order to remain competitive under a liberalised economic regime. On the other hand, the involvement of large foreign firms in the construction sector may induce technology transfer, given the wellplanned integrated project management schemes. For example, the transfer or the acquisition of technology within the construction sector entails training efforts designed for local personnel at operational, functional and management levels, the involvement of local contractors and the provision of employment to local staff (Semboja et al., 1997b). The participation of local construction firms as subcontractors to foreign firms is an important element in enhancing innovation and the acquisition of technology. However, in general, production technology and market orientation predominantly focus on domestic demand, rather than on target export sectors. The case studies in ESRF-PSRC (1996) show that adaptive technological change resulting in the improvement of performance occurred least frequently and was relatively the most limited. Local manpower involvement in carrying out feasibility studies, choice and design of production methods, choice of product, participation in the installation of plant and machinery is limited. This deficiency has diminished the opportunity to kick-start serious innovativeness by imitation, adaptations or skill transfer by local personnel. Product technology improvements were substantial in leather (shoemaking), metal and engineering industries (ESRF-PSRC, 1996). Designs of a number of products had been made in collaboration with local institutions. In terms of the size of firms, many small- and medium-scale enterprises have been identified with product-centred innovative


capabilities, mainly owing to the flexible nature of their production and the significant role of the customer in product design. However, larger firms are more capable of carrying out innovativeness, given their relative abilities to purchase and adopt technological progress. Some studies (Niosi and Rivard, 1988) show that there has been an increasing tendency for the industrial nations to adapt and modify their exported technology to suit the conditions of host countries. The ESRF-PSRC (1996) study on the divestiture impact on industrial performance corroborates this phenomenon in the case study of Moshi Leather Industries, where most of the technological innovations in the company were the initiatives to adapt the foreign technology (from the parent company) to fit the local raw materials (skins and hides). In the case study of the G&T (shoemaking) Company, the management recruited a full-time technician with the ability to modify and adopt the machinery to suit the local conditions. Therefore, the need for major technical adaptations by foreign firms (for example, multinational companies) may be reduced. The acquisition of technological capabilities was restricted to the importation of foreign technology, adaptations to suit local conditions, that is, gaining mastery, making minor improvements. General key aspects of technology appear to be specifically relevant to product innovation. Small firms’ dependence on a number of larger suppliers of raw materials, components, and equipment or on subcontractors reduces their control over their production process, resulting in difficulties in maintaining product quality and the level of service. Evidence from the garment sub-sector shows that firms’ responsiveness to changing market conditions and policy changes (for example, import liberalisation) was reduced by the long periods of protection they had enjoyed before liberalisation. As a result firms were unable to develop the abilities and culture to respond dynamically to challenges, based on their own innovations and initiatives. In this regard, employee participation is considered to influence their commitment to firm initiatives for innovation. This article supports the thrust of recent trade and growth models which have focused more explicitly on the micro-foundations of innovation by addressing firm-level decisions to invest in product or process innovations. Some surveys and studies on manufacturing firms in Tanzania (ESRF-PSRC, 1996; Semboja et al., 1997b) have shown that the ceaseless search for improvements in technology (especially product quality and cost-lowering process innovations) has been instrumental in improving productivity. Productivity growth, in turn, was a most important factor enabling exporting firms to succeed in the changing technological and market conditions. Exporting firms maintained and improved their market position by investing in technology and continuing to improve on it. Improvements were made not only in the firms’ products but also in the processes of production, in order to cope with pressure to keep costs at competitive levels or to


improve product quality (level and consistency). These responses were derived from signals given in export markets. Experience in some large-scale and high-growth firms (Tanzania Breweries Ltd, Afro-Cooling, and so on) shows that productivity has steadily improved because of improvements in technology and management techniques. In areas where labour accounts for around 90 per cent of value added there have been many improvements in labour productivity as a result of managerial innovativeness. Both the large scale textile mills said that the wage cost of the conversion of raw materials into finished products has been kept low while volumes produced have increased. This has been made possible by a number of factors, including improved training of labour, improvements in production programming, and reducing the varieties and styles produced in the production runs.

4

Factors influencing innovativeness for productivity

As a result of liberalisation, Tanzania presently operates a virtual ‘open door’ policy towards foreign investment and technology. Capital goods, raw materials and technical knowledge can now be imported without hindrance. It is crucial to examine whether there are factors which determine the extent to which the previous import substitution regime or the current export promotion strategies result in absence of innovativeness or positive innovative efforts which contribute to technological capacity building. This section examines some of the critical factors influencing innovative behaviour of firms. The purpose of this examination is to identify these factors and reflect on the extent to which innovative efforts have been limited in the Tanzanian manufacturing sector. These factors include: human resource development, R&D efforts, finance, competitive market turbulence, institutional framework and linkages, access to technical information, and the legal and regulatory framework. 4.1

Human resource development

The various ESRF manufacturing studies show that the educational background of owners/managers of firms has been identified as an important factor in enhancing firms’ productivity. Interaction with other persons and educational experience can make management more receptive to change, and promotes its imaginative role in innovation. Significant numbers of firms in these surveys pointed to financial constraints as the cause of insufficient formal training.8 Most of the firms prefer in-house training as a cheaper option, mostly dealing with specific problems or process improvements. In this respect, innovativeness envisioned as a result of external training and exposure of technical workers is rather far-fetched.9 In-house training is done continuously to improve upon the initial skills and productivity of the workers. Some firms have in-house training


programmes for their employees while others do not. About 61 per cent of the surveyed firms in the ESRF Multi-Country Comparative Study on Private Enterprise Development survey (Semboja et al., 1997b) have indicated that they have in-house training programmes, whereas 39 per cent do not have them. The nature of such training is mainly on the job training, and its main objectives are to enhance/upgrade workers’ skills, equip them with new production techniques, and teach them better quality control and other machine operations. This is especially important for new recruits, and for improving productivity and the quality of goods and services. Many firms send their employees to local institutions, especially technical colleges or vocational training colleges, such as the Vocational Educational Training Authority (ESRF-VETA, 1998). Prior experience in business has been suggested as important in providing owners with knowledge of current operations. Training obtained in a successful innovative firm will make management responsible for innovation more progressive and open-minded (Rothwell, 1977). Without the support of previous managerial experience for innovation, it will be difficult to develop a successful innovation programme. In Tanzania, however, in most firms long experience of the workers or managers did not translate into significant innovation. This is because managers are preoccupied with routine tasks and responsibilities, where their work experience has made them capable of solving the day-to-day technical/managerial problems. However, the experience and exposure of the managers of highgrowth firms (for example Tanzania Breweries Ltd) are shown to be a dynamic source of innovation, high productivity and competitiveness (ESRF-PSRC, 1996). The Tanzanian manufacturing sector is characterized by a critical shortage of technical skills necessary to enhance firms’ innovativeness. As a result, some of the high-growth firms have resorted to hiring foreign experts. Besides foreign technical experts, some firms engage foreign consultants. The Semboja et al. (1997b) survey data showed that 22 per cent of the firms engage foreign consultants while 78 per cent do not. Those that engage foreign consultants seek expert advice in the technical and managerial fields. 4.2

Research and development efforts

Several empirical studies on innovation suggest that ‘effective coordination’ of R&D is required to enhance firms’ innovations (Souder and Chakrabarti, 1980). R&D initiatives include efforts to improve existing products, develop new products, develop new production processes/ methods or equipment and improve product quality. Tanzanian experience in terms of R&D activities is gloomy. While the institutional framework for R&D has been established, little has been achieved in the productive firms in terms of significant innovation. This problem has been exacerbated by


lack of finance to carry out scientific activities, and failures in the commercialization of R&D initiatives. The basic infrastructure for R&D is underdeveloped, and links between productivity/innovation centres and the users are weak (see also Chapter 7). The government role in this aspect is not very effective, given the modest expenditure allocations for such institutions/centres. In addition, firms have a tendency to invest in short-term profit-intensive plans, and shy away from long-term investment in innovativeness and R&D activities. Further analysis of the R&D efforts reveals a critical shortage of research and technological personnel, not only in the technological centres but also in the firms themselves. The impact of this shortage is aggravated by three other factors. First, there is an unfavourable ratio between the number of engineers and the number of technicians who are supposed to support and complement the work of those engineers. Most of the engineers have been offered administrative tasks in many technology institutions and have not been preoccupied with innovation activities. Second, there is a shortage of research inputs and facilities with which R&D activities can be done. Most of the funds allocated to research are used to pay staff salaries (which are themselves low), leaving little for basic inputs and facilities for research. Third, there is a weak link between innovative activities and production. One explanation of the weak link between R&D and production has centred around the problems of relevance of the R&D activities for problems facing production activities in the country (Mlawa and Sheya, 1990). The other side of the coin is the fragility of the productive sector and its limited financial capacity to engage in new investments, reinforced by the unfavourable economic environment in which enterprises have been operating. In general, however, there are but few industrial firms that engaged in R&D activities. The RPED (1994) survey categorized the number of firms engaged in R&D according to the size of firms: 3 per cent of smallscale firms, 12 per cent of medium-scale firms and 16 per cent of largescale firms engaged in R&D. Most scientists, engineers and technicians are found in the large firms (food industry). Of professionals employed by the firms, engineers are the most frequently engaged in R&D activities (RPED, 1994). 4.3

Competitive market turbulence and customer base

Successful implementation of technical change influences innovation and the productivity of a firm. The question is whether high levels of competition stimulate enterprises to innovate products more regularly in order to remain competitive. Experience with Tanzanian manufacturing enterprises shows mixed results on this, with some firms responding defensively and others offensively. Both strategies are innovative ways of surviving the intense competition. For instance, the garment firms (most hit by the


import competition) showed a defensive response where firms shifted from production activities to trading activities. Of course, such a shift does not really refer to innovation in the more strict sense of technological change. In a small- and medium-enterprise context, customers may play a profound role in influencing product design and development. This is more evident in the garment industry survey, and partly in the light engineering firms in the sample, where customer choice influenced the trend and intensity of product innovation. In most cases, customers (both foreign and local) were an important source of product design and innovation (Semboja and Kweka, forthcoming). Rothwell (1977) has argued that an effective marketing policy is essential for successful innovation. The choice by firms to expand their product supply to other markets or develop competitive products for emerging markets will effectively enhance innovations. This depends on whether or not firms are ready to take risks inherent in moving from traditional to new markets. However, Wangwe (1995) indicated that most of the exporting firms started by serving the domestic market. The strategy of exporting came later in response to developments in their domestic markets. The capabilities gained by serving the domestic market and developing competitive products helped them to be innovative and embark on the export market. The current ‘opening up’ market policies have to an extent contributed to a search for new markets and thus to firms’ innovativeness, in some cases a firm had only one major local customer, which collapsed. The response of the supplier firm was to turn to the export market (Wangwe, 1995).10 Nevertheless, various studies11 have indicated that the share of the export market in sales is usually small, around 5–15 per cent. 4.4

Financial capabilities

Maintaining an active product innovation process requires two basic things with regard to finances. First, the finances should be adequate to cater for such expenditures. Second, a careful monitoring of budgetary and cost control processes is required, so that financial resources are used effectively. Thus, an active product innovation process requires a constant flow of funds. Availability, adequacy and sources of funds of the firms were examined in most of the ESRF industrial studies, including assessments of the role of financial institutions in financing industrial investment. Enterprises’ own savings and bank loans are the major sources of funds. In all cases, funds are highly inadequate because of poor turnover and shrinking markets. In recent years, bank loans have become difficult to access, due to high interest rates and lack of development finance for the improvement of firms’ technological and innovative capabilities. In Tanzania firms differ much in their abilities and culture with regard to financing innovation. Firm size and financial strength are important variables in this respect. Lack of adequate financial resources has negatively


affected the extent of technological upgrading in firms. Firms in the engineering subsector, for instance, show that while there has been some upgrading of production technology following liberalisation, such upgrading has been rather limited in most other sub-sectors, and has mostly been confined to small groups of firms. At the national level the problem of underfunding of innovation and science and technology (S&T) activities has apparently been recognized. In 1984 it was proposed that S&T expenditure in GDP be raised from below 0.5 per cent in 1984 to 1.5 per cent in 1985/86, and further to over 3 per cent by the year 2000 (United Republic of Tanzania, 1985a). At present there is no sufficient evidence to show that this share has indeed changed significantly from that of 1984 (for example, Wangwe, 1994). This suggests the need for initiatives to mobilize non-government sources of finance to fund urgent S&T activities in the country.

4.5

Institutional framework

The institutional framework for S&T development is largely structured along sectoral lines, under the overall coordination of the Commission for Science and Technology. 12 R&D activities in Tanzania are mainly carried out by public institutions and are mostly agriculture oriented. Links between industrial R&D institutions and enterprises are weak, as indicated by the failure to commercialise many R&D results and the failure to tackle the major technological problems facing productive activities in the economy (Mlawa and Sheya, 1990, Chapter 7 in this volume). For productive sectors to carry out significant innovation, government support and commitment towards technological advances are crucial. In addition, institutional restructuring and reforms in the economy have also impacted on the innovation institutions. Some of the government-owned R&D institutions have been divested, and most of the others made to operate independently (or commercially) with little or no reliance on government support. As a result, the innovative efforts of these institutions have been severely hampered. In addition, the generally acknowledged weak link between the innovation centres (suppliers/supporters of innovation) and firms (users and developers of innovation) has been eroded even more by the weak interaction between the private and public sector institutions in Tanzania. Governments may support technological advance by providing incentives and support for companies to upgrade their technology. The ‘wait and take’ attitude towards technology acquisitions is common among enterprises in sub-Saharan African countries; many companies passively receive technology from foreign suppliers without controlling or trying to affect technology choice. The industrial/business supporting institutions are mainly state-owned. The links among and between these institutions and


between them and the final consumers (that is, the firms) are weak and rudimentary. 4.6

Access to technical information

One important aspect of promoting innovative activities is the access to technical information. As Semboja et al. (1997b) noted, the sources of technological information can be divided into following categories: information from other firms, educational/academic sources (bulletins, journals, books and scientific institutions/research and development centres) and finally information from the company’s own technological initiatives or R&D activities. The Semboja et al. (1997b) survey showed that other firms are important sources of technological information, via inter-firm linkages. This source is even more significant in special types of industrial organisation such as clusters, through collective efficiency and agglomeration economies (Musonda and Kweka, 1998). It is thus argued that relationships and cooperation between firms through some conventional linkage arrangements facilitate the flow of technological information, the net effect of which is an increase in innovative capability. Most firms obtained their information from two important sources, namely: copying from domestically made products (by other firms), and through local and foreign buyers/customers (Semboja et al., 1997b). Further details on the firms’ responses to the sources of information show that domestic and foreign consultants were important sources of technological information. In the specific case of product innovation, the sources of product technology are also important sources of technical information. There are various sources of product technology such as: copying domestically made products, copying imported products, foreign buyers, local buyers, own R&D, and foreign partners. The pattern of use of these sources has been changing over time. In the initial years after their establishment, metal and light engineering firms copied their product and process engineering technologies from foreign buyers (since most of the products were exported). After the adoption of economic reforms (for instance in 1990), product technology was acquired by copying from imports. Recently, most firms have diversified their sources of product technology. In general, all three sub-sectors covered under the ESRF-MCCPED (Semboja et al., 1997b) sample have shown that the major source of their product technology, particularly prior to trade liberalisation, consisted of copying from domestically made products. Inward-looking industrialization strategies coupled with restrictive trade policies discouraged linkages with foreign firms, and meant that local manufacturers had to link amongst themselves. After the reforms, firms had opened up to learning from foreign firms – hence copying from imported products became a significant source of product innovation. The percentage of firms with


domestic products as a source of their product technology declined from 31.3 per cent in the 1980s to 27.0 per cent in the early 1990s. The percentage of firms using imports as a source of product technology increased from 12.5 per cent in 1980s to 16.3 per cent in 1990s. 4.7

Legal and regulatory framework

There are few explicit laws governing and/or facilitating the development of innovative activities in Tanzania. As a product of the colonial legacy, most of these laws are outmoded. Existing laws are implicit in this regard, often regulating other socioeconomic and business activities or sectors, such as investment, international transactions, human resource development (education and training), labour market, industrial development and trade. Patents, trade marks and various licences used in technological innovations depend on the existing policy, the patents law and the possession of industrial licences. The patent system is often used as one method for rewarding innovators.13 In Tanzania, the patent law was formulated in 1930 (cap. 217 of Tanganyika laws). The law does not provide much incentive to local experts to develop innovative industrial ideas. For a long time, the Tanzania patent system was pegged to the UK system. This system was really meant to allow foreign companies to have access to the Tanzanian market rather than to cater to domestic innovators and investors. In fact, the analysis of patents issued has shown that the patents were largely granted to multinational companies which were mainly importing the products into the country while actual manufacturing was located in other countries. A new patent law endorsed by the parliament (Act no. 1 of 1987) was expected to give a challenge and much needed motivation to scientists and technologists in their efforts to develop industrial innovations. However, the law is not effective because of a lack of basic tools and scarcity of experts to implement the patent tasks (Sustainable Industrial Development Policy, 1996–2020; Ministry of Industry and Trade, 1996). Since the patent system started operating in Tanzania no more than six patents have been established in Tanzania under this act of parliament. The registration office only serves as a re-registration office for patents registered in London (Mlawa and Sheya, 1990), and still lacks the necessary facilities to search for patent information online. Therefore, the regulatory framework has not been effective in enhancing innovative efforts in Tanzania. An inefficient fiscal regime characterised by high tax rates and a multitude of taxes has implicitly deterred innovative efforts. According to the sustainable industrial development policy, 1996–2020 (Ministry of Industry and Trade, 1996; see Chapter 17 in this volume), the government will review the relevant regulations relating to industrial investment and business company management to take care of all the flaws that inhibit the process of industrial development.


4.8

The role of the government

Though economic reforms meant a significant retreat by the government from intervening in the economy and production, there are still important tasks for the government in enhancing innovation. The role of the government is to provide the conducive environment and initial conditions for the productive sector (firms) to carry out meaningful innovative activities. Such conditions are reflected in the government budget, the incentive structure, the provision of infrastructure, training, and human and nonhuman resources for the development of innovative capabilities of the firms. The other area for government intervention is the provision of an environment in which flexible adjustment of production structures in the face of changing demand conditions is made possible. The provision of efficient infrastructure in the form of functioning water and electricity supplies and efficient telecommunications and transportation facilities would reduce firms’ operating costs significantly. The government should also facilitate access to information about, and the acquisition of, technologies by domestic firms. It should promote contacts between domestic firms and foreign technology suppliers. Support in these efforts might significantly reduce search costs.

5

Conclusions and policy recommendations

As documented by the surveys discussed, innovative efforts and achievements in Tanzanian manufacturing are marginal. Most firms have not been able to carry out significant innovations related to R&D efforts. Where innovative efforts have taken place, they have taken the form of technological upgrading, introduction of new products, adaptations of the transferred technology to suit local conditions, gaining mastery over the process and making minor improvements. Such efforts have not translated into viable technological progress because of the lack of an institutional framework that would enhance the links between innovations and improvement of market performance at firm level. It is worth noting that innovative capabilities differ widely between firms of different size and ownership structure, on the one hand, and between various subsectors on the other. Large firms have achieved significant innovative efforts unlike the smaller ones. The light engineering subsector is shown to be more dynamic in terms of innovative efforts compared to other sectors such as garments and textiles. Various factors were identified as important in determining innovativeness at firm level. These include: human resource development, R&D efforts, financial capabilities, competitive market turbulence and customer


base, institutional framework and linkages, access to technological information, the legal and regulatory framework and the role of the government. Apparently, while innovativeness at firm level is crucial in enhancing productivity, the development of viable innovative activities depends much on the ability of the firm to carry out significant technological, organisational and human resource investment. It is thus recommended that appropriate institutional frameworks are sought for the mobilisation of various stakeholders (firms, government and related institutions) for the support of innovation efforts at firm and national levels. The major policy implication lies in the need to strengthen and promote sustainable industrial linkages among all units (institutions, and production and market systems) within the Tanzanian manufacturing sector. However, successful innovative performance also depends on the ability and commitment of the government to address the constraints related to these factors, and to create a technologically enabling environment for the innovativeness and competitiveness of the Tanzanian manufacturing sector.

Notes * Economic and Social Research Foundation (ESRF), Dar es Salaam. 1. There are no comprehensive studies on innovation in Tanzania. The few existing studies are based on micro-firm-level data and/or are founded on a sound theoretical framework (for example, innovative activity matrix). Recent industrial case- and macro-based studies show that both private firms and public innovation supporting institutions operate within an unfavourable economic environment (Wangwe, 1986; Watanabe, 1987; RPED, 1994). However, there are significant adaptive and incremental innovative activities conducted in some medium- and large-sized manufacturing firms. Direct foreign investment, expatriates and imported technology remain the major sources of acquisition of technology and industrial performance. 2. The article draws some insights from theoretical and empirical works on this subject and utilises data and information on various manufacturing sector surveys done by the ESRF and other institutions. The surveys of the manufacturing sector include the following: • the ESRF-PSRC (1996) study of the impact of divestiture on industrial performance (six case studies on Tanzania Breweries Ltd, ABB Tanelec, Morogoro Textile Mills, Moshi Tanneries Ltd, Tanzanian Parkers Limited, G&T Company) • the UNU-INTECH QEH-ESRF (1996) study (also referred to as Semboja and Kweka, forthcoming) on the impact of import liberalisation on the manufacturing sector, based on a survey of 47 light engineering firms and 17 garment firms • the ESRF Multi-Country Comparative Study on Private Enterprise Development, also referred to as the Semboja et al. (1997b). This study presents results of a survey of 74 firms from three sub-sectors, food, construction, and metal and light engineering, held in 1997.

Innovativeness and Productivity 169 • Regional Program on Enterprise Development (RPED, 1994); Development and Growth of Industrial Enterprise in Tanzania, focusing on textiles, wood products, food products and metal products. • RPED (1995), Dynamics of Enterprise Development in Tanzania, Final Report on Round II Survey Data, Helsinki School of Economics. 3. Cited in Romano (1990). 4. Research and Development (R & D) is defined here as any creative and systematic activity undertaken to increase the stock of knowledge, and use of this knowledge to device new applications. It includes fundamental research, applied research and experimental development work leading to new devices, products or processes. 5. See the studies listed in footnote 2. 6. The combination of the basic technology and the latest global technology from ABB National Transformers ensures the high quality of ABB Tanelec distribution transformers. ABB National Transformers has a long history in producing highquality switch gears. The combination of this with local expertise has resulted in high-quality medium- and low-voltage switch gears utilizing the latest technology. 7. This is the MCCPED study (Semboja et al., 1997b). 8. Innovative efforts at firm level are linked with training, because it is through training that technological information can be transferred. Thus, there is a need to improve training linkages to enhance innovative efforts. 9. Training requirements and availability of training programmes (in-house and external) were reflected in most of the surveys on the manufacturing sector. 10. Morogoro Canvas Mill was one of the cases in point. It had one customer, the Morogoro Shoe Company. The collapse of this company meant the loss of the entire market for the firm. The Canvas Mill adjusted swiftly by changing its product mix and entering the export market. The firm achieved this strategy by making the necessary investments in technology and innovation to meet the requirements of the export market. In addition, the firm employed foreign management consultants to manage the production. As a result the firm managed to raise its share of exports to output from less than 10 per cent to over 60 per cent in 1995. 11. Among others, Wangwe (1995), RPED (1995), Semboja et al. (1997b). 12. During much of the 1960s and 1970s the only R&D institutions in the country were the agricultural research stations. These focused on research on export crops such as cotton, coffee, tobacco and tea. Further developments in science and technology institutions came in the 1970s with the establishment of the Faculty of Engineering at the University of Dar es Salaam in 1973, and the creation in the late 1970s and early 1980s of R&D institutions such as the Tanzania Industrial Research and Development Organisation (TIRDO), the Tanzania Engineering and Manufacturing Design Organisation (TEMDO), the Institute for Production Innovation (IPI), the Centre for Agricultural and Rural Mechanization and Appropriate Technology (CARMATEC) and the Tanzania Bureau of Standards (TBS). In addition to the formally constituted R&D institutions there are some R&D activities that take place in institutions whose main function is not R&D but production or education. In this category one finds enterprises which have workshops or design activities as secondary activities complementing the principal production activities. In addition, universities and

170 The Industrial Experience of Tanzania other institutions of higher learning undertake R&D activities, besides their principal teaching function. 13. Innovative efforts may be rewarded with promotion, recognition and productivity bonuses. For example, rewards for innovative efforts are similar to and sometimes substitute those for productivity, that is, normally in the form of salary increments, with promotion coming only occasionally. Two companies in the garment sample report using an incentive bonus system to achieve higher levels of productivity. The management sets an output target for every section/department and for the entire firm. Any production above the target level is rewarded by a bonus system on a percentage basis. The second bonus system rewards the department that achieves the best results at the end of each year. A third is the general bonus, which is awarded to all the workers depending on the general performance of the firm.

7 Development and Diffusion of Technology: The Case of TIRDO Bartelt Bongenaar and Adam Szirmai*

1

Introduction

In the context of late industrialization the effective transfer and adaptation of technology is of great importance for economic development. International technology transfer is not a costless process, but requires considerable technological effort and investments in the development of technological capabilities (Lall, 1999). Ever since the Arusha declaration in 1967, the Tanzanian government considered a weak technological research and development base as one of the reasons for slow growth of the manufacturing sector (Kahama, 1982). To tackle this problem an aim of government policy was to set up centres of excellence in research and development. One of these institutes was the Tanzanian Industrial Research and Development Organization (TIRDO), which started operations in 1979. The basic functions of this organization are to carry out applied research, to provide technical services to industry and to manage a system of documentation and information designed to enhance industrial production. An important task of TIRDO is to adapt technology to local circumstances and to transfer it to domestic industrial firms. The technology projects of the institute make use of domestic resources and try to substitute for imported products. The aim of this article is to take TIRDO as a case study of technology development and technology diffusion in the context of an African developing economy. On the basis of an in-depth analysis of a large number of technology projects, the article tries to identify the factors contributing to or hampering the successful development and diffusion of technologies in the context of a developing economy.1

2

The Tanzanian R&D sector

The policies of the government research institutes are linked to wider government socioeconomic policies and policy goals (see Chapter 1). The 171


first national science and technology policy for Tanzania was formulated in 1985. The STP policy was derived from the general long-term Basic Industrial Strategy. The shift towards a more open and market-oriented economy since 1986 and the passing of the national investment act in 1990 necessitated the formulation of a new science and technology policy. Although a draft version of such a plan has been drawn up, a final version still awaits publication at the time of writing this article. R&D in Tanzania is mainly conducted within government organizations. R&D in firms is almost non-existent. The government of Tanzania is also the most important provider of funds to the R&D sector. Government spending on R&D is modest, varying between 0.99 and 1.34 per cent of total government expenditures (Mlawa and Sheya, 1990, table 3.4). Some institutes succeed in getting some form of external support from international donors or generate funds by offering consultancy, training and technical services to enterprises. Usually, the income thus generated is modest (Mlawa and Sheya, 1990, pp. 14–17).2 Eight institutes are primarily involved in industrial research, development and design. The organizations have similar or partly similar goals. Four of them have goals that overlap with those of TIRDO: the Tanzania Automotive Technology Centre (TACT), the Tanzanian Engineering and the Mechanical Design Organization (TEMDO), the Centre for Agricultural Mechanization and Rural Technology (CAMERTEC) and the Institute for Production Innovation (IPI).

The Tanzania Industrial Research and Development Organization (TIRDO) TIRDO was set up in 1979. With additional funding from the UNDP, the EEC and the World Bank laboratories and facilities were created between 1980 and 1990. The aim of TIRDO is to promote the manufacturing of industrial goods and to stimulate the use of local resources. Potential clients include small, medium-sized and large-scale enterprises. According to its statutes, TIRDO’s activities include: the training of technical personnel, development and adaptation of technologies, servicing domestic industries, monitoring technological developments, coordinating the execution of R&D and the promotion of new technologies. As a result of insufficient resources, and an overlap in goals with other organizations in Tanzania, TIRDO’s activities are more limited in practice. They focus on (TIRDO, 1993b, p. 2): consultancy studies for industry: the execution of feasibility studies for its own and other products to analyze economic viability of new products or processes • execution of technical (R&D) projects •

Development and Diffusion of Technology 173 • •

supplying specific technical information to industry technical services for industry.

In 1995 TIRDO employed 128 staff members of whom 26 were university graduates. Educational levels are represented in Table 7.1.

3

Theoretical background

The theoretical rationale for funding public research and development organizations is market failure. Given the semi-public nature and positive external effects of research and development, and the high degree of uncertainty involved in R&D, the volume of private investment will tend to be suboptimal. In the context of a low-income economy, this is compounded by imperfect information and lack of skilled personnel and financial resources in the private sector. This article examines whether a public research and development institute can fulfil its theoretically defined functions, through successful development and diffusion of appropriate technologies. In this section we present a simplified scheme of the technology development process to structure the empirical analysis of technology development projects in Section 4.3 The scheme involves two basic assumptions. The first is that a research and development institute in a low-income economy will not develop technology from scratch, but will use and adapt existing technology to fulfil technological needs: it innovates rather than invents,4 and it focuses more on specific techniques than on wider technologies.5 The second Table 7.1 Department

Educational levels of TIRDO staff in 1995a High High middle Low middle Low Unknown Total

Industrial technology department 11 Engineering department 7 Information department 2 DG/DRD/DAF officeb 3 Administration Accounts 1 Estates Dispensary Other 2

5 57

2

4 4 3 2 5 1 5 2

Total

19

29

26

2 3 2

6 3 3 23

1 2

9 2 1 51

3

26 23 8 10 32 6 14 6 128

Source: staff list, TIRDO (1/10/95). Note: a High: university; high/middle: higher vocational training; low/middle: form III and above; low: below form III or no education; DG: director general; DRD: directorate research and development; DAF: directorate administration and finance.


assumption, based on the charter of TIRDO, is that the target group of the research and development organization consists of domestic industrial enterprises. Figure 7.1

Phases in the technology development process

Identification ➛ Acquisition ➛ Adaptation ➛ Selection ➛ Technology ➛ Implementation and selection of of firms transfer of innovations of technology technology

Figure 7.1 presents the six main phases in a model of the technology development process, seen from the perspective of an R&D institute: identification and selection of technologies, acquisition of technology, adaptation, selection of firms, technology transfer to firms, and implementation of innovations. These steps have to be executed for each technology development project and will be analyzed separately. The phases are analytical rather than sequential. There are relationships of circular causation, from ‘later’ phases to ‘earlier phases’ (feedback) and from earlier to later phases, and activities in different phases can – and often should – be undertaken simultaneously. This article examines the innovation process from the perspective of the research institution. Thus, we examine the transfer of technology from the R&D institute to industrial organizations. It should be stressed that R&D is only one of the aspects of a more complex model of technological change and innovation developed by Kline and Rosenberg (Malecki, 1991, pp. 114–17). In that model the interactions between R&D, technology and markets are very important, as the market ultimately defines the technology needs and specifications, and is the customer for R&D outputs. Technological change only occurs if these different aspects are coordinated. The present analysis concentrates on the functioning of an R&D organization and tries to fill part of the black box of ‘research’ in the Kline and Rosenberg model of technological change. 3.1

Identification and selection of technology

The two main concepts in the identification and selection phase are assessment of needs and appropriateness. In order for a technology to be adopted by the R&D institute’s target group, there has to be a need for it. For a research and development institute the target group or market consists of domestic industrial firms and their needs. These needs, in turn, are based on the needs of the customers of the target group: consumers and other industries (Rogers, 1983, UNIDO, 1991, p. 167). Needs for a type of technology can be assessed through formal or informal needs assessment methods (Wissema and Euser, 1988, pp. 19–29).

Development and Diffusion of Technology 175

The second important criterion for technology assessment by an R&D Institute is the appropriateness of the technology. The chosen technology should be appropriate in terms of local market conditions, local resources, labour supply and quality of work force, environmental and geographic conditions, cultural features and national objectives and policies (Riedijk, 1987, p. viii; UNIDO, 1991, p. 167; van Egmond, 1995, p. 10). Two variables affecting appropriateness are technological distance and adaptability. Technological distance is defined as the technological sensitivity to differences in pertinent technological capabilities, and economic and physical circumstances. The greater the technological distance, the more difficult the transfer of technology becomes (van der Straaten et al., 1992; Westphal and Evenson, 1993, pp. 5–6; Caniëls, 1999). Adaptability refers to the question whether the research and development organization has the capabilities to engage in a given technology development process (Mourik et al., 1991). In the process of technology selection assessments have to be made by the management of the R&D institute concerning financial risks, technical risks and the availability of the necessary capabilities.

3.2

Acquisition of technology

Technology development as analyzed in this article involves the adaptation of already existing technologies. The technology will have to be acquired by the R&D organization on the international technology market, a market characterized by extreme imperfectness and a weak and dependent position of developing countries (van Egmond, 1993, p. 86).6 The R&D institute has to try to acquire the coveted technology in an imperfect market at the most favourable conditions. The direct and indirect costs of technology depend on the manner and the package in which it is transferred (United Nations Conference on Trade and Development, 1972, pp. 24–7). ESCAP (van Egmond, 1993, pp. 81–4) distinguishes five categories of technology flows: (1) free-flow of technology; (2) technology flows through the purchase of products; (3) sponsored flows, where third parties (for example, international agents, local governments) pay the costs of the technology transfer; (4) flows embedded in commercial contracts between parties, concerning the production of goods and services; and (5) flows through commercial technology acquisition contracts. The choice of flow mechanism depends on the unpacking possibilities of the technology and the unpacking capabilities of the organization. Unpacking involves the breaking down of a technology into its components and the separate purchase of every component, if possible from different suppliers (United Nations Conference on Trade and Development, 1978, pp. 11–13; van Egmond, 1995, p. 20). This strengthens the bargaining position of the R&D institute and increases the scope of choice between alternative technology flow channels.


3.3

Adaptation

The purpose of adaptation is to make a technology more appropriate to local conditions. The adaptation phase is an extremely important phase in the transfer of technology to developing economies (Dar, 1990, pp. 137ff.; van Straaten et al. 1992). Many innovations fail because of insufficient attention to differences in the socioeconomic and physical conditions in the originating and target environments. The basic goal of the adaptation process is to compensate for the differences between environmental conditions before and after the technology transfer. Therefore, the adaptation process depends on the characteristics of the environment of origin embodied in the design of the technology and the characteristics of the target environment for which the technology was designed. The differences between the circumstances in which the technology has functioned and those in which the technology has to function define the setting and need for adaptations. The adaptation of a technology not only requires the adaptation of knowledge to local factor conditions, but also the capability to modify and add products and processes to suit local preferences and requirements. Adaptation also involves measures to speed up absorption of a technology, by making it more appropriate for local use and specifying required methods for introduction (for example, training) (UNIDO, 1991, p. 177). In the adaptation phase the participating technical staff is the main input. But technical staff members cannot operate without non-technical capabilities within the team. Market knowledge, process control knowledge and knowledge of the social and economic implications of the technology also are essential for successful adaptation of technology (Mourik et al., 1991, pp. 111–112). Therefore, the characteristics of the adaptation team and its external partners are of considerable importance 3.4

Selection of firms

The interest of the R&D institute lies in successful transfers to firms of technologies developed at the institute. Therefore, it should select firms from the perspective of achieving the highest possible success rate in the transfer of the adapted technology. Pack (1987) emphasises that the successful functioning of a technology within a firm depends on the firm’s ability to incorporate the technology in its organization. The selection of firms should thus depend on the R&D institution’s assessments of the technological, organizational, information processing, managerial and educational capabilities necessary to incorporate a new technology into the organization (see also Laseur, 1989, 1991). The selection process also involves the process of getting firms interested in an innovation and persuading some of them to consider adopting it. In a model of innovation decisions, Rogers (1983, pp. 163–72) identifies two phases which determine the innovation decisions of firms: the knowledge


phase and the persuasion phase. In the knowledge phase a decision making unit (DMU) is exposed to the existence of the innovation and gains some understanding of how it functions. In the persuasion phase the DMU develops a favourable or unfavourable attitude towards the innovation, based upon the information available to it. To influence these processes communication instruments can be used (Rogers 1983, p. 318; Wissema and Euser, 1988, pp. 75–80). These include informal communication with opinion leaders in innovation networks, the use of change agents, the use of financial and non-financial incentives, and approaching gatekeepers. 3.5

Technology and knowledge transfer

During the technology transfer phase the technology must be transferred to the selected firms, that have decided to adopt it. Two aspects are important in this phase: the knowledge gap and the organization of the knowledge transfer. The knowledge gap is determined by discrepancies between the knowledge required for the innovation, and the existing knowledge within the firm. A small knowledge gap simplifies the transfer process. The organization of the transfer refers to the instruments used: communication instruments, change agents and gatekeepers. 3.6

Implementation and diffusion

The knowledge and technology transferred have to be effectively diffused throughout the organization and built into a functional system. In the implementation phase special attention must be paid to problems arising during the implementation of the innovation and its diffusion to managers, engineers and production workers within the organization. In this phase the tacit knowledge necessary to make a technology function in a given environment has to have a chance to develop. The implementation of new technologies requires active support, both from within and outside the organization. During the implementation phase, further adaptation and reinvention may be necessary (Rogers, 1983, pp. 16–17). The organization structure and firm capabilities need to be aligned with the production system in which the technology functions (Laseur, 1989, p. 39; Westphal and Evenson, 1993).

4

The field research

Between 1979 and 1996, 25 technology development projects were executed by TIRDO (see Bongenaar, 1997, appendix C). Using the analytical framework discussed in Section 3, 12 of these projects were examined in this study. The unit of measurement is the project. On the basis of the analysis of projects, we draw conclusions concerning the unit of analysis, the research institute.


The first author of this article spent eight months doing fieldwork at TIRDO (from November 1995 until June 1996). Two projects – national dyes and caustic soda – were investigated in great detail. Research methods included repeated open interviews with project officers and document research. Another ten projects were examined with help of a standardised questionnaire completed by the project leaders, again supplemented by documentary research. (For more detailed discussion of the technical aspects of the projects, their output and the research findings, see Bongenaar, 1997.) With the exception of three projects (castor oil, satellite receivers and turkey oil), all projects were well documented. The 12 projects are briefly described in Section 4.1. In the subsequent sections, we discuss the different analytical phases of technology development, making use of the theories and concepts introduced in Section 3.

4.1

The projects

The twelve projects included in this study are: • Natural dyes project: an investigation into the possibilities for the use of natural dyes in Tanzanian industry. The project was executed between 1981 and 1990. The goals for the project were to conduct laboratory tests on procedures and techniques and give small-scale demonstrations on the extraction and use of these natural dyes. • Caustic soda project: a research project focusing on the possibilities of local production of caustic soda by Tanzanian industry. The project started in 1984 and ended in 1994, when the project was stopped after the creation of a pilot plant. Initially the project focused on the batch production of caustic soda. Later it focused on continuous production of caustic soda. The goals for the project were to identify the procedures for the production of caustic soda, to design an appropriate production plant, and to develop and demonstrate domestic commercial production using locally available resources. • Aluminium sulphate project: a project on the production of aluminium sulphate for the use in local water purification. The project started in 1988 and ended in 1993. It was shelved because of the lack of interest of the NUWA (the Tanzanian water company), the organization for which the investigation was started. The goals for the project were to identify and adapt processes for the production of aluminium sulphate, and to demonstrate the possibilities of using locally available resources. • Dehydrated castor oil project: investigation of the production of castor oil using local castor seeds, for the use in, for example, the textile industry. The project was stopped because of a lack of materials and interest from both entrepreneurs and TIRDO management. Goals for the project were to identify and adapt the process for the production of dehydrated castor oil.

Development and Diffusion of Technology 179 •

•

•

•

•

•

Activated carbon project: a research project into the production of activated carbon using local waste materials, used within the food and beverages industries. The project was executed between 1992 and 1994, until it was shelved because of a lack of results from the laboratory investigation. Goals for the project were to identify and adapt the process for production and demonstrate the quality of the local produced activated carbon. Pectin project: a project investigating the production of pectin using local waste materials for use by the food manufacturers in Tanzania. The project was executed between 1989 and 1992, and was stopped because of the insufficient quality of the pectin produced and the lack of funding. The goals for the project were to identify and adapt the process for the production of pectin and to demonstrate the quality of the pectin. Refractories project: investigation of the production of refractories using local materials. Refractories are widely used in all industries involving heating processes and heating facilities. The project was executed between 1988 and 1993, and was shelved because of the lack of outside funding. The goal for the project was to develop a process using specific available raw materials. School chalk project: a project on the production of chalk using Morogoro Ceramic Ware waste gypsum and precipitated calcium carbonate from the caustic soda pilot plant. The project started in 1990, and was still going on in 1996. Goals for the project are the identification and adaptation of a process for production, and the demonstration of its technical and financial viability. Wood adhesive project: Research into the production of wood adhesives using cashew nut shells. After initial research was finished, the project was extended with research into an effective, locally available and safe preservative against insects and fungal decay for wood products. Research was started in 1989, and was still continuing in 1996. Though there was national and international interest in the project, no diffusion took place to industry. Goals set for project are the identification and adaptation of a process for production using cashew nut shells, identification of locally and naturally available fungicides, and the demonstration of the technical and financial viability of the production of the wood adhesives and fungicides. Solar thermal systems project: research project of thermal energy for drying (as main purpose). Within the project, the investigators try to combine specific needs of users with specific solar systems available. The project started in 1991, and was still not officially ended in 1996, though no further activities were being undertaken because of a lack of funding and lack of commitment from entrepreneurs. Goals for the project were to design, produce and stimulate the use of local systems of thermal drying.


Satellite receiver project: project for the local design of a satellite receiver for televisions. The project was started as a follow-up to a satellite dish project. The project started in 1995, and was still going on in 1996. The goal was to produce simple receivers using the organization’s equipment. • Turkey red oil project: a project on the local production of turkey red oil using locally available castor seeds, used within the textile industry to dissolve dyes. The project was shelved because of a lack of funding and lack of interest from the target group. The goal for the project was to identify and adapt a process for the production of the oil. •

4.2

Technology identification and selection

Central to the process of the technology selection is the question ‘why’. What are the reasons for TIRDO to start with each of the projects and what is the justification for the project itself? To assess these questions, one also has to look at who took the initiative for a project, whether a needs assessment was made and who performed the investigation. For ten of the twelve projects examined, the initiative is taken by a member of the organization or one of its governing bodies. Industrial firms play only a marginal role in the process of project selection. Table 7.2 gives an overview of the reasons given for undertaking a project. The identification of a project is most frequently motivated by import-substitution considerations and the availability of indigenous raw materials. In four cases, pressure from the management of the institute to produce a ‘viable’ proposal is mentioned as a motivation. In three cases only is the initiative (partly) the result of contacts with (representatives of) industry. During the whole selection process entrepreneurs are hardly involved. At best, entrepreneurs are contacted by the investigator for technical information concerning the use of certain goods, materials and techniques within its production processes, both in terms of quality and quantity. This lack of involvement of productive enterprises is characteristic of all activities of TIRDO in the development of new technologies. Table 7.2

Motives for initiating projects

Motive National shortages Substitution of imports Availability of raw materials Contacts with industry Follow-up research Pressure to develop proposal Note: n = 12.

Times mentioned 2 9 10 3 3 4


The project proposals are normally written by the principal researchers. They are responsible for the definition of the proposal, and they evaluate the technology and project as well. The quality of the proposals differs considerably because of the lack of standards for the writing up of projects. In some cases the principal researcher involves other members of the organization, such as members of the information department, but this is not the rule. All proposals give a short description of the technology involved, and occasionally discuss alternative technologies. The adaptability of the technologies by TIRDO – is TIRDO able to conduct the research? – is normally investigated. Taking into consideration the complexity and the type of technology and the capabilities of the organization, the time, equipment and funding needed to master and adapt the technology are estimated. Normally a distinction is made between laboratory investigations and pilot investigations. All proposals specify the tasks to be executed within the project. Normally only technical tasks concerning development, production and testing are included. With regard to the appropriateness of the project, two aspects are important: the technology’s desirability for solving certain problems, and its suitability to certain aspects of the environment. These aspects are hardly investigated by TIRDO. Using some characteristics of the technology, the expected effects are estimated by the principal researcher. It is rare for an economist, an entrepreneur or a social group from outside TIRDO to be involved in the investigation of appropriateness. Normally, only positive effects are mentioned; negative effects are simply ignored. Aspects of government policies and economic criteria such as import substitution, foreign exchange earnings and tax earnings are regularly considered. For some projects, socio-economic aspects such as income distribution and job opportunities are mentioned in the evaluation. Financial implications, for example, the earning and investments for future entrepreneurs, are sometimes evaluated. If a pilot plant is within the scope of a project, a division is made within the proposal (in the older projects the pilot plant is specified as a separate project). Since TIRDO does not have the budget to develop technologies itself, budgets are specified in terms of donor contributions and TIRDO contributions (with TIRDO providing the staff for a project). After the approval of the projects by the council of TIRDO, the proposals are used to secure funding from donor agencies. The projects start on a full scale only after funds have been made available; laboratory research normally starts earlier. 4.3

Acquisition of technology

The technologies involved in the projects have some typical characteristics. The technologies are mostly old, always above five years of age, already


studied worldwide and never considered to be of an advanced nature. In two cases, the technology are even not innovative for Tanzania. In the case of the caustic soda project, some components of the technology are found in the domestic economy. But this is an exception. In all other cases, acquisition of foreign technology is the rule. Normally, the aim of the project is process innovation. The technologies are mostly embodied in knowledge: knowledge concerning the process involved, its relation with the available materials and the desired output. Information needed for the project focuses on the process description, and information on the control of the process. Equipment consists of relatively simple locally available or locally producible equipment. In the acquisition phase it is mainly technical staff that is involved. Depending on the project and the activities within the acquisition process, chemical or mechanical personnel or a combination of both participate. The involvement of non-technical staff occurs in half the cases, and is mostly limited to members of the information department. Table 7.3 summarizes the technology transfer mechanism used in the acquisition phase. Two important sources of knowledge are TIRDO’s information department and public libraries. Foreign and domestic R&D institutes are also regularly used for the acquisition of knowledge and information, through training and documentation. The involvement of foreign or local companies is relatively low. When it occurs it is of a noncommercial nature.7 Regarding the technology flow mechanism, free flow is by far the most important. Product-embodied flow associated with the purchase of products occurs for projects in a pilot plant stage. Commercial technology acquisition occurs twice. Usually, the organization does not bargain with technology suppliers concerning the conditions for technology transfer, except in the case of Table 7.3

Suppliers of technology and flow mechanisms for transfer

Technology suppliera Information department Libraries Foreign companies Local companies Foreign R&D institutes Local R&D institutes Other

Times used 9 6 1 2 3 2 1

Flow mechanismb Free flow Sponsored flow Flow accompanied with products Commercial flow

Relative importance (%)c 71 14 10 4

Notes: a n = 12. b Importance of each flow mechanism measured on scale 1–5. c Relative importance based on average scores per mechanism (average score/summed average scores).


acquisition of equipment. Costs for the acquisition of the technology usually seem to be low. Twice, larger expenditures were made for the setting up of pilot plants. A relatively small part of total project costs for the organization is spent during acquisition. Costs identified include expenditures on travel, photocopying and in some cases equipment. In two cases, some indirect costs of an acquisition were indicated. Equipment, only acquired if funding is available, is mostly acquired locally. TIRDO tries to use or adapt local designs for its specific purposes. Only complex equipment, for example, measuring equipment, valves and motors, is purchased from foreign suppliers. The project team approaches several suppliers, and selects on the basis of both technical and price specifications provided by prospective suppliers. In conclusion it seems that the organization acquires technology effectively, using an unpacking strategy. The technologies are acquired from different sources and mostly through free-flow channels. The technologies involved are suitable for this approach: old, well studied and embodied in knowledge and information. The organization has the capabilities to unpack the technology. Costs in this phase are relatively low, compared to total project costs. Bargaining activities are applied in the acquisition of equipment.

4.4

Adaptation

The purpose of the adaptation phase is to make a technology appropriate for its target environment. TIRDO’s main aim in this phase is the adaptation of a technology to locally available raw materials. This has always been the main goal for projects executed by the chemical department. Aspects of the technological environment (available knowledge and equipment), the social environment (special social benefit groups) and the economic environment (financial implications) are sometimes mentioned as secondary goals of the adaptation process. The adaptive criteria tend to be related to physical characteristics of the technology, and reflect the capabilities and areas of interest of the project team. These interests are chiefly technical in nature. Most staff involved in the adaptation process come from within the organization. Besides the project team other departments are sometimes involved, especially during the pilot plant stage. The first part of the research, the laboratory investigation, is normally executed by the project team itself: 80 per cent of the staff involved are chemical and mechanical engineers. Within the organization non-technical involvement in the adaptive work is very low. In two-thirds of the cases, actors from the environment (companies and other institutes) are involved in the work. This contribution is normally limited to the testing of materials. Only R&D institutes cooperate in the execution of the work itself.


In our study the success of an adaptation is measured as the extent to which the output of the project meets the expectations of the staff involved. We distinguish technical and non-technical expectations. Table 7.4 presents the scores on these items for the different projects. Whether a project has been shelved, or whether its goals have been realised within the projected time and budget, is also indicated. As indicated in Table 7.4, the realised technical result often meets prior expectations. However, non-technical expectations are frequently not realised. Many researchers expect that there will be interest for the technology developed on the part of industrial enterprises and donors (industries are more interested in profitability, while donors are interested in the direct applicability). In practice, however, projects often have to be shelved because of lack of interest on the part of donors and industry. Other reasons for shelving projects are insufficient attention on the part of TIRDO’s management, and problems in purchasing raw materials. Insuperable technical problems hardly ever occur, except during pilot plant construction.8 Budget overruns hardly every occur, because of strict financial project control. Only for projects in a pilot plant phase have minor overruns been recorded. In half the cases the research (or part of the research) extended beyond the initially proposed period. Reasons given for this are researchers’ inability to focus sufficiently on a project, acute financial problems, and lack of results. Table 7.4

Success of the adaptation process

Project

Result met expectationsa

Shelved Budget overrun Time overrun

Technical Non-technical Activated carbon Aluminium sulphate Castor oil Caustic soda Natural dyes Pectin Refractories School chalk Solar systems Turkey red oil Wood adhesives Average

6 5 5 8 6 4 5 7 6 6 9 6.1

3 3 1 4 2 2 3 4 4 1 5 2.9

Yes Yes Yes No No Yes Yes No Yes Yes No

No No No Yes No No No Yes No No No

No No No Yes Yes Yes No Yes Yes No No

Note: a Project results meeting technical and non-technical expectations both scaled from 1–10, n = 11. In the case of shelved projects, the question whether results met expectations referred to the partial results. See Appendix A, variable IV, for details of scale construction.


In sum, TIRDO usually executes the adaptations from a narrow technical point of view. From this technical viewpoint, the participation of staff is sufficient, the selection of aspects for adaptations is in line with the staff’s capabilities and the results are good. However, as a result of this one-sided input of capabilities, the technologies are not sufficiently adapted in a broader sense. Financial, economic and social aspects are not included in this phase, and the technologies are not appropriate for domestic industry. 4.5 Selection of firms This section focuses on one of the activities most underrated by TIRDO’s management: the ‘selling’ of a technology to local industry. Within the R&D organization there is still a widespread belief that sooner or later adequate technologies should and will sell themselves. Only a few projects reach the phase where TIRDO searches for entrepreneurs interested in adopting the innovation. Most projects are shelved before this phase is reached. Only in five projects have diffusion activities actually taken place (natural dyes, caustic soda, school chalk, wood adhesives and solar systems projects). In three other projects (aluminium sulphate, pectin and refractories) some promotional activities have taken place during an initial stage of the project. We divide the selection activities into activities to interest entrepreneurs in an innovation and activities to persuade entrepreneurs to adopt the innovation. Table 7.5 summarizes the different methods used. To create interest in the projects, TIRDO uses a combination of general and selective communication instruments. Selective communication instruments are preferred by the organization. The target group of firms is based upon the expected use of certain goods within their current production processes.9 For two projects, Table 7.5

Interest creating activities and instruments for persuasion

(a) Methods for initial interest

(b) Methods for persuasion

Method

Method

Publication Radio & TV Trade fairs Workshops Direct communication through extension officer Direct mailing Other Note: n = 8.

Times used

Times used

2 2 3 3

Free consultancy Technology ownership Support in loan securing Financial stimuli Private guidance

2 1 2 1 2

5 3 2

Tailor-made (re)design Result insurance Pilot plant implementation

4 2 2


the school chalk and the wood adhesives projects, a market survey has been performed to analyze a target group for the innovation. Some activities are undertaken to persuade the enterprises to accept an innovation by reducing entrepreneurs’ uncertainties. The tailor-made design of production systems based upon the specific needs of the entrepreneurs is mentioned in four projects. One method, though not specifically mentioned, is subsidizing the use of a technology. The R&D organization does not try to recover the total development costs of a technology, only charging direct costs and (usually) some fixed fee for development costs. Besides the project team, the information department is involved in 60 per cent of the projects. In some cases, members of the information department participate in the work of the project team; in some other cases their involvement is on request and is limited to selectively approaching some entrepreneurs. The number of interested firms differs considerably across the projects. The number of instruments TIRDO uses to interest and persuade also differs from project to project, and a relation between the instruments used and number of firms expressing interest seems evident. As indicated, the total activities for technology diffusion are modest. The organization does not put enough effort into interesting and motivating entrepreneurs. So far, no firms have been persuaded to adopt an innovation. One may conclude that TIRDO’s execution of the selection phase is not very effective.

4.6

Transfer and implementation

In the survey study the transfer and implementation phases have not been formally included, as no project had progressed up to the stage that the innovation was actually being transferred to firms in the target group. For the two case studies investigated in more detail, natural dyes and caustic soda, some remarks regarding transfer and implementation were made by persons involved. These remarks focus on the plans the R&D organization had formulated for the transfer of the innovations. However, at the time of writing no transfer had actually been realised. For the natural dyes project, it is indicated by participants that TIRDO assumed that SIDO (the Small Industries Development Organization), the organization with which they cooperated in this project, would take the lead in the transfer and implementation phase. TIRDO would only be involved in the transfer of the technical aspects of the innovation to SIDO. It was also prepared to give support in the implementation of the pilot plant. SIDO was to serve as a medium for technology transfer and it was expected to have sufficient capabilities and experience in technology transfer processes. For TIRDO, the main focus within the knowledge transfer phase was on the design of equipment and plant. Production management was not included. Apart from the fact that SIDO was expected to have the


required management expertise, TIRDO itself certainly did not possess the knowledge of how to run a plant. In the case of the caustic soda project there was also no transfer of the innovation. The staff did indicate what the implications of technology transfer would be. They described the equipment and knowledge prerequisites for transfer in general terms. (for example, ‘are expected to have “feeling for chemicals”’). There was interest on the part of firms, a pilot plant was in operation and one entrepreneur seemed to be willing to adopt the technology. However, the worsening economic climate and decreasing prospects of protection of domestic industry prevented the entrepreneur from following up his interest. A few remarks are in order concerning two instruments of importance for the diffusion of technology in the last three phases of technology development projects: the use of change agents and use of contacts within industrial networks. Change agents are normally not used in TIRDO projects. Only in one case, the school chalk project, could a kind of change agent could be identified: one member of the project team 10 focused on supporting all entrepreneurs with the different aspects of the innovation decision. Contacts with firms and individuals within a well-developed industrial network can contribute positively to the selection of a technology for development and the subsequent success in the diffusion of the innovation.11 The involvement of networks in TIRDO’s projects turns out to be low. Direct influence of domestic industry on project content and on activities of project teams is lacking. Though there are some contacts with industrial firms in most of the projects investigated, the scope of these contacts is limited. 4.7

General characteristics of the technology development process

As no transfer or implementation of technology has taken place during the period studied, it is not possible to measure the overall success of the projects in terms of the numbers of successful adoptions. In this section, we use the interest expressed by entrepreneurs in a project, and the stage the negotiations between TIRDO and entrepreneurs have reached, as proxy measures of success. Table 7.6 summarizes five important variables characterizing the four evaluated phases of the technology development process: assessment of existing needs (I) and appropriateness (II) in the technology identification and selection phase, care exercised in acquisition (III), success in adaptation (IV) and use of diffusion instruments in the firm selection phase (V). The table also presents a rough proxy variable for project success (VI) based on the interest in an innovation and the stage the negotiations concerning the innovation have reached. The variables are based on the survey returns. The scaling of the variables is briefly described in the Appendix and in more detail in Bongenaar (1997). The scaling procedures are rough and

Table 7.6

Main variables for the success of the technology development process in different phases

Project Aluminium sulphate Caustic soda Natural dyes Castor oil Activated carbon Pectin Refractories School chalk Wood adhesives Solar systems Satellite receivers Turkey red oil Averagea

Assessment of existing needs (I)

Appropriateness (II)

Acquisition (III)

Adaptation (IV)

3 2.5 2 1.5 3 2 3 2 3.5 4 1.5 7.5 3

6 6 5 1.5 5 6 4 7.5 7.5 7.5 5 4 5.4

8 6 7 5 7 6 5 7 8 7 8 6 6.7

4 4 6 3 5 3 4 6 7 5 – 4 4.5

Activities for diffusion (V) 1.5 10 1.5 – – 1.5 1 8.5 6 1.5 – – 3.9

Success of project (VI) 2 7 6 1 1 2 3 8 7 7 1 2 3.9

Note: a Calculated with respectively n = 12, n = 12, n = 12, n = 11, n = 8, n = 12. The variables are roughly scaled from a negative pole of 1 to a positive pole of 10.


ready, and should not be given any precise mathematical interpretation. Nevertheless, they do provide useful summary indicators of activities in different phases of the technology development process. Only the acquisition phase has an average score of more than six. Two variables have very low scores: the assessment of existing needs (I) and the variable relating to the diffusion activities in the firm selection phase (V). This indicates that the main problems in project execution are concentrated within these two phases. Both phases refer to the interactions between the institute and its environment and indicate that this interaction does not taken place as often and as intensively as required.12 Inspection of the individual projects indicates that their scores in the different phases tend to be related. ‘Good’ projects tend to score higher on most variables. The school chalk, wood adhesives and solar systems are examples of such projects. The castor oil, activated carbon and refractories projects are examples of ‘bad’ projects. This suggests that the inter-phase relationships are also very important. Nevertheless, irrespective of the project, specific phases of the technology development process tend to have much lower scores than the other phases.

5

Conclusions

The research and development organization TIRDO is successful in technology development to the extent that members of the organization succeed in developing technologies that meet their expectations in a technical sense. However, a measure of success in the technical sphere has not resulted in any transfer of the technologies to industrial enterprises, as originally intended. One possible explanation for this phenomenon is the general lack of innovativeness of Tanzanian industrial enterprises, struggling to survive in a difficult environment. During the period studied, many parastatals were being privatized, industrial production was stagnating and there were serious financial constraints which hampered innovativeness (see Chapter 6). Small-scale enterprises in particular lacked the financial and human resources to innovate. The innovativeness of enterprises was not the focus of this investigation, though it is obviously a relevant factor. However, the research reported on in this article indicates that lack of success in technology diffusion is also caused by insufficient efforts to align TIRDO’s activities to the expectations and needs of domestic industrial organizations. To start with our positive findings, both the acquisition and the technical adaptation activities of TIRDO are fairly successful. The acquisition of technology is performed reasonably well. Technologies involved in the technology development processes are normally old, and processes are described in publicly accessible information. The technical staff members involved are well educated in the field of their technical specialization and the


information department of the organization is experienced in the search for information on the international market. The technologies acquired are unpackable and the acquisition teams have the capacity to unpack technologies. In the adaptation phase, in most of the projects the implemented adaptations functioned according to technical expectations of the organization. Many projects can be regarded as technologically successful. One could argue that the experience of successful acquisition and adaptation of technologies in TIRDO has contributed to the building up of technological capabilities within the institute. These technological capabilities are of potential value in future stages of industrial development. Nevertheless, a successful technology development path involves more aspects than those mentioned above. The selection of projects is usually not based on formal or informal assessment of the technological needs of domestic industry. It is seldom related to specific requests from industrial firms. Initiatives for the identification of technologies are invariably taken by members of the TIRDO staff. They are based on national indicators and policy documents, or on needs as perceived by members of the staff. An effective R&D institute should use industry’s needs for technology as a starting point in the development process. The organization should take measures to intensify communication with industry and its management should formulate procedures to translate the needs of industry into viable projects. Direct influence of entrepreneurs on the definition and selection of projects should be encouraged. The organization does little to evaluate the appropriateness of technology for the Tanzanian environment. TIRDO does not make much use of instruments to interest entrepreneurs or to persuade them to adopt an innovation. Few efforts are made to interest and persuade entrepreneurs and there is no consistent policy underlying these efforts. As a result, there is little interest on the part of entrepreneurs, and few entrepreneurs are persuaded to adopt new technologies. The lack of activities in this phase is partly due to the institution’s assumption that good technologies should sell themselves. TIRDO does not consider it necessary to improve the process of diffusion, since, in its opinion, the lack of diffusion is mainly a problem of the domestic entrepreneurs. On the basis of the analysis in this article, the organization should focus on intensifying diffusion activities both in terms of quantity and quality. Diffusion activities should not begin at a late stage after the technology has been adapted. They should start at very early stages of the technology development process. Needs assessment and the selection of firms should be seen as part and parcel of the diffusion process. Research and development institutes should regard the marketing of technologies as an essential element of the research and development effort. An effective R&D organization should step up its efforts to extend its industrial network (cf. Meeus and Oerlemans, 1993). Frequency and intensity of contacts with enterprises


should be increased. Technologies should not be regarded as successful on their own merits, but should be judged to the extent they succeed in serving even a technologically conservative industry’s needs and expectations. Several of the issues and problems discussed in this article are of wider relevance for technology development in research and development institutes in developing economies. Though our limited data do not allow for strong inferences, our results are consistent with findings in the literature. The importance of industrial networks, needs assessments, appropriateness of technologies and marketing are mentioned in many studies as essential to the successful diffusion of innovations. Giving higher priorities to these aspects of technology development could lead to a more effective use of scarce resources invested in research and development.

Appendix Construction of the variables The variables I–V in Table 7.6 are composite variables. The score of each project is the average of the composite variable scores calculated from the questionnaires completed by the principle project researchers and a questionnaire completed by the first author, based upon his analysis of the project files. For more details see Bongenaar (1997), appendix B. I: Assessment of existing needs (scale 1–10) The composite variable score is the average of scores on three variables, each scaled from one to ten: • initiative taken by the organization versus initiative taken by industry (scale 1–10; 1: TIRDO – 10: industry) • initiative motivated by expressed needs of firms or motivated from within the organization (scale 1–10; 1: no needs assessment – TIRDO internal need assessment – government – 10: expressed needs) • degree of influence of target group in this phase (scale 1–10; 1: no influence; 10: maximal influence). II: Appropriateness (scale 1–10) The composite variable score is the sum of scores for two variables: • extensiveness of the investigation of appropriateness, measured in terms of the number of appropriateness criteria mentioned (scale 0–5; number of criteria mentioned divided by 4, maximum number mentioned 20) • indicated appropriateness of the technology, measured by average indicated appropriateness of the technology on a maximum number of 20 criteria of appropriateness (scale 1–5, after rescaling of average scores to a 1–5 range). III: Careful acquisition of technology (scale 1–10) The composite variable score is the sum of scores for two variables: • unpackability: the possibilities of unpacking a technology (scale 1–5; 1: difficult to unpack – 5: easy to unpack). The unpackability scale scores are the average of


•

the scores on the following three items: age of technology (1: young – 5 old), familiarity of technology (1: hardly studied – 5: widely studied) and accessibility (1: mainly private – 5 mainly public). unpacking capabilities: the capabilities of the project team to unpack a technology reflecting technical capabilities and non-technical capabilities (scale 0–5, after rescaling of scores to a 0–5 range).

IV: Degree of success in adaptation (scale 1–10) The composite variable score is the average of two variables: • project result fulfilled technical expectations (scale 1–10; 1: did not fulfil expectations – 10: completely fulfilled expectations); these scores are based upon the scores of 8 items, with a scale of 1–4, which are averaged and rescaled to a 1–10 range) • project result fulfilled non-technical expectations (scale 1–10: 1: did not fulfil expectations – 10: completely fulfilled expectations); the score of this variable is based upon 8 items with a scale of 1–4, averaged and rescaled to 1–10. V: Use of diffusion instruments (scale 1–10) The composite variable score is the sum of scores for two variables: • use of awareness-creating instruments in a project, to create awareness of the technology amongst members of its target group (initial interest) (scale 0–5, number of instruments used divided by maximum number of instruments, rescaled to 0–5) • use of persuasion instruments in a project, to persuade members of the target group to adopt a technology (scale 0–5, number of instruments used divided by maximum number of instruments, rescaled to 0–5). VI: Success of the project (scale 1–10) This variable reflects the judgement of the investigator concerning the degree of interest of firms in an innovation (1: no interest expressed by any firm – 10: adoption by at least one firm). The score is in turn based on • the number of firms showing some interest in an innovation • the stage of negotiations reached regarding an adoption decision.

Notes * Section of Technology and Development Studies and Eindhoven Centre for Innovation Studies, Faculty of Technology Management, Eindhoven University of Technology. Bartelt Bongenaar is now working at IBM, The Netherlands. We thank Emilia van Egmond, Leon Oerlemans and Henny Romijn for valuable comments. 1. This article is based on the MSc thesis of B. Bongenaar, ‘Analyzing Technology Development’, Eindhoven, Technology and Development Studies, March 1997. This thesis was based on eight months of fieldwork at TIRDO. We thank TIRDO and its former director Dr G. Njau for the opportunity to execute this research project. We should like to note the impressive degree of openness on the part of the staff of TIRDO to outside examination of their projects. The aim of this paper is not to offer facile criticisms, but to contribute to better understanding of the factors which may contribute to or hamper the success of technology development.

Development and Diffusion of Technology 193 2. Income from services in the R&D sector is estimated at 15 per cent of total income, external assistance at 30 per cent of total income. 3. For a more extended discussion of the relevant theoretical framework see Bongenaar and Szirmai (1999). 4. Invention is the output and the process by which a new idea is discovered or created. An innovation is an idea or object perceived as new by a group (Rogers, 1983). 5. There is much refined discussion of concepts and definitions in the literature, which we will try to avoid in this article. Technique generally refers to the total of means and procedures for the production and marketing of an existing good. Technology is the wider frame of reference (knowledge, ways and means) within which techniques can be applied to a set of objectives (van Egmond, 1993, pp. 15–16). However, the difference between the two concepts is a matter of degree. In this paper we will use the term technology in a rather rough sense, referring both to specific techniques and the knowledge required to make these specific techniques work. Where the wider concept of technology is referred to, this will be clear from the context. 6. TIRDO projects do not involve the acquisition of domestic technologies. 7. No costs being charged for acquisition, or knowledge being provided without profit. 8. Both pilot projects had problems with the designs and the implementation of the designs. 9. For example, in the case of caustic soda, TIRDO expected the producers of soap to use the technology in their production process; wood adhesives were expected to be used by the producers of particle board. 10. A Dutch student temporarily involved in the project. 11. Better contacts with firms and clients will also have positive effects in the acquisition and adaptation phases. 12. This problem is not limited to TIRDO. It is also mentioned in the literature on the Tanzanian R&D sector in general (Sinda, 1991; Kisimbo, 1994; Mageni, 1994).

8 Technological Capabilities: A Core Element for National Development Opportunities? A Study of Technological Capabilities in the Dwelling Construction Sector for Lower-Income Households in Tanzania Emilia van Egmond-de Wilde de Ligny*

1

Introduction

Only two decades ago the search for factors determining national economic competitiveness led to the introduction of the concept ‘technological capability’. Like technology, technological capability is a complex concept comprising both the utilisation and development of technologies, either through indigenous efforts or through international technology transfers. The search for useful definitions and operational indicators has yielded an extensive body of literature. First, the theoretical views underlying the definition of the concept of technological capabilities are reviewed. Then the preliminary results of investigations regarding the technological capabilities in the dwelling construction sector for lower-income households in urban Tanzania are presented.

2 Emergence of new theoretical views on technological capabilities Research findings on international development differentials, the laggard position of developing countries and the production technologies they use have lately become more or less common knowledge. It is only in the last two decades, however, that the question of which factors determine the competitiveness of nations is seriously addressed. 194

Technological Capabilities 195

By the late 1970s neoclassical economic theories began to lose much of their appeal in the field of technology-related development studies. An important reason for this was that neoclassical economics failed to explain convincingly the role of technologies in production and the processes and causes of technology development (Stewart, 1978; Lall, 1982, 1985, 1987, 1990; Nelson and Winter, 1982). Authors such as Nelson, Winter and Rosenberg even went so far as to claim that, based on neoclassical theories, highly unrealistic and misleading assumptions were made about the processes of technology development (Fransman, 1984). The new views that emerged were largely based on the common denominators in the history of the industrialization of the European countries in the 19th century, Japan after World War II and South Korea in the last decades of the 20th century. The successful industrial development of these countries was attributed to a number of interconnected factors, of which the accumulation of skilled human resources in the form of craftsmen, engineers, scientists, technicians and managers was thought to be of primary importance. These indigenous capacities, it was argued, were used to adapt and improve imported technologies and know-how to suit local conditions, and to develop entirely new and more appropriate ways of problem solving. This specific competence was attributed to a central element existing in the countries concerned: technological capabilities. By the end of the seventies it became clear that countries with hardly any endogenous science and technology also suffered from less favourable social and economic conditions (Stewart, 1978; Sagasti, 1979). Following these ideas, the bottleneck in the formulation and implementation of national development policies to support and improve production is formed by the absence of capabilities, capabilities that should exist in a network of interrelated organisations, institutions and enterprises. The ideas that emerged thereafter were based on the assumption that the conditions under which technology utilisation and technology development take place, either through indigenous research and development efforts or through international technology transfer, play an extremely important role (Lall, 1987). Technological capabilities, it was found, include the abilities of nations, organisations and enterprises to: (1) expand production output, (2) meet the demand for local and/or new products, (3) switch to new production lines, (4) make new investments, (5) upgrade managerial organisational skills, (6) eliminate deficiencies in skills and know-how in the labour force, (7) facilitate access to information and documentation, (8) enhance technological advancements and (9) make use of and upgrade existing (small-scale) production. In other words, the competitiveness of a nation and its ability to follow the rapid pace of social, economic and technological development depend on the status of its technology and its ability to maintain and upgrade this technology.


Owing to a lack of technological capabilities, embedded in a strong technology infrastructure, a country may fail to use its scarce resources efficiently, resulting in high costs to enterprises and the national economy. Inefficiency of utilisation of the existing facilities, declining productivity over time, a high and continuing degree of dependence on imported inputs and technologies, and a lack of local technological infrastructure are all indicators of a lack of technological capabilities. Consequently, it was recommended that nations should dedicate their efforts to the development of a creative technology system in order to decrease the technology dependency and create a technological self-reliance situation (Stewart, 1978; Lall, 1982, 1985, 1987, 1990; Fransman, 1984; Rosenberg, 1986, 1990). International competitiveness in technology-based industries, relying on indigenous technological capabilities, appeared to have become the new measure of national economic performance.

3 3.1

Technological capabilities: a review of the literature Definition of technological capabilities

In most publications the definition of technological capabilities is restricted to mentioning a number of activities without indicating the particular nature of the capabilities. Fransman, Lall and Stewart, for example, refer to technological capabilities in terms of the potential to efficiently and effectively carry out processes of technology utilisation and technology development. These become evident in the efficiency and effectiveness of (1) the technology utilisation in the execution of production processes, (2) the search for available and alternative technologies, (3) the selection of most appropriate technologies, (4) the adaptation of technologies to suit specific production conditions, (5) the improvement of technologies (incremental developments) and (6) the execution of basic research and development for technologies, either in-house or through institutionalised facilities. In the literature no precise indication can be found regarding the particular nature of the capabilities that are needed for technology acquisition, selection, utilisation and development. The only definition indicating which specific components constitute the complex of technological capabilities is the one given by Bell et al. (1984). They refer to (1) the stock of disembodied technical knowledge and information, (2) the human embodied knowledge and experience and (3) the institutional resources as the components of industrial technological capacity. The role of technology learning in the technology development process features quite prominently in literature. Improved production performance through technology development is undoubtedly related to sufficient skills and experience of the labour force. The level of human resource development, however, is certainly not the only indicator of technological capabilities. Technology development in itself should be seen as a production


process in which at least four elements play a role: (1) technology itself, or ‘technoware’, (2) people and their skills and knowledge, or ‘humanware’, (3) information of all sorts, or ‘infoware’, and (4) the ways and means by which production is organised, or ‘orgaware’ (ESCAP, 1989). 3.2

Levels of technological capabilities

Rosenberg (1986) argued that technological capabilities appear to be embedded in the complex network of the social system: in institutions and their structural and cultural characteristics. He failed, however, to make the various concepts fully operational. At the same time he pointed at the existence of a number of levels of technological capabilities in countries to generate technological innovations suitable to their needs, and made an attempt to indicate the existence of a direct relationship between technological capabilities and the performance of production systems. He also mentioned the extreme variability in the willingness and ease with which technologies are adopted and utilised, as well as maintained and improved. In other words, technological capabilities will be positively affected by a national setting with (1) a physical infrastructure to support the production system and (2) national policies dedicated to technology development in support of productive activities. In addition, various authors have tried to determine the levels of technological capabilities in international perspective by associating them with: • • • • • • • •

the nature of the so-called national economic development status (Weiss, 1990) international trade and trade policies (Dore, 1984) trade orientation (Ranis, 1984) restrictions on trade (Lall, 1985) international competitive pressure (Stewart, 1984; Porter, 1985, 1986, 1990) international technology transfers (Dore, 1984) international technology status (Lall, 1982; Katz, 1987) technology gaps (Freeman, 1982; Kaplinsky, 1984).

3.3

Technological capability building

Most authors agree that, by taking the importance of technological capability as a starting point, capability building is the most obvious route to development. Also here, human and institutional capabilities feature prominently. This is why technological capability building is often defined as the accumulation of human and institutional capabilities and is also equated with learning. In this sense it is considered to include the set of processes by which firms accumulate technical knowledge, know-how and experience


relevant to the planning, construction, operation, adaptation, improvement and replacement of production processes, with the ultimate objective to increase production output, both qualitatively and quantitatively (Maxwell, 1977). Bell et al. (1984) indicate that technological capability building should take place in the form of learning by doing, changing, learning from performance feedback, training, hiring of expertise and searching for better solutions. They state that ‘technological capability building enables a firm or economy to produce change in the technical characteristics of the industrial production systems or at least play an active role in that process of technical change’. They make a distinction between the forms of knowledge, expertise and organisation – ‘technological capacity’ – and the knowledge, skills and organisation required for a production process operation and maintenance of an existing production system – ‘production capacity’. Last but not least, Bell et al. emphasise that to build up a significant capacity for generating technological change, additional investments in knowledge and expertise will be required, along with different kinds of institutions and organisations. 3.4

Assessment of technological capabilities

To be able to streamline and enhance technological capability building in the desired direction, the current state of the art of technological capabilities should be known. Among others, authors such as Lall (1982, 1985, 1987), Nelson (Nelson and Winter, 1982; Nelson, 1990), Weiss (1990), Westphal (1990) and Romijn (1996, 1999) have attempted to measure technological capabilities. The complexity of the concept, including the multidimensional aspects incorporated in activities of utilisation and development of technology, requires a multidisciplinary approach, an approach that makes the assessment of technological capabilities a cumbersome but interesting exercise. All definitions found include a number of activities involved in the process of technology utilisation and technology development. In most cases only the output of these activities was measured. 3.5

Conclusion

There is consensus with respect to the fact that improved production performance and further social development can be attributed to the status of technological capabilities and technology utilisation in production processes. In spite of this, there is no consensus yet on an all-comprising and workable definition for the various concepts. Neither is there consensus with respect to a successful way to make them operational.


4

New definition of technological capabilities

In our view, technological capabilities refer to the totality of national resources which can be committed to the production system and provide the necessary inputs for efficient and effective production. This stock of national resources should supply the country with the required means, skills, know-how and knowledge. The stock is not only necessary to select, master and adapt the technologies that are needed and most appropriate, but should also enable the country to develop and generate its own new technologies (self-reliant technology generation). This definition differs from the majority of definitions formulated by other authors. In our definition the human and institutional capabilities are considered to be incorporated in the capabilities to produce, select, maintain and develop technologies, and are represented in the complex of four components: 1. Technology stock. This is defined as the total range of technologies available in the technology subsystem. A distinction is made between product technologies and process technologies. Product technologies refer to the features of the components that constitute the products produced by the sector. Process technologies refer to the features of the production-process components used to transform the raw materials and intermediate products into the required and specified products in the sector. The status of the technology stock indicates the level of technology advancement. 2. Human resources stock. The status of the human resources stock is taken to indicate (a) the availability of human resources relevant for the sector, (b) educational backgrounds and (c) occupational status. 3. Natural resources stock. The status of the natural resources stock relevant for the sector is taken to indicate (a) the availability of exploited natural resources relevant for and applicable in the production processes in the sector, and (b) the capabilities to exploit these natural resources judiciously. 4. Technology infrastructure of institutionalised research and development, education and documentation facilities, technology and intermediate products producing and supplying enterprises and organisations that supply the financial resources to the production system to support productive activities. The status of the technology infrastructure gives information on the complex and the outcomes of interactions between the technology-utilising organisations and firms of the sector and the other technology-supporting and -promoting agents and groups. All this implies that technological capability building means the increase of the quantity and quality of these components. The creation and


accumulation of technological capabilities can be reached through many mechanisms. Next to local development and improvement of the existing capabilities is the acquisition of industrial production systems from abroad, the international transfer of technology. Also for the execution of these processes of transfer and dissemination, technological capabilities are needed in order to search, select and acquire the desired technologies on equitable terms and conditions.

5 5.1

Dwelling construction industry in Tanzania Economic importance of the construction industry

The construction industry has a crucial role to play in a country’s socialeconomic development. First of all the industry is unique in its potential to provide one of the basic human needs: shelter. In addition, backward and forward linkages with other production sectors make the industry an engine of growth for the economy of a country as a whole. The linkages, however, work both ways and the construction industry is vulnerable with respect to developments in other sectors. With the objective to provide insight into the performance of the Tanzanian construction industry and its present and potential contribution to national development, studies were carried out on the technological capabilities of the industry. The focus in these studies was on the dwelling construction for the lower-income households in urban areas. This was justified by the fact that 75 per cent of the construction activities in Tanzania take place in this sector of the construction industry (van Egmond, 1996). In the research at hand, the construction industry embraces the formal and informal enterprises, institutions and individuals that are involved in the process of constructing civil and/or building works. Since small-scale and informal-sector construction activities are excluded from official statistics, data were collected by means of a number of field studies.1 These data were then compared with those found in the literature. 5.2 Current social-economic performance of the Tanzanian construction industry and its (potential) contribution to national development From a historical perspective, the construction sector went through an impressive development considering its state of affairs in colonial times. Foreign dominance, however, is still strong particularly in the fields of building materials and equipment. Over time the social-economic situation in the country affected the performance of the industry. In the early 1980s the industry was at the point of collapse. This was not only caused by the grim overall economic situation, but also by the long-term lack of policy attention for the construction


industry. If policies were formulated at all, such as those formulated after the Arusha declaration (1967), they were not based on any factual insight into the actual situation in the construction sector. At present, lack of confidence in future developments prevents contractors from investing in process and product developments. The lack of investment capacity prevents the further development of the indigenous construction industry. The economic indicators on the current performance of the construction industry show that the potential of the construction sector to contribute to national economic growth has not yet been fully exploited. There is no structural link between the growth of the national economy (GDP) and the share of the construction industry in this growth (see Tables 8.1–3). The construction industry shows an unstable trend, and this points to activities in the informal sector. Small-scale (private) construction units fulfil a large part of the demand for construction resulting from the overall economic growth. Over time the percentage of investment in buildings has been low (8 per cent to 15 per cent of gross fixed capital formation) compared to investments in equipment and works. However, when we take into consideration the estimated average annual gross output of the informal construction industry and add this to the official GFCF figures, then the share of buildings in total GFCF rises from 10 to 42 per cent (in 1990). This figure is more in line with the 40 to 70 per cent figures mentioned in the literature (Moavenzadeh, 1987). The sector’s official role in employment, at less than 1 per cent, is more than modest. The informal construction sector, however, accounts for 64 per cent of the informal sector employment in the country (see Table 8.4). This makes the sector’s contribution to employment and income generation important in a country that has to face high official unemployment rates and an unfavourable income situation for a large part of the population. To a large extent the sizeable role played by the informal construction sector has, so far, been neglected. Official statistics still exclude the productive activities of the informal sector. Since our investigations have shown that the informal construction sector, in terms of the size of its contribution to GDP, is more or less equal to that of the formal sector, this results in a distorted picture. Also, the contribution to GFC is considerable, as is the contribution to employment. All this indicates the existence of a potential that can be better exploited and expanded to benefit the social-economic development in the country (see Tables 8.1–4). Our investigations also indicate that the construction sector has to deal with a number of deficiencies, particularly regarding the informal sector activities. The production output is of a rather low quality and is even formally considered to be of semi-permanent and temporary nature, with a lifetime of, on average, five years. Any investment in construction realized in the current way can thus be seen as a form of capital destruction.

202

Table 8.1

Relation between GDP and construction 1983–93

Average annual GDP growth (%) Share of constr. in GDP (%) Growth in constr. GDP (%) Contribution of constr. to GDP growth (%)

1983

1984

1985

1986

1987

1988

1989

1990

1991

–2.4 2.4 –41.0

3.4 2.8 20.2

2.6 2.5 –8.9 –0.1

3.3 2.8 17.3 0.6

5.1 4.0 49.2 –0.2

4.2 4.3 11.9 0.5

4.0 3.3 19.2 2.0

4.8 5.4 69.8 0.5

3.9 4.4 14.1 –0.6

1992 3.5 4.6 9.4 0.4

1993 3.7 4.0 –10.3 –0.4

Source: Bureau of Statistics. National accounts of Tanzania 1976–1993. 1994c, p. 9/10. The data in the table are based on constant 1976 prices. The contribution of construction to GDP growth has been calculated as ‘growth of construction (%) × share of construction sector in GDP’.

Technological Capabilities 203 Table 8.2

Total GFCF in Tanzania 1983–91 in million current TSh 1983 1984

1985

1986

1987

1988

1989

1990

1991

GFCF Buildings 1 847 2 510 3 096 4 339 6 057 6 802 11 618 20 423 22 101 Works 1 687 2 203 2 762 4 665 13 020 28 089 20 492 48 039 56 871 Equipment 4 218 7 260 11 014 19 675 45 998 62 410 96 912 149 942 183 406 Source: Bureau of Statistics. Selected Statistical Series 1951–1991. 1994e, p. 35.

Our investigations of the dwelling construction units revealed that they face a number of internal problems that prevent them performing their tasks efficiently and effectively. These include (1) the lack of availability of equipment and tools; (2) lack of access to capital and loans, particularly for small-scale and informal contractors; (3) lack of training facilities; (4) lack of assistance in acquiring new jobs; (5) the lack of availability of building materials, and (6) the prevailing government regulations. Also, external factors cause a weak performance of the construction industry. In general the Tanzanian operating environment imposes many bottlenecks on contractors, such as the unfavourable economic situation, the high and ongoing increasing costs of building materials and equipment, and the lack of skills and knowledge among the labour force (see Maro, 1991; Mwaiselage, 1992; NCC, 1992a, 1992b, 1995). The existence of these bottlenecks contributes to the comparatively bad performance of local contractors, and the presence of foreign contractors as a consequence. The literature indicates that the linkages in the actor network of the construction industry are not strong. Moreover, the performance of the various actors in the network leaves much to be desired (Maro, 1991). The Tanzanian government has an important task in alleviating the bottlenecks and in creating an environment to further exploit the potentials of the construction industry to contribute to the social-economic development in the country. The Tanzanian government acknowledges this (NCC, 1995). The overall development objective for the sector was formulated ‘to develop an efficient and effective, self-sustaining construction industry that is capable of meeting the diverse needs for construction, rehabilitation and maintenance of all building and civil works’. Most policy objectives mentioned were never met. Many strategies were formulated in a vague way, and implementation of the strategies was therefore hardly possible. A good example is the policy objective ‘increasing the level of technology’. It is not indicated what this technology should be, which level would be most appropriate and how exactly the current level could best be increased. That most policy objectives were never met is also partly due to external factors, such as increasing costs of imported building materials and tools. A positive aspect is the growing

204

Table 8.3 1983–93

Share of different type of construction activities in GFCF/growth rate of construction activities and total GFCF,

Growth in total GFCF (%) Growth in GFCF buildings (%) Growth in GFCF other works (%) Share of buildings in GFCF (%) Share of other works in GFCF (%)

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

–33.2 –32.7 –17.0 19.9 18.3

45.7 17.2 14.1 16.0 14.3

22.6 3.2 6.5 12.7 10.9

3.0 10.1 27.6 14.4 14.4

29.4 13.1 98.3 9.7 22.0

2.3 19.4 1.3 11.3 21.8

5.1 21.4 –25.3 9.3 17.2

6.5 29.5 136.2 12.9 43.4.

32.0 49.7 32.1 14.5 22.2

11.0 3.33 9.1 16.9 27.2

5.4 16.6 12.3 13.7 30.0

Source: Bureau of Statistics. Selected Statistical Series 1951–1991. 1994e, p. 35; Statistical Abstract: 1993. 1995, p. 31. The data in the table are based on constant 1976 prices.

Table 8.4

The formal and informal construction industry in Tanzania Informal sector

No of peoples employed (× 1000)

160 000 (TIS 1991)

% of total Annual gross output

64 14 577 million TSh 1991

Annual value added

10 864 million TSh 1991

Formal sector 91 649 (LFS 1991) 22 113 (WB 1994) 36 14 416 mill TSh 1991 88 554 mill TSh 1994 4 324 mill TSh 1991 26 420 mill TSh 1994

Total 251 649 182 113 100 28 993 15 188

Sources: Statistical Abstract 1991, Labor Force Survey 1990/1991, National Informal Sector Survey 1991 (Planning Commission, MLYD, 1991), Survey of Construction, Trade & Transport, Dar es Salaam1994.

205


attention now paid in policy documents to the potential of the informal sector. It is clear that an improved performance of the construction industry in Tanzania requires a number of actions to be taken. Judicious stimulation, application and further development of the available technologies should be based on more detailed, thorough and comprehensive insights in the technological capabilities available in the sector. Better insight and knowledge of the potential resources available for exploitation in the country will contribute to a better operating construction industry.

6 Technological capabilities in the subsector of dwelling construction for lower income households in urban areas The four components of the complex of technological capabilities were investigated by executing a literature study, a sequence of field studies in the construction industry and interviews with local experts. 6.1

Status of the technology stock

The data collected regarding the product technologies in the construction industry reveal that the production output and the level of product technological quality do not meet the basic terms of reference (see Tanzanian Building Regulations, BRU 1974–95). The current dwelling-construction sector, for example, provides for only 20 per cent of the actual demand for dwellings. A majority of houses are built in the informal sector. The number of houses built by the public housing sector dwindled to practically zero (Kyessi, 1995). The quality of the production output of the dwelling construction subsector in Dar es Salaam only met the basic requirements in at most 75 per cent of cases. In most cases houses are still built with traditional construction systems (Fereira, 1995; van Egmond, 1998). The characteristics of the complex of the process technology components were used as technology indicators to get an insight in the actual status of the process technologies utilised in the construction industry. Not many detailed data on these technology indicators were readily available. As a preliminary proxy for the level of production process technologies that are utilised within the construction industry, use was made of data on the quantity and type of equipment that contracting organisations owned. Such data were available at the National Board of Architects and the Quantity Surveyors and Building Contractors (NBAQS and BC, 1992). Formal sector contractors in Tanzania need to be registered with this Board. One of the registration criteria is the minimum amount and type of equipment a contractor owns. See Table 8.5. The criteria mentioned in Table 8.5, combined with the number of contractors registered per class, were used to get a first indication

Table 8.5

Criteria for registration of building contractors: the equipment component

Equipment item

Class I

Class II

Class III

Class IV

Class V

Class VI

Class VII

– 1 1 – – – – – – –

– 1 – – – – – – – –

– – – – – – – – – –

– – – – – – – – – –

70

157

146

650

Tower crane Concrete mixers Block making mixers Steel bending machines (set) Water pumps Concrete dumpers Heavy duty motor vehicles Compactors Compressors Concrete vibrators

1 3 2 1 set 2 2 3 2 1 2

1 2 1 1 set 1 2 2 1 1 1

– 1 1 1 set 1 – 1 – – 1

Total number of contractors registered (1990)

45

20

56

Source: NBAQS&BC, 1993 Dar es Salaam, Tanzania.

207


of the level of technology available among contractors. Table 8.5 only lists the minimum amount of equipment available at a construction company. The actual technology level may show a positive deviation from this. Only 10 per cent of the contractors formally registered in 1990 met the equipment criteria for registration. Of these, a considerable number were foreign contractors using foreign equipment. It appears that Tanzanian contractors, formal as well as informal, generally keep little equipment themselves. This can also be concluded from the statistics on capital investments in the construction industry (NBS, various issues). The major reason is that Tanzania has to import the majority of equipment. Only the higherclass contractors are able to finance the purchase of foreign equipment. Moreover, contractors work in project-based organisations and deal with the uncertainty of continuity of project assignments. The risk of investing in equipment is considered too high. Informal contractors do not have the capital to buy equipment at all. From the foregoing, a first conclusion can be drawn: the technology level of contractors in Tanzania, using the equipment criteria set by the NBAQS and BC, appears to be comparatively low. When taking the relatively large number of foreign contractors within the highest three classes into consideration, the indigenous technology level is even lower. The field studies on the process technologies in the dwelling construction industry revealed that construction projects are generally carried out with hand tools. The construction projects, in all cases, are built with hired craftsmen from the informal sector. The amount of information and documents on site is negligible, and the project organisation takes place on an ad hoc basis. The inputs in the construction processes involve, in the majority, raw materials. These findings confirm the earlier conclusion with respect to the low indigenous technology level. 6.2 6.2.1

Status of human resources stock relevant for the sector2 Availability of human resources

The population of Tanzania (30 million) shows the characteristic structure of most of the developing countries. The majority of the labour force, 84 per cent, is working in the agricultural sector. Only 1 per cent of the total population is officially employed in the construction industry. Urban population growth (20.7 per cent between 1991 and 1995), which was partly caused by migration from rural areas, contributed to a high rate of unemployment. The failure to create sufficient jobs in the public and private sectors was, to some extent, neutralized by the creation of employment in the informal sector: 95 per cent of dwelling construction takes place in this sector. With 40 per cent of the population younger than 15 years, the population is relatively young. This not only means an additional burden for the


majority of households, but also puts pressure on social services such as education. If the educational system could meet the future demand of the various production sectors, however, the large percentage of young people would also be an asset. 6.2.2

Educational backgrounds

A positive effect of the policy focus on primary education is that Tanzania experienced a major increase in its literacy rate: from 46 per cent in 1978 to 76 per cent in 1993 (UNESCO, 1995). Unfortunately the quality of primary education deteriorated because of a lack of qualified teachers, instruction materials and schools. Tanzania is among the two countries with the lowest secondary education enrolment in the world: 4.7 per cent in 1990 (World Bank, 1995c). In the period 1986–8, enrolment in tertiary education was only 0.3 per cent (UNESCO, various publications). Training of project managers and entrepreneurs in the sector of dwelling construction for lower-income households mainly takes place in the informal sector. Here people are either trained on the job (34 per cent) or in a kind of apprentice system (57 per cent) (MLYD, 1991). Only a small percentage (9 per cent) is formally trained: 2 per cent in government service, 3 per cent in private enterprises and 4 per cent in training institutes. Our investigations revealed that 91 per cent of the labour force need training in the field of technical skills. In the fields of commerce and finance, 6 per cent of the labour force indicated a training need and 3 per cent needed literacy training (Mwaiselage, 1992; van Egmond, 2000). It was remarkable that none of the contractors or foremen expressed a need for training in managerial skills, whilst the lack of these skills was found to be a major bottleneck. A possible reason for this is that the project management, progress control, quality control (if any), and selection of materials, equipment and labour are normally in the hands of the owner of the house. The training needs show that the performance of the sector depends strongly on the country’s educational system. Only with an appropriate education system can the local construction capacity improve, both quantitatively and qualitatively. 6.2.3

Occupational status

Most of the labour force (84 per cent) in the construction industry work as a craftsman or technician. In the informal sector this percentage is even higher. Some 16 per cent of the labour force in the construction industry are employed as a professional or project manager. With regard to the available research and development staff, no comprehensive data were available. The available research and development statistics show a limited number of researchers in Tanzania. The research and development institutes are almost entirely staffed by Tanzanian nationals.


Mlawa and Sheya (1990) investigated a selection of research and development institutes, and their findings showed that the number of people involved stayed relatively constant over time. Government institutes under direct responsibility of the various ministries predominantly carry out the research and development activities. Relevant for the construction industry are staff carrying out research under the Ministry of Works, Ministry of Science and Technology and Higher Education, and the Ministry of Land and Urban Development. In 1990 some 84 persons were employed as research and development staff in these institutes. This is only 14 per cent of the total research and development staff in the country (Mlawa, 1990). Most of the persons were employed by National Construction Council (20), Building Research Unit (17) and the Ardhi institute of the University of Dar es Salaam (17). These numbers have most probably decreased after the liberalisation policies were introduced. In view of the construction GDP as a percentage of total GDP, and the construction labour force as a percentage of the total labour force, the potential in terms of human resources in research and development institutes appears to be reasonable. It can, therefore, be expected that research and development human resources are an enhancing factor in the improvement of the construction performance in the country, provided, of course, that this potential is fully exploited and that the technology infrastructure is adequate.

6.3

Status of natural resources stock relevant to the sector

The indicators used for the assessment of natural resources for the construction industry are the available quality and quantity of the major inputs in a construction process. It should be noted that the importance of the stock of certain natural resources is derived from the need for building materials. This, in turn, is derived from the demand for building and construction for the other economic, social and infrastructure sectors of the economy. In this context, the importance of the stock of natural resources can be considered as a direct function of the performance in the sectors of the economy mentioned earlier. The following review of the natural resources stock is based on a literature study which was complemented by interviews with a number of local experts. The share of the value of building materials in gross construction output averages 60 per cent. By value, some 70 per cent of the principal building materials used are purchased from Tanzanian producers. Some 30 per cent are imported. The import content, however, of the locally produced materials is rather high. Some 57 per cent, for example, of the costs of the locally produced cement consists of imported components. With some 90 per cent, this is even more for metal products (Kisanga, 1990).


The availability of land is of basic importance to the construction industry. Land use in Tanzania is rather modest, and this is related to a relatively low population density of 33 persons per square kilometre. Natural resources in Tanzania include a variety of metallic and nonmetallic minerals, such as phosphates, tin, iron ore, coal, diamonds, gold and natural gas, as well as water, agro-based resources and forestry resources. Natural resources particularly relevant to the construction industry include clay, with an estimated production of 2000 metric tons in 1992; gypsum with an estimated production of 35 000 metric tons in 1992; limestone, pozzolanic materials, vermiculites, lateritic soils and stones such as marbles, basalt and granites. Especially lime can be of great importance for the Tanzanian construction industry. The current production is, however, inadequate, and uses primitive technologies. Kimambo (1984) listed several problems faced by the Tanzanian mining industry in the early 1980s. These, among others, include a lack of financial resources, a lack of skilled high and middle management, a lack of transport facilities, a very small internal market and a long distance to the main world mineral markets. After a decline in the 1980s, mining GDP increased during the early 1990s. A survey among 43 local building-material producers by the National Construction Council (NCC, 1992a) revealed a shortage of local raw materials as the major constraint for their production. This confirms the present inability of the mining sector to supply local building-material producers with the necessary raw materials. Forests and woodlands constitute 47 per cent of the total land area. The government reserved one third of the total forest area for industrial production. Currently, the share of the construction industry in total wood use is only 2.2 per cent. The local wood industry is able to fulfil the demand for wood from the construction sector in a quantitative sense, but not in a qualitative sense. Since natural regeneration is slow and replanting rare, the area under forest decreases. Particularly around settlements, fuel-wood consumption leads to serious deforestation (Iwaarden, 1997). 6.4

Status of the technology infrastructure (actor network)

Indicators used for the assessment of the status of the technology infrastructure are the type of activity of the technology-supporting and -promoting agents and groups, the policies, objectives and strategies, the source and availability of capital and the nature of the relations with the construction companies. Our investigations concerning the major actors of the technology infrastructure for the subsector of dwelling construction in urban areas led to the following conclusions. The technology infrastructure of the construction industry is rather diffuse by nature because of its fragmented character of operations, the presence of a diversity of different actors in the network and the


heterogeneity of inputs and output. Many bottlenecks in the operating environment contribute to the comparatively low-level performance of the construction units in the execution of construction projects. (Maro, 1991; Mutagahywa and Materu, 1992). The linkages within the actor network of the sector of dwelling construction are rather weak. Communications and relations of the major actors with the construction units that carry out the dwelling construction projects for lower-income households are practically non-existent. Materials and equipment are generally bought in the informal sector. Moreover, the performance of the various actors in the network leaves much to be desired (Maro, 1991). The most important causes mentioned were a lack of capital, skills, experience, equipment and management. 6.5 Conclusions with respect to the status of technological capabilities The technology stock shows a low level of technological advancement. Indicators are the low quantity and quality of the product technologies, and the limited quantity and traditional nature of the process technologies. The human resources potential for the construction industry is relatively large, provided that the education and training opportunities improve. At present the majority of the labour force is employed and trained in the informal sector, and has a lack of knowledge and skills. The number of research and development staff is promising and might form a reasonable basis for a further development of the construction industry. The Tanzanian natural resources stock has considerable potential to provide the necessary raw materials. The exploitation of natural resources by the mining industry is limited, given the high import content in the majority of building materials. The technology infrastructure is weak and has a number of bottlenecks (Maro, 1991; NCC, 1995). These are mainly related to the functioning and nature of the actors in the infrastructural network itself, and the communications and relations between the actors and the construction units building houses for lower-income households in particular.

7

Conclusions

The conclusion is that optimal performance of the sector of dwelling construction for lower-income households in urban Tanzania requires considerable efforts with respect to technological capability building. This implies an upgrading of the technology stock, the development and more efficient use of the available stock of human and natural resources, and an improvement of the technology infrastructure. See also Table 8.6. The work presented is based on an extensive study of the available literature and a number of field studies. The adoption and application of the

Table 8.6

Requirements for technological capability building Requirements for TC building

1. quantity of output of bottlenecks in the construction process

investment in the housing sector and its backward linkages for the alleviation

2. supply of facilities in the house

investment in physical infrastructure: access roads, water, electricity and sewerage

3. application of appr. materials and constr.systems

investment in establishment of local production of building materials with high local content

4. application of proper phys.-techn. constr. details

investment in training and education, information and documentation systems

5. simplification of production processes on site

investment in r&d of construction technologies and systems; investment in proper diffusion of available technologies

6. cost reduction

investment in r&d for the determination of solutions for improvements and lower costs of construction technologies and systems

Process technologies

Requirements for TC building

1. access to proper equipment and tools with high local content

investment in the establishment of local production of equipment and tools

2. knowledge and skills of project managers, general foremen, craftsmen, labourers

investment in the establishment of proper training and education systems adjusted to the needs with regard to: project management, supervision and technical control, crafts, training possibilities on the job and literacy

3. avail. of and access to proper info. and doc. regarding technical specific. and doc., planning and control systems, material and equip. data bases

investment in: (a) establishment of well functioning and accessible consultancy and advisory organizations (b) establishment, accessibility and diffusion of planning and control systems (c) establishment, accessibility and diffusion of materials & equip. databases

213

Product technologies

(continued)

214

Table 8.6

Process technologies


4. project organisation and management

investment in: (a) training and education of enterprise management including planning, administration, cost control, personnel management (b) improvement of technology infrastructure, access, communication and relations (c) establishment and accessibility of financing system

Human resources


quantity of required labour force

investment in education and training of an additional 4000–5000 peoples professionals (10% of total additionally required) supervisors & technicians (40% of total additionally required) craftsmen and labourers (50% of total additionally/required)

quality of required labour force

investment in: 1. improvement of existing educational system 2. (on-going) education and training of the existing manpower in (a) enterprise management, (b) project management, (c) supervision and technical control, (d) crafts, (e) specific training on the job and (f) literacy

Natural resources

Required TC building

land delivery system

investment in surveying and delivery system of dwelling construction plots

energy supply

investment in the establishment of an adequate energy supply system

mineral resources

investment in adequate r&d and production of local building materials with high local content of inputs

agro-based and forestry resources

investment in adequate r&d and production of local building materials with high local content of inputs

Table 8.6

(continued)

Technology infrastructure


clients

investment in establishment and accessibility of info. and doc. systems for clients and house owners

consultants

investment in: (a) training and education (b) establishment and accessibility of financing investment in: (a) mining, quarrying, transport and infrastructure (a) r&d on product and process technologies (c) training and education of manpower (d) availability of equipment and tools

materials & equipment suppliers

r&d institutions

investment in: (a) training and education of r&d manpower (b) facilities and equipment (c) info. and doc. system for a proper diffusion of the results

information and documentation centres

investment in establishment and accessibility of info. and doc. facilities for the construction industry including all actors of its network.

educational institutions

investment in the establishment and accessibility of education and training programmes adjusted to needs.

financing organisations branch organisations

investment in establishment accessibility of financing systems for the construction industry investment in improvement and accessibility of branch organisations

labour organisations

investment in establishment of incentive systems 215

216

Table 8.6

(continued)

Technology infrastructure

Requirements for TC buildinggovernment

investment in:

(a) improved insight in the actual needs for an optimal performance of the construction industry through further research (b) establishment of an enabling environment for the improvement of the performance of the construction industry by judicious application of policies and strategies in terms of legal, financial or fiscal measurements. These refer, for example, to the establishment of adapted regulations, standards and norms, financing and saving systems, tax regulations on imports and investments


concept of technological capabilities in technology and development studies appear to be useful. Though not yet elaborated on in detail, the assessment of technological capabilities, by using the indicators as applied in our studies, provides the opportunity to get a comprehensive and integral view of the opportunities, problems and constraints for the further development of a sector. More precise information is needed to be able to determine the details of the input that should be given by specific actors, and the sequence in which this input should be given. The sequence should be based on the assessment of possible costs and the effects of the efforts in the short, medium and long term. Also, here an integral multidisciplinary approach should be applied in which the social-economic, as well as the technological aspects, are determined, analysed and evaluated.

Notes * 1.

2.

Section of Technology and Development Studies, Faculty of Technology Management, Eindhoven University of Technology. The field studies were executed from 1993 to 1996 by Dankers, Rijkenberg, Tegelaars and Treffers, MSc alumni of the Eindhoven University of Technology and supervised by van Egmond and Gaillard, senior lecturers in Technology and Development Studies, Eindhoven University of Technology. The data for this component of the technological capabilities are derived from Statistical Abstracts (National Bureau of Statistics, formerly Bureau of Statistics, various issues), UNESCO (1995) and World Bank (1995c).

9 Technical Education, Knowledge and Skills in the Metalworking Industry in Tanzania Raymond Duijsens and Paul Lapperre*

1 1.1

Education and technical education Importance of education for development

Although the nature and extent of the role of education in the successful transition from a dominantly agricultural society to a dominantly industrial one are still debated (Szirmai, 1997a), in general one can say that knowledge and skills – primarily gained by various forms of education – were among the key variables in achieving the second phase of the Western transition. Against the back ground of a dramatic increase in the pace of technological innovation in the industrial, agricultural and service sectors of the economy in the highly industrialized countries since the late 1980s, and the fact that many developing countries, particularly in Africa, cannot keep up that pace by far (Castells, 1997), knowledge and skills are probably more important than ever in developing countries. Here, only a well-educated and well-trained labour force will be able to help to adopt modern technologies and to adapt these successfully to local production processes. In this context, Tanzania is no exception. Education is, of course, not only important for the adoption and adaptation of modern technologies. Education can induce changes in the values and attitudes towards development and the environment. Education can be geared to help to improve people’s ‘social carrying capacities’. Schooling can enhance the level of tolerance required for living in an increasingly multiform world. Improved health, lower fertility and better nutrition depend, among other factors, on greater literacy (Graham-Brown, 1991). The UNESCO Statistical Yearbook 1995 (UNESCO, 1995) enables us to obtain a bird’s-eye view with respect to the state of education worldwide. Access to education and enrolment is increasing in almost all developing countries, and it is expected that it will continue to do so in the 21st century. It is encouraging that almost all the boys in the world nowadays 218

Technical Education, Knowledge and Skills 219

get some form of primary education. In Asia and Africa, however, enrolment for girls is much lower than for boys, and is likely to hamper the development process in the decades to come. Despite the growth in primary education, illiteracy continues to rise in terms of absolute numbers. A large gap still exists between industrialized and developing countries in enrolment rates beyond primary level. With respect to secondary education, most developing countries were not even expected to reach the 1960 industrial country levels by the year 2000. In the field of tertiary education, the situation is even more a reason for concern. In particular, enrolment in technical education stays behind, both at the secondary and tertiary levels. In most developing countries, education is provided publicly (by the government), and the determinants of demand (private costs versus expected private benefits: the demand is a derived demand) are generally more important than the determinants of supply (determined largely by political processes unrelated to economic criteria other than limitations to public expenditure) (Lapperre, 1993). UNESCO forecasts for the next century suggest a continuation of these trends. 1.2

Demography and education

Although Tanzania’s government expenditure on education per student at the primary, secondary and tertiary levels compares favourably to that of other countries in the ‘low-income economy’ bracket, its enrolment performance at all levels of education, but particularly at the secondary and tertiary levels, is poor. Its drop-out rates at all levels are high – between 25 and 60 per cent – and the percentage of students (in the specific age brackets) with the highest diploma is comparatively low (UNESCO, 1989, 1994, 1995). Tanzania’s educational system is under great pressure of a relatively large and rapidly growing population and a large and growing number of people in the age bracket of 5 to 24 years. Figure 9.1 presents Tanzania’s population (millions) in the period 1960–95 and the corresponding Dutch population as a comparison. Figures 9.2a–d present the age group division in respectively Tanzania in 1957 and 1990 and the Netherlands in 1960 and 1990 (Bos, 1995; Mitchell, 1995). From Figures 9.2a–d some relevant observations can be made with respect to the (volume) problems that Tanzania encounters in education now and will encounter in the years to come. In the Netherlands the population is still growing, but in the next decade is likely to begin to decrease. Tanzanian population growth exceeds the Dutch one by a factor of 3, and the growth is likely to continue. Hence the total number of people eligible for education in Tanzania will continue to grow rapidly. The age group division figures for Tanzania show that the percentage of the population under 25 years increased in the period between 1957 and 1990. The

220 The Industrial Experience of Tanzania Figure 9.1 Population in absolute numbers (millions) in Tanzania and the Netherlands (1960–1995)

Tanzania Netherlands

Figure 9.2a

Age groups in Tanzania in 1957 in absolute numbers

1 500 000

1 000 000 Tanzania

Figure 9.2b

70–74

60–64

50–54

40–44

30–34

0–4

10–14

0

20–24

500 000

Age groups in the Netherlands in 1960 in absolute numbers

1 500 000 1 000 000 Netherlands

Figure 9.2c

70–74

60–64

50–54

40–44

30–34

20–24

10–14

0

0–4

500 000

Age groups in Tanzania in 1990 in absolute numbers

5 000 000 4 000 000 3 000 000 Tanzania

2 000 000 1 000 000 70–74

60–64

50–54

40–44

30–34

20–24

10–14

0–4

0

Technical Education, Knowledge and Skills 221 Figure 9.2d

Age groups in the Netherlands in 1990 in absolute numbers

1 500 000 1 000 000 Netherlands

500 000

70–74

60–64

50–54

40–44

30–34

20–24

10–14

0–4

0

absolute number of potential pupils/students (age bracket 5–24) almost tripled. In the Netherlands the percentage of the population under 25 years, between 1960 and 1990, decreased significantly and the absolute number of potential pupils/students (age bracket 5–24) remained virtually constant. In view of Tanzania’s financial constraints, bringing the educational infrastructure (buildings, teaching materials, teachers) in line with the demographic developments will raise serious problems indeed. 1.3

Educational structure and enrolment

Formal education in Tanzania is controlled by the state through three ministries: the Ministry of Education and Culture, covering primary and secondary education, the Ministry of Science, Technology and Higher Education, covering higher education, and the Ministry of Labor and Youth Development, covering vocational training (Ministry of Education and Culture, 1993). There are some private schools, mainly at the secondary level, established according to conditions set by the ministry. Private universities are a recent phenomenon (1997). Policy issues and curricula for all schools (government and private) are the responsibility of the respective ministries. Administration of primary and adult education is decentralized to local authorities. Except for private schools, where ministry-controlled school fees are paid, all primary education in Tanzania is paid for by the government, which allocates about 20 per cent of its recurrent budget to education. Increasingly, however, also a parent contribution is demanded. In secondary schools a government-stipulated fee is to be paid by the students. Furthermore, each school is meant to meet 25 per cent of its catering bill (expenses for food, boarding facilities, and so on) by school-owned workshops. This is typically a relic of socialist times. Primary education starts at the age of seven and lasts seven years, that is Standard I to VII. The teaching related to technology or preparatory subjects is restricted to calculus and (natural) sciences. At the end of Standard VII, pupils sit for a Primary School Leaving Examination. After this, a pupil may choose post-primary technical education, national vocational training or secondary school.


In post-primary technical education pupils are trained to be artisans for two years. After this, they may decide to work or to enter national vocational training (NVT) for two years. During the period at one of the NVT centres, there is a close interaction with industry. After finishing NVT successfully, pupils will have completed Trade Test 3. After several years of working experience, they may return to NVT to pass their Trade Test 2 and eventually even their Trade Test 1. Entrance to secondary education (O-levels) is open only to those selected on the basis of a specially designed formula after Standard VII. O-level education lasts four years. Students can choose the following orientations: agriculture, home economics, commerce and technology. Students who pass well the National Form 4 Examination at the end of their O-level education may enter Form 5: secondary education (A-level). The number of students admitted depends on the number of places available. The orientations are the same as for the O-levels. After two years, the students sit their National Form 6 Examination, which leads to the National Higher School Certificate, which is equivalent to ‘A-level’. Students who pass Form 4, after following technically oriented secondary education, may enter a technical college. When they pass after three years they obtain a Full Technician Certificate. After Form 6, or after technical college, students can apply for university admittance. Enrolment requirements include high academic qualifications and good character references. Technical courses (BSc) have a duration of four years. Figures 9.3a–c present the gross enrolment (percentage of age group) in primary, secondary and tertiary education in Tanzania and the Netherlands. Note that gross enrolment generally overestimates the actual enrolment in developing countries by some 25 per cent and in

Figure 9.3a Gross enrolment in primary education (% age group)

Figure 9.3b Gross enrolment secondary education (% age group)

120

140

100

120 100

80

80

60

60

40

40

20

20



1995

1990

1985

1980

1975

1970 1975 1980 1985 1990 1995

1970

0

0

in

Technical Education, Knowledge and Skills 223 Figure 9.3c Gross enrolment in tertiary education (% age group)

Figure 9.4 Technical students in secondary education (% total secondary)

45 40 35 30 25 20 15 10 5 0 1980 1982 1984 1986 1988 1990 1992 Tanzania Netherlands

industrialized ones by some 10 per cent. Figure 9.4 presents the technical students in secondary education in Tanzania and the Netherlands as a percentage of the total of students in secondary education (Centraal Bureau voor de Statistiek, 1995; Duijsens, 1995; Mitchell, 1995). Compared to the Netherlands, Tanzania has a low gross enrolment in primary education. The net enrolment is even more revealing: between 1980 and 1995, in the Netherlands, net enrolment was approximately 95 per cent. In Tanzania, between 1980 and 1995, it dropped from 70 to 50 per cent. Compared to the Netherlands, gross enrolment in secondary and tertiary education in Tanzania is extremely low. In the Netherlands, from 1980 onwards, there is an upward trend. In Tanzania enrolment was almost constant from 1980 onwards. Between 1960 and 1995, the percentage of technical students in secondary education (as a percentage of total secondary education) in the Netherlands dropped from almost 22 per cent to a mere 11 per cent. In Tanzania the percentage of technical students in secondary education remained almost constant between 1 and 2 per cent. 1.4

Demand and supply of technically educated manpower

Although it is beyond the scope and intention of this paper to make an analysis of the supply and demand of (technically) educated manpower at the secondary level, some preliminary calculations can be made. In 1988, approximately 120 000 students were enrolled in secondary education in Tanzania’s 19 regions (Bureau of Statistics, 1990). Assuming (see Figure 9.4) that approximately 2 per cent of these were receiving some form of technical education, this would bring the number of technical students to approximately 2400. This number – even without making allowances for drop-outs – is less than the unsatisfied demand for approximately


3500 qualified technical personnel as calculated from various sources, most of the demand being in communications, construction and manufacturing (Misanga, 1993). There is little reason to believe that the situation has changed since the early 1990s. In 1990 Tanzania was one of the least industrialized countries in the world, with only 5 per cent of its total labour force employed in industry. Assuming that industrialization will proceed and that Tanzania will reach the average industrialization level of the ‘low-income economy’ group of countries with 15 per cent of its labour force employed in industry somewhere early in the 21st century, the demand for technically qualified personnel at the secondary level will at least triple. In view of the present structure and capacity of the educational system, some 10 000 vacancies would remain unfulfilled. This, of course, tells us little about the levels of knowledge and skills industry needs, and the ways in which the technical educational system caters for these. This issue is pursued further in the next paragraphs for the case of the metalworking industry.

2 Nature and causes of shortcomings in knowledge and skills in the metalworking industry 2.1

Introduction

Metalworking enterprises play a small but significant role in the Tanzanian industry (Bureau of Statistics, 1994d). A state-of-the-art survey of 168 enterprises in the Dar es Salaam region was conducted in 1990 through the Development of Improved Production Systems project (DIPS) (Masuha et al., 1992). The main goal was to make an inventory of available technologies in the metalworking industry, and little attention was paid to knowledge and skills of employees. Not only did the researchers of the University of Dar es Salaam want to rectify this omission, also in the literature increasingly doubts were voiced with respect to the quality of (vocational) technical education (Psacharopoulos, 1991). At the request of the University of Dar es Salaam, the Eindhoven University of Technology agreed in 1994 to a research effort to investigate the nature and causes of shortcomings in knowledge and skills of employees with formal technical (vocational) education in the metalworking industry. Only employees who had recently completed their education were to be included in the research. The preparation of the research took place in the second half of 1994, and the fieldwork was executed from January to September 1995. 2.2 2.2.1

Methodology Identification of the nature of shortcomings in knowledge and skills

In order to identify the nature of shortcomings in knowledge and skills, these were defined in such a way that they could be judged by means of a


survey employing closed-question questionnaires. Irrespective of a worker’s educational level, knowledge was defined as ‘information stored in a person’s mind’, and skills as ‘actions and reactions (physical and intellectual) which a person performs in a competent way to achieve a goal’ (Romiszowski, 1984). For three reasons it was decided not to judge knowledge and skills separately, but to focus on skills. For applying skills one needs a body of knowledge and when measuring skills one will at the same time, though indirectly, measure knowledge (Romiszowski, 1984). In addition, skills contribute more to the improvement of the functioning of employees at the levels included in this research than knowledge. Last but not least, a focus on skills in vocational education will enhance a student’s job opportunities, which is particularly important in developing countries (Kessels, 1992). Once it was decided to focus on skills, four categories were distinguished: cognitive skills (decision making, logical thinking, planning), psychomotor skills (physical action, head–hand–foot coordination), interactive skills (attitudes, feelings) and reactive skills (dealing with others purposively). Each of these skills can be either reproductive (applying procedures or algorithms to known situations) or productive (depending on a present body of knowledge, built up through instruction or experience, which is composed of relevant general principles and which is structured into specific strategies of thought or action). In order to judge the nature of shortcomings in skills of employees in the metalworking enterprises, a questionnaire was drawn up. For this purpose a list, based on a breakdown of a flow chart applicable to any metalworking process, was compiled, including all actions which a person has to perform. For these actions all necessary skills were listed to be judged by scores on a scale ranging from ‘very capable’ to ‘not capable’. Each of the skills was judged both for its productive and reproductive state. One open question was included, referring to major shortcomings in the employee’s formal technical education. From the 60 enterprises (ISIC 3.8) included in DIPS, 28 were selected for a survey. These represented the size categories small (100 employees), the metalworking processes forming, removing, joining and coating, and the production workers’ educational levels vocational training centres (VTCs), technical secondary schools (TSSs) and technical colleges (TCs). In the 28 enterprises, 31 managers (small- and medium-scale enterprises) and foremen (large enterprises) were interviewed, using the questionnaire, with respect to the knowledge and skill levels of their immediate subordinates. 2.2.2

Identification of the causes of the shortcomings in knowledge and skills

For each of the selected institutes the shortcomings as found in the enterprise survey were checked against the goals of these institutes with the


purpose of determining shortcomings not relevant to the institutes (Goodlad, 1979). For this, each of the selected institutes was asked to rank its top five goals out of a total of eight. From these rankings it appeared that each of the indicated shortcomings was relevant with respect to the educational goals of the institutes and, therefore, each of them was included in the remainder of the research. Three types of institutes were covered: VTCs, TSSs and TCs. The total national number of these institutes, focusing on the teaching of metalworking processes, was 45, 12 and 3 respectively. With the assistance of the respective authorities and ministries, eight institutes were selected for further investigation, taking into consideration size (large or small with respect to number of students), ownership (public or private) and location (urban or rural). This resulted in the inclusion of three VTCs, three TSSs and two TCs. Additionally, teacher training colleges (for VTCs and TSSs), the respective ministries (TSSs and TCs) and authorities (VTCs), the Tanzania Institute for Education (responsible for curriculum development for TSSs) and the National Examinations Council (for TSSs and TCs) were visited to collect information. To relate the shortcomings found in the enterprise survey with characteristics of educational institutes, a distinction was made between institutional and instructional characteristics (Hammond, 1973). Aspects of the former are teachers, students, family and community and school management, while instructional characteristics comprise facilities, costs, organisation and contents and methods. The cost aspect was investigated separately and in more detail to see whether the level of financial means affects the level of knowledge and skills of the students. Private VTCs receive relatively generous donor funding and are financially better off than public ones (King, 1988; Syrimis, 1988). Private TSSs receive little or no donor funding and are worse off than public ones. Taking examination outcomes and pass rates as an indicator for the level of knowledge and skills, these were linked to ownership – private and public – and thus to the level of available finance. For the VTCs the pass rates of the 1994 Trade Test 3 examinations were analysed (all subjects included), for the TSSs the CSSE (Certificate of Secondary School Examination: O-level) examination results for 1993 and 1994 (mathematics, English and the combined technical subjects). Since there are no private TCs in Tanzania, a comparison could not be made for these institutes. 2.3 2.3.1

Results Nature of shortcomings in knowledge and skills

Table 9.1 presents the outcomes of the survey in 28 metalworking enterprises. From the table one may conclude that, for all educational levels, reproductive skills are consistently less developed than productive skills,

Table 9.1

Nature of shortcomings in knowledge and skills Skills (scores, 1–4 range)a based on closed questions

Educational level of employee Cogn

Psym

Knowledge and skills based on single open question

Inter

Rep

R

P

R

P

R

P

R

VTC

2.5

3.1

2.8

3.4

2.6

3.3

3.0

TSS

2.4

3.0

2.6

3.2

2.6

3.2

3.0

TC

2.6

3.2

2.8

3.2

2.5

3.3

3.0

Low understanding of English; low knowledge of basic metal working processes; low knowledge of modern metalworking processes; low attitude of striving for excellence; low awareness of safety-related aspects Low understanding of English; low knowledge of modern metalworking processes; low attitude of striving for excellence; low awareness of safety related aspects; low awareness of preventive maintenance Low understanding of English; low knowledge of modern metalworking processes; low attitude of striving for excellence; low awareness of safety-related aspects; low management and supervision capabilities

Note: aAll skills range from 1 (not capable of executing the specific type of skill) to 4 (very capable); Cogn = cognitive skills, Inter = interactive skills, Rep = reproductive skills, Psym = psychomotoric skills, R = reactive skills and P = productive skills.

227


and that cognitive reproductive and productive skills tend to be less developed than psychomotoric, interactive and reactive reproductive and productive skills (for reactive skills, the productive variant was not measured). In addition, and again for all educational levels, the understanding of English, the knowledge of modern metalworking technologies, the attitude to strive for excellence and the awareness of safety-related aspects in the working environment were found to be low. School leavers from VTCs, TSSs and TCs appeared to have respectively a limited knowledge of basic metalworking processes, a limited awareness of preventive maintenance, and a lack of management and supervision capabilities. 2.3.2

Causes of shortcomings in knowledge and skills

Table 9.2 relates the identified and confirmed shortcomings in knowledge and skills to the institutional and instructional characteristics of the various types of institutes. The relationships between observed shortcomings in knowledge and skills and the institutional characteristics are summarized as follows: Students. Many of the indicated shortcomings find their origin in the educational background of the students. Tanzanian education, especially primary education, is known for its high teacher/pupil ratios (1:40 and higher) (Ministry of Education and Culture, 1988; Dar es Salaam Technical College, 1995; Vocational Education and Training Authority, 1995). This forces teachers to use the most basic expositive methods and ‘tight rule’ over a class. This, in turn, makes students passive and less receptive to other methods in subsequent education. Furthermore, and although the selection of new students is, among others, based on their mastery of English, their knowledge of English is insufficient. This situation is even worse at VTCs, where the student population in all classes is a mixture of graduate O-level students and primary school leavers. Family and community. Generally, the student’s family and community do not foster the use of English: as soon as a student sets foot outside the classroom, the environment is predominantly Swahili. Teachers. At the TSS and particularly the VTC level, the mastery of English of the teachers is, in general, poor. In many cases, their educational background is hardly higher than that of their students, and family and community characteristics apply to them in a same way. Furthermore, at teacher training colleges emphasis is put on (technical) contents, rather than on the mastery of the medium of instruction. In addition, the teacher training colleges encounter the same instructional and institutional difficulties as the VTCs and TSSs. As far as TCs are

Table 9.2

Relation between shortcomings in knowledge and skills and institutional and instructional characteristicsa Educational levelb

Indicated shortcoming

VTC Reproductive skills Cognitive skills Understanding of English Knowledge of modern metalworking technologies Striving for excellence Awareness of safety-related aspects Theoretical knowledge (basic metalworking processes) Awareness of preventive maintenance Management and supervision

TSS

Institutional characteristicsc TC

t

√ √ √ √

√ √ √ √

√ √ √

x x

√ √

√ √

√ √

x

s x x x

f,c

sm

Instructional characteristicsd f

c

o,c

m

x x

x x

x x x

x

x

x x x x

x x

x

x x

x

x

x

x

x

√

x

x

√

x √

x

x

x x

x

Notes: a √ shortcoming related to and investigated at the particular educational level; x relationship (either direct or indirect) b VTC, vocational training centre; TSS, technical secondary school; TC, technical college. c t, teachers; s, students; f,c, family, community; sm, school management. d f, facilities; c, costs; o,c, organization, contents; m, method.

229


concerned, there is no teacher training course: all teachers either hold a BSc degree or are graduates from a technical college (advanced diploma in engineering). Most of them, however, have ample working experience in industry, and often combine teaching with working. Their experience, therefore, compensates for the lack of a formal teacher training course, and their mastery of English as well as their knowledge of metalworking technologies appear to be sufficiently high. School management. This has virtually no direct influence on the teaching of knowledge and skills to students. At all levels it is very restricted in its ability to set priorities according to the institute’s needs. The only influence it has on attracting financial means is by acquiring projects from third parties. The relationships between observed shortcomings in knowledge and skills and the instructional characteristics are summarised as follows: Facilities and costs. The educational facilities to teach the various kinds of knowledge and skills (books and machines) are desperately lacking in most institutes, and this is a direct result of a lack of financial means. At the VTC level, private institutes, depending fully on foreign donor organisations, have more financial means available per student than their public counterparts, which rely exclusively on government support. At the TSS level, donor organisations are hardly involved and here the private institutes are worse off than the public ones. Organisation, contents and methods. Other than by limiting attention and time for other subjects, tackling many of the indicated shortcomings by devoting extra attention and time will be difficult under the present conditions. An important reason is that the profoundness and scope of the teaching of skills are hampered by the language of instruction, English, which is not the native language for any of the persons involved in the teaching. As a result, the method of teaching is mainly expositive, a situation that is aggravated by a shortage of educational means. To relate the indicated causes of shortcomings in knowledge and skills to the availability of financial means, the examination outcomes of private institutes were compared with those of public ones (for VTCs and TSSs only: all TCs in Tanzania are exclusively public). At the VTC level, private institutes, cooperating with and sponsored by foreign donor organizations, are in general able to allocate more financial means per student than their public counterparts, which are exclusively dependent on government financing. At this level, the research included the examination results of 494 students at 19 public VTCs and 209 students at 26 private ones.


Contrary to the VTC level, at the TSS level private institutes in general hardly receive any donor funding and they can only partially compensate for this by student fees, since these are limited to a maximum. With respect to the per capita amount of money, private TSSs are at a disadvantage compared to public ones. At this level, the research included the examination results of 10 073 students at 8 public institutes, and 1749 examination results at 4 private institutes. The results of the comparison are presented in Tables 9.3 and 9.4. Although the average pass rates are higher for private VTCs than for public ones, the differences are not significant (–z95%55% Prob >90%

Total nitrogen –

Total phosphorus

Most likely

s

7 600 48 200 81 000

1 800 18 300 17 700

2 100 40 400 15 200

136 800

25 500

43 200

111 000–180 000 86 000–223 000

s

+

s–

s+

920 5 710 9 120

280 1 440 3 650

1 650 2 740 1 010

15 750

3 900

3 400

Most likely

11 900–19 200 8 000–23 000

Industry and Environment 249 Figure 10.2

Most likely BOD loads to Lake Victoria Industrial Domestic

8000

BOD (ton/year)

7000 6000 5000 4000 3000 2000 1000 0 Kenya

4.2

Uganda

Tanzania

TN and TP loading estimates

The results of the assessment of TN and TP loads penetrating to Lake Victoria are summarised in Table 10.9. The waste loads discharged into the lake are about 70 per cent of the generated loads, mainly as a result of wetland filtration. The role of domestic treatment in reducing nutrient loads is insignificant. The estimated most likely total nitrogen load is 136.8 × 103 t yr–1. The most likely P input is 15.75 × 103 t yr–1. Table 10.9 reveals that wet deposition is the most important source of both N and P. Domestic liquid waste plays only a minor role in nutrient loading. The most likely values for N loads are presented in Figure 10.3. For P loads, the picture is similar. Clearly, Tanzania, occupying the largest part of the catchment area, generates the largest agricultural nutrient load. Nutrient loads calculated from livestock population and artificial fertilizer application allow for verification of the presented loads estimated from land use (as described before). The verification shows that 66 000 t yr–1 of N originate from manure, while only 2900 t yr–1 originate from artificial fertilizer application. For P, these figures are 4800 t yr –1 and 800 t yr–1 respectively. For the entire lake basin, therefore, artificial fertilizer application plays a minor role in nutrient loading. Further data analysis revealed, apart from some Kenyan districts, that artificial fertilizer use is insignificant. As a final verification, the basic model, represented by Equation 4, was applied. Average lake water concentrations (CA) of 0.640 mg l–1 for N and 0.074 mg l–1 for P were applied, based on reported sampling studies from Gophen et al. (1995) and Lehman and Branstrator (1994). Combined with

250 The Industrial Experience of Tanzania Figure 10.3 by country

Most likely nitrogen loads, by source, and agricultural nitrogen loads,

Burundi 4%

Wet Deposition 59%

Agriculture 35%

Urban Domestic 6%

Tanzania 44%

Rwanda 7% Uganda 13%

Kenya 32%

a lake volume (V) of 2760 km3, an average depth (H) of 40 m, a river outflow (QE) of 23.5 km yr–1 (Crul, 1993) and settling velocities (VAS)1 of 2.3 m yr–1 for both TN and TP, the nutrient input (WA) was calculated. The resulting WA are 117 000 and 14 000 t yr–1 for N and P respectively, which is only about 15 per cent lower than the rapid assessment results presented in Table 10.13.

5 Rapid assessment evaluation of the Tanzanian industry based on the 1989 industial census The rapid assessment methodology presented above can also be used to get a first impression of the environmental situation on a national scale. The data necessary can be extracted from national censuses or more specific surveys when available. In this study we attempted to analyse the contribution of different industrial sectors to the total environmental pressure on aquatic systems. The data used are those of the 1989 Industrial Census, referring to establishments with ten or more persons employed. As with the analysis in the Lake Victoria region, we limit ourselves in this study to organic, nitrogen and phosphorus pollution. As a departure for the analysis, we start with the 29 ISIC three-digit sectors as presented in Table 10.10, together with their total and proportional contribution to the value added in 1989. According to WHO

Table 10.10

Basic data extracted from the Tanzanian industrial census, 1989 Value added of polluting industries with complete data (000 TSh)

% of value added omitted because of incomplete data

Value added industries for which WHO conversions are available (000 TSh)

% of value added taken into account in the assessment

57 995 567 3 213 597 1 760 820

56 222 869 2 925 230 821 185

3.06 8.97 53.36

43 363 464 1 885 467 97 774

74.77 58.67 5.55

1.43

3 795 996

3 738 471

0.57

3 721 712

98.98

6 003 916

1.31

5 218 226

4 160 039

20.28

1 085 634

20.80

3 855 234 134 708 655

0.84 29.47

2 366 671 5 305 363

2 350 033 4 024 467

0.70 24.14

2 234 182 4 024 467

94.4 75.86

ISIC code

Sector

Value added (000 TSh)

Proportional contribution

311–12 313 331–32

Food Beverages Wood/wood products/furniture Paper/paper products/printings publishing Chemicals/ petroleum/petroleum & coal products Rubber/plastic All metal and metal products

58 526 087 3 213 597 32 161 429

12.81 0.70 7.04

6 531 744

341–42

351/354

355–6 371–2,381

Value added of polluting industries (000 TSh)

Source: Bureau of Statistics, data files of the 1989 Census of Industrial Production (10+ establishments).

251


guidelines (Economopoulos, 1993), not all four-digit ISIC subsectors contribute to the pollution of aquatic systems with organic material (BOD 5), total nitrogen and total phosphorus. Those sectors that do not contribute to any of the loads looked at are not included in the table. The total value of polluting industries adds up to 29 per cent of the total value of industry. The industrial census of 1989 contains many missing data and non-standardised information. These had to be omitted from further analysis. The industries that could be assessed together represent 68.7 per cent of the total value added of all polluting industries. The WHO guidelines introduce a further limitation to the assessment. Conversion values are not available for all subsectors. Sectors for which polluting data are not available are 321, textiles, and 323–4, leather products/footwear. This leads to a further reduction of industries. The remaining industries together produce 43 per cent of the total value added of all polluting industries. With the conversion values from the WHO report the contribution of different sectors to BOD5, total suspended solids, nitrogen and phosphorous can be determined. From Table 10.11 it becomes clear that from the industries taken into account in this assessment, the paper/paper products/printing/publishing subsectors account for almost all organic pollution (BOD5). The subsector food is, within the sample, responsible for most of the emissions which cause eutrophication (P and N). The wood/wood products/furniture subsector contributes significantly to environmental problems because of the nitrogen emitted. The other subsectors, including chemicals/petroleum and coal products and rubber/plastic, are not important for the environmental problems reviewed here. Of course, they could be responsible for important point pollution which could cause other important environmental problems (see next section). One step further in the analysis would be to analyse individual firms within a subsector with respect to their contribution to important emissions. With the data available from the 1989 census this proved to be possible. Because of the small number of firms within a subsample, the presentation of the results of such an analysis would be in conflict with the privacy that has to be taken into consideration.

6 Information on industrial pollution in urban areas and point pollution sources Although information on industrial pollution in urban areas is as scattered as environmental data in general, there is more relevant information on pollution in two of the major cities of Tanzania, Dar es Salaam and Tanga. In these urban areas industries are considered to be heavy polluters. This assumption is now assessed for water and air pollution.

Table 10.11 ISIC code

311–12 313 331–2 341–2 351/354 355–6 371–2 381 Total

Waste volumes and contribution to pollution for different sectors Sector

Food Beverages Wood/wood products/ furniture Paper/paper products/ printing/publishing Chemicals/petroleum/ petroleum & coal products Rubber/plastic All metal and metal products

Waste volume

BOD5

m3

%

[kg]

34 041 272 1 525 445 17 954

0.44 0.02

The Industrial Experience of Tanzania

Tanzania

Tanzania

Tanzania

Tanzania

Tanzania

Historical Dictionary of Tanzania

Tanzania: agriculture

Industrial restructuring: World Bank experience, future challenges

The Experience of Eternity

The Anthropology of Experience

The puzzle of experience

The Experience of Freedom

The Anthropology of Experience

The Experience of Freedom

The Anthropology of Experience

The Puzzle of Experience

The Puzzle of Experience

The Anthropology of Experience

The primitive edge of experience

The Unity of Philosophical Experience

The Neuroscience of Religious Experience

The long transition: building socialism in Tanzania

Lonely Planet Tanzania (Country Guide)

Colitis (The Experience of Illness)

The Puzzle of Conscious Experience

Adorno: The Recovery of Experience

The Content of Perceptual Experience

The Varieties Of Religious Experience

The Industrial Experience of Tanzania

Tanzania

Tanzania

Tanzania

Tanzania

Tanzania

Historical Dictionary of Tanzania

Tanzania: agriculture

Industrial restructuring: World Bank experience, future challenges

The Experience of Eternity

The Anthropology of Experience

The puzzle of experience

The Experience of Freedom

The Anthropology of Experience

The Experience of Freedom

The Anthropology of Experience

The Puzzle of Experience

The Puzzle of Experience

The Anthropology of Experience

The primitive edge of experience

The Unity of Philosophical Experience

The Neuroscience of Religious Experience

The long transition: building socialism in Tanzania

Lonely Planet Tanzania (Country Guide)

Colitis (The Experience of Illness)

The Puzzle of Conscious Experience

Adorno: The Recovery of Experience

The Content of Perceptual Experience

The Varieties Of Religious Experience

Recommend Documents