The Industrial Experience of Tanzania Adam Szirmai and Paul Lapperre

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 x

Introduction
Adam Szirmai and Paul Lapperre 1
Part I: Long-run Economic Performance 9

1 Public Policy and the Industrial Development of Tanzania,
1961–95
Donné van Engelen, Adam Szirmai, Paul Lapperre 11
2 Is African Manufacturing Skill Constrained?
Howard Pack and Christina Paxson 50
3 Measuring Manufacturing Performance in Tanzania: GDP,
Employment and Comparative Labour Productivity 1961–95
Adam Szirmai, Menno Prins and Wessel Schulte 73
4 The Role of Technological Factors in the Early Stages of
Industrial Exports: A Note
Charles Cooper 114
Part II: Innovation, Technological Capabilities and Choice of

Techniques 133
5 Public Choice, Technology and Industrialization in Tanzania.
Some Paradoxes Resolved
Jeffrey James 135
6 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
7 Development and Diffusion of Technology:
The Case of TIRDO
Bartelt Bongenaar and Adam Szirmai 171
8 Technological Capabilities: A Core Element for National
Development Opportunities? A Study of Technological
v
vi Contents
Capabilities in the Dwelling Construction Sector for

Lower-Income Households in Tanzania
Emilia van Egmond-de Wilde de Ligny 194
9 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 informa-
tion 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 Herman Gaillard

Eindhoven University of Eindhoven University of
Technology Technology
IBM Deployment of New The Netherlands
Technologies
Amsterdam Jeffrey James
Tilburg University
Charles Cooper The Netherlands
The United Nations
University/INTECH Josephat P. Kweka
Maastricht, The Netherlands Economic and Social Research
Foundation
Jan van Cranenbroek (ESRF), Dar es Salaam,
Bureau for International Affairs, Tanzania
Eindhoven University of
Technology Paul Lapperre
The Netherlands Eindhoven University of
Technology
Raymond Duijsens The Netherlands
Technology Lex Lemmens
Netherlands Red Cross Centre for Technology for
The Netherlands Sustainable
Development (TDO)
Emilia van Egmond-de Wilde de Eindhoven University of
Ligny Technology
Eindhoven University of The Netherlands
Technology
The Netherlands A.K. Maziku
Ministry of Industries and Trade
Donné van Engelen Tanzania
Technology A.V.Y. Mbelle
The Netherlands University of Dar es Salaam
Tanzania
viii
Conference Participants ix
T.S.A. Mbwette Wessel Schulte

University of Dar es Salaam Eindhoven University of
Dar es Salaam, Tanzania Technology
The Netherlands
Hubert Meena UNIDO/UNDP
University of Dar es Salaam Uganda
Dar es Salaam,
Tanzania Haji H. Semboja
Economic and Social Research
Foundation
Humphrey P.B. Moshi
(ESRF), Dar es Salaam,
Economic Research Bureau
Tanzania
University of Dar es Salaam
Tanzania
Adam Szirmai
Eindhoven Centre for Innovation
Gregory Njau
Studies
Environmental Association of
Tanzania
Technology
(ENATA), Dar es Salaam
The Netherlands
Tanzania
F. van der Vleuten
Howard Pack Eindhoven University of
World Bank/University of Technology
Pennsylvania Free Energy Europe
USA Eindhoven, The Netherlands
Menno Prins Samuel M. Wangwe

Eindhoven University of Economic and Research Foundation
Technology (ESRF), Dar es Salaam
ERICSSON, Sweden 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,

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),

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,

The Netherlands
Peter Scheren, Centre for Technology for Sustainable Development,

Wessel Schulte, Section of Technology and Development Studies, Faculty

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

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,

Introduction
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 devel-
opment, 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 experi-
ences, contributes to a better understanding of both Tanzanian and African
economic development.
Common features of African development include the lack of an indus-
trial 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 econ-
omic 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 coun-
tries. 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 inde-
pendence, 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 instabil-
ity, corruption, military conflict and ethnic strife were major factors con-
tributing 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 insti-
tution, 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 invol-
untarily, 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, struc-
tural 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 econ-
omic 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, liberal-
ization and structural adjustment.
This book is structured into four parts. Part I focuses on long-run econ-
omic performance. It consists of two chapters on Tanzania and two chap-
ters which place Tanzanian industrialization in a comparative international
perspective. Part II brings together five articles dealing with issues of tech-
nology 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 differ-
ent 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 con-
tinued 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 underesti-
mated 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 medium-
and 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 edu-
cation 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 develop-
ment. In a non-competitive, inward-looking and stagnant setting, invest-
ment 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 preoccupa-
tion with technological upgrading. He argues that it is much more impor-
tant for sub-Saharan Africa to find out how and why Asian NICs
transformed themselves into labour-intensive exporters, and why non-
NICS 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 transfor-
mation. 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 compar-
ative 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 inter-
national 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 part-
ners, 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 coordina-
tion 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. Operation-
alizing 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, market-
ing 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 edu-
cation in the metalworking sector. At all educational levels, major short-
comings 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 analy-
sis 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 activi-
ties. This article develops an ingenious methodology to assess environmen-
tal 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 produc-
tion. Though some might argue that in a poor country such as Tanzania
economic development and industrialization take precedence over envir-
onmental considerations, van der Vleuten, Lemmens and associates argue
persuasively in Chapter 11 that there are ample opportunities for leapfrog-
ging 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 privati-
zation. As various authors indicate, Tanzania is swallowing the IMF and
World Bank medicine of deregulation, export orientation and privatiza-
tion 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 industrializa-
tion 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 his-
torical 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 industrializa-
tion 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 insti-
tutions, 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 con-
straints 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 con-
sensus concerning the shortcomings of the model of state-led import-
substitution 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 manu-
facturing 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 manufac-
turing firms.
On the basis of econometric analyses Mbelle concludes that manufactur-

ing 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 macro-
policies.
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 com-
petitive 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 iden-
tify the main factors which need to be tackled by policy. They include:
improving institutional capacity to provide supportive technological ser-
vices 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 pri-
orities: policies aimed at technological learning and building competitive-
ness, 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 con-
tained 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 empha-
sizes that one of the important new elements in policy making is a consul-
tative 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.
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 sub-
jected 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 manufac-
ture 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 enter-
prises, the colonial government provides fiscal advantages (tax relief on
capital expenditures, refunds of custom duties) and some measure of pro-
tection (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 process-
ing 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 prod-
ucts 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 enter-
prises 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 enter-
prises 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 develop-
ment 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 moni-

toring 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 short-
coming 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, devel-
opment 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 distin-
guishing periods of coherent policy: by empirically monitoring the use of
policy instruments over time. As there are a large number of policy instru-
ments 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 manufac-
turing 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 regula-
tory control, the relative roles attributed to the public and the private
sector, and the nature of dependence on foreign finance. 1 Policy instru-
ments 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. Industrial trade strategy 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
15
16
Table 1.1 (continued)
Aspect of industrial strategy Scaling Policy instruments
III. Relative roles attributed to the a. Private sector Privatisation, investment incentives (tax reductions,
private and the public sector 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
IV. Inflow of foreign finance: relative a. direct foreign investment Subsidies and tax incentives for foreign investors,
roles attributed to (DFI) 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
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 pro-
motion. 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 sub-
stantially 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 indus-
trial 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 indica-
tions 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 invest-
ment 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 protec-
tion (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 instru-
ments 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 self-
reliant in industrial production. Amongst others, this results in the nation-
alization 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 indus-

trial 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 so-
called 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 manu-
facturers 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 appreci-
ation (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.
Table 1.2 Exchange rates: nominal, parallel, real, 1967–94
Year Nominal exchange rate Parallel market rate Ratio parallel/ Real exchange rate
(TSh/US$) (TSh/US$)a nominal (%) (1966 = 100)b
1967 7.14 8.68 1.22 90.24

1968 7.14 8.25 1.16 76.77
1969 7.14 9.10 1.27 68.23
1970 7.14 10.45 1.46 69.58
1971 7.14 15.00 2.10 70.02
1972 7.14 15.40 2.16 71.14
1973 6.90 13.45 1.95 74.68
1974 7.14 14.00 1.96 78.21
1975 8.26 25.00 3.03 70.86
1976 8.32 20.40 2.45 74.78
1977 7.96 15.15 1.90 72.27
1978 7.41 11.75 1.59 68.01
1979 8.22 13.50 1.64 70.33
1980 7.96 23.80 2.90 59.60
1981 8.18 26.00 3.12 51.54
1982 9.52 30.90 3.24 47.57
1983 12.46 39.80 3.19 46.14
1984 18.40 80.00 4.42 42.16
1985 16.50 125.00 7.58 41.88
1986 51.72 160.00 3.09 60.01
1987 83.72 180.00 2.15 87.97
1988 125.00 210.00 1.68 103.82
1989 192.30 245.08 1.27 131.00
1990 196.60 309.08 1.57 148.70
1991 233.90 384.58 1.64 n.a.
1992 335.00 422.92 1.26 n.a.
1993 479.87 506.23 1.05 n.a.
1994 523.45 560.07 1.07 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 sub-
stantial 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 pre-
sented in Table 1.3, which shows that from 1979 onwards the manufactur-
ing sector can on average count on being allocated less than one fifth of
the requested amounts of foreign exchange for the purchase of raw materi-
als, machinery and spares from abroad. In 1981 no foreign exchange at all
is made available for the importation of machinery and spares for manufac-
turing purposes. The same is true for raw material imports for manufactur-
ing 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: 1977 1978 1979 1980 1981 1982 1983 1984a
Industrial raw materials 43 52 24 14 17 16 n.a. n.a.
Raw materials 60 51 22 41 14 11 15 0
(manufacturing)
Petroleum 100 100 93 71 67 89 n.a. n.a.
Machinery 26 23 1 0 7 3 n.a. n.a.
Spares 44 43 14 17 20 1 n.a. n.a.
Machinery & spares 29 29 12 16 0 5 11 16
(manufacturing)
Transport 37 35 13 10 8 2 n.a. n.a.
% of total requests 50 52 38 32 31 33 n.a. n.a.

allocated
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 intro-
duced 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 domes-
tic producers the financial profitability is taken care of by the use of a cost-
plus 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 inter-
mediate 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 mar-
keting costs’.
The three policy instruments mentioned above not only allow for
increased control of the manufacturing sector. They also allow the govern-
ment 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 Price controlled items, 1973–91 (product groups and separate prod-
ucts)
Controls 1973 1979 1982 1983 1986 1987 1988 1991
Total number of items n.a. 329 72 56 21 22 12 2

locally produced n.a. 154 54 50 21 22 n.a. n.a.
imported n.a. 175 18 6 0 0 n.a. n.a.
Total number of products 1 000 1 031 333 225 107 83 n.a. n.a.
locally produced n.a. 456 241 216 107 83 n.a. n.a.
imported n.a. 575 92 9 0 0 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 effec-
tive 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 protec-
tion 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 perform-
ing 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 con-
clude 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 Branch 1984 1986
Canned fruits and vegetables 184 335 Food products 53 65

Soft drinks –24 5 Beverages and tobacco 172 84
Beer 187 1 300
Tobacco 528 317
Textiles 269 240 Textiles 811 55
Sisal and jute bags 1 inf
Tanning and leather 130 inf Tanneries and leather inf 41
Footwear 123 inf
Pharmaceutical products 0 2 952 Plastics and pharmaceuticals 1 762 45
Soap 151 5 258
Chemicals and fertilizers inf 2
Tyres and tubes 270 59 Rubber, glass, wood, paper and cement 309 28
Glass products 31 424
Paper products 26 6 682
Cement 12 101
Metal products 95 inf Iron, steel and metal products 3 780 28
Machinery and transport equipment 1 347 25
Sample average 134 526 Sectoral average 470 n.a.
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 magni-
tude 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 intro-
duced. 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 Own Export Loans/ Import OGLc Barter

resourcesa fundsb retention grants support
1984 35.3 19.4 – 33.4 6.7 – 5.2

1986 26.1 37.2 0.1 27.6 8.2 – 0.8
1987 32.5 34.6 0.2 23.6 8.2 – 0.9
1988 18.6 35.0 0.4 33.5 9.2 2.7 0.6
1989 29.1 26.3 1.1 23.0 11.1 9.1 0.3
1990 22.3 26.1 n.a. 24.3 9.1 18.0 0.2
1991 13.7 28.7 4.6 20.8 4.8 27.0 0.4
1992 * 27.3 n.a. 22.2 2.6 47.4 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 (par-
tially) 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 exemp-
tions 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 effec-
tively 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 man-
ufacturing is still protected. Table 1.5 shows what the effective rates of pro-
tection 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 prod-
ucts to exchange their goods for raw materials and spare parts from abroad.
Finally, the export duty drawback scheme (DDS) is established for manufac-
turers 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 trans-
fers, 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
Table 1.7 External assistance to Tanzania and the Tanzanian manufacturing

sector 1986–94 (US$ m)
Year Total external Assistance to Ratio (%) Technical assistance Ratio (%)
assistance industrya to industryb
(1) (2) (2/1) (3) (3/2)
1986 670.1 52.6 7.8 25.9 49.3

1987 814.9 81.8 10.0 25.0 30.6
1988 905.5 112.1 12.4 24.8 22.0
1989 905.0 224.6 24.8 26.0 11.6
1990 956.2 163.5 17.1 14.8 9.0
1991 1 059.9 171.7 16.2 29.8 17.4
1992 1 112.7 109.0 9.8 32.2 29.5
1993 905.4 48.9 5.4 8.1 16.5
1994 895.0 10.7 1.2 3.2 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 impor-
tance 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 safe-
guards against expropriation and guarantees against nationalization
without compensation, the government attempts to encourage foreign and
local private investment in industry. Additionally an institutional invest-
ment 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 indus-
tries. 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 regard-
ing credit allocation. Since the banking system is forced to operate on com-
mercial 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 lib-
eralisation 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, imply-
ing 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 strat-
egy. 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-tra-
ditional 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 environ-
ment 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 environ-
ment 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 dramati-
cally. When changes in the external environment of manufacturing are fol-
lowed 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 1961–67 1967–73 1973–84 1984–90 1990–95
Import substitution (IS)/export promotion (EP) IS IS IS IS → EP EP

Degree of direct control (low/moderate/high) low moderate high moderate low
Private ownership (priv.)/public ownership (publ.) priv. priv. → publ. publ. publ. publ. → priv.
Relative inflows of foreign DFI DFI → aid aid aid aid → DFI
investment (DFI) and aid
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 negoti-
ations regarding industrial investment between Tanzania and the other
partner states of the East African Community are an example of this rela-
tively benign environment. Thus, in 1964 it is agreed that Tanzania has the
sole right to the manufacturing of radios, motor tyres and tubes, and alu-
minium 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 manage-
able. Reasons for the second balance of payments crisis include the dissolu-
tion 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 macro-
economic 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
Table 1.9 Indicators of changes in the external environment of the Tanzanian manufacturing sector, 1970–94
Year GDP growth rate Inflation Curr. acc./GDP Terms of Trade Debt/GDP GDP growth
(%)a (%)b (%)c (1987 = 100)c (%)c 9 African Economiesd
1970 4.0 3.1 –3.1 n.a. n.a. 9.1

1971 3.5 5.0 –8.0 n.a. n.a. 5.3
1972 6.3 7.6 –4.6 n.a. n.a. 2.5
1973 2.8 10.6 –6.5 145.7 34.2 5.0
1974 1.3 18.4 –14.3 173.6 43.1 6.5
1975 5.1 27.0 –9.5 142.0 48.3 1.4
1976 6.1 6.9 –1.3 151.7 56.2 5.8
1977 0.4 11.4 –2.3 182.0 60.5 4.3
1978 2.1 6.6 –12.7 151.9 62.2 1.4
1979 2.4 12.9 –8.8 138.7 64.3 4.6
1980 3.0 30.3 –11.4 142.0 65.0 5.1
1981 –0.5 25.7 –7.7 129.0 58.3 2.0
1982 0.6 28.9 –8.7 127.1 55.4 2.1
1983 –2.4 27.1 –5.5 128.0 64.1 –1.3
1984 3.4 36.1 –7.4 131.0 77.2 3.7
1985 4.6 33.3 –6.6 126.1 74.0 4.1
1986 6.6 32.4 –8.1 140.5 123.6 3.9
1987 5.9 30.0 –16.1 100.0 211.4 1.9
1988 4.8 31.2 –11.6 107.2 188.0 5.3
1989 2.2 30.4 –10.5 103.2 171.1 3.2
1990 7.1 35.9 –12.5 92.6 202.1 2.2
1991 2.8 28.8 –11.9 93.8 188.5 1.4
1992 1.6 21.9 –12.2 84.7 210.9 –0.6
1993 –0.6 25.2 –12.8 n.a. 240.9 n.a
1994 1.2 34.1 n.a. n.a. n.a. n.a
1995 3.2
1996 4.6
1997 3.2
33
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.
(1980–84) reflect the drying up of external financial flows due to disagree-

ments 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 neg-
ative real growth rates of GDP and external and internal imbalances.
Growth rates pick up from the mid-1980s onwards, but external and inter-
nal imbalances persist well into the 1990s. Calling to mind the periods dis-
tinguished 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 manufac-
turing development is characterized by six periods with differing combina-
tions 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 analy-
sis presented in Table 1.10. As may be expected, there is not always a one-
to-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 manu-
facturing 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 1961–67 1967–73 1973–80 1980–84 1984–90 1990–95
Industrial strategy
IS/EP IS IS IS IS IS → EP EP
Control low moderate high high moderate low
Priv./public priv priv. → publ. publ. publ. publ. publ. → priv.
DFI/aid DFI DFI → aid aid aid aid aid → DFI
Macroeconomic
situation relatively relatively first signs crisis internal and internal and
stable stable of imbalance external imbalance external imbalance
35
Figure 1.1 The share of manufacturing value added in GDP (%)
14
12
10
8
MVA/GDP (%)
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
100
80
Index
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.
Figure 1.3 Index of manufacturing labour productivity, 1965–90 (establishments

with 10 or more persons engaged, 1976 = 100)
120
100
80
Index
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, com-
parative 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
Figure 1.4 Comparative labour productivity in Tanzanian manufacturing, 1965–90

(USA = 100)
12
10
Tanzania/USA (%)
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-à-
vis 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 protec-
tion.
As pointed out in Section 3.1, tariff protection results in high effective
rates of protection, especially for the manufacturers of consumer and inter-
mediate 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 sub-
stitution 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 pro-
tection (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 devel-
opment 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 employ-
ment 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 invest-
ment 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 ven-
tures are established’.
40
Table 1.11 Import substitution of manufactured goods, 1961–90
Year Total manufactured supply Domestic market/total manufactured supply
Domestic market (%) Exports (%) Imports (%) Consumer (%) Intermediate (%) Capital (%)
1961 29.6 8.2 62.2 50.0 25.7 16.4

1965 36.0 7.8 56.2 59.5 32.1 16.6
1973 42.5 6.4 51.1 72.2 39.8 21.4
1980 50.9 6.6 42.5 81.2 55.2 24.3
1984 58.6 5.7 36.3 87.4 58.6 23.7
1990 46.6 8.5 44.9 89.4 47.1 27.7
Sources: 1961–73 data from World Bank (1987a); 1980–90 data calculated from United Republic of Tanzania (1995c).
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 32.4 23.0 10.4 6.2 1.5 26.5

1967 26.4 33.2 9.9 11.5 4.6 14.3
1973 26.8 36.1 7.2 10.6 3.3 16.0
1980 21.4 33.9 9.9 15.4 3.3 16.1
1984 31.9 25.4 9.5 14.4 2.1 16.7
1990 44.2 14.8 9.9 13.1 3.1 14.7
1995 51.4 12.2 12.8 11.9 3.7 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 sub-
sequent 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 impor-
tance 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 contin-
ued 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 combi-
nation 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 capac-
ity 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 pub-
licly 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 sub-
Saharan Africa inefficient production in public manufacturing results to a
considerable extent from an ‘acute scarcity of skilled managers, administra-
tors 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 establish-
ment 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 import-
intensive 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 combi-
nation 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] mana-
gers to focus their efforts on seeking additional import allocations and per-
missions 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 import-
substituting policies have generally been associated with a loss of efficiency
in the longer run (Page, 1990; Krueger, 1992). Protection allows manufac-
turers 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 through-
out 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 produc-
tion. 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 poten-
tial, 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 neg-
ative 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 dif-
ference 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 combi-
nation 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 per-
formance 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 manu-
facturing 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 man-
power, 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 work-
force by the Tanzanian government.
In conclusion, the weaknesses of the oversized and already inefficiently
operating manufacturing sector are revealed by the 1980–84 macroecon-
omic 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 com-
patible 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 fre-
quent 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 substitu-
tion 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 sta-
bilize between 1984 and 1989. So does the trend in the share of manufac-
turing 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 manu-
facturing 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 interna-
tional 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 allo-
cated 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 manufac-
turers can obtain raw materials, machinery and spare parts necessary to
improve capacity utilization. As a consequence of these new import possi-
bilities, 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 struc-
tural changes occurring during the period (Table 1.12). Another disincen-
tive 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 ori-
ented 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 man-
ufacturing 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 treat-
ment of the public sector, which is still cherished by the government
between 1984 and 1990. In 1989, 56 per cent of the publicly owned manu-
facturing 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 large-
scale 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 grad-
ually 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 invest-
ment 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 liber-
alisation 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 posi-
tive 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 contra-
dict the commercial principles banks operate on from 1991 onwards. As
such, credit has replaced foreign exchange as the major production con-
straint 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 con-
tinued 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 envi-
ronment 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 estab-
lishment 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 weak-
ness, 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 develop-
ment. 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
* Section of Technology and Development Studies and Eindhoven Centre for
Innovation Studies, Faculty of Technology Management, Eindhoven University
of Technology
1 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.
2 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.
3 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 sub-
Saharan 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 underly-
ing theoretical framework a view that the national economy can be mod-
elled 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 cross-
country 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 pro-
ductivity growth in sub-Saharan Africa.1 The first holds that pervasive gov-
ernment 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 prob-
lems 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 pre-
sents empirical results on the analysis of the determinants of productivity
in a large group of firms in the three countries. Section 5 offers an interpre-
tation 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 guar-
antee 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 cost-
reducing 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 pro-
duces 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 pro-
duction 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 train-
ing, 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 posi-
tive 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 tech-
nological 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)   (2.1)
 A(t ) 
The extent to which local LDC technology improves is a positive func-

tion, ’(h) > 0, of the level of human capital and proportional to the mag-
nitude of the differential between current and ‘best practice’ technology. As
the technology T(t) does not have to be invented, the potential productiv-
ity gain from the transfer of this technology is the benefit of relative back-
wardness (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 tech-
nologies 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 equip-
ment, new material inputs, new knowledge obtained from consultants,
licensers and foreign owners. The potential level of technology characteriz-
ing 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 avail-
able to the LDC firm. If foreign exchange shortages or arbitrary rules pro-
hibiting 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 conven-
tional inputs in the production function, Equation 2.4 implies that human
capital will have no effect on the level of output obtained with conven-
tional 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 tech-
nology, 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 com-
bined 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 interme-
diate inputs that may embody new technology and (2) the sectoral struc-
ture 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 con-
stant 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 struc-
ture, 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 Imports of goods embodying new technology
Country Growth rate of Share of machinery and transport Share of other manufactured
imports, 1980–93 equipment in total imports imports in total imports
1985 1993 1985 1993
Kenya –0.8 23 25 28 29
Ghana 2.7 40 26 28 31
Zimbabwe 0.2 65 36 26 31
Korea 11.4 34 34 23 29
Malaysia 9.7 46 54 28 30
Source: World Development Report (various issues).
55
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 differ-
ent 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. Empirical results at the firm level
4.1 The data

We analyse the relationship between skills and production costs using firm-
level 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 sam-
pling procedures. All collected information on the value of output, produc-
tion 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 non-
registered) 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 informa-
tion 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 sta-
tistics that follow, but the means of all firms in the sample. It is thus possi-
ble 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 esti-
mate 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, part-
nerships 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 Total Vocational Technical Professional
None 6 5.4% 0 0.0% 0 0.0% 0 0.0%

Primary 54 51.0% 5 7.6% 3 1.7% 0 0.0%
Secondary 87 40.0% 7 2.0% 13 4.2% 10 4.4%
University 46 3.7% 0 0.0% 3 0.0% 8 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)
Education level Total Vocational or technical training
None 3 2.4% 0 0.0%

Primary 32 45.3% 14 8.8%
Secondary 64 46.5% 30 20.2%
University 35 5.8% 6 1.0%
of firms (with owners) in the population, and are calculated using population weights provided
by the survey study.
Table 2.3c Ghana: education of firm’s owners (153 firms)
None 17 11.1% 0 0.0% 0 0.0% 0 0.0%

Primary 12 7.8% 1 0.7% 3 2.0% 0 0.0%
Secondary 107 69.9% 13 8.5% 18 11.8% 4 2.6%
University 17 11.1% 0 0.0% 0 0.0% 3 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) Production manager (N = 58)
None 2 0.1% 1 15.2%

Primary 8 33.4% 3 1.2%
Secondary 47 40.2% 25 57.7%
University – nontechnical 25 3.4% 5 1.4%
University – technical 21 17.8% 21 23.8%
Postgraduate (Kenya) 3 0.3% 1 0.3%
Postgraduate (abroad) 4 4.9% 2 0.5%
study.
Table 2.4b Zimbabwe: distribution of managers’ education
None 3 2.4% 1 0.9%

Primary 27 42.4% 16 14.8%
Secondary 86 47.1% 49 45.4%
Postgraduate (local) 2 0.1% 3 2.8%
study.
Table 2.4c Ghana: distribution of managers’ education
None 4 8.5% 0 0.0%

Primary 4 8.5% 3 5.9%
Secondary 10 21.2% 11 21.6%
Postgraduate (local) 1 2.1% 1 2.0%
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 inter-
national markets. The vast majority of all firms import none of their raw
materials, export none of their output, and do not have any foreign owner-
ship. 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
Table 2.5 Average education of managerial employees
Kenya (126 firms)

None 1 0.2% 0 0.0% 0 0.0% 0 0.0%
Primary 8 14.2% 0 0.0% 2 8.0% 0 0.0%
Secondary 82 74.2% 7 6.4% 8 14.3% 11 1.5%
University 35 11.4% 1 0.0% 0 0.0% 9 1.6%
Ghana (64 firms)
None 2 3.1% 0 0.0% 0 0.0% 0 0.0%
Primary 1 1.6% 0 0.0% 0 0.0% 0 0.0%
Secondary 45 70.3% 4 6.2% 8 12.5% 14 21.9%
University 16 25.0% 0 0.0% 0 0.0% 2 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 indi-
cated 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 func-
tion 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
None 0–25% 25–50% 50–75% 75–100%
Kenya
Percentage of raw materials 121 (76.1%) 32 (11.4%) 15 (3.3%) 25 (3.3%) 25 (5.9%)
imported
Percentage of output exported 170 (78.0%) 29 (6.6%) 7 (0.1%) 4 (0.2%) 8 (0.3%)
Percentage of foreign 184 (96.5%) 4 (0.2%) 6 (1.7%) 10 (0.3%) 14 (1.3%)
ownership of the firm
Zimbabwe
imported
Ghana
imported
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 (mea-
sured by F), the two entering the cost function interactively.
We start by estimating basic cost functions that do not include any vari-
ables 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 differ-

ences 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 sim-
plest specification, in which the logarithm of costs is regressed on the loga-
rithm 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 vari-
ables 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 man-
agerial employees were coded in the same way, and we construct dummies
that indicate whether the owner (or managerial employees) had no school-
ing, 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
Table 2.7 Basic cost functions for formal sector firms or (for Ghana) large
firms (absolute values of t-statistics in parentheses)
Kenya Zimbabwe Ghana
Wood –0.11 –0.09 –0.23 –0.22 –0.32 –0.32

(0.92) (0.70) (1.55) (1.60) (1.58) (1.56)
Textiles –0.03 –0.02 0.01 –0.04 –0.18 –0.15
(0.23) (0.20) (0.08) (0.38) (0.96) (0.79)
Metal 0.06 0.04 –0.02 –0.01 –0.18 –0.19
(0.48) (0.32) (0.12) (0.09) (0.95) (0.98)
ln(q) 0.915 0.75 0.88 0.42 1.00 1.33
(44.94) (5.68) (43.20) (3.87) (27.10) (6.45)
ln(q)2 0.02 0.05 –0.022
(1.57) (1.90) (1.64)
ln(k) 0.24 0.37
(1.65) (3.65)
ln(k)2 –0.01 –0.01
(0.70) (.73)
ln(q)ln(k) –0.01 –0.02
(.43) (0.36)
Obs 125 125 142 142 80 80
R2 .95 .95 .93 .94 .93 .93
Full sample of firms
Formal 0.35 0.28 0.32 0.31 0.12 0.10
(3.56) (2.15) (2.40) (2.26) (0.72) (0.65)
Wood –0.18 –0.17 –0.11 –0.06 –0.14 –0.11
(1.59) (1.50) (0.83) (0.51) (0.73) (.55)
Textiles –0.10 –0.09 0.04 0.03 –0.14 –0.11
(0.85) (0.79) (0.42) (0.27) (0.77) (0.63)
Metal –0.08 –0.09 0.07 0.04 0.02 0.02
(0.70) (0.79) (0.56) (0.30) (0.09) (0.12)
ln(q) 0.90 0.78 0.89 0.69 1.109 1.19
(47.93) (9.94) (46.29) (9.46) (25.78) (8.21)
ln(q)2 0.022 0.06 –0.01
(2.03) (3.06) (1.17)
ln(k) 0.10 0.13
(1.54) (2.69)
ln(k)2 0.003 0.016
(0.33) (1.42)
ln(q)ln(k) –0.02 –0.06
(1.11) (2.09)
Obs 172 172 170 170 132 132
R2 .96 .96 .96 .96 .92 .92
with owner/managers with no education, and so we combine the ‘no

schooling’ and ‘primary education’ categories. For all countries we experi-
mented 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 correla-
tion 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 reduc-
tions 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
Table 2.8a Cost functions with skill measures included, Kenya
66
Kenya: formal firms (obs = 125) Sample means
Owner/manager’s education:
Primary school –0.038 (0.32) –0.017 (0.15) –0.053 (0.45) 0.32
Secondary school –0.062 (0.47) –0.065 (0.49) –0.009 (0.07) 0.18
University –0.056 (0.47) –0.041 (0.33) 0.032 (0.23) 0.31
% raw materials imported 0.0006 (0.46) 0.0000 (0.04) 26.8
% output exported –0.0034 (1.78) 0.0014 (0.59) 8.9
% foreign ownership –0.0019 (1.16) 0.0001 (0.04) 10.2
Owner/manager education is
university, times:
% raw material imported 0.0010 (0.43) 16.5
% output exported –0.0115 (3.20) 4.3
% foreign ownership –0.0023 (0.72) 6.5
R2 .95 .95 .96
Kenya: all firms (obs = 172)
Primary school –0.054 (0.56) –0.034 (0.35) –0.042 (0.42) 0.31
Secondary school –0.005 (0.05) 0.004 (0.03) 0.033 (0.26) 0.15
University 0.016 (0.15) 0.039 (0.34) 0.076 (0.60) 0.24
% raw materials imported 0.0000 (0.02) –0.0008 (0.47) 22.3
% output exported –0.0026 (1.26) 0.0027 (1.09) 6.5
% foreign ownership –0.0020 (1.27) –0.0021 (0.84) 8.2
university, times:
% output exported –0.0123 (3.33) 3.1
% foreign ownership 0.0011 (0.32) 5.3
2
R .96 .96 .97
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
Zimbabwe: formal firms (obs = 142) Sample means
Secondary school 0.0216 (0.15) .0107 (0.08) –0.041 (0.29) 0.48
University –0.0373 (0.25) –0.0321 (0.21) 0.102 (0.63) 0.42
% output exported 0.0030 (1.46) 0.0033 (1.33) 13.0
% foreign ownership 0.0011 (0.80) 0.0032 (1.62) 14.0
university, times:
% raw material imported –0.0044 (1.50) 8.7
% output exported –0.0025 (0.61) 5.7

R2 .95 .95 .95
Zimbabwe: all firms (obs = 170)
Secondary school 0.0500 (0.44) 0.0376 (0.33) 0.0008 (0.01) 0.46
University –0.0256 (0.20) –0.0204 (0.16) 0.1104 (0.77) 0.37
% output exported 0.0031 (1.48) 0.0034 (1.37) 11.5
university, times:
% raw material imported –0.0043 (1.45) 7.2
% output exported –0.0023 (0.56) 4.8
R2 .96 .96 .97
Table 2.8c Cost functions with skill measures included, Ghana
68
Ghana: large firms (10 or more employees) (obs = 80) Sample means
Primary school 0.043 (0.22) 0.038 (0.19) 0.044 (0.22) 0.39
Secondary school –0.087 (0.42) –0.091 (0.44) –0.135 (0.61) 0.33
University 0.427 (1.71) 0.426 (1.70) 0.340 (1.22) 0.16
% raw materials imported –0.0006 (0.29) –0.0041 (1.29) 20.0
% output exported –0.0023 (0.78) –0.0034 (0.77) 6.7
university, times:
% output exported 0.0022 (0.38) 3.7
R2 .93 .93 .94
Ghana: all firms (obs = 132)
Primary school 0.389 (2.29) 0.386 (2.24) 0.389 (2.23) 0.45
Secondary school 0.216 (1.16) 0.215 (1.15) 0.190 (0.96) 0.27
University 0.526 (2.12) 0.519 (2.07) 0.465 (1.72) 0.11
% raw materials imported –0.0002 (0.09) –0.0019 (0.57) 12.8
% output exported –0.0027 (0.75) –0.0038 (0.72) 4.0
university, times:
% output exported 0.0024 (0.35) 2.3
R2 .92 .92 .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/man-

agers 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 elec-
tricity. 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 prob-
lems’ 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 government-
imposed handicaps affecting them, it would be surprising if their techno-
logical 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 liberaliza-
tion 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 economy-
wide 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 sec-
toral 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 environ-
ment limits these gains. Much of the empirical testing of the role of educa-
tion in adjusting to technical change has been in the agricultural sector,
which is usually much more competitive than the industrial sector, particu-
larly 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 inter-
nal rivalry is reduced by the relatively small number of firms producing a
specified product. The shift in short-run supply curves induced by abnor-
mal profits in competitive markets is not forthcoming, as there are a
limited number of firms capable of entering most industries. Firms possess-
ing technical abilities may choose not to utilise them to increase productiv-
ity 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 forgo-
ing 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 con-
straint on African industrial development is an insufficient supply of tech-
nologically 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 techno-
logical 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 sug-
gests that as the economy becomes open to flows of international knowl-
edge, 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
* University of Pennsylvania and Princeton University. We are grateful to Tyler
Biggs of the Africa Technical Department of the World Bank for making avail-
able 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.
1 Meier and Steel (1989) contains a good selection of representative views.
2 For a general discussion of the technological requirements for industrial devel-
opment see Lall (1990). Pack (1993) provides an overview of industrial develop-
ment in Africa and extensive references to the literature.
3 For studies demonstrating the interaction of education and technological
change in agriculture see Welch (1970), Foster et al. (1995a, 1995b), Rosenzweig
(1995).
4 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 intermedi-
ate 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.
5 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 bever-
age 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.
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 equip-
ment, 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 gen-
erates 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 oper-
ation in 1961, only 101 predated 1945. Including small scale establish-
ments, 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 manufactur-
ing 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 large-
scale 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 mea-
surement 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
Table 3.1 Indicators of Tanzanian manufacturing performance, 1946–89a
Year Number of establishments Persons Value

engaged added
In production Increase c
(thousands) (% of GDP)
(number) (annual average)
1946 101 b
1957 293 b 17
Independence (1961) 380 b 22 22 3.5
1965 569 38 28 9.2
1978 1 282 51 110 12.3
1989 886 –33 124 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 pro-
ductivity 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 manufac-
turing sector as a whole and for major branches of manufacturing
(Sections 2 and 3)
• to present consistent indexes of industrial output and labour productiv-
ity 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 productiv-
ity in Tanzania vis-à-vis the USA in 1989, using the methodology devel-
oped 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 under-
standing 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 pro-
duction 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 com-
mission, 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 sum-
marizes these sources and compares their coverage.
Figure 3.1 Comparison of coverage of statistical sources
IS- IS-
Coverage a b
Census I/O table Census surveyd surveyd
1+
Census
5+
10+ ASIP Ec. surveyc ASIP ASIP
1961 1965–74 1976 1978 1979–88 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 question-
naires 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 large-
scale 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 cover-
age of medium- and large-scale establishments in the directory varies over
time and does not adequately reflect changes in the real volume of activi-
ties. 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 direc-
tory includes 711 establishments; this jumps to 866 in the census year
1989. In some cases establishments, such as furniture making and tailor-
ing, have deliberately been omitted from the annual surveys (omitted
establishments). In most cases, the differences are unintentional (under-
coverage).
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 cover-
age 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 establish-
ments 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 previ-
ous 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 understate-
ment 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 establish-

ments which are not responding for a given number of years (for example,
two successive years), weighted by the price indexes over the period of suc-
cessive non-response.5 In formula, the degree underestimation of value
added in a size class is calculated as follows:
St
tnrt nrt (t − i )  CPI t 
UEt = ∑ ue
i= 1
t ,i and uet ,=
i
Nt tnrt 
 CPI t − i
− 1

(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 produc-
tion 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 calcu-
lated 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 ques-
tionnaire 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 cate-
gory. 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 dona-
tions.
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
60
Intermediate inputs (billion TSh)
50
40
30
20
10
0
Electr. & Water
Transport
Chem. & Pack.
Miscellaneous
Bank charges &
Raw materials
All other costs

Industrial
Professional &
Cost of resales
services
Fuel, Lubr.,
office exp.
materials
insurance
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 adjust-
ment is the 241 per cent increase in value added for ISIC category 32 (tex-
tiles 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
MVAa after adjustments for: Change of MVAa due to adjustment for:

Unadjusted
MVA OC-adj.b NR-adj.c OE-adj.d UC-adj.e Extrapolated OC-adj.b NR-adj.c OE-adjd UC-adj.e Total
(Millions TSh) (%)
1961 243
1962 272
1963 276
1964 350
1965 267 282 307 405 6 9 32 52
1966 295 297 327 441 1 10 35 49
1967 319 323 344 460 1 6 34 45
1968 378 390 402 522 3 3 30 38
1969 475 497 659 5 33 39
1970 561 571 766 2 34 37
1971 643 657 874 2 33 36
1972 806 1 087 35 35
1973 914 1 244 36 36
1974 1 157 1 537 33 33
1975 1 246 1 863 50 50
1976 1 480 2 143 45 45
1977 2 075 2 678 29 29
1978 2 186 2 926 34 34
1978* 2 842 2 842 2 842 2 926 0 0 3 3
1979 2 927 2 927 2 985 3 238 0 2 8 11
1980 2 891 2 900 3 279 3 622 0 13 10 25
1981 3 108 3 148 3 983 4 555 1 27 14 47
1982 3 204 3 333 4 862 5 873 4 46 21 83
1983 3 620 3 994 5 706 6 812 10 43 19 88
1984 4 417 5 269 7 241 8 094 19 37 12 83
81
82
MVAa after adjustments for: Change of MVAa due to adjustment for:

Unadjusted
MVA OC-adj.b NR-adj.c OE-adj.d UC-adj.e Extrapolated OC-adj.b NR-adj.c OE-adjd UC-adj.e Total
(Millions TSh) (%)
1985 5 112 6 373 8 915 9 812 25 40 10 92

1986 6 412 8 525 11 714 12 360 33 37 6 93
1987 11 062 14 545 20 790 21 679 31 43 4 96
1988 11 358 15 866 25 849 26 642 40 63 3 135
1989 21 474 32 443 32 653 32 653 51 1 0 52
1990 23 956 37 478 38 681 38 681 56 3 0 61
1991 48 128
1992 53 473
1993 63 535
1994 82 159
1995 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)
Adjusted Unadjusted
ISIC 31 32 33/34 35 36 37/38/39 3 3

Branch Food, Textiles Wood, Chemicals, Non-metallic (Basic) metals, Total Total
beverages & leather furnitures & petroleum, minerals machinery & manufacturing manufacturing
& tobacco fixtures, rubber & equipment and other
paper, printing plastics manufacturing
& publishing
1961 243 439

1962 271 745
1963 275 520
1964 350 061
1965 131 318 93 178 42 032 25 154 5 916 107 189 404 787 266 701
1966 135 661 144 031 42 794 26 787 11 749 79 507 440 530 295 162
1967 12 627 152 849 45 803 52 882 21 394 65 912 460 466 318 625
1968 180 020 143 557 47 739 53 640 23 361 73 202 521 519 378 324
1969 202 029 193 533 55 176 62 789 28 923 116 294 658 743 475 411
1970 235 746 278 753 70 593 65 486 30 090 84 841 765 510 560 616
1971 286 292 300 707 64 198 81 270 33 817 107 580 873 864 642 871
1972 310 905 368 850 61 144 121 236 49 636 174 880 1 086 650 806 328
1973 334 000 448 992 89 133 131 646 41 022 199 329 1 244 123 914 327
1974 390 272 481 125 122 775 249 184 48 491 245 445 1 537 293 1 156 652
1975 403 249 572 779 150 762 286 029 96 173 354 489 1 863 480 1 245 622
1976 477 801 675 596 177 831 337 360 60 140 414 570 2 143 298 1 480 345
1977 627 467 900 910 213 552 339 244 62 566 534 159 2 677 899 2 074 758
1978 694 642 892 085 287 409 347 488 97 185 606 824 2 925 632 2 842 316
1979 800 919 970 942 323 294 471 689 99 560 571 271 3 237 676 2 927 333
1980 774 318 1 227 974 359 672 558 216 120 741 581 366 3 622 286 2 890 897
1981 1 134 567 1 345 426 494 176 676 847 189 175 714 357 4 554 547 3 108 206
83
84
Adjusted Unadjusted
ISIC 31 32 33/34 35 36 37/38/39 3 3

Branch Food, Textiles Wood, Chemicals, Non-metallic (Basic) metals, Total Total
beverages & leather furnitures & petroleum, minerals machinery & manufacturing manufacturing
& tobacco fixtures, rubber & equipment and other
paper, printing plastics manufacturing
& publishing
1982 2 026 862 1 198 534 499 515 982 171 246 378 919 253 5 872 714 3 203 832
1983 2 067 669 1 812 035 549 270 889 370 326 831 1 167 309 6 812 485 3 619 760
1984 2 582 369 2 022 511 760 200 1 228 673 170 774 1 329 517 8 094 045 4 417 219
1985 3 148 163 2 238 412 870 414 1 398 033 384 734 1 772 483 9 812 238 5 111 606
1986 3 099 117 2 577 458 1 107 314 2 189 489 917 068 2 469 293 12 359 739 6 412 236
1987 5 157 851 5 331 612 1 546 226 4 903 191 1 237 203 3 503 219 21 679 302 11 062 008
1988 7 666 631 6 865 440 2 044 644 4 960 480 1 359 874 3 744 791 26 641 860 11 357 863
1989 12 441 829 5 826 626 3 017 332 4 901 342 1 422 835 5 042 905 32 652 870 21 474 018
1990 17 149 464 5 597 233 3 755 861 5 432 448 1 176 377 5 570 034 38 681 417 23 955 853
1991 21 232 538 6 459 244 4 328 090 6 944 192 2 059 175 7 105 139 48 128 378
1992 25 522 856 7 033 218 4 138 842 7 650 103 1 809 982 7 317 953 53 472 955
1993 30 743 364 7 567 179 7 639 361 8 002 358 2 427 974 7 154 866 63 535 101
1994 43 202 958 8 666 806 6 716 699 10 917 273 2 781 754 9 873 702 82 159 192
1995 53 931 263 12 257 957 8 543 564 12 284 511 3 770 135 8 030 217 98 817 647
Source: Prins and Szirmai (1998).

3.4 Summary of adjustments

Table 3.2 provides a summary of all the adjustments to aggregate 10+ man-
ufacturing 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 rela-
tive 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 cloth-
ing 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 com-
modities), 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 quan-
tity 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) pro-
vides a sophisticated index of industrial production for the years 1965–72.
Utilizing an impressive amount of data derived from the QSIP question-
naires, 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 pro-
duction, 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 consis-
tent long-run series of manufacturing employment is available at branch
level from published statistical sources. We have reconstructed an employ-
ment 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 undercover-
age in the employment series, under the assumption that the labour pro-
ductivity of the covered (or responding) establishments is equal to the
labour productivity of the non-covered (or non-responding) establish-
ments. Thus we can apply the ratio of value adjusted for coverage and non-
response 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 pre-
sents the employment figures for six manufacturing branches and for total
manufacturing. Table 3.6 presents the corresponding labour productivity
indices.
88
Table 3.4 Indexes of industrial production 1965–95
ISIC 31 32 33/34 35 36 37/38/39 3

Branch Food, Textiles Wood, Chemicals, Non-metallic (Basic) metals, Total
beverages & leather furnitures petroleum, minerals machinery & manufacturing
& tobacco & fixtures, rubber & equipment and
paper, printing plastics other manufacturing
& publishing
(1976=100)
1961 22
1962 24
1963 26
1964 28
1965 37 24 100 26 28 24 31
1966 37 30 114 55 57 27 37
1967 41 31 101 57 59 24 38
1968 43 45 116 62 113 33 48
1969 51 47 124 62 118 29 52
1970 51 63 123 65 124 29 57
1971 59 62 145 78 114 48 65
1972 84 80 141 91 130 61 84
1973 86 90 159 88 150 81 93
1974 81 94 142 90 106 91 93
1975 78 92 123 95 115 92 91
1976 100 100 100 100 100 100 100
1977 92 105 151 94 115 112 105
1978 103 97 124 98 105 133 107
1979 90 104 100 80 111 138 103
1980 77 99 108 77 84 130 96
1981 67 90 97 75 107 114 88
ISIC 31 32 33/34 35 36 37/38/39 3

Branch Food, Textiles Wood, Chemicals, Non-metallic (Basic) metals, Total
beverages & leather furnitures petroleum, minerals machinery & manufacturing
& tobacco & fixtures, rubber & equipment and
paper, printing plastics other manufacturing
& publishing
(1976=100)
1982 64 83 91 71 92 132 83
1983 67 62 68 70 68 167 77
1984 62 60 58 120 102 194 88
1985 57 57 62 77 103 137 72
1986 54 58 83 71 121 152 73
1987 48 75 114 70 126 115 71
1988 57 82 111 64 133 102 75
1989 55 76 109 76 136 110 75
1990 62 77 101 84 157 141 84
1991 64 73 90 84 214 133 84
1992 62 65 71 78 154 100 74
1993 65 68 101 72 164 88 75
1994 66 66 72 79 139 125 79
1995 69 58 66 75 153 69 72
Source: Prins and Szirmai (1998), table K-2.
89
90
Table 3.5 Persons engaged in 10+ manufacturing, by branch (1965–90)
ISIC 31 32 33/34 35 36 37/38/39 3

Branch Food, Textiles Wood, furnitures Chemicals, Non-metallic (Basic) metals, Total
beverages & leather & fixtures, paper, petroleum, minerals machinery & manufacturing
& tobacco printing & rubber & equipment and
publishing plastics other manufacturing
1965 10 038 12 102 9 203 955 652 6 204 39 154
1966 9 767 14 504 9 597 2 330 1 744 7 116 45 058
1967 12 666 14 265 8 235 1 767 1 422 5 977 44 332
1968 14 930 19 033 9 596 2 507 2 380 5 402 53 847
1969 13 478 21 770 9 737 1 997 2 622 5 538 55 142
1970 16 340 22 791 8 885 2 208 2 572 5 719 58 515
1971 17 652 24 245 9 635 2 782 3 511 7 582 65 407
1972 21 334 24 795 10 896 4 876 2 866 8 277 73 044
1973 18 880 28 609 11 117 4 769 3 343 9 304 76 022
1974 20 035 33 200 12 219 5 309 3 061 10 316 84 139
1975 22 026 34 140 13 125 5 235 2 942 11 253 88 721
1976 22 333 34 661 13 302 5 304 2 982 11 402 89 984
1977 24 793 38 168 15 308 5 604 3 476 13 947 101 295
1978 28 111 43 926 14 781 7 900 3 773 14 194 112 685
1979 28 455 43 072 14 909 7 480 3 333 12 836 110 084
1980 30 291 45 407 14 637 8 890 3 218 13 566 116 009
1981 32 437 48 670 14 513 9 388 4 247 14 090 123 345
1982 38 041 45 308 14 852 9 772 3 300 14 242 125 515
1983 34 885 49 642 15 793 9 439 3 687 14 546 127 991
1984 30 458 44 545 14 262 8 085 3 614 13 156 114 120
1985 29 138 42 191 13 722 6 447 3 821 12 402 107 721
1986 38 641 43 499 14 452 6 930 5 344 11 303 120 168
1987 39 858 43 426 14 211 7 042 4 517 11 041 120 096
1988 43 200 41 069 15 229 6 741 4 484 10 919 121 642
1989 45 282 38 128 16 930 6 677 5 069 13 793 125 879
1990 47 397 37 674 24 360 6 534 4 237 14 211 134 413
Source: Prins and Szirmai (1998), table M-10.

Table 3.6 Labour productivity index for six branches of manufacturing, 1965–90 (1976 = 100)
ISIC 31 32 33/34 35 36 37/38/39 3

Branch Food, Textiles Wood, furnitures Chemicals, Non-metallic (Basic) metals, Total
beverages & leather & fixtures, paper, petroleum, minerals machinery & manufacturing
& tobacco printing & rubber & equipment and
publishing plastics other manufacturing
1965 82 68 145 146 128 44 71
1966 85 71 158 126 98 43 73
1967 73 75 162 171 123 47 76
1968 65 83 160 131 141 71 81
1969 84 75 169 165 135 60 84
1970 70 95 184 155 144 59 88
1971 75 89 200 148 97 72 90
1972 88 112 173 99 135 84 104
1973 102 109 191 98 134 99 110
1974 90 98 155 90 103 101 100
1975 79 93 124 96 117 93 93
1976 100 100 100 100 100 100 100
1977 83 96 131 89 99 91 93
1978 82 76 112 66 83 106 86
1979 70 84 89 57 99 123 85
1980 57 75 99 46 78 109 74
1981 46 64 89 42 75 92 64
1982 38 64 81 38 83 106 60
1983 43 43 57 39 55 131 54
1984 45 47 54 79 84 168 70
1985 44 46 60 63 81 126 60
1986 31 47 77 54 68 153 55
1987 27 59 107 52 83 119 53
1988 30 70 97 50 88 107 55
1989 27 69 86 60 80 91 53
1990 29 71 55 68 111 113 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 informa-
tion 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 con-
verters are required. In this paper, industry-of-origin unit value ratios
(UVRs) are calculated for branches of manufacturing, using the industry-of-
origin 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 pro-
duction 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 corre-
sponding 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 S
∑Q X
ij ∗ PijX ∑ (Q U
ij ∗ PijX )
UVR XU ( X )
j = i= 1
S
UVR XU (U )
j = i= 1
S
(3.2)
∑
i= 1
Q ijX ∗ PijU ∑
i= 1
(Q Uij ∗ PijU )
where
UVRjXU(X) 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 manu-
facturing branch level by taking the weighted average of sample industry
UVRs using 1989 sample industry gross output as weights:
∑ [GO
j= 1
U (U )
j ∗ UVR XU
j ]
(U )
UVR XU (U )
k = O
(3.3)
∑ GO
j= 1
U (U )
j
∑ GO
j= 1
X (X )
j
UVR XU ( X )
=
∑ [GO ]
k O
X (X )
j / UVR XU
j
(X )
j= 1
where
GOjU(U)is gross output in US sample industry j in dollars
GOjX(X)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 manu-
facturing – the UVRs can be used to convert 1989 value added into the
currency of the other country for purposes of real value added compar-
isons. In binary comparisons one gets two UVRs at every level of aggre-
gation, 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 man-
ufacturing. For these branches we used the calculated UVRs for total manu-
facturing 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 Data sources for the 1989 benchmark comparison

5.2.1 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 pro-
ducing 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, espe-
cially 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 ser-
vices, 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 prod-
ucts. For most, though not all, products the US census provides both
quantity and value information for 1987. First, unit value ratios were com-
puted 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
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 102 45 68 0.5

Beverages 69 81 75 0.5
Tobacco products 51 51 51 0.4
Textile mill products 105 110 107 0.7
Wearing apparel 42 28 35 0.2
Leather products & footwear 36 20 27 0.2
Wood products, furniture & 90 86 88 0.6
fixtures
Paper products, printing & 160 139 149 1.0
publishing
Chemical products (incl. oil) 470 110 227 1.6
Rubber & plastic products 333 325 329 2.3
Non-metallic mineral products 90 143 114 0.8
Basic & fabricated metal productsa 177 78 117 0.8
Machinery & transport equipmenta 177 78 117 0.8
Electrical machinery & equipmenta 177 78 117 0.8
Other manufacturing industriesa 177 78 117 0.8
Total manufacturing, census 177 78 117 0.8
weightsb
Total manufacturing, 198 73 120 0.8
implicit UVRsc
Exchange rate 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 low-
quality 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 low-
income 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 pro-
ductivity in Tanzanian manufacturing in 1989 is 3.7 per cent of that in US
manufacturing. There is substantial variance in branch productivity perfor-
mance, 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 oper-
ating 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 list-
ings 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 devel-
oping 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 At US prices
prices (in TSh) (in US$) Geometric
average,
Tanzania USA Tanzania/ Tanzania USA Tanzania/ Tanzania/
USA (%) USA (%) USA (%)
Food manufacturing 203 370 8 387 897 2.4 4 511 82 572 5.5 3.6
Beverages 816 907 10 413 211 7.8 10 027 150 548 6.7 7.2
Tobacco products 712 352 15 691 221 4.5 13 871 305 552 4.5 4.5
Textile mill products 205 282 4 229 796 4.9 1 871 40 281 4.6 4.7
Wearing apparel 48 669 1 306 552 3.7 1 734 30 798 5.6 4.6
Leather products & footwear 120 675 1 305 577 9.2 6 111 36 648 16.7 12.4
Wood products, furniture & 150 251 3 828 677 3.9 1 754 42 331 4.1 4.0
fixtures
Paper products, printing 321 831 11 517 612 2.8 2 318 72 141 3.2 3.0
& publishing
Chemical products 845 705 66 647 074 1.3 7 711 141 811 5.4 2.6
Rubber & plastic products 726 943 17 537 943 4.1 2 234 52 669 4.2 4.2
Non-metallic mineral products 390 347 5 703 193 6.8 2 737 63 071 4.3 5.4
Basic & fabricated metal products 470 994 10 680 022 4.4 6 072 60 344 10.1 6.7
Machinery & transport equipment 419 871 12 819 273 3.3 5 413 72 431 7.5 4.9
Electrical machinery & equipment 778 159 11 234 739 6.9 10 032 63 478 15.8 10.5
Other manufacturing industries 267 095 12 307 000 2.2 3 443 69 537 5.0 3.3
Total manufacturing 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 1978 1989
ISIC Branch Unadj. Adj. Unadj. Adj. Unadj. Adj.
31 Food, beverages & tobacco 42 31 24 24 42 38

32 Textiles & leather 25 33 30 30 8 18
33–4 Wood, furniture & fixtures, paper, 12 10 10 10 11 9
printing & publ.
35 Chemicals, petroleum, rubber & 8 6 12 12 17 15
plastic products
36 Non-metallic mineral products 3 3 3 3 5 4
37–9 (Basic) metal products, mach. & 10 18 21 21 17 15
equipm. & Other man.
3 Total manufacturing 100 100 100 100 100 100
Source: Prins and Szirmai (1998), table G-2.

Figure 3.3 Unadjusted and adjusted nominal value added for 10+ manufacturing,
1965–90
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
0
1972
1973
1974
1975
1976
1977
1978
1965
1966
1967
1968
1969
1970
1971
Unadjusted and adjusted nominal value added,1978–1990

45 000 000
Nominal value added in 1000 TSh
40 000 000
Unadjusted
35 000 000 Adjusted
30 000 000
25 000 000
20 000 000
15 000 000
10 000 000
5 000 000
0
1981
1984
1986
1978
1983
1985
1988
1989
1990
1979
1989
1982
1987
Source: Table 3.3.

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 com-
pared 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 col-
lapse 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 repre-
sents 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
120
100
Index 1985 = 100
80
RMVA series BoS 1995b
RMVA series QSIP
60 New IIP
40
20
1987
1989
1991
1993
1995
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
Sources: see text.

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
45 000
NA&QSIP
Million constant 1989 TSh
New IIP
40 000
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
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 pro-
duction. 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 tex-
tiles and leather more or less correspond to the pattern for total manufac-
turing, 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 manufac-
turing. 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
100
Index, 1976 = 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 pro-
ductivity in Tanzanian manufacturing, but the absolute decline evens out
Figures 3.7 Comparative labour productivity in Tanzanian manufacturing, 1965–90

(USA = 100)
12
11
10
9
8
Tanzania/USA (%)
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 con-
stant 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.
Appendix

Table A.1 Basic data on output and employment for Tanzania, 1989 (establishments with 10 or more persons engaged)
Number of Gross value Gross value Gross value Gross value Employment Gross value
establishments of output at added at added at factor added in (person) added per
factor cost factor cost cost, US census branch as % person
(mill. TSh) (mill. TSh) concepta of total employed
(mill. TSh) TSh
1 Food manufacturing (311/12) 146 24 531.6 5 630.9 7 177 18.24 35291 203 370
2 Beverages (313) 19 8 959.6 3 003.1 4 116 10.46 5039 816 907
3 Tobacco products (314) 3 6 237.2 3 359.7 3 528 8.97 4952 712 352
4 Textile mill products (321) 83 17 701.0 5 256.4 6 367 16.18 31014 205 282
5 Wearing apparel (322) 54 519.5 105.8 136 0.35 2803 48 669
6 Leather products & 21 1 883.2 401.1 520 1.32 4311 120 675
footwear (323/324)
7 Wood products, furniture & 214 4 399.7 1 115.4 1 454 3.69 9674 150 251
fixtures (331/2)
8 Paper products, printing & 62 7 411.8 1 865.7 2 335 5.94 7256 321 831
publishing (341/2)
9 Chemicals products (351–3) 51 11 680.3 3 515.8 4 201 10.68 4967 845 705
10 Rubber and plastic 15 4 900.7 954.0 1 243 3.16 1710 726 943
products (355/6)
11 Non-metallic mineral 22 5 608.1 1 413.7 1 979 5.03 5069 390 347
products (361–9)
12 Basic & fabricated metal 94 13 583.2 2 231.6 2 882 7.32 6118 470 994
products (371–81)
13 Machinery & transport 72 8 062.1 1 768.3 2 234 5.68 5320 419 871
equipment (382/4)
Table A.1 (continued)
Number of Gross value Gross value Gross value Gross value Employment Gross value
establishments of output at added at added at factor added in (person) added per
factor cost factor cost cost, US census branch as % person employed
(mill. TSh) (mill. TSh) concepta of total TSh
(mill. TSh)
14 Electrical machinery & 6 2 303.4 710.0 830 2.11 1067 778 159
equipment (383)

15 Other manufacturing 24 1 126.4 278.9 344 0.87 1288 267 095
Industries (385–90)
Total manufacturing 886 118 907.7 31 610 39 345 100.00 125 879 312 562
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.
Table A.2 Basic data on output and employment in manufacturing, USA, 1989 (establishments with 10 or more persons
engaged)a
Annual Survey of Manufactures

Gross value Gross value Gross value Employmentb GVA/person
of output added added in (1 000s) (US$)
(mill. US$) (mill. US$) branch as
% of total
Food manufacturing 310 109.7 106 053.8 8.31 1 284.4 82 572.0

Beverages 49 695.9 24 103.4 1.89 160.1 150 547.9
Tobacco products 25 789.9 18 916.2 1.48 61.9 305 552.2
Textile mill products 67 072.6 27 123.1 2.12 673.4 40 280.7
Wearing apparel 61 451.2 31 361.7 2.46 1 018.3 30 798.2
Leather products & 9 757.4 4 543.6 0.36 124.0 36 648.4
footwear
Wood products, furniture 108 550.5 47 791.0 3.74 1 129.0 42 331.0
& fixtures
Paper products, printing 271 117.7 152 708.2 11.96 2 116.8 72 140.8
& publishing
Chemicals, incl. petrol. 415 580.0 168 705.0 13.21 1 189.6 141 811.3
refining
Rubber & plastic 96 725.5 46 850.0 3.67 889.5 52 668.9
products
Non-metallic mineral 60 938.3 32 895.4 2.58 521.6 63 071.2
products
Basic & fabricated metal 308 697.4 132 557.2 10.38 2 196.7 60 344.0
products
Machinery & transport 618 891.6 282 524.5 22.13 3 900.6 72 431.1
equipment
Table A.2 Basic data on output and employment in manufacturing, USA, 1989 (establishments with 10 or more persons
engaged)a (continued)
Annual Survey of Manufactures

Gross value Gross value Gross value Employmentb GVA/person
of output added added in (1 000s) (US$)
(mill. US$) (mill. US$) branch as
% of total
Electrical machinery 190 906.6 105 044.9 8.23 1 654.8 63 478.2

& equipment
Other manufacturing 150 519.0 95 611.5 7.49 1 375.0 69 536.7
industries
Total manufacturing 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.
Table A.3 Number of UVRs, coverage rates and reliability

Number of Coverage, USA Coverage, Reliability Reliability
UVRs Tanzania PPP at US PPP at Tanzanian
quantity weights quantity weights
Food manufacturing 22 20.3 46.6 0.10 0.27

Dairy products 4 38.6 39.0 0.1 0.0
Preserved fruits & vegetables 4 31.1 3.2 0.5 1.4
Fats and oils 4 51.5 53.4 0.1 0.2
Grain mill products 4 59.2 80.3 0.1 0.1
Bakery products 1 34.4 16.3 0.0 0.0
Sugar & confectionary, food 5 41.0 34.1 0.2 0.7
not elsewhere specified
Beverages 2 28.0 15.7 0.18 0.79
Malt and malt beverages 2 89.5 37.1 0.1 0.7
Tobacco products 1 8.6 21.4 0.00 0.00
Tobacco stemming and redrying 1 86.5 42.1 – –
Textile mill products 6 15.4 39.5 0.09 0.08
Textile mill products 6 48.4 45.9 0.1 0.1
Wearing apparel 7 20.3 17.3 0.26 1.20
Wearing apparel 7 61.0 19.1 0.2 1.2
Leather products and footwear 2 39.9 2.1 0.43 0.11
Leather footwear 2 90.5 7.8 0.0 0.1
Wood products, furniture & 13 27.6 41.8 0.13 0.13
fixtures
Wood products & furniture 13 57.7 41.9 0.1 0.1
Paper products, printing & 9 11.6 40.0 0.16 0.15
publishing
Paper, printing & publishing 9 24.1 40.2 0.1 0.2
Table A.3 (continued)
Number of Coverage, USA Coverage, Reliability Reliability

UVRs Tanzania PPP at US PPP at Tanzanian
quantity weights quantity weights
Chemicals, incl. petrol. refining 9 3.4 19.4 0.67 0.86

Chemical products 9 44.1 38.1 0.5 0.8
Rubber & plastic products 2 6.1 24.1 0.00 0.02
Rubber tyres & tubes 2 50.0 69.2 0.0 0.0
Non-metallic mineral products 3 6.5 38.3 0.41 0.01
Cement & bricks 3 67.1 109.0 0.2 0.0
Basic & fabricated metal products
Machinery & transport equipment
Electrical machinery & equipment
Other manufacturing industries
Total manufacturing 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
* 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.
1 Unless indicated otherwise, GDP always refers to the factor cost concept.
2 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.
3 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.
4 Note that this information is not available for the period 1978–90 for the full
population of establishments in the census.
5 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.
6 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.
7 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.
8 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 indus-
try level. The CPI and the PPI are spliced in 1992 to construct a series for
1990–95.
9 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 1989-
based intra-industry weights.
10 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 under-
lying UVR for the category. Reliability is measured as variation of unit value
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 calcu-
lated 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 manu-
facturing.
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 devel-
oping 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 export-
promoting 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 latter-
day 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 differ-
ent 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 technologi-
cal 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 atten-
tion. 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 sub-
Saharan 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 under-
stand 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 conclu-
sions 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 long-
run 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 econ-
omic 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 pre-
vailing 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 accu-
mulation 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 move-
ments 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 accumula-
tion in an open economy and apply their framework to the (early) develop-
ment 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 acceler-
ated. 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 sub-
stitution phase it was expected to move rapidly to the next turning point,
called by Ranis and Fei the ‘commercialization point’. At the commercial-
ization 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 produc-
tivities? 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 com-
mercialisation point, the idea is that the institutionally determined low real
wage will rule. Once labour is fully absorbed, that is, once the commercial-
ization point is past, the real wage will naturally rise. In the first, ‘pre-
commercialisation’ 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 pro-
ductivity 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 fol-
lowed by Malaysia also seems similar to the conventional anticipation.
Over the whole period 1970–90 Malaysia had a low growth of productiv-
ity (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 avail-
able.) This Malaysian pattern is very close to the expectation that tech-
nology 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 labour-
intensive 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 deter-
mined 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 impor-
tant ‘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 unex-
plained, 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 sub-
stitution 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 interven-
tions, which certainly appear to have been positively influential in industrial-
ization; 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 ‘pre-
dictions’ 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 tech-
nological 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 evi-
dence 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 charac-
terized 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 assump-
tion 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 tech-
nology. 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 gener-
ated 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 explana-
tion 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 produc-
tion that make it possible to adopt new technologies. So, if the technolo-
gies 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 compara-
tively 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 necessar-
ily 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 phenom-
ena 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 some-
thing out.
Essentially the Fei–Ranis framework says very little about the implications
of technological change for the process of entry into world markets in manu-
factures. 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 frame-
work 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 pro-
ductivities 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 manu-
factured 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 inten-
sive 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 labour-
intensive 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 technolog-
ical 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 gener-
alization 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 appli-
cation 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 techno-
logical 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 them-

selves. Others will try to exploit the advantages of being a follower, by imi-
tating 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 con-
tinue to use the old methods of production – if there is a process innova-
tion. 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 innova-
tive 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 industriali-
sation, such as textiles and garments production, for example.
It does not follow that firms in developing countries have to become inno-
vators in order to compete; nor do they necessarily have to adopt new tech-
nologies 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 competi-
tiveness 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 pro-
nounced 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 coun-
tries 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 develop-
ments 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 differ-
ences in factor productivities arising from differential learning, or from dif-
ferences 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 deter-
mining 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 productiv-
ity 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 ‘auto-
matic’ in the Arrow tradition (Arrow, 1962) or the result of conscious manage-
rial 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 productiv-
ity 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 depen-
dent. 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 accu-
mulate ‘spontaneously’ through the experience of production, for the most
part the accumulation of technological capabilities depends on the alloca-
tion 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 enter-
prises 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 con-
cerned 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 impor-
tance 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 institu-
tions – 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 long-
term 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, espe-
cially 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 inno-
vation 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 manu-
factured 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 import-
ance 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 accom-
plishments 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 techno-
logical 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 techno-
logical 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 discus-
sion is whether an initial entry into international markets through labour-
intensive 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 partici-
pants 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-
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.
3. 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.
4. 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.
5. Although for these countries the matter is somewhat different since – as their
export data indicate – they have not really entered the ‘export substitution’
phase.
6. The data available are too weak to support tests of causality.
7. The terminology ‘generic technological change’ is used very frequently in a
certain part of the ‘technology’ literature, often in an unclear way. The dictio-
nary 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 pro-
duction, but are widely relevant to many. They are technologically and not
industrially defined.
8. 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.
9. 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.
10. 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.
11. 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.
12. The recognition that learning processes involve important resources has a sub-
stantial 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 develop-
ing 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
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 behav-
iour 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 dom-
inated 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 technologi-
cal 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 consider-
able emphasis on small-scale, labour- and local-input-intensive technolo-
gies, 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 tech-
nological 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 declara-
tion, 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 industrializa-
tion strategy in the post-Arusha period (see Chapter 1). The second five-
year plan, covering the years 1969 to 1974, for example, not only provided
explicit encouragement of labour-intensive techniques, but also empha-
sized 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 employ-
ment 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 small-
scale (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 alterna-

tives usually presented themselves.
Table 5.1, for example, shows that the available techniques for rice-
milling 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 invest-
ment 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 tech-
niques: 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 cor-
poration chose instead to expand its capacity on the basis of the large-
scale (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 aware-
ness 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 Total capital TSh Benefit–cost ratio

cost (TSh 000) per worker hour (shadow prices,
full capacity)
Rice huller
4–8 tons per day 93.0 2.4 7.94a
Rice roller
24 tons per day 1 752.5 14.3 2.52
60 tons per day 4 410.0 11.9 1.69
120 tons per day 7 927.0 16.7 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 techno-
logical 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 tech-
nological behaviour from those that have just been described. And, espe-
cially 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 enter-
prises, 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 techno-
logical 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 incon-
sistent – 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 develop-
ing countries. It is often argued in this literature, for example, that techno-
logical differences between firms reflect differences in the types of products
they manufacture (where differences refer to variations in the characteris-
tics 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 expla-
nations in the choice of technology, namely, those that impute various
types of non-economic preferences to decision makers in developing coun-
tries. Perhaps the best known of these explanations is the ‘engineering-
man’ hypothesis advanced by Louis Wells (1975). His contention, in brief,
is that under conditions of imperfect competition the preference of ‘engi-
neering-man’ for sophisticated automated technology and modern prod-
ucts 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 lit-
erature, such as, for example, the sense of ‘national pride’ which is some-
times 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 low-
cost 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 appreci-
ably lower amount of capital. In circumstances where labour is compara-
tively 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 supply-
ing 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 labour-
Table 5.2 A comparison of two textile plants
Friendship Mwanza
Capital cost up to 1969 (million shillings) 61.5 106.5

Production of woven fabrics in 1975 (million linear metres) 24.0 22.5
Number of employees in 1975 5057 2486
Profit in 1975 (million shillings) 2.8 2.3
Cost of carded cotton in 1973 (shillings per tonne) 2512 2910
Labour hours per tonne of carded cotton in 1973 247 98
Source: Coulson (1982b).

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 sup-
plier’s credit rather than by bilateral foreign aid. The most important differ-
ence, 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 techno-
logical 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 sophisti-
cated, ‘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 man-
agers seek to maximize external project finance – with a historically
oriented view of how this behaviour leads to particular technological out-

comes, 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 maxi-
mizing 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 par-
ticular 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 empha-
sizing in part because most of the gains that accrue to the budget maximiz-
ing 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 inter-
ests of its own’ and with considerable opportunities for ‘personal aggran-
disement’, 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 econ-
omic 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 maximisa-
tion 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 extraor-
dinarily 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 conven-
tional government methods’ (ibid.) It was already all too plain, for
example, that the parastatals were behaving less like vehicles in the transi-
tion to socialism and more like ‘bastions of capitalism’, with only a very
tenuous connection to the political institutions of the state. What particu-
larly concerned the ‘political stratum’ was the National Development
Corporation (NDC), which accounted for no less than half the total invest-
ment 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 par-
ticular. 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 Public sector employment in Share of public sector in Share of public sector in
in manufacturing manufacturing manufacturing employment manufacturing value
(million shillings) (absolute numbers) (per cent) added (per cent)
1966 21a 2 330 5

1967 5 300 15.5 14
1971 971 20 113 46.4 29
1979 2 851 53 000 50.0 31
1980 50.0 37.1
1981 3 278.4 47.7 48.2
1982 52.7 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 frag-
mentation 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 descrip-
tion 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 man-
agement). Size showed ‘umuhimu’ or importance relevance; c) strength-
ened 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).
Table 5.4 The growth of the public sector in general
Established posts in the Total number of parastatals

civil service
1966 65 708 43
1967 80 239 73
1971 99 564
1980 295 342 380
Sources: (col. 1) Mukandala (1985), (col. 2) de Valk (1992).
From the perspective of a sector engaged in a struggle for its very sur-
vival, 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 notwith-
standing. … This in turn led to a faster increase in the importation of
capital rather than intermediate goods and over-installation of new indus-
trial capacity’ (Mukandala, 1988, pp. 128–9). Tanzania’s increasing reliance
on imported capital goods over this period was to a large extent made pos-
sible 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
Table 5.5 The growth of external sources of finance
Total aid Total aid as percentage Share of foreign

(million US dollars) of monetary gross financing in government
investment expenditure (%)
1960s 30–35
1970 29.9 11.4 40
1975 114.9 25.3 45
1980 493.6 46.6
1982 514.9 54.3 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 intro-
duced 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 Share of state-owned Share of state-owned Share of state-owned

enterprises in GDP enterprises in enterprise investment enterprises in
non-agricultural in gross domestic employment
economic activity investment
1978–85 10.8 18.1 24.6 22.3

1986–91 13.7 23.8 30.0 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 invest-
ment 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 assump-
tions about bureaucratic behaviour in that country: the first is that bureau-
crats 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 them-
selves significantly and favourably from the pre-Arusha companies’
(Clark, 1978, p. 140), and concluded that ‘There has been as yet no developmen-
tal 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-inten-
sive technologies and ‘Despite the rhetoric, Tanzania’s industrialization pro-
gramme 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 under-
gone 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 institu-
tions 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 inno-
vation 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 technol-
ogy rather than indigenous technological change. It is upon this under-
standing 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 institu-
tions can make important contributions to importing, modifying and dif-
fusing 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 con-
straints 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 com-
petition, growth and development. Competition by technology is an
important force for a firm and an industry in the context of both industri-
ally 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 techno-
logical 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 produc-
tion 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 ‘demand-
pull’. 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 ini-
tially 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 activ-
ity. Firms must market their products in order to survive.
In the pursuit of innovation, firms need to lay down consistent strate-
gies 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 innova-
tion strategy is aimed at achieving technical and market leadership by
being ahead of competitors in the introduction of new products and invest-
ing 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 rou-
tines 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 possibi-
lities 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 learn-
ing 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 per-
formance was limited.
Evidence from various ESRF surveys of industrial performance 5 has
shown that innovation-relevant types of linkages, such as inter-firm link-
ages, 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 innova-
tion. Major sources are established business channels, personal contacts, in-
house 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 man-
agement 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 initi-

ated 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 differ-
ent 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 well-
planned 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 pre-
dominantly 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 fre-
quently 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 perfor-
mance 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 suppli-
ers 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 liberali-
sation. 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 export-
ing firms to succeed in the changing technological and market conditions.
Exporting firms maintained and improved their market position by invest-
ing in technology and continuing to improve on it. Improvements were
made not only in the firms’ products but also in the processes of produc-
tion, 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 manage-
ment 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 vari-
eties 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 hin-
drance. 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 behav-
iour 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 turbu-
lence, institutional framework and linkages, access to technical informa-
tion, and the legal and regulatory framework.
4.1 Human resource development

The various ESRF manufacturing studies show that the educational back-
ground 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 improve-
ments. 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 indi-
cated 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 ser-
vices. 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 provid-
ing owners with knowledge of current operations. Training obtained in a
successful innovative firm will make management responsible for innova-
tion 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 preoccu-
pied 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 high-
growth 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 short-
age 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 con-
sultants. 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 manage-
rial fields.
4.2 Research and development efforts

Several empirical studies on innovation suggest that ‘effective coordina-
tion’ 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 commer-
cialization of R&D initiatives. The basic infrastructure for R&D is underde-
veloped, 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 institu-
tions/centres. In addition, firms have a tendency to invest in short-term
profit-intensive plans, and shy away from long-term investment in innova-
tiveness 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 prob-
lems 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 small-
scale firms, 12 per cent of medium-scale firms and 16 per cent of large-
scale 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 competi-
tion 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 pro-
found 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 inten-
sity 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 compet-
itive products helped them to be innovative and embark on the export
market. The current ‘opening up’ market policies have to an extent con-
tributed 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 exam-
ined 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 vari-
ables in this respect. Lack of adequate financial resources has negatively
affected the extent of technological upgrading in firms. Firms in the engi-

neering subsector, for instance, show that while there has been some
upgrading of production technology following liberalisation, such upgrad-
ing 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 gov-
ernment 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 innova-
tion) and firms (users and developers of innovation) has been eroded even
more by the weak interaction between the private and public sector institu-
tions in Tanzania.
Governments may support technological advance by providing incen-
tives and support for companies to upgrade their technology. The ‘wait and
take’ attitude towards technology acquisitions is common among enter-
prises 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 con-
ventional 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 informa-
tion show that domestic and foreign consultants were important sources
of technological information.
In the specific case of product innovation, the sources of product tech-
nology 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 tech-
nologies 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 industrializa-
tion 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 percent-
age 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 develop-
ment (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 inno-
vators.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 indus-
trial 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 man-
agement 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 govern-
ment 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 non-
human resources for the development of innovative capabilities of the
firms.
The other area for government intervention is the provision of an envi-
ronment 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 achieve-

ments 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 technolog-
ical upgrading, introduction of new products, adaptations of the trans-
ferred 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 frame-
work 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 inno-
vative 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 innovative-
ness 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 indus-
trial 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 theo-
retical framework (for example, innovative activity matrix). Recent industrial
case- and macro-based studies show that both private firms and public innova-
tion supporting institutions operate within an unfavourable economic environ-
ment (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, expa-
triates and imported technology remain the major sources of acquisition of tech-
nology 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 manufactur-
ing sector include the following:
• the ESRF-PSRC (1996) study of the impact of divestiture on industrial perfor-
mance (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 manufac-
turing 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 pre-
sents results of a survey of 74 firms from three sub-sectors, food, construc-
tion, and metal and light engineering, held in 1997.
• Regional Program on Enterprise Development (RPED, 1994); Development and

Growth of Industrial Enterprise in Tanzania, focusing on textiles, wood prod-
ucts, 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 system-
atic 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 high-
quality 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 man-
agement 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 cre-
ation 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 institu-
tions 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
other institutions of higher learning undertake R&D activities, besides their prin-
cipal teaching function.
13. Innovative efforts may be rewarded with promotion, recognition and productiv-
ity bonuses. For example, rewards for innovative efforts are similar to and some-
times 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/depart-
ment 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 con-
siderable technological effort and investments in the development of tech-
nological 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 Develop-
ment 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 devel-
oping 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 govern-
ment 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, develop-
ment 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 person-
nel, development and adaptation of technologies, servicing domestic
industries, monitoring technological developments, coordinating the ex-
ecution 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 prod-
ucts 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 uncer-
tainty 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 develop-
ment 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 Educational levels of TIRDO staff in 1995a
Department High High middle Low middle Low Unknown Total
Industrial technology department 11 5 4 6 26

Engineering department 7 57 4 23
Information department 2 3 3 8
DG/DRD/DAF officeb 3 2 2 3 10
Administration 3 5 23 1 32
Accounts 1 2 1 2 6
Estates 5 9 14
Dispensary 2 2 2 6
Other 2 1
Total 26 19 29 51 3 128
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 implemen-
tation 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 causa-
tion, 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 technol-
ogy needs and specifications, and is the customer for R&D outputs.
Technological change only occurs if these different aspects are coordi-
nated. 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 assess-
ment 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 insti-
tute 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).
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 dif-
ferences 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 ques-
tion whether the research and development organization has the capabili-
ties 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, techni-
cal 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 tech-
nology 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, con-
cerning the production of goods and services; and (5) flows through com-
mercial 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 compo-
nents and the separate purchase of every component, if possible from dif-
ferent suppliers (United Nations Conference on Trade and Development,
1978, pp. 11–13; van Egmond, 1995, p. 20). This strengthens the bargain-
ing 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 differ-
ences 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 technol-
ogy 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 tech-
nologies 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 technolog-
ical, organizational, information processing, managerial and educational
capabilities necessary to incorporate a new technology into the organiz-
ation (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 impor-
tant in this phase: the knowledge gap and the organization of the knowl-
edge 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 man-
agers, 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 organiza-
tion 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 exe-

cuted 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 indus-
try. 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.
• Activated carbon project: a research project into the production of acti-

vated carbon using local waste materials, used within the food and bev-
erages 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 pro-
duced 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 car-
bonate from the caustic soda pilot plant. The project started in 1990,
and was still going on in 1996. Goals for the project are the identifica-
tion 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 assess-
ment 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 identifica-
tion 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 informa-
tion 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 Times mentioned
National shortages 2
Substitution of imports 9
Availability of raw materials 10
Contacts with industry 3
Follow-up research 3
Pressure to develop proposal 4
Note: n = 12.
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 con-
siderably because of the lack of standards for the writing up of projects. In
some cases the principal researcher involves other members of the organ-
ization, 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 tech-
nologies 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, produc-
tion and testing are included.
With regard to the appropriateness of the project, two aspects are impor-
tant: the technology’s desirability for solving certain problems, and its suit-
ability 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 some-
times 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, acqui-
sition 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 infor-
mation department and public libraries. Foreign and domestic R&D insti-
tutes 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 non-
commercial nature.7 Regarding the technology flow mechanism, free flow is
by far the most important. Product-embodied flow associated with the pur-
chase 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 Times used Flow mechanismb Relative

importance (%)c
Information department 9 Free flow 71

Libraries 6 Sponsored flow 14
Foreign companies 1 Flow accompanied
Local companies 2 with products 10
Foreign R&D institutes 3 Commercial flow 4
Local R&D institutes 2
Other 1
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 equip-
ment, 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 effec-
tively, using an unpacking strategy. The technologies are acquired from dif-
ferent 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 en-
vironment (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 orga-
nization. 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 adap-
tive work is very low. In two-thirds of the cases, actors from the environ-
ment (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 technol-
ogy 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 6 3 Yes No No

Aluminium sulphate 5 3 Yes No No
Castor oil 5 1 Yes No No
Caustic soda 8 4 No Yes Yes
Natural dyes 6 2 No No Yes
Pectin 4 2 Yes No Yes
Refractories 5 3 Yes No No
School chalk 7 4 No Yes Yes
Solar systems 6 4 Yes No Yes
Turkey red oil 6 1 Yes No No
Wood adhesives 9 5 No No No
Average 6.1 2.9
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 ade-
quate technologies should and will sell themselves.
Only a few projects reach the phase where TIRDO searches for entrepre-
neurs 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 adhe-
sives and solar systems projects). In three other projects (aluminium sul-
phate, 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 innova-
tion. Table 7.5 summarizes the different methods used. To create interest in
the projects, TIRDO uses a combination of general and selective communica-
tion 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 Times used Method Times used
Publication 2 Free consultancy 2

Radio & TV 2 Technology ownership 1
Trade fairs 3 Support in loan securing 2
Workshops 3 Financial stimuli 1
Direct communication Private guidance 2
through extension
officer 5 Tailor-made (re)design 4
Direct mailing 3 Result insurance 2
Other 2 Pilot plant implementation 2
Note: n = 8.
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 entrepre-
neurs 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 approach-
ing 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 trans-
fer 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 prerequi-
sites 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 develop-
ment 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 innova-
tion.11 The involvement of networks in TIRDO’s projects turns out to be
low. Direct influence of domestic industry on project content and on activ-
ities of project teams is lacking. Though there are some contacts with
industrial firms in most of the projects investigated, the scope of these con-
tacts 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 pro-
jects 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 Assessment of Appropriateness Acquisition Adaptation Activities for Success of

existing needs (I) (II) (III) (IV) diffusion (V) project (VI)
Aluminium sulphate 3 6 8 4 1.5 2

Caustic soda 2.5 6 6 4 10 7
Natural dyes 2 5 7 6 1.5 6
Castor oil 1.5 1.5 5 3 – 1
Activated carbon 3 5 7 5 – 1
Pectin 2 6 6 3 1.5 2
Refractories 3 4 5 4 1 3
School chalk 2 7.5 7 6 8.5 8
Wood adhesives 3.5 7.5 8 7 6 7
Solar systems 4 7.5 7 5 1.5 7
Satellite receivers 1.5 5 8 – – 1
Turkey red oil 7.5 4 6 4 – 2
Averagea 3 5.4 6.7 4.5 3.9 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 dif-
ferent 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 concen-
trated within these two phases. Both phases refer to the interactions
between the institute and its environment and indicate that this interac-
tion does not taken place as often and as intensively as required.12
Inspection of the individual projects indicates that their scores in the dif-
ferent 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 technol-

ogy 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 orig-
inally 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 tech-
nology diffusion is also caused by insufficient efforts to align TIRDO’s activi-
ties 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 tech-
nology is performed reasonably well. Technologies involved in the technol-
ogy 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 imple-
mented 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 adapta-
tion of technologies in TIRDO has contributed to the building up of tech-
nological 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 technol-
ogy 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 per-
suaded 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 develop-
ment institutes should regard the marketing of technologies as an essential
element of the research and development effort. An effective R&D organ-
ization 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 insti-
tutes 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 com-
pleted 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 assess-
ment – 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 men-
tioned 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 appro-
priateness (scale 1–5, after rescaling of average scores to a 1–5 range).
III: Careful acquisition of technology (scale 1–10)

• 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 technol-
ogy 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 expecta-
tions – 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)

• 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: adop-
tion 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.
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 acqui-
sition 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 capa-
bility’. Like technology, technological capability is a complex concept com-
prising 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 inves-
tigations regarding the technological capabilities in the dwelling construc-
tion 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 denom-
inators 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 abili-
ties 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 tech-
nological 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 tech-

nology 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 rec-
ommended that nations should dedicate their efforts to the development
of a creative technology system in order to decrease the technology depen-
dency 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 Technological capabilities: a review of the literature
3.1 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 effec-
tively carry out processes of technology utilisation and technology devel-
opment. 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 particu-
lar 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 capa-
bilities is the one given by Bell et al. (1984). They refer to (1) the stock of
disembodied technical knowledge and information, (2) the human embod-
ied 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 develop-
ment, however, is certainly not the only indicator of technological capabil-
ities. 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 exis-
tence of a number of levels of technological capabilities in countries to gen-
erate technological innovations suitable to their needs, and made an
attempt to indicate the existence of a direct relationship between techno-
logical 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 techno-
logical 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 capabil-
ity 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 accu-
mulation 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, improve-

ment 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 insti-
tutions 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 capabili-
ties 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 tech-
nological capabilities. The complexity of the concept, including the multi-
dimensional aspects incorporated in activities of utilisation and
development of technology, requires a multidisciplinary approach, an
approach that makes the assessment of technological capabilities a cumber-
some 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 per-
formance 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 consen-
sus 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 pro-
duced by the sector. Process technologies refer to the features of the pro-
duction-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 technol-
ogy 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 rele-
vant 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 develop-
ment, education and documentation facilities, technology and interme-
diate 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 Dwelling construction industry in Tanzania
5.1 Economic importance of the construction industry

The construction industry has a crucial role to play in a country’s social-
economic 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 per-
formance 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 situa-
tion, 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 pre-
vents 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 consid-
eration the estimated average annual gross output of the informal construc-
tion 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 employ-
ment 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 produc-
tive 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 con-
tribution 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 for-
mally 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
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993
Average annual GDP growth (%) –2.4 3.4 2.6 3.3 5.1 4.2 4.0 4.8 3.9 3.5 3.7
Share of constr. in GDP (%) 2.4 2.8 2.5 2.8 4.0 4.3 3.3 5.4 4.4 4.6 4.0
Growth in constr. GDP (%) –41.0 20.2 –8.9 17.3 49.2 11.9 19.2 69.8 14.1 9.4 –10.3
Contribution of constr. to GDP growth (%) –0.1 0.6 –0.2 0.5 2.0 0.5 –0.6 0.4 –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’.
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 equip-
ment, 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 con-
struction 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 bot-
tlenecks and in creating an environment to further exploit the potentials of
the construction industry to contribute to the social-economic develop-
ment 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 ‘increas-
ing 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 Share of different type of construction activities in GFCF/growth rate of construction activities and total GFCF,
1983–93
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993
Growth in total GFCF (%) –33.2 45.7 22.6 3.0 29.4 2.3 5.1 6.5 32.0 11.0 5.4
Growth in GFCF buildings (%) –32.7 17.2 3.2 10.1 13.1 19.4 21.4 29.5 49.7 3.33 16.6
Growth in GFCF other works (%) –17.0 14.1 6.5 27.6 98.3 1.3 –25.3 136.2 32.1 9.1 12.3
Share of buildings in GFCF (%) 19.9 16.0 12.7 14.4 9.7 11.3 9.3 12.9 14.5 16.9 13.7
Share of other works in GFCF (%) 18.3 14.3 10.9 14.4 22.0 21.8 17.2 43.4. 22.2 27.2 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 Formal sector Total
No of peoples employed (× 1000) 160 000 (TIS 1991) 91 649 (LFS 1991) 251 649
22 113 (WB 1994) 182 113
% of total 64 36 100
Annual gross output 14 577 million TSh 1991 14 416 mill TSh 1991 28 993
88 554 mill TSh 1994
Annual value added 10 864 million TSh 1991 4 324 mill TSh 1991 15 188
26 420 mill TSh 1994
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 knowl-
edge 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 tech-
nological 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 practi-
cally 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 con-
tractor 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
Tower crane 1 1 – – – – –
Concrete mixers 3 2 1 1 1 – –
Block making mixers 2 1 1 1 – – –
Steel bending machines (set) 1 set 1 set 1 set – – – –
Water pumps 2 1 1 – – – –
Concrete dumpers 2 2 – – – – –
Heavy duty motor vehicles 3 2 1 – – – –
Compactors 2 1 – – – – –
Compressors 1 1 – – – – –
Concrete vibrators 2 1 1 – – – –
Total number of contractors 45 20 56 70 157 146 650
registered (1990)
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 invest-
ments in the construction industry (NBS, various issues). The major reason
is that Tanzania has to import the majority of equipment. Only the higher-
class 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 consid-
eration, the indigenous technology level is even lower.
The field studies on the process technologies in the dwelling construc-
tion 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 docu-
ments 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 Status of human resources stock relevant for the sector2

6.2.1 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 employ-
ment 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 popula-
tion 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 infor-
mal 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 per-
centage (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 manage-
ment, 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 quan-
titatively 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 insti-
tutes are almost entirely staffed by Tanzanian nationals.
Mlawa and Sheya (1990) investigated a selection of research and develop-

ment 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 lib-
eralisation 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 insti-
tutes appears to be reasonable. It can, therefore, be expected that research
and development human resources are an enhancing factor in the improve-
ment 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 con-
struction 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 con-
struction 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 litera-
ture 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 materi-
als 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 indus-

try. 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 non-
metallic 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 indus-
try include clay, with an estimated production of 2000 metric tons in 1992;
gypsum with an estimated production of 35 000 metric tons in 1992; lime-
stone, 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 trans-
port 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 mate-
rials 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 produc-
tion. 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 con-
sumption 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 infra-
structure are the type of activity of the technology-supporting and -pro-
moting agents and groups, the policies, objectives and strategies, the source
and availability of capital and the nature of the relations with the construc-
tion 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 construc-
tion are rather weak. Communications and relations of the major actors
with the construction units that carry out the dwelling construction pro-
jects 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 communi-
cations and relations between the actors and the construction units build-
ing houses for lower-income households in particular.
7 Conclusions
The conclusion is that optimal performance of the sector of dwelling con-

struction for lower-income households in urban Tanzania requires consid-
erable 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 improve-
ment of the technology infrastructure. See also Table 8.6.
The work presented is based on an extensive study of the available litera-
ture and a number of field studies. The adoption and application of the
Table 8.6 Requirements for technological capability building
Product technologies Requirements for TC building
1. quantity of output investment in the housing sector and its backward linkages for the alleviation
of bottlenecks in the construction process
2. supply of facilities in the house investment in physical infrastructure: access roads, water, electricity and sewerage
3. application of appr. materials and investment in establishment of local production of building materials with
constr.systems high local content
4. application of proper phys.-techn. investment in training and education, information and documentation
constr. details systems
5. simplification of production processes investment in r&d of construction technologies and systems; investment in
on site 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 investment in the establishment of local production of equipment and tools
with high local content
2. knowledge and skills of project managers, investment in the establishment of proper training and education systems
general foremen, craftsmen, labourers 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 investment in:
doc. regarding technical specific. and doc., (a) establishment of well functioning and accessible consultancy and advisory
planning and control systems, material organizations
and equip. data bases (b) establishment, accessibility and diffusion of planning and control systems
213
(c) establishment, accessibility and diffusion of materials & equip. databases
214
Process technologies Requirements for TC building
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 Requirements for TC building
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
Technology infrastructure Requirements for TC building
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
materials & equipment suppliers 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
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 investment in establishment accessibility of financing systems for the
construction industry
branch organisations investment in improvement and accessibility of branch organisations
labour organisations investment in establishment of incentive systems
215
216
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 technologi-
cal aspects, are determined, analysed and evaluated.
Notes
* Section of Technology and Development Studies, Faculty of Technology
Management, Eindhoven University of Technology.
1. 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.
2. 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 Education and technical education
1.1 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 indus-
trial 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 adapta-
tion 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, enrol-
ment 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 sec-
ondary 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 par-
ticular, enrolment in technical education stays behind, both at the sec-
ondary and tertiary levels.
In most developing countries, education is provided publicly (by the gov-
ernment), 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 popula-
tion (millions) in the period 1960–95 and the corresponding Dutch popula-
tion 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 popu-
lation 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
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
500 000
0
10–14
20–24
30–34
40–44
50–54
60–64
70–74
0–4
Figure 9.2b Age groups in the Netherlands in 1960 in absolute numbers
1 500 000
1 000 000
500 000 Netherlands
0
50–54
70–74
20–24
30–34
10–14
0–4
40–44
60–64
Figure 9.2c Age groups in Tanzania in 1990 in absolute numbers
5 000 000
4 000 000
3 000 000
2 000 000 Tanzania
1 000 000
0
20–24
30–34
40–44
50–54
60–64
70–74
10–14
0–4
Figure 9.2d Age groups in the Netherlands in 1990 in absolute numbers
1 500 000
1 000 000
Netherlands
500 000
0
0–4
10–14
20–24
30–34
40–44
50–54
60–64
70–74
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 min-
istries: the Ministry of Education and Culture, covering primary and sec-
ondary 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 min-
istry-controlled school fees are paid, all primary education in Tanzania is
paid for by the government, which allocates about 20 per cent of its recur-
rent 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 sub-
jects 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 eventu-
ally 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 educa-
tion lasts four years. Students can choose the following orientations: agricul-
ture, 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 techni-
cally 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 Figure 9.3b Gross enrolment in

education (% age group) secondary education (% age group)
120 140
100 120
100
80
80
60
60
40 40
20 20
0
0
1970
1975
1980
1985
1990
1995
1970 1975 1980 1985 1990 1995
Tanzania
Tanzania
Netherlands
Netherlands
Figure 9.3c Gross enrolment in tertiary Figure 9.4 Technical students in sec-
education (% age group) ondary 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 percent-
age 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 manu-
facturing (Misanga, 1993). There is little reason to believe that the situa-
tion 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 enter-
prises 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 technolo-
gies in the metalworking industry, and little attention was paid to knowl-
edge 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 increas-
ingly doubts were voiced with respect to the quality of (vocational) techni-
cal 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 employ-
ees 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 Methodology
2.2.1 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 intellec-
tual) 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 sepa-
rately, but to focus on skills. For applying skills one needs a body of knowl-
edge and when measuring skills one will at the same time, though
indirectly, measure knowledge (Romiszowski, 1984). In addition, skills con-
tribute 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 opportuni-
ties, which is particularly important in developing countries (Kessels,
1992). Once it was decided to focus on skills, four categories were distin-
guished: cognitive skills (decision making, logical thinking, planning), psy-
chomotor skills (physical action, head–hand–foot coordination), interactive
skills (attitudes, feelings) and reactive skills (dealing with others purpo-
sively). 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 (<10 employees), medium
(11–99 employees) and large (>100 employees), the metalworking processes
forming, removing, joining and coating, and the production workers’ educa-
tional 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 enter-
prise 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 metal-
working 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 character-
istics of educational institutes, a distinction was made between institu-
tional 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 gener-
ous 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 Results
2.3.1 Nature of shortcomings in knowledge and skills
Table 9.1 presents the outcomes of the survey in 28 metalworking enter-
prises. 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
Educational Skills (scores, 1–4 range)a Knowledge and skills based on single
level of employee based on closed questions open question
Cogn Psym Inter Rep

R P R P R P R
VTC 2.5 3.1 2.8 3.4 2.6 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
TSS 2.4 3.0 2.6 3.2 2.6 3.2 3.0 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
TC 2.6 3.2 2.8 3.2 2.5 3.3 3.0 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 devel-
oped than psychomotoric, interactive and reactive reproductive and pro-
ductive 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 shortcom-
ings in knowledge and skills and the institutional characteristics are sum-
marized as follows:
Students. Many of the indicated shortcomings find their origin in the edu-
cational 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 expos-
itive 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 class-
room, 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 back-
ground is hardly higher than that of their students, and family and com-
munity 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 instruc-
tional 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
Indicated shortcoming Educational levelb Institutional characteristicsc Instructional characteristicsd
VTC TSS TC t s f,c sm f c o,c m
Reproductive skills √ √ √ x x x x x
Cognitive skills √ √ x x x x x
Understanding of English √ √ √ x x x x x
Knowledge of modern √ √ √ x x x x
metalworking technologies
Striving for excellence √ √ √ x x x
Awareness of safety-related √ √ √ x x x x
aspects
Theoretical knowledge (basic √ x x x x x
metalworking processes)
Awareness of preventive √ x x
maintenance
Management and supervision √ 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 en-
gineering). 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 counter-
parts, 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 condi-
tions. 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 situ-
ation 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 com-
pared 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%<zt<z95% for all pass/fail
combinations). Average pass rates for all subjects shown are significantly
higher for public TSSs than for private ones (zt<–z95%). In view of the exam-
ination results and the financial constraints the respective institutes face, it
stands to reason to assume a direct and positive relationship between the
amount of money spent per student and his or her examination results.
2.4 Analysis of results

From the confrontation of the shortcomings in knowledge and skills with
the institutional and instructional characteristics of the various institutes, it
has become plausible that shortcomings in knowledge and skills are to a
large extent caused by the use of English as a medium of instruction. In
addition, from observations made during the fieldwork, it became apparent
that financial means also constituted a serious shortcoming.
When comparing examination results of institutes on the basis of own-
ership (public versus private), the importance of adequate financing of
institutes became all the more apparent. When considering future policies
with respect to technical education in Tanzania, the two main causes of
shortcomings in knowledge and skills will face policy makers with several
dilemmas.
The use of English as a medium of instruction affects both students and
teachers and influences the teaching methodology (directly as well as via
Table 9.3 Pass/fail combinations at VTCs
Pass/fail combination Public VTCsa Private VTCsa Test z1
Theory Practical p q n p q n
Pass Pass .490 .510 243 .550 .450 115 –1.064

Pass Fail .091 .909 45 .081 .919 17 0.047
Fail Pass .337 .667 167 .282 .718 59 0.796
Fail Fail .082 .918 39 .086 .914 18 –0.050
Source: Vocational Education and Training Authority, Tanzania (1994).

Note: a p, proportion meeting a specific pass/fail combination; q, proportion not meeting a
specific pass/fail combination; n number of applicants meeting a specific pass/fail combination.
232
Table 9.4 Average pass rates at technical secondary schools
Subject Public TSSsa Private TSSsa Test z1
p q x s n p q x s n Pass rate CSSE score
Basic mathematics .603 .397 3.855 1.488 2,123 .136 .864 4.788 .372 382 22.777 –22.798
English .796 .231 3.686 .946 2,108 .468 .532 4.403 .516 380 11.685 –16.869
Technical subjects .630 .370 3.746 1.592 5 842 .502 .498 4.175 .788 987 7.485 –6.393
Source: National Examinations Council, Tanzania (1994, 1995).

Note: a p, proportion passing a specific subject; q, proportion failing a specific subject; x, mean score of the population; n, number of applicants per
subject. s, estimate population’s standard deviations.
the teaching material), which in turn has its effect on the performance of
students. Changing the medium of instruction from English to Swahili
would seem to be a logical decision. However, there are a number of argu-
ments against such a step. (1) Most (if not all) literature is in English.
Cooperation with foreign countries in production requires a mastery
of English. Introducing Swahili would frustrate this cooperation.
(2) Introducing Swahili would make it necessary to translate almost all lit-
erature. Apart from the fact that this would take considerable time, the
costs would be prohibitive. (3) Although Swahili is a complex and precise
language it unfortunately has an inadequately developed technical vocab-
ulary and developing a suitable technical vocabulary would again take
considerable time and be costly. There are, however, also arguments
against the present use of English as a medium for instruction: (1) teachers
and students have insufficient mastery of the language, and, as a result,
(2) the teaching methodology remains mainly expositive. Considering the
wide use of English in science and technology, and the importance of
these fields with respect to industrial growth and development, and con-
sidering the immense problems to overcome when switching to Swahili,
the Tanzanian government should stimulate the use of English, inside as
well as outside the educational institutes, much more intensively.
The financial means of the educational institutes come from three differ-
ent sources (government, donor organisations and school fees), and
influence the quality of the teachers and the availability and state of
books and machines, of which the latter is directly related to the most
appropriate teaching methodology. Furthermore, the comparison of
examination results of public and private institutes indicated that stu-
dents at institutes which have ample financial means obtain better results
than students at institutes which are financially less well off. At the voca-
tional level, the private institutes have more ample spending possibilities,
while at the secondary level the public institutes have more financial
means than their private counterparts. For the latter these differences are
aggravated by the fact that, with respect to selection of students, primary
school leavers with the highest examination results are selected by the
government to enter public secondary schools, while others (that is, with
lower examination results) are free to enter private institutes. Also, with
respect to the selection of teachers, private schools are at a disadvantage:
teachers at public secondary schools are appointed, trained and allocated
by the government, whereas teachers at private institutes have to apply
on their own initiative – their employer (private school) and/or the teach-
ers themselves having to pay for the bulk of their training. Since also at
the teacher training college persons with the highest results at formal
educational levels are allowed to enter, the private schools, unless they
have ample financial means, are left with the less qualified teachers to
teach the less bright students.
Because of these relationships it would seem a logical suggestion to raise

the schools’ financial means by increasing funding from one or more of the
sources mentioned earlier. However, for a number of reasons, this will
cause severe complications. (1) The government’s ability to raise its contri-
butions to the educational sector hardly exists. (2) Under present condi-
tions, donor funding in general contributes positively to a school’s
performance, and, therefore, both the Tanzanian government and the edu-
cational institutes themselves should be pressed to stimulate donor
financing. This, however, should be accompanied by a more even-handed
approach towards private secondary schools by the Tanzanian government,
with respect to regulations concerning fees, appointment and training of
teachers and the selection of students. (3) School fees should be collected
and spent at the level where this is most efficient. Furthermore, although it
could make the educational system more elitist, school fees could be raised,
considering the fact that the number of applications outstrips the number
of vacancies by far, at all levels included in the research. At the technical
colleges an increase of fees is quite feasible. At secondary level, an increase
of fees should refer to private institutes only, thus leaving entry to public
institutes in essence meritocratic. Finally, at the vocational level, the
number of public training centers is quite substantial. An increase of fees at
these institutes, where deemed necessary, will probably not lead to
restricted access at this level.
Notes
Management, Eindhoven University of Technology.
Part III
Environmental and Energy Aspects
of Industrialization
10
Industry and Environment:
Methodologies for Environmental
Assessment in Data-Poor Situations
Lex Lemmens*, Peter Scheren**, Harro Zanting*, Gregory Njau***
and David van Horen*
1 Introduction
The key environmental issues for Tanzania are: land degradation, lack of
accessible water supply and poor water quality, deterioration of aquatic
systems, loss of wildlife habitats and biodiversity, deforestation and envir-
onmental pollution (Division of Environment, 1995). These environmental
problems have evolved over a long period of time throughout the country.
The contribution of industry to these problems is relatively recent.
Industry can be a cause for all of the problems mentioned, but is most
often associated with environmental pollution. Indeed the impact
of industry in the other problem areas is often through pollution.
The determination of the relative importance of industry as a threat to
the quality of the environment is one of the topics of this paper. The
major problem when monitoring or assessing the environment
in Tanzania is the structural lack of data (DANIDA, 1989; Lemmens
et al., 1996).
It is, therefore, of utmost importance to develop methodologies that
provide the best possible environmental assessments in view of the limited
information available. This article presents the rapid assessment methodol-
ogy as a tool for making an environmental pollution inventory in data-
poor situations. The tool was first developed, evaluated and applied to the
case of Lake Victoria. The results of this research will be presented first.
Then it will be shown that the methodology can also be used to get a first
insight into the environmental situation of the entire country using data
from the national census. To complete the available information on indus-
try and environment, the last sections focus on industrial pollution in the
two major cities, Dar es Salaam and Tanga, and on pollution caused by
some mining activities.
237
2 The rapid assessment methodology
Most rapid assessment procedures can be generalised by the following

formula:
Waste Load = Functional Variable * Pollution Intensity (10.1)
in which the pollution intensity represents the amount of waste produced

per unit of a given functional variable. Examples are listed in Table 10.1.
While data on the size of functional variables need to be gathered ad hoc,
for example, through field surveys, the pollution intensities derive from
literature studies.
This study elaborates on the generally applied procedures. First, to com-
plete the balance of nutrient input into a given water body, an evaluation
of atmospheric deposition of nutrients is presented. Penetration factors are
introduced to account for human (for example, treatment plants)
or natural purification (for example, water purification in rivers and
wetlands).
In addition, any rapid assessment methodology obviously requires
assumptions and generalizations introducing uncertainty in its results. In
the proposed methodology a method for error analysis is introduced
(Lemmens et al., 1998b). To put the results in their right perspective in this
article, only the results of the error analysis are presented. In order to test
the results, a simple verification model is applied.
2.1 Waste loads

On the basis of an extensive literature survey and the assumption of a
linear relation between waste load and production size, the World Health
Organization (1982) and Economopoulos (1993) report pollution intens-
ities for quantification of industrial waste loads from the most polluting
industrial branches. This information, together with locally acquired infor-
mation on industrial production, government statistics on industries, pre-
vious industrial (pollution) surveys, discussions with civil servants and
visits to industrial sites, led to an estimate of the waste load of the most
polluting industries in the research area.
Table 10.1 Functional variables and pollution intensities for the principle
pollution sources
Pollution source Functional variable Pollution intensity
Industries Annual production Waste production per unit product

Households Population number Annual waste production per person
Agriculture Area of (non) cultivated Annual export coefficients of nutrients
land
Industry and Environment 239
As with industrial pollution, the World Health Organization (1982) and

Economopoulos (1993) have defined pollution intensities for domestic
waste loads. Additional sources for pollution intensities are Jørgensen
(1980) and Vighi and Chiaudani (1987). It is necessary to distinguish
between waste loads from urban and rural populations. Liquid waste from
urban households ends up in watercourses via discharge of sewage, run-off,
drainage, and so on. Especially for those towns bordering the lakeshore,
this waste load directly infiltrates the lake. Rural households, on the other
hand, are generally scattered over a vast area, in the case of Lake Victoria
extending up to 250 km from the lake.
Only a small fraction of household waste reaches water courses via run-
off and the numerous seasonal and non-seasonal rivers (penetration factor
≈ 0). Waste loads from the rural population can, therefore, be considered
part of land run-off as assessed under agricultural pollution. Per capita
waste loads for urban inhabitants vary according to sanitary conditions. Is
there a sewage system, are there septic tanks or only pit latrines?
The method applied to determine agricultural nutrient loads uses the
export coefficient approach of Jørgensen (1980), an approach widely used
in many studies (for example, Scheren et al., 1995; Mattikalli and Richards,
1996). The loss of nitrogen and phosphorus to watercourses is estimated
from the area of different land usage in the catchment area, multiplied by
typical export coefficients for each land-use type. Based on Thomann and
Mueller (1987), the identification of different land-use types was limited to
two categories: cultivated arable land and non-cultivated (forest, pasture
and fallow) land. In addition to the type of cultivation, other characteris-
tics also control the magnitude of nutrient flux. These include precipitation
run-off, soil texture, slope and fertilizer application.
Livestock production is an important activity in terms of land-use
pattern in many agricultural areas. It can, therefore, be expected that
animal droppings, either left on the field during grazing or spread out over
croplands as organic fertilizer, are an important factor affecting nutrient
loading. An assessment of both organic and artificial fertilizer use in the
catchment area served as a verification of the nutrient loading estimate
based on export coefficients (Scheren et al., 1995). Data on artificial fertil-
izer use (or alternatively its distribution) and livestock population were
used as a basis for calculations.
In the case of Lake Victoria, artificial fertilizers are applied before or in
the rainy seasons, which are also the main crop-growing seasons.
Therefore, the total amount applied is subject to loss to watercourses. As
leaching and run-off of nutrients from manure occur during the rainy
seasons only (approximately five months per year for the Lake Victoria
area), the load is estimated at 5/12 of the total amount of manure. The frac-
tion of nutrients from (natural and artificial) fertilizers running off and
leaching to watercourses is taken at 13 per cent for phosphorus and
30 per cent for nitrogen compounds (Orlova and Yaroshenko, 1976;

Santschi and Schindler, 1977; Ministry of Health and Environmental
Protection, 1980).
2.2 Wet deposition

During rainfall all kinds of dust particles are washed down to the surface of
the lake. The relevance of this wet deposition lies in the presence of sub-
stances containing nitrogen and phosphorus in the air, which can con-
tribute significantly to the total nutrient load. Human activities in a wide
area around the water body can very well be responsible for an important
part of the nutrient contents of rain water. Biomass burning, for example,
is known to be an important contributor. The load of nitrogen and phos-
phorus as a result of wet deposition is estimated by multiplying average
concentrations of N and P in rain with the average amount of annual rain-
fall over the lake. Selection of average concentrations can be based on local
analyses (as done for Lake Victoria) or on typical values from the literature
(see, for example, Jørgensen, 1980; Thomann and Mueller, 1987; Ritter,
1988).
2.3 Penetration factors

The annual industrial, domestic and agricultural waste loads calculated in
the manner indicated earlier are the loads generated at the different pollu-
tion sources. In many cases these are higher than the loads eventually dis-
charged into the lake. The difference in loading can be expressed by one or
more penetration factors, each with a value between one and zero. This
leads to the following adaptation of the general formula:
Waste Load = Functional Variable * Pollution Intensity * Penetration

Factor(s) (10.2)
The most common penetration factors are the effectiveness factors of

domestic and industrial waste-water treatment facilities. The efficacy of
treatment facilities in reducing waste loads can be estimated from typical
performances of normally operating plants, as reported by Economopoulos
(1993). The efficiency of badly maintained or overloaded treatment plants
(observed widely in the Lake Victoria basin) is lower than these typical per-
formances. In such cases, approximate treatment efficiencies may be
defined, based on effluent analysis data and selected site visits. The latter
has been done for most domestic and industrial facilities in the case of Lake
Victoria.
Next to treatment facilities, there are other, often natural, conditions
that reduce the waste load released by industries, cities or agriculture before
reaching the water body of the lake. This study proposes the incorporation
of the self-purification potential of rivers and streams, and the liquid waste
filtration of wetlands into the method. The biochemical oxygen demand

(BOD) load from a distant source can be wholly or partly broken down
when it is moving from the source to the lake. A first-order decay process
can describe the breakdown of BOD5 in rivers:
L  − K× D 
= exp   (10.3)
L0  86.4 × 
in which L represents downstream BOD5, L0 represents BOD5 at the point

of discharge, K is the decay rate (day–1), D is the discharge distance from the
lakeshore (km) and v is the flow velocity (m s–1).
The penetration factor L/L0 can be defined for each distance D from the
lakeshore if K and v are known. K depends on many different river charac-
teristics, such as bacterial composition, water temperature, water turbu-
lence and stream velocity, and will be different for individual (parts of)
rivers. Typical values for K vary from 0.2 to 0.8 day –1 (Thomann and
Mueller, 1987; Metcalf and Eddy, 1991; Zanting, 1996), with 0.3 as a ‘most
likely’ value. Thomann and Mueller (1987) present an equation that relates
the river flow-through velocity to characteristics such as river volume flow,
slope, drainage area and length. Based on this relation and the characteris-
tics of several Kenyan rivers (Ministry of Water Development, 1992), the
velocity v was estimated to average between 0.3 and 1.0 m s –1 for rivers
around Lake Victoria. Calibration of Equation 3 with these ranges for K and
v, and for several values of the distance D, results in BOD penetration
factors. For five ranges of distances from the lakeshore these are presented
in Table 10.2.
Many rivers, and even sewage outlets, drain through swamp areas before
discharging on the open waters of a lake. In this way the concentration of
nutrients and oxygen demanding waste is reduced. The impact of wetlands
on the reduction of waste loads before reaching the lake is incorporated by
defining wetland penetration factors. The waste-water treatment capacity
of wetlands is closely tied to various aspects of wetland hydrology, such as
water depth, retention time and degree of channelization.
Typical BOD penetration factors through natural wetlands, as reported
by Richardson and Nichols (1985), range from 0.30 to 0.05. A rounded
Table 10.2 Selected distance penetration factors for BOD5
Distance (km) Penetration factors
0–10 1.00
10–50 0.90
50–100 0.75
100–150 0.60
150–250 0.45
mean factor of 0.20 is selected as the most likely value. The performance of
wetlands in removing nutrients from waste loads depends highly on the
average nutrient loading per wetland surface unit. Using studies of several
(North American) wetlands, Richardson and Nichols (1985) present graphi-
cal relations between P and N removal rates as a function of average loading.
These relations have been applied to estimate nutrient filtration by swamps
in the Lake Victoria basin. Selected values are presented in Table 10.3.
2.4 Reference nutrient load from modelling

In this section a brief presentation is given of a theoretical model for the
nutrient balance of a lake. A more detailed description and evaluation are
presented by, for example, Scheren (1995). The model assumes complete
mixing of the water, thus averaging concentrations over a lake or water.
Although local situations or fluctuations are not explained in this way, the
procedure is very useful in clarifying the overall nutrient balance. The three
main assumptions of the model are: (1) nutrient concentrations are the
same throughout the lake; (2) the volume and outflow of the lake are con-
stant and (3) all processes involving nutrients (for example, sedimentation)
are linearly related to the nutrient concentration. In a steady state situ-
ation, in which there is no accumulation of nutrients in the lake water, the
total nutrient input in the lake is assumed to be equal to the sum of nutri-
ent outflow through river(s) and the net loss of nutrients to the sediment.
This mass balance results in Equation 10.4:
 v SA 
WA = C A×  Q+E V  (10.4)
 H
in which WA represents the input of nutrient A (g yr–1), CA is the concentra-

tion of nutrient A (g m–3), QE is the river outflow (m3 yr–1), V is the lake
volume (m3), VAS is the sedimentation velocity for nutrient A (m yr–1), and H
is the average lake depth (m). Average data of these parameters for Lake
Victoria result in a value for the nutrient load WA (for both nitrogen and
phosphorus) that serves as a verification of the rapid assessment nutrient
load estimate.
Table 10.3 Nutrient penetration factors through wetlands (selected most

likely values)
Loading, g (N/P) m–2yr–1 N P
2 0.4
10 0.3 0.6
50 0.7 0.75
500 0.9 0.8
3 Lake basin statistics
With a surface area of 68 800 km2, Lake Victoria is the largest lake in Africa.
Its adjoining catchment area (Figure 10.1) of 194 000 km2 stretches out
over five countries: Tanzania, Kenya, Uganda, Rwanda and Burundi. Only
the first three border on (and share a part of) the lake.
The relatively shallow lake (average depth is 40 m) serves its riparian
population in the provision of fresh water and fish, as well as a depository
for waste from different sources. The only outflow is the Victoria Nile,
leaving the lake at Jinja in Uganda. Evaporation and precipitation domi-
nate the water balance of Lake Victoria.
The features of the catchment area of Lake Victoria that determine
the functional variables for the waste load assessment are discussed in the
following sections.
Figure 10.1 The catchment area of Lake Victoria (source: Crul, 1993)
Victoria
Nile
Victoria
Nile
UGANDA Kampala KENYA

Kisumu
Lake
Kagera Victoria
Bukoba Mara
Musoma
RWANDA
Muanza
BURUNDI
TANZANIA
3.1 Industrial characteristics

Most industrial activity (Table 10.4) is located in the major towns border-
ing the lake, such as Kampala and Jinja in Uganda, Mwanza and Musoma
in Tanzania, and Kisumu in Kenya. Important exceptions are the large
sugar factories located at some distance from the lake on the Kenyan side.
The discussion on industrial liquid waste is limited to BOD loads, because
the discharge of N and P is negligible compared to other sources.
For several industrial pollution sources, only penetration factors for dis-
tance have been applied as, with a few exceptions, industrial waste-water
treatment is generally absent or inefficient because of poor maintenance.
However, much of the Ugandan industrial effluent drains through wetlands
before reaching the lake surface water. Wetland penetration factors have,
therefore, been applied. The waste load from those urban industrial sites
discharging to the domestic sewer system was corrected for the defined
treatment efficiency of the municipal plants.
3.2 Population
With a population of about 30 million, the Lake Victoria catchment area is
one of the most densely populated parts of Africa. Annual population
Table 10.4 Estimated 1995 production of major industries in the catchment

area of Lake Victoria by country and industrial (sub)sector
ISIC Industry Shared production of major factories, t yr–1

(number of major factories)
Kenyaa Ugandab Tanzaniac
3111 Slaughterhouses 3 000 (1) 14 000 (3) –

3112 Dairy factories – 25 000 (1) –
3114 Fish factories 18 000 (6) 12 000 (4) 6 630 (2)
3115 Vegetable oil refineries – 3 300 (2) 31 400 (4)
3118 Sugar factories 640 000 (6) – 3 500 (1)
3131 Distilleries 18 000 (1) 460 (1) –
3133 Breweries 60 000 (1) 17 000 (1) –
3134 Bottleries 41 000 (1) 40 000 (2) 11 100 (2)
3211 Cotton mills 1 500 (1) – 1 320 (2)
3231 Leather tanneries 270 (1) 60 (1) 220 (1)
3411 Paper mills 94 000 (2) – –
3523 Soap factories – 40 000 (2) 15 000
(1)
Notes: a Sources: Oerlemans (1985); Calamari (1994); Ministry of Planning and National
Development (1994); Kirugara and Nevejan (1996), selected industrial site visits.
b
Sources: Droruga (1990); Ministry of Finance and Economic Planning (1993); Ministry of
Natural Resources (1995); selected site visits.
c
Source: Scheren et al. (1995).
growth is 2–4 per cent in most parts of the lake basin, but urban popula-
tion growth is over 5–10 per cent per year in most of the larger towns.
Table 10.5 summarizes the population statistics of the lake basin. For each
of the 40 towns larger than 10 000 inhabitants (held as a threshold for
urban characteristics) in the Lake Victoria basin, information on sanitary
conditions, domestic waste-water treatment, wetland drainage and distance
from the lakeshore was gathered.
3.3 Agriculture
There are many densely cultivated areas in the Lake Victoria basin, espe-
cially in Kenya, Rwanda and Burundi. Important staple crops are maize
and bananas. The main cash crops are sugar cane, coffee and tea.
Application of artificial fertilizers is significantly higher in Kenya than in
other parts of the catchment area. Agricultural characteristics are summa-
rized in Table 10.6. Most data in Table 10.6 are extrapolations based on
surveys from the early to mid-1990s. An annual growth rate of 1 per cent
is assumed for the area of cultivated arable land (Ministry of Water
Development, 1992). Wetland penetration factors have been applied for
areas draining through wetlands, requiring considerable judgement of the
investigators.
3.4 Rainfall
Average rainfall over Lake Victoria is 1450 mm yr–1 (±10 per cent), equiva-
lent to approximately 100 km 3 yr–1. Regional rainfall varies between
895 mm yr–1 in Musoma (Tanzania) and 2216 mm yr –1 in Kalangala
(Uganda) (Crul, 1993).
Table 10.5 Domestic characteristics of the Lake Victoria catchment area

(1995)
Total population Urban population (1 000 people) Number of towns

(1 000 people)
Sewered Unsewered
Kenyaa 10 200 390 630 18

Ugandab 5 600 210 870 9
Tanzaniac 5 200 27 340 4
Rwandad 5 900 – 400 5
Burundid 2 800 – 140 4
Total 29 700 627 2 380 40
Notes: a Source: Ministry of Water Development (1992).

b
Sources: Ministry of Finance and Economic Planning (1992), Ministry of Natural Resources
(1993), National Water and Sewerage Corporation (1995).
c
d
Source: Verlinden (1996).
246
Table 10.6 Estimated agricultural statistics of the Lake Victoria catchment area, 1995
Catchment land area (1 000 ha) Livestock population (1 000 heads) Fertilizer use (t y–1)
Cultivated Non-cultivated Total Cattle Sheep & goats Chicken N P 2O 5

a
Kenya 1 470 3 400 4 870 4 890 3 140 11 400 9 000 14 900
Ugandab 1 400 2 100 3 500 1 630 1 590 3 750 420 210
Tanzaniac 1 500 5 540 7 040 2 800 1 900 1 300 420 130
Rwandad 930 1 130 2 060 480 310 – 300 80
Burundid 670 640 1 310 200 180 – 1 000 800
Total 5 970 12 810 18 780 10 000 7 120 16 450 11 140 16 120
Notes: a Sources: Lake Basin Development Authority (1987); Kenya Grain Growers Cooperative Union (1991); Ministry of Water Development (1992);
Kirugara and Nevejan (1996); Ministry of Natural Resources (1996).
b
Source: Ministry of Agriculture (1992, 1995); World Bank (1993e), Bank of Uganda (1995).
c
d
Source: Bullock et al. (1995); United Nations (1995).
Only limited data are available on nutrient concentrations in rainwater

in the region. Average concentrations of total nitrogen (TN) and total phos-
phorus (TP) reported by Bootsma and Hecky (1993), Scheren (1995) and
ICRAF (1995) range from 0.58 to 0.92 mg l –1 for N and from 0.07 to
0.11 mg l–1 for P (see Table 10.7). The N concentrations are similar to
values reported from larger, North American, data sets. On the other hand,
phosphorus concentrations from Lake Victoria rainwater are surprisingly
high in comparison with reported values (for example, Ritter, 1988).
A possible explanation for the observed high phosphorus concentrations
in rain could be that, in all reported cases, samples were taken relatively
close to the lakeshore. Here, dust and air pollution might be considerably
higher than further offshore. It is debated, however, as to whether the high
P concentration in rainwater might be caused by the extensive forest
burning going on in the adjacent Rwanda and Burundi highlands (Bootsma
and Hecky, 1993; Scheren, 1995b).
3.5 Wetlands
The largest wetland area in the Lake Victoria basin is in western Uganda.
The entire drainage of the Katonga and Ruizi-Kibali river basins is through
river channels that used to be more rapidly flowing in the past. Now, these
channels are ridden with swamps and small swampy lakes (Beadle and
Lind, 1960). Other important wetlands in Uganda are the small swamps on
the lakeshore, receiving the domestic and industrial waste-water from
Kampala and Jinja (Ogaram and Kalema, 1995). Furthermore, the Kagera
river basin, draining through Rwanda, Burundi, Tanzania and Uganda, is
covered with large areas of wetlands. Important swamps on the Kenyan
side are at the mouths of the rivers Yala and Nyando. For all these wet-
lands, information on size and average nutrient loads where combined to
determine the nutrient penetration factors on which Table 10.3 is based.
4 Results from the catchment area of Lake Victoria
4.1 BOD loading estimates

Organic waste loads, represented by a BOD load, were assessed for domestic
and industrial (point) sources. As the study focuses on the pollution of the
Table 10.7 Selected average nutrient concentrations in direct rainfall on Lake

Victoria
Average nutrient concentrations (mg l–1)
TN 0.80
TP 0.09
lake, the estimated generated BOD loads were corrected for purification in
treatment plants, rivers and wetlands. Table 10.12 presents the results.
The probability intervals presented in Table 10.8 and 10.9 result from an
error analysis (Lemmens et al., 1998b). They account for the roughness of
the data in presenting, more or less, a best- and worst-case scenario. The
authors have chosen the values in such a way that they represent the most
likely case.
Breweries, sugar-cane factories and soap and oil factories display the
largest amount of industrial BOD load. The most likely BOD loads are pre-
sented in Figure 10.2. The conclusion is that domestic pollution accounts
for most of the BOD load. An interesting note here: 75 per cent of Uganda’s
domestic BOD load originates from its capital, Kampala, while in Kenya
50 per cent originates from Kisumu. Further calibration of the Kenyan
results shows that, if all the present treatment plants could perform opti-
mally and meet the theoretical reduction efficiency, BOD loads could be
brought down to 50 per cent of the presented most likely value.
Table 10.8 Results of the Lake Victoria BOD5 loading assessment
Unit: t yr–1 Kenya Uganda Tanzania
Source Most likely Most likely Most likely
Industrial 1 810 540 820

Domestic 5 700 4 000 3 100
Total 7 510 4 540 3 920
Prob >55% 5 100–10 700 3 800–6 000 3 400–4 500

Prob >90% 2 700–13 800 3 000–7 500 2 900–5 000
Table 10.9 Results of the Lake Victoria nutrient loading assessment (t yr–1)
Source Total nitrogen Total phosphorus

– +
Most likely s s Most likely s– s+
Urban domestic 7 600 1 800 2 100 920 280 1 650
Agriculture 48 200 18 300 40 400 5 710 1 440 2 740
Wet Deposition 81 000 17 700 15 200 9 120 3 650 1 010
Total 136 800 25 500 43 200 15 750 3 900 3 400
Prob >55% 111 000–180 000 11 900–19 200

Prob >90% 86 000–223 000 8 000–23 000
Figure 10.2 Most likely BOD loads to Lake Victoria
Industrial
Domestic
8000
7000
6000
BOD (ton/year)
5000
4000
3000
2000
1000
0
Kenya Uganda Tanzania
4.2 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 deposi-
tion 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 fertil-
izer 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 fertil-
izer application. For P, these figures are 4800 t yr –1 and 800 t yr–1 respect-
ively. 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
Figure 10.3 Most likely nitrogen loads, by source, and agricultural nitrogen loads,
by country
Burundi 4%
Wet Deposition 59% Tanzania 44%

Agriculture 35%
Urban Domestic 6% 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 indus-
trial sectors to the total environmental pressure on aquatic systems. The
data used are those of the 1989 Industrial Census, referring to establish-
ments 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 propor-
tional contribution to the value added in 1989. According to WHO
Table 10.10 Basic data extracted from the Tanzanian industrial census, 1989
ISIC Sector Value added Proportional Value added Value added % of value Value added % of value added
code (000 TSh) contribution of polluting of polluting added omitted industries for taken into account
industries industries because of which WHO in the assessment
(000 TSh) with complete incomplete conversions
data (000 TSh) data are available
(000 TSh)
311–12 Food 58 526 087 12.81 57 995 567 56 222 869 3.06 43 363 464 74.77
313 Beverages 3 213 597 0.70 3 213 597 2 925 230 8.97 1 885 467 58.67
331–32 Wood/wood 32 161 429 7.04 1 760 820 821 185 53.36 97 774 5.55
products/furniture
341–42 Paper/paper 6 531 744 1.43 3 795 996 3 738 471 0.57 3 721 712 98.98
products/printings
publishing
351/354 Chemicals/ 6 003 916 1.31 5 218 226 4 160 039 20.28 1 085 634 20.80
petroleum/petroleum
& coal products
355–6 Rubber/plastic 3 855 234 0.84 2 366 671 2 350 033 0.70 2 234 182 94.4
371–2,381 All metal and 134 708 655 29.47 5 305 363 4 024 467 24.14 4 024 467 75.86
metal products
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 con-

tribute 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 guide-
lines 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 phospho-
rous can be determined. From Table 10.11 it becomes clear that from the
industries taken into account in this assessment, the paper/paper prod-
ucts/printing/publishing subsectors account for almost all organic pollu-
tion (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 emis-
sions. With the data available from the 1989 census this proved to be possi-
ble. 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 Waste volumes and contribution to pollution for different sectors
ISIC code Sector Waste volume BOD5 Tot N Tot P
m3 % [kg] % [kg] % [kg] %
311–12 Food 34 041 272 0.44 97 654 796 16.35 5 908 84.90 1 207 100.00
313 Beverages 1 525 445 0.02 845 081 0.14 0 – 0 –
331–2 Wood/wood products/ 17 954 <0.005 17 516 <0.005 1 051 15.10 0 –
furniture
341–2 Paper/paper products/ 7 603 316 996 97.56 446 015 025 74.70 0 – 0 –
printing/publishing
351/354 Chemicals/petroleum/ – – 1 902 827 0.32 0 – 0 –
petroleum & coal products
355–6 Rubber/plastic 64 142 <0.005 693 <0.005 0 – 0 –
371–2 381 All metal and metal products 154 296 491 1.98 50 664 519 8.49 0 – 0 –
Total 7 793 262 300 100.00 597 100 457 100.00 6.959 100.00 1 207 100.00
253
6.1 Water pollution

Industrial development in Tanzania was pursued without environmental
regulation, resulting in importation of industries without waste-water
treatment facilities. In Dar es Salaam, as in all other towns with industries,
liquid effluent is discharged directly into water systems or drainage
leading into them. A study carried out in 1997 showed that 68 per cent of
all industries surveyed in Dar es Salaam discharge into the Indian Ocean,
either directly or indirectly through a total of 12 streams draining through
heavily inhabited areas of a city with 2.3 million inhabitants (Mgana and
Mahongo, 1997). Table 10.12 shows the effluent composition discharged
directly into the Indian Ocean. The effluent consists of BOD5 2715 t yr–1
and SS (suspended solids) 15 454 t yr–1 representing 19 per cent and 55 per
cent of all BOD5 and SS respectively. The rest comes from domestic
sources.
Tables 10.13 and 10.14 show the effluent composition discharged
through Msimbazi stream, either directly or through the Vingunguti waste
stabilization ponds. Msimbazi stream passes heavily inhabited areas.
Tanzania Breweries Ltd, which accounts for about 60 per cent of the BOD5
and SS, discharges its effluents to the stream untreated. Most of the nitrate
and phosphorus materials come from slaughterhouses.
The situation is not so different in Tanga, where 675 tons of effluent are
discharged directly into the Indian Ocean. Tanga is endowed with many
big rivers, and most of the rural sisal estates discharge their industrial
effluents into these rivers. A fertilizer company was spewing 700 m 3 of
effluent per hour, containing acids, fluoride, phosphorus, nitrogen and
cadmium, before it was closed down a few years ago. Together with other
effluents from soap factories and textile mills, this resulted in an accumula-
tion of anoxic sediments 2 mm below the surface in the area around the
city of Tanga (USAID/Tanzania 1996).
6.2 Air pollution

Sources of air emission are mainly from industries, transport and domes-
tic cooking. The survey carried out in the cities of Dar es Salaam
and Tanga revealed that domestic combustion of biomass is the major
source of air pollution, followed by industries and road transport (Mgana
et al. 1997).
As shown in Tables 10.15 and 10.16, 93 per cent of all the TSP and 96 per
cent of all the CO and VOC emission in Dar es Salaam came from house-
holds. Cement factories in both Dar es Salaam and Tanga were leading
industrial air pollution. The Dar es Salaam cement factory accounts for
40 per cent (2185 tons) of total TSP, 51 per cent (323 tons) of SO 2 and
60 per cent (323 tons) of NOx annually. The Tanga Cement factory annual
emissions were 37 per cent (2354.6 tons) of TSP, 73 per cent (342 tons) of
SO2 and 49 per cent (720 tons) of NOx.
Table 10.12 Pollution sources discharging into sea, Dar es Salaam
Industry Pollution loads (t yr–1)
BOD5 SS Oil Putrescible N P Phenol S Cr Fe Cl
Tiper Industry 2.0 6.5 4.8 0.7 0.02 0.03 0.004

Twiga Chemicals 1.4 0.4 4.7 0.3
SAPA Chemical Ltd 9.8 3.9 32.9 2.1
Kioo Ltd 31.9 9 378.6
Bobby Soap 122.0 20.3
Banco Products Ltd 72.3 2 680.9
Henkel 1 057.5 1 797
Mifuku Ltd 28.8 30.6
Tanganyika Tegry Ltd 0.1 0.3 2.7
Simba Plastic Ltd 1.4 3.9
Kwanza Bottlers 67.0 22.3 241.9 40.5
Costal Oil Industry 214.3 211.8 454.4 76.0
Tradeco Oil Industry 402.6 397.8 337.2 56.4
Murzha Oil Mill Ltd 298.8 295.2 236.0 39.5
Transtech Ltd 209.2 206.6
Tanpack Industry 2.4 0.8
Colgate Palmolive 0.1 0.1 0.6
Medicare Industry 5.0 8.5
Fahari Bottlers Co. 81.5 31.0
Blanket Manf. Co. 37.0 179.0
H.K. Foam 0.11 0.4
Pan Africa Ltd 0.1 0.18
Kibo Paper Mill 173.2 40.4
JEJE Industries Ltd 48.2 0.2
Twiga Paper Prod. Ltd 18.2 36.3
Subtotal 2 714.71 15 454.48 1 335.5 212.4 0.7 2.42 0.03 0.004 48.2 0.2
Sewerage system 441.0 716.0 887.0 221.0 68.0

Septic tank system 277.0 482.0
Pit latrine 10 963.0 11 443.0 3 099.0 410.0
Subtotal 11 681 12 641 887.0 3 320 478
255
Grand Total 14 395.71 28 095.48 2 222.5 212.4 3 320.7 478 2.42 0.03 0.004 48.2 0.2
Table 10.13 Sample of industries directly discharging into Msimbazi stream,

Dar es Salaam
Industry Pollution Loads (t yr–1)
BOD0 SS Oil N P
Tanzania Breweries Ltd 857.4 332.9

Dar Brew Ltd 259.9 100.9
Friendship Textile 50.4 20.7
Kimara Slaughter – Slab 2 29.6 27.6 10.35 3.5 0.25
Subtotal 1221.26 504.43 18.75 6.34 0.449
Source: Mgana et al., 1997.
Table 10.14 Sample of industries discharging liquid wastes into Msimbazi

stream through Vingunguti ponds, Dar es Salaam
Industry Pollution loads (t yr–1)
BOD0 SS Oil N P Cr Zn Fe
ALUCO Ltd 2.6

PIPECO Ltd 23.2 0.2 0.2 0.1 1.9
GALCO Ltd 92.8 0.8 0.6 0.3
STEELCO Ltd 17.4 276.9 29.9
Steel Cast Ltd 6.5
Vingunguti Abattoir 0.7 0.9 0.3 1.2 0.13
Subtotal 18.1 400.3 32.8 1.2 1.13 0.8 0.4 1.9
Grand total 1 239.56 904.73 51.55 7.54 1.579 0.8 0.4 1.9
Source: Mgana et al. (1997).
6.3 Point pollution sources because of mining activities

A pollution source which has to be mentioned because of its high toxical
threat is gold mining. Small-scale gold mining is associated with mercury
usage. The mines are found in many parts of the country, including areas
around Lake Victoria. Between 1990 and 1994, the Bank of Tanzania
encouraged small-scale miners to sell gold through the bank by providing
them mercury. The Bank bought 17.8 tons of gold during that period (Bank
of Tanzania, 1997). It takes two parts by weight of mercury to produce one
part by weight of gold.
Table 10.15 Sample of stationary and mobile air emissions, Dar es Salaam region
Pollution loads (t yr–1)
TSP SO NO CO VOC H 2S HC Pb SO3
A1: Stationary sources, industry

Tiper 0.2 5.7 1.6 0.4 0.25 0.1
Tanganyika Tegry 0.0036 0.0021
Simba Plastic Ltd 1.14 0.498
Berger Paints Ltd 4.95 98.96
BANKO Products Ltd 6.31 126.2
Ubungo Spinning 1.3
Organic Chemicals Ltd 2.28 4.3
Galasy Ltd 0.37 0.6
Sadolins Ltd 6.6 9.8
Kioo Ltd 23.6 44.4 98.6 2.8
Kibo Paper Mill 105.3 4.1 6.0 7.0
Kiuta (NCP) 400.00
Aluco 12.4
Tanzania Brew Com. Ltd 36.5 11.4
Dar-Brew Con. Ltd 11.1 3.5
Ubungo Garments 0.7
Tanzania Fishnet 0.6
Tanzania Portland Cement Ltd 2 185.1 323.0 323.0
Ubungo Power Station 8.73 245.8 118.2 26.9 9.7
Steel Cast 124.2 8.1 1 313 5.8
DSM Agricultural Activities 3 038 15 714 1 672.2
Subtotal 5 569.3836 631.1 541.4 17 054.3 2 346.2101 7 5.8 0.1
257
258
TSP SO NO CO VOC H2S HC Pb SO3
A2: Stationary sources domestic combustion

Kerosine 19.8 11.0 142.8 39.0 1.9
LPG 0.4 0.04 12.6 2.6 0.2 0.2
Charcoal Manufacturing 77 780.9 7 017.8 100 588.9 91 816.6
Charcoal 3 275 46.8 397.7 198 83.8 10 058.96
Subtotal 81 076.1 57.84 7 570.9 120 514.3 101 877 0.2

B: Mobile sources
Land transport 581.6 813.8 3 216.3 16 757.5 2 265.4 37.3
Grand total 87 227.1 1 502.7 11 328.6 154 327.3 106 489.2 7.0 43.1 0.2 0.1

Table 10.16 Sample of stationary and mobile air emissions Tanga municipal
TSP SO2 NO CO VOC
A1: Stationary sources industry

Amboni Plastic Co. Ltd 9.8 4.4
Tanga Cement Factory 2 354.6 341.7 720.2
Wood fuels (for salt 429.8 6.1 52.2 2 609.5 1 320.1
production)
Subtotal 2 794.2 347.8 772.4 2 609.5 1 324.5
A2: Stationary sources, domestic combustion

Kerosine 4.4 2.5 31.9 8.7 0.4
LPG 0.0014 0.00016 0.05 0.0097 0.0039
Wood fuel 3 511.3 50.2 426.4 21 318.4 10 784.6
(domestic)
Subtotal 3 515.7014 52.70016 458.35 21 327.1097 10 785.0039
B: Mobile sources
Land transport 52.8 69.1 253.0
Grand total 6 362.7 469.6 1 483.8 23 936.6 12 109.5
In Tanzania, gold recovery using mercury is estimated to be about 40 per

cent by weight of the total production, the rest being mined by other
methods such as cyanidation or pure nuggets from reef mines. Therefore,
the amount of mercury released in the environment surrounding the gold
mining areas, producing 17.8 tons of gold, is about 14.3 tons. These are the
official figures. Gold being sold in the black market, often in exchange for
mercury, is unrecorded. It is, however, estimated to be between 50 and
70 per cent of the total sold. Thus the amount of mercury pollution is
higher than that officially recorded.
The first stage of mercury poisoning occurs during amalgamation. The
mercury is dispersed in the slurry, containing gold particles, with bare
hands to form the gold alloy known as amalgam. The amalgam is then
heated in an open container to vaporise the mercury at around 400–420 °C.
The mercury vapour released is even more dangerous than other forms of
mercury, because of the high reactivity with the body and other living
organisms. It is easily absorbed by the body through the human skin and
also inhaled through the respiratory tract. The rest of the mercury finds its
way to animal and plant life as well as marine life, since panning is done
near the rivers or dams. The study of mercury levels in humans in mining
areas near Lake Victoria has shown that urinary mercury levels of those fre-
quently exposed to mercury vapours when heating mercury–gold amalgam
were quite high (mean 241 ng ml –1) compared to others not exposed to
mercury (mean 2.6 ng ml–1). Soil around the mining areas had an average
mercury concentration of 3.4 ug g–1. This in contrast to the regional back-
ground concentration of 0.05–0.06 ug g –1. The study showed low mean
levels of mercury in fish (8.9 ppb) in Lake Victoria.
According to the WHO environmental health criteria (WHO, 1976),
mercury concentrations in freshwater fish from non-polluted areas are
commonly in the range of 10–100 ppb.
7 Conclusions
This paper proposes techniques for rapid environmental assessment of

sources of water pollution. The suggested methodology elaborates on gen-
erally applied rapid assessment guidelines. Incorporation of penetration
factors in the methodology not only refines the assessment, but also pro-
vides information on the current and potential reduction of total waste
loads by domestic and natural treatment processes. By presenting the
results of a basic error analysis, insight is given into the reliability of the
results. The scarcity of information on (industrial) pollution intensities and
treatment efficiencies, typical for developing economies, still generates
major uncertainty that can only be overcome by multiple-year monitoring
schemes.
The Lake Victoria case study determines the main sources of organic pol-
lution of this largest of African lakes. Oxygen demanding waste loads
(BOD5) are highest on the Kenyan side of the lake, it being the most
densely populated industrialized part of the catchment area. Industrial
BOD loads create local problems, but their overall contribution is subordi-
nate to domestic loads. The contribution to the total nutrient input of
direct rainfall over the lake is surprisingly high. This outcome corresponds
to the observed increases in chlorophyll-a concentration and algal blooms
during rainy seasons, which can only partly be linked to a nutrient
upsurge from sediments (Ochumba and Kibaara, 1989). The highest uncer-
tainty in the overall nutrient balance relates to the uncertainty with
respect to the land-use nutrient-export coefficients and average rainwater
concentrations.
The verification of the nutrient loads through the application of a simple
nutrient-mixing model results in loading estimates only about 15 per cent
below the most likely rapid assessment results. This close proximation
enhances the reliability of the results of our rapid assessment estimations.
The rapid assessment based on the national census shows that the instru-
ment yields very valuable information, even though the results of this
assessment have to be used with care since the data of the census are ques-
tionable. The results of the analyses are robust enough to justify the con-
clusions that the most important subsectors to include in a environmental
policy concerning the Tanzanian industry are the paper/paper products/
printing/publishing sector for organic pollution and the food and furniture
sectors for their contribution to the eutrophication problem.
An interesting outcome of the case study is that an environmental policy
directed at reducing nutrient loading, in order to eradicate eutrophication,
should not concentrate on domestic liquid waste reduction. Rather, policies
should be developed to reduce the leaching of nutrients from agricultural
land (for example, the introduction of improved manure management and
enhanced farm systems). Also, the exact causes for the observed high con-
centrations of nutrients in rainwater should be investigated (for example,
forest and biomass burning, volatilization of N compounds in manure and
wind erosion), and policies should be directed to counteract these.
Although uncertainties remain, the outcome of the rapid assessment
methodology has proven very useful. Also, the introduction of a basic error
analysis, and several verification steps, has increased the method’s credibil-
ity. In order to further reduce the probability interval, several steps in the
methodology should be investigated. In the case of Lake Victoria, research
on waste loads from pit latrines and septic tanks, on run-off and leaching
of nutrients from different types of cultivated and non-cultivated land, and
on nutrient concentrations in rainwater would be the major focal points.
Also, the evaluation of some other potential nutrient sources, such as that
of dry deposition and nitrogen fixation by cyanobacteria, could be intro-
duced. Especially the latter should be considered for the case of Lake
Victoria, where algal blooms of cyanobacteria occur frequently, introducing
a large potential source of nutrient input. Further evaluation would,
however, require more advanced modelling techniques, incorporating also
other nutrient processes such as nitrification and denitrification. However,
such would be outside the scope of rapid assessment studies in general.
Available data for industries in Dar es Salaam and Tanga justify the con-
clusion that urban industry contributes considerably to pollution. Gold
mining is an important and dangerous source of point pollution.
Notes
* Centre for Technology for Sustainable Development, Eindhoven University of
Technology
** UNIDO, Côte d’Ivoire
*** Environmental Association of Tanzania
1. The settling velocity is calculated by applying: VSA = rS*SA/CA. The average values
for the sedimentation rate rS (g m–2 yr–1) and the concentration of nutrient A in
the sediment SA (mg g–1) (for both N and P) were taken from Hecky (1993).
11
Energy Conservation in the
Industrial Sector in Tanzania
Frank van der Vleuten*, Lex Lemmens**, Otto Bos***, Caspar
Samplonius***, Dick Toussaint*** and Michel Yhdego****
1 Introduction
The Energy Policy of Tanzania (MWEM, 1992) clearly specifies the impor-
tant role energy plays in the development process: ‘Energy is a prerequisite
for the proper functioning of nearly all subsectors of the economy. It is an
essential service whose availability and quality can determine the success or
failure of development endeavors’ (par. 2). ‘There cannot be sustainable
development and the satisfaction of basic needs of society without
sufficient and efficient supply and use of energy’ (par. 26). Since Tanzania
does not (yet) have at its disposal many indigenous energy sources, the use
of energy, however, is a heavy burden on the balance of payments and thus
on the national economy as a whole. While the long-term strategy of the
National Energy Policy is the reduction of dependence on external energy
sources and the exploration and rational management and utilisation of
the country’s own resources, the short- and medium-term strategies
include, among others, more efficient use of energy in the transport and
industry sector.
The Tanzanian industrial sector offers ample scope for energy efficiency
improvement, in particular in case studies involving the cement and beer
industries. However, in general opportunities have also been identified to
‘leapfrog’ to a cleaner production path, where industrial development can
be better tuned to the available resources and the natural environment.
This contribution discusses the energy dimension of the industrial sector of
Tanzania and shows where and how energy efficiency can be enhanced.
2 Case for energy conservation in the industrial sector
The National Energy Policy of Tanzania recognizes the importance of

energy conservation and mentions more efficient use of energy in the
industrial sector as one of the main short- and medium-term strategies
(MWEM, 1992). In the industry sector, a major policy objective of the
262
Energy Conservation in Industry 263
National Energy Policy is the progressive reduction of dependence on

imported petroleum products.
There are, however, more reasons to conserve energy:
• Reducing the cost of energy by conservation provides an attractive eco-

nomic option at the macroeconomic level. For Tanzania as a whole, the
cost of energy conservation may very well be significantly lower than
the cost of energy production. In terms of delivering energy services for
economic development in a cost-effective way, energy conservation may
well be able to compete with the chain of energy production, conversion
and distribution.
• At the company level, energy conservation may offer a viable invest-
ment opportunity. Pay-back times for such investments have been
found to be as short as three to six months. This was confirmed by a pre-
investment study addressing the industrial energy conservation poten-
tial, implemented by Environmental Resource Consultants in Dar es
Salaam (van der Vleuten, 1995).
• In several other countries, investments in energy-efficient technology
have been found to have additional benefits in production efficiency,
production quality or volume.
• In many industries energy conservation is one of the priorities to attain
cleaner production. This includes reducing environmental impact from
direct energy use within the industry – CO 2, dust, NOx, SO2, etc. – as
well as reducing emissions from energy production in, for example,
power generation.
• Although anthropogenic emissions of greenhouse gases in Tanzania are
relatively small, when compared to the emissions of more industrialized
societies, the country offers an attractive potential for climate change
mitigation. The Framework Convention on Climate Change (FCCC)
emphasizes that international cooperation should be promoted in order
to prevent climate change in the most cost-effective way. Cleaner pro-
duction technology programmes within industry present the opportu-
nity of a win–win situation in which greenhouse gas emissions are
reduced in a cost-effective way. By enhancing the efficiency of the tech-
nology used in industrial processes and stationary combustion systems,
the industrial sector of Tanzania can attract foreign investments. The
Kyoto protocol to the FCCC includes this option under the Clean
Development Mechanism.
3 Energy conservation in practice
The central concept of industrial energy conservation is to reduce the

quantity of energy needed to produce a certain industrial output. Energy
conservation necessarily takes into account the whole energy system:
energy sources, energy carriers (or fuels) and conversion technologies. The
efficiency of the energy system should be enhanced in such a way that the
same energy service can be supplied with less primary energy source/fuel,
or that more energy services can be supplied with the same quantity of
fuel.
The cement industry, which will be discussed in more detail in Section 5,
illustrates the various approaches towards energy efficiency. Table 11.1 pre-
sents the figures for an average cement plant in Africa (van der Vleuten,
1995). The cement industry is one of the most widespread energy-intensive
industries in Africa. For the production of cement, raw materials are ground
and burned in a kiln to produce clinker. The clinker is ground with
gypsum, and occasionally other additives, to produce cement. The con-
sumption of primary heat to produce one ton of cement not only depends
on the efficiency of the technology and the management of the clinker
manufacturing process, but also on the quantity of clinker used to produce
cement. This illustrates that the energy and the materials used in produc-
tion have strong interactions.
From Table 11.1 it is apparent that:
• significant energy saving can be achieved by better process management
and good housekeeping without significant cost
• energy efficiency of a cement plant not only depends on the energy
efficiency of the kiln, but to a large extent on the materials used in
grinding; this underlines that material efficiency (of energy-intensive
materials) can be very important to attain energy efficiency.
Table 11.1 Heat efficiency in cement production
Primary
heat 4200 MJ 3600 MJ 3600 MJ 3000 MJ
Clinker Typical dry Efficient Efficient State-of-

Technology management management the-art process
Product 1 t clinker 1 t clinker 1 t clinker 1 t clinker
Additives 5% gypsum 5% gypsum 5% gypsum, 5% gypsum

30% fly ash
Product 1.05 t cement 1.05 t cement 1.35 t cement 1.05 t cement
Process 4 000 MJ/t cement 3 429 MJ/t cement 2 667 MJ/t cement 2 857 MJ/t cement
efficiency
Added costs Baseline Very low Operational Investment

In view of the strong case for energy conservation, it is surprising that it

gets almost no attention in the industrial sector in Tanzania, particularly
since studies have shown that a substantial saving potential exists in tech-
nically and economically viable projects. The first efforts to increase energy
efficiency in Tanzanian industry, based on technical energy audits, date
back to 1982. At present more than 70 firms have been audited by the
Tanzanian Industrial Research and Development Organization (TIRDO)
(Tarimo, 1995). Generally the audits were performed by outside experts
who made recommendations to the management of the companies
concerned.
Apart from five industries that requested self-funded audits, all audits
were financed by international donors. The various audits showed a vast
potential for energy conservation, and TIRDO estimated the overall energy
conservation potential at 30 per cent of industrial consumption. Despite
the forecasts of good economic returns, however, very few companies have
been able to implement the recommendations (Mwandosya and Luhanga,
1993). To analyse why Tanzanian industry has not yet capitalized on the
vast potential for energy conservation, two case studies are discussed: the
beer and the cement industry.
4 Beer industry
4.1 Tanzania Breweries Limited

Owned by East African Breweries and under British management, the Dar
es Salaam brewery was established in 1933. In 1966 a new brewery was
commissioned on the same site. East African Breweries was nationalized in
1970 and became part of Tanzania Breweries Ltd. The capacity of the
brewing and bottling facilities of the Dar es Salaam plant were increased in
1984 and 1991. Starting in the 1980s, however, the production process
gradually deteriorated because old equipment was not replaced in time,
and maintenance was insufficient. In the early 1990s the management
started looking for a foreign partner. In November 1993 this partner was
found in Indol International Holding, a Dutch investment organisation
owning South Africa Breweries (SAB). SAB became directly involved in the
management of Tanzania Breweries Ltd (TBL). The energy situation at TBL
is analysed at the end of 1993, when the new management came in.
4.2 Energy efficiency at Tanzanian Breweries Limited

A general impression of the energy efficiency of the brewery can be
obtained by looking at the overall specific energy consumption and com-
paring the results with performances in other breweries. Table 11.2 shows
the specific energy consumption: the consumption of fuel oil and electric-
ity per hectolitre (hl) of beer in 1992 and 1993. Since water consumption is
related to energy consumption too, it is included in the table. The table
Table 11.2 Specific fuel oil, electricity and water consumption at TBL
1992 1993
Fuel oil
l/hl brewed beer 8.0 8.8
l/hl bottled beer 10.5 9.4
l/hl sold beer 11.0 10.1
Electricity
KWh/hl brewed beer 23.1 27.4
KWh/hl bottled beer 30.4 28.9
KWh/hl sold beer 31.6 31.4
Water
hl/hl brewed beer 34.9 44.8
hl/hl bottled beer 46.1 47.4
hl/hl sold beer 47.5 51.2
shows large differences in specific energy and water consumption between

brewed, bottled and sold beer, implying significant losses in the production
process.
In Table 11.3 the specific energy consumption of TBL (1993) is compared
with that of other breweries. TBL apparently consumes much more energy
than the reference breweries. It should be taken into account, however,
that TBL operates in a warmer climate than the European breweries listed.
This implies that more energy, mostly electricity, has to be used for
cooling. On the other hand, the total time cooling is needed – during fer-
mentation and conditioning – is shorter. At TBL total cooling time is about
24 days per batch of beer. For most of the European breweries the cooling
time is 4 to 12 weeks. From Table 11.3 it is obvious that TBL uses too much
energy and water. The causes, and thus the possibilities for improvement,
are analysed in the following paragraphs. First, however, an impression is
given of the economic importance for energy efficiency improvement at
the plant.
4.3 Energy costs

The importance of controlling energy costs is illustrated by looking at these
costs as a percentage of total costs. Table 11.4 shows that the so-called
semi-fixed costs (electricity, fuel oil, water, refrigeration and steam) add up
to 21.6 per cent of the total direct production costs. If energy costs at TBL
could be reduced by 30 per cent (including costs for water), direct produc-
tion costs would drop by 6 per cent.
Another way of assessing the importance of energy-cost reduction is the
effect on the profitability of the company. In order to include all costs,
including overheads at the TBL head office, the results are calculated for all
Table 11.3 Specific energy and water consumption at TBL compared to reference breweries
Brewery Beer output (hl/year) Specific fuel oil Specific electricity Specific total Specific water
consumption consumption energy consumption consumption
(l/hl beer) (kWh/hl beer) (MJ/hl beer) (hl/hl beer)
TBL (1993) 300 000 10.1 31.4 579 51.2

Germany 155 000 5.3 11.3 257 –
Germany 100–500 000 3.2 10.3 168 6.5
Germany 100–250 000 5.3 13.2 264 8.5
Ireland 361 000 6.2 11.7 296 –
Turkey 865 000 4.4 9.0 212 –
Sources: Ehrhorn (1982); Schu (1990); TIRDO (1990); Mayer (1992).
267
Table 11.4 Cost attribution of direct production costs per crate of beer
(TBL, 1994)
Cost attribution TSh (1994) %
Brewing materials 692 50.3

Bottling materials 269 19.6
Direct labour 36 2.6
Semi-fixed costs 297 21.6
Electricity 113 8.2
Fuel oil 103 7.5
Water 62 4.5
Steam/refrigeration 18 1.3
Repair & maintenance 82 6.0
Total direct costs 1 377 100.0
subsidiaries of TBL rather than for the Dar es Salaam branch only.
Table 11.5 shows that a net reduction of energy costs of 30 per cent causes
a gross profit increase of 43 per cent.
It is interesting to compare the specific energy costs of TBL with those of
Western breweries. The average specific energy consumption and costs for
German breweries in 1991 are used as a benchmark (Mayer, 1992). These
breweries had a production between 100 000 and 500 000 hl per year. The
specific energy consumption is based on 1993 figures, but to get a more up-
to-date profile energy prices of January 1994 are used. Figure 11.1a shows
that specific energy costs for TBL exceed those of Western breweries by a
factor of two. Energy prices for the reference breweries, however, are
higher, and when this is accounted for the energy costs of TBL exceed
those of Western breweries by a factor four. See Figure 11.1b.
All this shows that the reduction of energy use by conservation is impor-
tant from a financial point of view.
4.4 Energy conservation opportunities

A systematic inventory was made of direct and indirect factors affecting the
energy use (Toussaint, 1994) and thus the energy costs. The factors with
Table 11.5 Effect of cost reduction by energy conservation on value added

and gross profit (TBL, 1994)
Cost reduction by Increase of value added (%) Increase of gross profit (%)
energy conservation (%)
10 2 14
20 4 29
30 6 43
Figure 11.1 Specific energy costs at TBL in comparison with Western breweries
Western
Western brewery
brewery US$ 3.1/hl
US$ 3.1/hl extra
at TBL extra
+ US$ 3.5/hl at TBL
+ US$ 8.7/hl
a: b:
at equal energy prices at equal energy prices
the highest potential, expressed in high profitability, low investment

requirements and a relatively small foreign currency need, were analysed. A
review of these most important energy conservation opportunities (ECOs),
together with their financial indicators (when available), is presented in
Table 11.6.
It is obvious that increasing the capacity utilization and decreasing the
production losses are the most promising options for saving energy. Apart
from these two ECOs the total energy-cost savings of the other opportunities
could add up to 300 million TSh (US$590 000). This is an energy cost reduc-
tion of 1000 TSh per hectolitre, which equals a 30 per cent reduction. All
measures together could reduce specific energy costs by around 50 per cent.
Not all implementation costs and pay-back periods could be assessed. For
the measures for which costs could be assessed, the average pay-back period
is two months. The average foreign currency component of quantified costs
is about 75 per cent. This percentage is quite high, but foreign-currency
requirements as a percentage of total energy-cost savings is as low as
13 per cent.
4.5 Causes of energy inefficiency

The energy conservation opportunities are, in a way, also indicative of the
problems at the Dar es Salaam brewery. Productivity has obviously been
very low at TBL. In 1993 the Dar es Salaam plant produced about 300 000
hectolitres with 950 employees. A typical Dutch brewery, with a com-
parable production process, produces about 3.5 million hectolitres with
650 employees.
One of the reasons for the inefficiency is a problem with the bottling
lines. Two new bottling lines were installed at the end of 1991. Since their
installation they never reached their expected output. Though they were
bought as new, it was suspected that they were second-hand. Many break-
downs occurred, which delayed production seriously. Besides, the new lines
consumed more water than the old ones. All this coincided with raw water
supply shortages, resulting in even more production stops. Also, electricity
power cuts decreased the beer output.
270
Table 11.6 Overview of quantified feasible energy conservation opportunities
Energy conservation opportunities Savings Costs Foreign currency Payback

(000 TSh per year) (000 TSh) required (% of costs) period
1 Capacity utilization increase 169 000 000 – – –

2 Switch off air cooling of rooms if not in use 32 000 0 0 0
3 Mothballing old vertical fermenters 8 000 0 0 0
4 Air cooling load reduction 95 000 p.m. 0 –
5 Decrease of production losses 16% to 8% 37 600 000 p.m. p.m. –
6 Tightening leaks in refrigeration system 48 000 p.m. 0 –
7 Improvement of power factor 21 400 – 100 –
8 Insulation of water pipe lines and tanks 14 800 1 700 100 1 month
9 Water reuse in bottling hall 18 000 3 000 20 2 months
10 Improvement of compressed air system 16 300 3 300 30 3 months
11 Water from Wort cooling also to mash water tank 12 400 210 50 <1 month
12 Insulation of refrigeration piping 8 600 600 100 4 months
13 Improvement of load factor transformers 8 900 3 000 100 4 months
14 Removal of heat exchanger fermentation room 1 7 900 ? 0 ?
15 Improvement of water supply Wort cooler 6 700 680 65 1 month
16 Condensate return mash vessel 835 220 70 3 months
17 Tightening steam leaks 120/mm2 20/mm2 0 2 months
18 Insulation steam pipe lines 2 200 4 000 100 2 years
19 Insulation condensate pipe lines 2 720 5 420 100 2 years
The brewing section had other problems. The refrigeration system was
not able to keep all sections at the prescribed temperature. The main
reasons were old age and bad maintenance of the refrigeration system. In
particular, the compressors often broke down. Other problems related more
to the performance quality. Filled bottles often contained too much air,
reducing the shelf life of the beer and hence provoking customer com-
plaints. Improvements requiring investments were often not possible
because of a continuous liquidity problem. In addition, as can be seen in
Table 11.2, beer disappeared during the production process.
The conclusion with respect to this case study is that energy is an inher-
ent aspect of the functioning of the entire plant. Energy use is closely
linked to use and handling of water, steam and the product. The waste of
energy in the brewery can, therefore, be largely attributed to the perfor-
mance of the management. The major improvements that took place soon
after South African Brewers took over the management support this conclu-
sion (see also Chapter 15).
5 Cement industry
5.1 Tanzania Portland Cement Corporation Limited

Until 1966 Tanzania imported all its cement. In 1959 the Tanzania
Portland Cement Company Ltd (TPCC) was registered with the objective of
importing bulk cement, and bagging it at its Malindi Packing Plant. This
plant was situated at the Dar es Salaam harbour. In 1962 the Tanzanian
government empowered the TPCC to put up a cement plant of its own.
Subsequently, the first cement factory was constructed at Wazo Hill, some
20 km north of Dar es Salaam. Production started in 1966.
The plant was owned and managed by Cementia Holding AG in
Switzerland. It had one kiln with a production of 50 000 tons per year (tpy)
in 1966 (Figure 11.2). This increased to 177 500 tpy in 1971. On the basis
of a forecasted increase in the demand for cement, a second kiln was
installed in 1972, increasing the production capacity to 270 000 tpy. In
1973 the Tanzanian government took over the ownership of the plant and
transferred it to the State Mining Corporation. The Cementia management
experts withdrew in 1973. The management of the plant was taken over by
Indian expatriates. At the same time the international and thus national
economic situation changed drastically, affecting also the foreign exchange
availability for the TPCC. As a result, it became harder to buy spare parts
and consumables. The capacity utilization started to drop, and by 1979 had
reached 76 per cent.
In 1979 a third kiln was added and the production capacity increased to
520 000 tpy. In 1980 the requirements with respect to foreign exchange for
capital goods replacement, spares and consumables were around
US$2.5–3.0 million per year. Hardly half of this amount was available
Figure 11.2 Cement production of TPCC, 1966–1993
500 Total cement production 1966–1993
400
tonnes as ( x1000)
300
200
100
0
66 68 70 72 74 76 78 80 82 84 86 88 90 92
67 69 71 73 75 77 79 81 83 85 87 89 91 93
through the Bank of Tanzania and the aid from the Danish development
organization DANIDA. Performance consequently dropped to 55 per cent
of the rated capacity and further to 26 per cent in 1983. Early in 1983 the
last installed kiln (kiln 3) suffered a major breakdown. A solution for these
problems was found in contracting Cementia International AB (nowadays
Scancem) from Sweden and foreign exchange support from the Swedish
development aid organization SIDA. The capacity utilization slowly
improved to about 80 per cent in the early nineties. The present energy
analysis was made during the second half of 1993.
5.2 Energy efficiency at the TPCC

In view of the history of the factory, it can best be described in three units
linked to kilns 1, 2 and 3, installed in respectively 1966, 1972 and 1979.
The power use in the characteristic production stages of the different units
is presented in Table 11.7. With respect to the table it should be noted that
the collection of information was seriously hampered by the unreliability,
Table 11.7 Power use of the three units in kWh/ton cement
Unit 1 Unit 2 Unit 3
Crushing 1.8 1.8 1.8

Grinding 39 22 36
Blending 1.5 1.5 1.5
Burning 24 17 27
Finish grinding 44 25 38
Packing 1.6 1.6 1.6
Overhead 1.4 1.4 1.4
Total 113.3 70.3 107.3
sometimes even absence, of kilowatt-hour meters. Therefore, the figures in

Table 11.7 are not suitable for detailed conclusions. What can be con-
cluded, however, is that unit 2 is performing much better than the other
two. This can be explained by the fact that the mills of unit 2 were over-
hauled only a few years ago. Grinding and burning are using most of the
energy. The analysis of the grinding stage is rather straightforward. The
assessment of the burning stage needs a heat balance for the kiln. By com-
bining production data and our own measurements, the heat balance for
each kiln is calculated. The results are compared with the data of two refer-
ence kilns from the literature in Table 11.8 (Samplonius, 1994).
From Table 11.8 it is clear that most significant heat losses occur in the
pyro-processing department. The heat losses consist of exhaust gases
leaving the system, radiation and convection losses, and the heat content
of the clinker when leaving the system. The loss of heat in exhaust gases is
Table 11.8 The heat balance for the kilns compared to two reference kilns in
kJ/kg clinker
Kiln 1 Kiln 2 Kiln 3 Reference Reference

kiln 1 kiln 2
Input
Fuel
From sensible heat 21 20 21 13 14
From combustion 4 396 4 208 4 360 3 150 3 050
Raw meal
From sensible heat 116 116 116 54 92
From sensible heat of water 2 2 2 – –
Combustion air
From sensible heat of primary air 5 5 5 6 8
Total input 4 540 4 351 4 504 3 223 3 164
Output
Heat of formation 1 750 1 750 1 750 1 750 1 789
Evaporation of water from raw meal 12 12 12 13 17
Exhaust gas sensible heat 1 854 1 295 1 353 636 595
Dust sensible heat – 23 23 18
Incomplete combustion – – – –
Clinker sensible heat 62 62 62 63
Cooler exhaust gases – 0 0 423 520
Losses radiation & conv. kiln 862 903 759 297 243
Losses radiation & conv. cooler – 65 62
Losses radiation & conv. Pre-heater – 241 483
Rest 23
Total output 4 540 4 351 4 504 3 223 3 164

by a factor of two higher than in the reference kilns. The losses through
radiation and convection are two to three times higher. The heat remain-
ing in the clinker is comparable for all kilns. The reference kilns are a kiln
with a four-stage pre-heater and a grate cooler with a capacity of 2000–
3000 tpd (tons per day) and a kiln with a four-stage pre-heater, a pre-
calciner and a grate cooler with a capacity of 2800 tpd. The difference in
size of these reference kilns, compared to those of the TPCC, accounts for
part of the difference in radiation and convection losses. However, consid-
erable savings can be expected from reducing heat losses in this part of the
cement production process.
In the course of processing, the factory loses energy. Much more energy
is lost when cement production has to be interrupted. Whenever the kiln
has to be shut down, heat is wasted in the product since this has to be
rejected because it has not gone through the whole process. In addition
heat is wasted in starting the process up again. Production will remain low
until temperatures throughout the whole system are brought up to
normal.
It takes from thirty minutes to several hours to balance the system after
an interruption. Factors that cause kiln interruption can be internal or
external. For the TPCC, which gets its power from TANESCO, the external
factor which causes interruption is power supply. These power interrup-
tions affect not only energy efficiency but also production capacity use. In
1993 the kilns were down for about 600 hours because of power failure or
power reduction, which accounted for the major part of the inefficiency.
Internal factors that cause interruptions are unplanned maintenance and
lack of raw materials. These were not so important, but since they can be
directly affected by the management they provide important energy-saving
opportunities. In Table 11.9 the most important reasons for unplanned
stops (except power cuts) are presented.
5.3 Energy costs

Also in Tanzania, controlling energy costs is a major issue in the cement
industry. This can be seen from Table 11.10, which indicates that 46 per
cent of the total costs of the TPCC are energy costs. If one considers the
part energy costs play in the actual production process (Table 11.11), the
importance of energy savings becomes even clearer.
Table 11.9 Unplanned interruptions in 1993 (except power cuts)
Kiln 1 Kiln 2 Kiln 3
Loss of feed 8 7 8
Cyclone blockage 6 8 5
Waste gas fan tripped 3 8 4
Table 11.10 Breakdown of costs of TPCC (million TSh)
Personnel 1 180 10.3%

Electricity 2 020 17.6%
Fuel oil 3 050 26.6%
Diesel, petrol, lubricants 260 2.2%
Paper bags 780 6.8%
Gypsum 480 4.2%
Maintenance 500 4.4%
Interest 100 0.9%
Local freight 860 7.5%
Export freight 470 4.1%
Fees 610 5.3%
Other 1 150 10.0%
Total 11 460
Table 11.11 Breakdown of TPCC production costs (million TSh)
Mines Raw mill Kiln Cement mill Packing plant Total
Production costs 805 1 240 4 040 1 320 1 085 8 490

Fixed costs 530 280 455 265 260 1 790
Variable costs 275 960 3 585 1 055 825 6 700
Variable costs:
Energy costs 175 920 3 340 550 45 5 030
Other 100 40 155 505 780 1 580
Direct manufacturing costs in 1992 amounted to 8940 million TSh. The

fixed costs were 1790 million TSh and the variable costs 6700 million TSh.
The variable costs consisted for the larger part of energy costs: 76 per cent.
The main production costs were made in the kiln section and consisted for
the largest part of energy costs. It is clear that an energy saving of a few
percent would already make an important contribution to the improve-
ment of the economic performance of the TPCC.
5.4 Energy conservation opportunities

5.4.1 Individual power station
The major problems of the TPCC consist of power cuts. To reduce
unplanned kiln interruptions it is almost essential to acquire an individual
power station. The reported 600 hours’ down time because of electricity
problems can be converted to a production loss of 10 500 tons of cement
or 155 million Tsh (Samplonius, 1994). The extra fuel, which is needed for
the start-ups, costs an additional 122 million Tsh. An individual power
plant of 6 MWh would cost 1500 million Tsh. The anticipated pay-back
period is five years. In view of the high energy prices of TANESCO, an
additional advantage could be a lower kWh price for the self-generated
electricity.
5.4.2 New measurement systems for the kiln

Fuel combustion systems in the kiln are major contributors to energy
inefficiency and refractory damage. All three kilns have defective meters, of
which the gas analysers are essential to control the kiln. With new measur-
ing equipment an energy saving of 20 per cent would be easily feasible. The
savings in one year would be 320 million Tsh, many times the investment
needed.
5.4.3 Reduction of radiation and convection losses

To counteract the radiation and convection losses, insulation bricks can be
used. The expected energy savings would be 122 million Tsh. The costs for
the bricks would be 80 million Tsh. Taking into account the labour costs,
the pay-back period for this measure would be close to one year.
5.4.4 Replacing electrostatic filters

The cement factory emits enormous quantities of dust through the exhaust
pipes of the kilns. For kiln 3, the emission is estimated at 2.2 tons per hour.
The other two kilns add another ton per hour to this discharge. Dust is
neither crushed raw material nor clinker, but it has a value since it has been
mined, crushed, milled and partly burned. If the value of crushed raw
material is used to estimate the losses, the missed proceeds are at least
40 million Tsh per year. The costs for replacement or reparation of the elec-
trostatic filters could not be assessed, but the pay-back period will not
exceed a couple of years. Stopping the dust emissions is necessary in the
first place because of the environmental pollution.
5.4.5 Long-term options

Since early 1970, more and more cement factories have had a pre-cal-
ciner. The essential factor of a pre-calcining process is the addition of a
separate combustion chamber to a conventional pre-heater. The main
accomplishment of the pre-calciner is to provide direct heat in the most
effective manner possible where it is most needed: in the calcining stage
of the clinker production process. Pre-calciners increase the production
capacity of an existing kiln by a factor of two. They increase refractory
life, reduce the production of NOx and make it possible to replace part of
the energy input by waste fuels. The investment is considerable and the
layout of an existing factory needs to be changed. However, the expecta-
tion is that for the TPCC the installation of a pre-calciner is financially
viable.
5.5 Causes of energy inefficiency

The energy conservation opportunities presented in the last section
already indicate that there are three major reasons for the energy
inefficiency at the TPCC. First and foremost the basic instruments and
management structure necessary to run the plant in an optimal way are
not present. Better maintenance and a better attitude of the staff towards
the use of energy would also increase efficiency. The technology used for
processing is the second reason for energy losses. There is a need for the
replacement and introduction of modern equipment. An important exter-
nal reason is the dependence on an unreliable electricity supply by
TANESCO, which is causing many kiln interruptions, with the reported
consequences.
The reasons for the lack of implementation of new technologies and the
poor operation and maintenance are both internal and external. These
reasons coincide with those for general inefficiency of public enterprises.
The lack of foreign currency in the country is the major bottleneck to
implement even those measures that have a very short pay-back period.
6 Conclusions
Both cases confirm the high potential for energy conservation suggested by
earlier energy audits in Tanzania. Both in energy-intensive industries and
less energy-intensive plants, energy conservation presents an opportunity
to enhance the profits of the industry, to improve the manufacturing
process and to improve the environmental performance.
At the technical level, the main constraints with respect to more energy-
efficient industrial operations include the following:
• Poor plant management results in inefficient operation and poor main-

tenance of the present production technology. In addition, underutilisa-
tion of the production capacity causes a relatively high energy intensity
of the production.
• Production technology is generally relatively old and inefficient. Many
cost-effective and indeed highly profitable energy conservation options
exist. As a result of limited access to investment capital, the plants
cannot capitalize on these opportunities. This results in a missed oppor-
tunity to enhance the energy efficiency and the profitability of the
plant.
• The energy infrastructure poses a serious obstacle to industrial devel-
opment, especially for industries where energy is a crucial input.
This is clearly illustrated by the case study in the cement sector, where
unreliable power supply is a main obstacle on the road towards
a stable, that is, energy-efficient and profitable, operation of the
plant.
The case studies illustrate that energy conservation demands an institutional

structure that supports improved operations and investments in energy
efficiency. Such a structure should include a mechanism of incentives for all
people involved in production. The studies show that the inefficiency of
energy use is, apart from the technical reasons mentioned, to a large extent a
management issue. This explains why the vast potential for profitable energy
conservation, as identified by previous audits, has hardly been exploited.
7 What can be done?
The previous outcomes of energy audits, as well as the case studies, under-
line the large potential for financially and environmentally profitable
investments in energy efficiency existing in the industrial sector in
Tanzania. This potential can be seen as a resource, the exploitation of
which is in line with the sustainable development of Tanzania.
7.1 What can industries do?

The conclusions suggest that energy conservation should not start from
investments in technical measures. Rather, the first point of attention
should be to develop the management capacity and to put into place the
incentives needed to enhance the efficiency of present operations, and to
make use of the economies of scale offered by a full-capacity utilization.
This is not to say that energy conservation can only take place after a full
restructuring of the management of a plant has taken place. On the con-
trary, energy conservation may be a vehicle for enhancing the efficiency of
a production plant, by primarily focusing on the efficiency of this crucial
input to all phases of the production process.
A first priority would then be to establish a functional unit within the
institutional setting of a production plant that has the mandate and the
capacity to enhance energy efficiency throughout the plant. This means
that capacity building of sufficient staff is very important. Furthermore, the
staff should receive the mandate and support of the plant management to
identify, implement and evaluate changes to the operation. In this way, a
sufficient energy-efficiency improvement can be achieved, mainly from
better operation and maintenance.
In the second stage, the industry should then focus on the identification
of investment opportunities in energy conservation. Using an analysis of
the energy flows, as well as the material flows of the production system, the
plant management should identify the most attractive investment options
in terms of profitability of the investment, contribution to energy
efficiency and contribution to the total plant efficiency. The plant manage-
ment should allocate an internal budget for low investment options. For
high investment options, external investment capital sources have to be
identified and developed.
7.2 What can policy makers do?

Policy makers have a variety of economic and environmental interests to
promote industrial energy efficiency. Although the prime capacity for
enhancing industrial energy efficiency in Tanzania lies with the manage-
ment of the production plants, policy makers can certainly stimulate this
process. In line with the priorities for industry, the following elements are
suggested for a government policy on energy conservation:
1. The first is to create awareness with respect to the benefits of energy

conservation. As a main target group, the production plant manage-
ment should be involved in discussions on the benefits and costs of
energy conservation, both from the perspective of the public interest
and the perspective of the production itself. In this way, the right pri-
orities can be set for the further cooperation between the public and
the private sector (or semi-public sector) to attain higher energy
efficiency.
2. The second aspect in which the government should have a role to play
is the stimulation of capacity building of the people that have been
empowered to implement the energy conservation programme within a
production plant. This step should build on the experience that has
been gained with the CEPITA (Cleaner Environment Production in
Industry in Tanzania) cleaner production project in Tanzania.
3. A third aspect in which policy makers have a role to play is in stimulat-
ing the development of financial channels for industries to be able to
access capital for investments in energy efficiency. To allow for effective
financial channels, this should primarily focus on increasing the accessi-
bility of already present financial institutions. In addition, policy makers
should contribute to setting up new financial arrangements, such as an
energy-efficiency revolving fund.
4. It is important to notice that foreign funding of an energy conservation
programme may be very welcome in view of highly limited government
resources. Policy makers should play a role here by enhancing coopera-
tion with non-governmental organisations and by taking the initiative
to acquire funding for this type of project. The climate change mitiga-
tion and technology transfer programmes (such as Activities
Implemented Jointly and the Clean Development Mechanism) are very
suitable instruments for this.
If industry and policy makers succeed in developing an approach to

energy conservation based on the elements of empowerment, capacity
building and technology investments, industrial energy conservation
offers a win–win–win situation. It will increase the economic and financial
performance of the industrial sector, lower the dependence on expensive
external energy sources and enhance environmental performance. Energy
conservation can be an important element in the sustainable development

of the industrial sector of Tanzania.
Notes
* Free Energy Europe
** Centre for Sustainable Development, Einhoven University of Technolog
*** Section of Technology and Development Studies, Faculty of Technology
Management, Eindhoven University of Technology
**** Environmental Resource Consultants, Dar es Salaam
Part IV
Lessons from Past Experiences,
Economic Reform, Prospects for
the Future
12
Industrialization of Tanzania: Can
Tanzania Learn from European
History?
Paul Lapperre*
1 Industrialization in historical perspective
Industrialization refers to the course of transition from a preceding dom-

inantly agricultural society towards an industrial one. In the course of this
transition large-scale industry becomes the most characteristic form of pro-
duction, a form in which the simplification of work to make it better man-
ageable (rationalization), the replacement of manual labour by that of
machines (mechanization) and the shift of control, correction and feedback
activities from peoples to machines (automation) become increasingly
important. Rationalization, mechanization and automation usually go hand
in hand with the concentration of labour in larger working units. Concurrent
with the shift from agriculture towards large-scale industry, far-reaching
changes take place in almost all other spheres of life (Lapperre, 1992).
When Britain entered the industrial age in the late 18th century, and
rationalization, mechanization and automation took hold, this heralded
the end of the era of the traditional agrarian societies, societies which first
came into being in the flood plains of the Euphrates and Tigris some
5000 years earlier. Although Britain was the first to make the transition,
industrialization soon spread to the Continent and North America. Later,
other parts of the world followed.
Now, at the beginning of the 21st century, approximately one-quarter of
the world’s population lives in highly industrialized societies and generally
enjoys a level of material wealth and well-being never experienced in
history before.
Although most of the remaining three-quarters of the world’s population
is affected by the industrial mode of production in one way or another, the
transformation from agrarian to industrial society has been imperfect and
is not, or only to a lesser extent, accompanied by a rise in material wealth
and well-being.
283
2 Present industrialization worldwide
An idea of the present state of industrialization worldwide can be obtained

by grouping countries, on basis of income per capita, in four categories –
low-income, lower-middle-income, upper-middle-income and high-income
economies – and by comparing these categories with respect to the percent-
ages of the labour force employed in agriculture, industry and services, and
the value added contribution to GDP in agriculture, industry and services
(see Table 12.1).
Table 12.1 shows that a rising per capita income goes hand in hand with
a significant decline of the labour force in agriculture and an increase of
that in industry and particularly services. When the per capita income
grows, the value added percentage of GDP for agriculture declines. For
industry it remains comparatively constant and for services it rises. In
1990, 47 out of 49 countries of the low-income economies employed more
people in agriculture than in industry. Out of the 41 countries of the
lower-middle-income and the 17 countries of the upper-middle-income
economies, respectively 22 and 4 employed more people in agriculture
than in industry. In all of the 26 countries of the high-income economies,
the agricultural labour force was far smaller than the industrial one.
Based on the labour force employed in agriculture, in 1990 a total of
73 countries were classified as dominantly agricultural and 60 as domi-
nantly industrial. There is no reason to believe that the situation has
significantly changed since.
3 Industrialization in Tanzania
The coming to power of conservative governments in many indust-

rialized countries in the 1980s was accompanied by a free-market
Table 12.1 Population, GDP per capita and economic structure for
133 countries, by income category
Pop. GDP Labor force, Value added,

(mil.) per cap., % (1990) % GDP (1995)
1995 US$
(1995) Agr. Ind. Ser. Agr. Ind. Ser.
Low-income econ. (49) 3 180 290 69 15 16 25 38 35

Lower-middle income econ. (41) 1 153 1 670 36 27 37 13 36 49
Upper-middle income econ. (17) 438 4 260 21 27 52 9 37 53
High-income econ. (26) 902 24 000 5 31 64 2 32 66
Tanzania 30 170 84 5 11 58 17 24
Source: World Bank (1997).

Lessons from History 285
counter-revolution in economic theory and policy, a counter-revolution

that drew additional credibility from the dramatic collapse of the economy
of the Soviet Union and its allies. The renewed emphasis on free-market
policies also led to a call for the dismantling of public ownership, central
planning and excessive government regulation of economic activities in
the low-income and the lower-middle-income economies. The central argu-
ment of the advocates of free-market policies was, and is, that problems
encountered in the transition from agrarian to industrial society are the
direct result of poor allocation of resources due to incorrect pricing policies
and too much government intervention. Free-market economists insist that
countries do not get bogged down in their development because of preda-
tory actions by industrialized countries and international agencies, but
because of an excess of state control, gross inefficiency, serious corruption,
mismanagement and lack of economic incentives. What is needed, in their
opinion, is certainly not a reform of the international economic system, a
restructuring of dualistic development economies, more central planning
or increased foreign assistance. Rather, it is a matter of promoting free
markets and of governments creating preconditions for the ‘invisible hand’
of the market to do its work.
In the mid-1980s Tanzania was forced to join the bandwagon of the free-
market counter-revolution. It adopted and implemented a series of changes
in its approach to enhance economic development and speed up the transi-
tion from an agrarian to an industrial society.
Now the 20th century has drawn to an end, it has become increasingly
clear that the changes in economic policy have not yielded the expected
results, and that the transition process is proceeding at an excruciatingly
slow pace. Tanzania finds itself in the league of the 10 countries with the
largest percentage of the labour force involved in agriculture. Its value
added percentage of GDP for agriculture is the highest in the world. It is
among the 12 countries with the smallest percentage of the labour force in
industry, and among the 10 countries with the lowest value added percent-
age of GDP for industry.
Tanzania, in other words, ranks among the least industrialized countries.
There is little comfort in knowing that Tanzania shares its predicament
with many other countries in sub-Saharan Africa.
4 Pre-requisites for industrialization: early European

experience
In order to industrialize certain prerequisites have to be met. So far, it has

mainly been economists who identified and analysed the principles behind
the transition from a dominantly agrarian to an industrial society, and who
designed, with varying degrees of success, ‘development strategies’ for the
laggards.
Using a literature survey on the early European transition, this section

sets out to identify the structural and cultural prerequisites for industrial-
ization, not only in the institutional sphere of the economy, but also in the
spheres of politics, kinship and religion. In addition, education and health care
are given some attention. The prerequisites found are then confronted with
the prevailing structural and cultural aspects in relevant institutional
spheres in contemporary Tanzania.
4.1 Politics
Achieving internal integration and external adaptation are the two funda-
mental functions of the institutional sphere of politics in any society, func-
tions that can be summarized as ‘how can society, in face of a multitude of
internal contradictions, conflicting interests and threats from outside, be
kept together?’. Internal integration and external adaptation are achieved
through system maintenance, conversion and government functions
(Bertholet and Lapperre, 1991). The system maintenance function refers to
aspects such as socialization, recruitment, mobility and retrenchment of
the members of the political system – or, in other words, to the ways in
which the allocation of positions takes place. The conversion function
refers to aspects such as the articulation and aggregation of interests, and to
communication. The government function refers to aspects such as rule
making, rule application and rule adjudication. A number of aspects of the
functions mentioned have played an important role in the transition
process from agricultural to industrial society.
In traditional European societies it was particularly aspects related to the
allocation of positions and the articulation and aggregation of interests that
hampered the transition towards an industrial society. The medieval
European political system, embodied in the Three Estates (nobility, clergy
and common citizens), was characterized by a relatively low social mobil-
ity. Position allocation almost exclusively took place on the basis of kinship
ties: ‘one’s birth was more important than personal merits. By the end of
the 18th century the political system gradually ceased to obstruct social
mobility, and a new trading, commerce and manufacturing elite gained a
say in the affairs of the state. In addition, be it reluctantly at first, the state
also allowed alliances between various new groups active in the economic
sphere (Jansen, 1991).’
The degree of political development of a society can be judged on the
basis of a number of political system capacities (Lapperre, 1997). The extrac-
tive capacity of the traditional agrarian societies – the extent to which the
political system is capable of mobilizing material and human resources as
inputs – was relatively low. For an industrial take-off, the state has to be
increasingly capable to manipulate taxes, subsidies and markets in favour
of trade, commerce and industry, and to extract the necessary services. This
was exactly what was achieved in a number of European countries,
including Britain, in the early 19th century. The regulating capacity – the
extent to which the thinking and acting of the members of a social system
are governed by power and coercion, or the threat of coercion – in tradi-
tional societies was also low. A lot of coercion was usually necessary to keep
society together. When transition gets on its way, the political system
should be able to enforce its will without unduly alienating economically
important groups and thus rendering them less valuable. For this purpose it
is important that the symbolic capacity of the political system is high. The
latter was achieved by most of European nations during the early stages of
industrialization. As far as the internal and external response capacities – the
measure and way in which the system serves its interests and the measure
in which the political system succeeds to adapt to external pressures and to
maintain its internal integrity – are concerned, all political systems
throughout history have always tried to find the most favourable balance.
At the onset of the industrialization process, the political system should be
able to maintain its territorial integrity and to solve conflicting interests
with the least possible damage to the emerging new economic system. In
the late 18th century these prerequisites were already fulfilled in Britain.
With respect to the distributive capacity – the way in which the system
allocates the extracted goods, services, power, prestige and opportunities to
persons and groups in society – the traditional situation was one in which
the goods and services produced were allocated almost exclusively to the
kinship-based elite. At the take-off of the transition, the political system
has to be able to affect a significant transfer of extracted goods and services
to new economically important individuals and groups. This was exactly
what happened in Britain and Germany.
In traditional agrarian societies, the productive capacity – the measure in
which the system is effective in mobilizing and regulating the economic
sphere for its own political aims – was relatively low. For the take-off of the
industrialization process, it is imperative that the production capacity
increases. For this purpose the political system has to provide, among
others, collective facilities in the fields of transport, communication, trade,
commerce and education infrastructure. It has to do away with obstruc-
tions such as, for example, toll barriers and unfavourable tariffs. It has to
remove monetary and fiscal bottlenecks, and to allow a free movement of
labour. In addition, it is important that it is able and willing to act as entre-
preneur when no other possibilities exist. Already at the onset of the indus-
trial revolution, Britain, Germany and France had taken extensive measures
to improve their productive capacity by the means mentioned. The pre-
requisites for industrialization in the political sphere mentioned so far
were all structural.
The cultural prerequisites, among others, concern new skills and tech-
nologies, a high level of knowledge, individualistic, utilitarian and merito-
cratic ways of thinking and acting, thinking in terms of social classes
instead of estates, thinking in terms of the right to own property and to

make profit, a positive attitude towards work and thinking in terms of
economic systems. By the close of the 18th century, all these cultural pre-
requisites were already fulfilled in most of the European leading countries.
In short, the major political prerequisite for a successful transition is a
state that is both strong and flexible at the same time.
4.2 Economy
In a situation where four peasants can hardly maintain one person not
involved in agriculture (Grigg, 1982) and shortages of staple foods are a
recurrent phenomenon, there is little money available to buy all but the
most essential manufactured commodities. The overwhelming majority of
the population is bound to the land, and struggles to wrest a living from it
and to satisfy the demands of land owners. For the industrialization process
to take off, a productive agriculture and low staple food prices were among
the most important prerequisites that were regionally fulfilled by the end of
the 18th century and even much earlier (More, 1997).
In the cottage industry that had already developed in the 14th century,
production of, for example, textiles took place jointly with agricultural
work. When making hay or at harvest time, spinning and weaving just
took second priority.
When production eventually started to concentrate in places where
much was invested in machinery, buildings and infrastructure, part-time
labour, as in the cottage industry, no longer sufficed, and manufacturers
required and attained a steady and guaranteed supply of labour of sufficient
quality.
Capital is understood to mean produced means of production – capital
goods – and the power of disposition, expressed in money, over these
goods. In a metaphorical sense, it even means the disposition over people.
When, at the end of the 18th century, the demand for manufactured con-
sumer goods increased, this created problems of supply for the traditional
manufacturing sectors, and brought about a fertile environment for tech-
nological change (Cameron, 1991). To finance innovations for mass pro-
duction, in the long run, additional capital was needed over and above
family capital and profit from one’s own enterprise. New forms of
banking systems and enterprises developed, and were made legitimate by
new laws and government regulations (Rosenberg and Birdzell, 1986;
More, 1997).
Mass production uses immense amounts of raw materials and energy,
and it is more than convenient if raw materials and energy sources are at
hand locally. If they are not, there should be a good transport and trade
infrastructure. Before the end of the 18th century, prices of raw materials,
agricultural produce and handicrafts were strongly affected by the distance
to be covered between producers and consumers.
Prices of grain, for example, tripled with every 150 km of transport over
land (Thirsk, 1982). Manufactured goods, in so far as they were not made
locally, were only available to the elite. Mass production of consumer
goods, on the other hand, relies on a steady demand and this demand, cer-
tainly in the take-off phase of the industrialization process, depended to a
large extent on relatively low prices. To achieve these low prices, a well-
established network of roads, canals, harbours and, later, railroads was
established and maintained.
Goods, services and money can, however, only be made really mobile by
a good financial infrastructure. By 1720 some 300 years of slowly expand-
ing markets had been accompanied by a corresponding expansion of agri-
cultural and craft production. The large-scale cottage industry and
ironworks, which first emerged in the 15th century and expanded in the
17th, were indicative of the growing pressure expanding markets exerted
on traditional production processes. In the cottage industry the entrepre-
neurs were usually merchants, and they supplied the materials to cottages
and workshops, bought and distributed the final products, and, sometimes,
provided the necessary equipment. In the iron industry the entrepreneurs
were often connected to the landed gentry. In both the cotton and iron
industries entrepreneurship had its roots in trade, commerce and agricul-
ture (Pieterson, 1987). Although these traders and gentry were keen on
profit and avidly looked for market opportunities, they were generally not
quite prepared for a new situation in which they had to perform a host of
functions at the same time. Products had to be designed and tested, and
raw materials acquired and tooled. Parts had to made and assembled to
yield a final product.
Large numbers of labourers had to be trained in factory discipline, new
regional, national and international markets had to be found and distribu-
tion networks established. By the end of the 18th century a class of entrepre-
neurs had emerged with the realization that manufacturing was not simply
equivalent to efficiently operating a small workshop or factory, but involved
the creation of changes – in product, distribution and organisation – that
would increase the margins between costs and revenues (Righart, 1991).
The transition from agrarian to industrial, of course, also required a
demand for industrial products. Such a demand, however, could only
develop in an environment in which there was enough purchasing power
and in which consumers were willing and able to change their consump-
tion patterns. When the first industrial revolution took off, new skills and
technologies were required to make the change from cottage industry to
the mass production of consumer and capital goods. Around 1760 the
level of knowledge and skills was already high in many fields. Science,
however, did not yet play a significant role in the industrialization
process. In the beginning this process was predominantly fed by craft
skills and knowledge.
Even in the late Middle Ages the world was still conceived as relatively
static. The notion prevailed that God had given every human being its
place in the world already at birth, and that the Three Estates were a God-
given order. Private property and interest were guilt-ridden concepts
(Evans, 1985). Work was considered more a curse than a blessing.
Renaissance, Reformation and Enlightenment eventually contributed to the
emergence of individualistic and meritocratic ways of thinking and acting,
thinking in terms of classes instead of estates, thinking in terms of the right to
own private property, the pursuit of profit and a more positive approach to
work, cultural aspects in the economic sphere which can all be considered
prerequisites for industrialization. Last but not least, in an environment which
became increasingly complex, entrepreneurs had to think in terms of econ-
omic systems. They learned how to view the world as something which can
be empirically researched and manipulated on the basis of research findings.
4.3 Kinship
Both overall population growth and urban population growth in an envir-
onment of declining rural populations were stimuli for developments that
created preconditions for industrialization. Particularly urban population
growth in an environment of declining rural populations, a common phe-
nomenon in parts of 16th and 17th century Europe, brought about innova-
tive behaviour of small farmers and increased their productivity (de Vries,
1984). In view of the importance of the size of the market for manufac-
tured products, countries with a relatively large population and a high and
evenly distributed population density had an advantage over countries
which did not have these features.
At a certain point in the development process, ascriptive attitudes were
replaced by more individualistic and meritocratic ways of thinking and
acting, and the paramount importance of kinship as the integrating
element in society diminished (Hillebrand and Jansen, 1991). Estate
society, with its reliance on ascription, gave way to forms of class society
(Appelbaum and Chambliss, 1995). The production-oriented extended
family gradually became a consumption-oriented nuclear family.
4.4 Religion
The zeal to convert the infidels to Christianity and religious faith in general,
in different ways, had their role in the ‘slow revolution’ that brought about
industrial society in Europe. When, during the crusades, European rulers
tried to rid the Holy Land of the Arabs, this also meant the re-establishment
of trade links with the Middle East – trade which contributed, in turn, to the
revival of the economy in the 13th century (Runciman, 1994). Religious zeal
was one of the motives behind the European expansion in the 15th century,
and during the religious wars, in the 16th and 17th centuries, religion was
instrumental to the coming into being of a variety of independent and
competing nation states with their own social identity (Wright, 1984).
Religion, and particularly Calvinism, it is argued by a number of scholars,
was instrumental in the development of labour ethics which were particu-
larly suited for industrial production (Romein, 1954).
4.5 Education
At the onset of the European transition from a traditional agrarian society
to an industrialized one, education had been fully institutionalized in
almost all European countries. Primary education was available to many
but certainly not all, and literacy rates generally remained low. Practical
training in crafts and early manufacturing usually took place during (long)
apprenticeships and on the job.
There were hundreds of universities in which a great variety of subjects
could be studied. With a few exceptions, however, technology, industry
and science were not yet linked in joint endeavours to increase productiv-
ity. The availability of skilled artisans, trained in apprenticeships and on
the job, in almost all fields of economic activity was probably the only
direct significant prerequisite for the transition to take place.
4.6 Health care

At the end of the 18th century, European health care was certainly not yet
fully institutionalized. Medical science was still in its infancy and care for
the sick was more often than not in the hands of the family, the churches
or charity organisations. Life expectancy at birth was still relatively low. It
is, therefore, hard to imagine any prerequisites in the sphere of health care
for the industrialization process to take off.
5 Prerequisites for industrialization: early European

experience and contemporary Tanzania
The identified prerequisites for the early European transition in the institu-
tional spheres of politics, economy, kinship, religion, education and health
care can now be checked against relevant aspects of the same institutional
spheres in contemporary Tanzania. The Tanzanian aspects are derived from
literature and research findings of 14 postgraduate students of the
Department of Technology and Development Studies of the Faculty of
Technology Management, Eindhoven University of Technology, in the
Netherlands. All these students conducted their research in Tanzania
between 1991 and 1997.
5.1 Politics
Table 12.2 presents an overview of identified prerequisites for the early
European (E) transition in the institutional sphere of politics, and checks
these against the state of relevant aspects in contemporary Tanzania (T).
Table 12.2 Overview of identified prerequisites in the sphere of politics
Prerequisites E T
Structural
State able and willing to:
• give the new (economic) elite a say in the political system + –
• let the old and new (economic) elite form alliances + –
• manipulate taxes, subsidies and markets in favour of trade,
commerce and industry + –
• enforce its will without unduly alienating economically important
groups + +
• maintain its territorial integrity and solve conflicting interests with
the least possible damage to the emerging new economic system + +
• affect a significant transfer of extracted goods and services to new
economically important individuals and groups + +
• provide collective provisions in the field of transport,
communication, trade and education infrastructure + –
• do away with obstructions such as toll bars and unfavourable
tariffs + +
• remove monetary and fiscal bottlenecks + +
• allow free movement of labor + +
• act, when necessary, as entrepreneur + +
Cultural
State able and willing to promote:
• new skills, technologies and a high level of appropriate know-how + –
• individualistic, utilitarian and meritocratic ways of thinking and
acting + –
• thinking in terms of classes in stead of estates or other rigid
entities + +
• thinking in terms of private property and profit + +
• thinking in terms of economic systems + +
In the institutional sphere of politics, Tanzania does not (fully) fulfil the
structural prerequisites of a system able and willing (1) to give the new
economic elite a say in politics, (2) to let the new economic elite form
strategic alliances, (3) to manipulate taxes, subsidies and markets in favour
of trade, commerce and industry and (4) to provide collective provisions in
the field of transport, communication, trade and education infrastructure.
Certainly before the economic reforms of the 1980s, but also thereafter,
Tanzania’s political system lacked the necessary flexibility to fully incorpor-
ate economically important individuals and groups, notably from the
Asian community, in the various institutes of power. Although the state in
theory advocates the free development of political alliances of (economi-
cally) important groups, in practice the possibilities are limited by the
biased way in which parties are funded and the unequal access to the
media. Both aspects are detrimental to the national economy, since they
push actors into the illegal informal sector from which it is difficult to
extract revenues (Schulte, 1996). Mainly because of its bureaucratic charac-
ter and a widespread culture of bribery – in spite of the efforts to curb the
latter – Tanzania’s political system lacks the strength to fully exploit an
effective system of levying taxes in favour of important economic activities
such as, among others, the provision and maintenance of an adequate
infrastructure in the fields of transport, communication, trade and, to a
lesser extent, education (Newberry and Stern, 1987). Contrary to many
other African countries, Tanzania has been remarkably successful in its
peaceful endeavour to forge a nation state from what in essence is a multi-
ethnic society. It therefore meets the prerequisite of a system able and
willing to enforce its will without alienating important groups. In the past
decades it has also met the prerequisite of a system able and willing to
maintain its territorial integrity with the least possible damage to the eco-
nomic system, and a system able and willing to solve internal conflicting
interests with the least possible damage. An exception was the military
intervention in Uganda in the early 1980s, which caused prolonged grave
harm to the economy.
A cultural prerequisite that is not (fully) fulfilled is a state willing and
able to promote (1) new skills, technologies and a high level of appropriate
know-how, and (2) individualistic, utilitarian and meritocratic ways of
thinking and acting.
Although, in theory, the state promotes new skills, technologies and high
levels of appropriate know-how, in practice the situation is less favourable.
The budget limitations to enhance education and technological develop-
ment are severe, and the actual priorities set are often at odds with formal
policies. In the field of education, for example, vocational secondary tech-
nical education is high on the priority list but receives inadequate funds to
fulfil the demand (Duijsens, 1996). Since the mid-1980s, educational pro-
grammes, among others, promote individualistic, utilitarian and merito-
cratic ways of thinking and acting (see also Section 5.3 on kinship) but the
attitude of the state remains ambivalent, in the sense that the African
Socialism ideologies (‘Ujamaa’) from the 1960s and 1970s are still pervasive
in many government laws and regulations.
5.2 Economy
European (E) transition in the institutional sphere of the economy, and
checks these against the state of relevant aspects in contemporary Tanzania
(T). In the institutional sphere of the economy, Tanzania does not (fully)
fulfil the structural prerequisites of (1) a highly productive agricultural
system, (2) relatively low real prices of staple foods, (3) a continuous and
guaranteed supply of labour, (4) labour with sufficient skills and knowl-
edge, (5) capital over and above family capital, (6) new banking systems,
Table 12.3 Overview of identified prerequisites in the sphere of the economy
Prerequisites E T
Structural
Highly productive agricultural systems + –
Relatively low prices of staple foods + –
Continuous and guaranteed supply of labour + –
Labour with sufficient skills and knowledge + –
Capital over and above family capital + –
New banking systems + –
New forms of organization of production + +
Locally available raw materials + –
Locally available energy + –
Good trade infrastructure + –
Good network of roads, canals and harbours + –
Entrepreneurs with the conviction that mass production is not
simply equivalent to efficiently operating a small workshop or
factory, but concerns the creation of changes – in product,
distribution and organization – that increase the margins between
costs and revenues + –
Demand for industrial products + +
Sufficient purchasing power + –
Environment in which consumers are able and willing to change
their consumption patterns + +
Cultural
New skills, new technologies and a high level of know-how + –
Individualistic, utilitarian and meritocratic ways of thinking and
acting + –
Thinking in terms of classes instead of estates or other rigid
units + +
Thinking in terms of private property and profit + +
Positive view on work + +
Thinking in terms of economic systems + +
(7) locally available raw materials, (8) locally available energy, (9) a good
trade infrastructure, (10) a good network of roads and harbours, (11) a class
of innovative and cost-conscious entrepreneurs and (12) sufficient purchas-
ing power of the majority of the people.
The absence of a highly productive agricultural system – with the related
relatively high prices for staple foods and their irregular supply – is one of
the most notable bottlenecks in Tanzania for industrialization to proceed
(FAO Yearbooks). This is particularly significant since the European experi-
ence shows that countries which seriously fell behind in the productiveness
of their agriculture in the 18th and 19th centuries – for example Russia and
to a lesser extent France – were also late to industrialize. A highly productive
agricultural system releases labour for industrial employment and guarantees
a regular supply of staple foods at affordable prices. Since less money has to
be spent on food, more money can be spent on manufactured commodities.
Although unemployment in Tanzania is high, the prerequisite of a con-
tinuous supply of labour with sufficient skills and knowledge, essential for
industrial enterprises, is not always met. Most workers have extensive
(extended) family obligations and, forced by the low wage levels, hold
more than one job at a time. Absenteeism, therefore, is high and endangers
the necessary continuity of work (ILO, 1992). Nationwide there is an
insufficient supply of well-educated workers, particularly at the technical
vocational levels (Duijsens, 1996). If the recent acceleration of the innova-
tion rate in almost all sectors of the economy in industrialized countries is
also going to affect developing countries, the availability of well-trained
workers will become ever more significant.
The lack of sufficient infrastructure in the fields of banking, trade, roads,
railways, harbours and energy supply – although recently significant
improvements have become visible – is probably the second most impor-
tant bottleneck for further industrial development. Although part of the
banking system has been modernized and liberalized recently, loans and
bank guarantees at reasonable conditions are hard to get for those without
‘connections’ (Zijl and Lassche, 1997).
Roads, railways and harbours have been significantly improved during
the last decade, but funds for proper maintenance are lacking (Fleisheuer,
1994; Gnoth, 1994; Koenders, 1994; van der Ven, 1996). Hence, these facil-
ities are deteriorating again. Hydropower-generated electricity is available
in all major towns, but the supply suffers from frequent and prolonged
cuts, notably in the dry season (Kers, 1997).
Tanzania has a great variety of important raw materials, ranging from
good-quality coal to different types of ore. Owing to a lack of capital and
expertise, and transport difficulties, these resources have so far not yet been
(fully) exploited.
Tanzania has a small class of innovative and cost-conscious entrepre-
neurs, many of whom originate from the Asian community. The majority
of indigenous African entrepreneurs have their roots in small-scale manu-
facturing and artisanal enterprises. Generally, and for a variety of reasons,
they find it difficult to increase their margins between costs and revenues
by means of innovations in product, organisation and distribution. They
have little access to capital, lack managerial and technical skills, and carry
the burden of extended family ties and communal attitudes connected
with the long period of African Socialism (Zijl and Lassche, 1997).
Last but not least, the per capita income of the overwhelming majority of
the Tanzanian population is so low (US$170 in 1995) that the overall pur-
chasing power is seriously impaired.
Cultural prerequisites that are not (fully) fulfilled are (1) new skills, tech-
nologies and a high level of appropriate know-how and (2) individualistic,
utilitarian and meritocratic ways of thinking and acting (see comments in

the sections dealing with the political and kinship institutions).
5.3 Kinship
European (E) transition in the institutional sphere of kinship and checks
these against the state of relevant aspects in contemporary Tanzania (T). In
the institutional sphere of kinship, Tanzania does not fulfil the prerequi-
sites of (1) relatively high population densities, (2) a population relatively
equally distributed over the territory, (3) an urban population growth in an
environment of a declining rural population, (4) consumption-oriented
nuclear families and (5) individualistic, utilitarian and meritocratic ways of
thinking and acting.
Tanzania’s overall relatively low population densities and the unequal
distribution of this population over the country are likely to impede indus-
trialization. A transport and communication infrastructure – among others
effectively linking market centres for specialized agricultural produce and
manufactured products – is (too) costly to establish and particularly to
maintain. The same applies to electricity grids and the electronic commu-
nication infrastructure. Although modern computer systems are available
in Tanzania, the degree of their interconnectedness is extremely low, and
hence the role of modern information technology in industry is small.
Urban population growth in an environment of a declining rural popula-
tion – bringing about agricultural innovation, increased agricultural pro-
ductivity and an impetus to the development of ‘cottage industries’ –
appears to be a specific European phenomenon, and is of little consequence
for contemporary Tanzania.
Table 12.4 Overview of identified prerequisites in the sphere of kinship
Prerequisites E T
Structural
Relatively large populations + +
Relatively high population densities + –
Population relatively equally distributed over the territory + –
Population growth + +
Urban population growth in an environment of declining rural
populations + –
Some kind of class society + +
Consumption-oriented nucleus families + –
Cultural
Individualistic, utilitarian and meritocratic ways of thinking and
acting + –
The extended family still features prominently in Tanzania, and nuclear

families are usually only found in high-income groups in urban areas. Even
there, however, family links are still extensive. Extensive family obligations
in a low-income environment hamper the possibilities of individuals to
save, and are therefore detrimental to the economy as a whole.
Extensive family obligations, in addition, are accompanied by significant
expenditures of social time, and the latter is often in conflict with the
requirement of a continuous labour input in industry. As a result of the
extremely low average income per capita, Tanzanian families are as yet sub-
sistence and not consumption oriented.
In the cultural sphere, in present-day Tanzania, communal ways of think-
ing and acting still have, generally, preference over individualistic, utilitar-
ian and meritocratic ways. Communal ways of thinking and acting have
their roots deep in the ancient cultures of the (dominantly) subsistence
peasant societies of East Africa. These societies were not classless, but lacked
classes, certainly until the colonial domination began, in the capitalist sense
(Shivji, 1975). Communal aspects dominated over class aspects at the
family, village and tribal levels. After independence communal aspects were
re-emphasised in the philosophy of ‘traditional ujamaa living’, as outlined
for the first time in the pamphlet ‘Socialism and Rural Development’ by the
first president of Tanzania, J. Nyerere. According to Nyerere, there are three
basic assumptions underlying ‘traditional ujamaa living’: (1) respect – each
member of the family recognizing the place and rights of the other
members, (2) common property – acceptance that whatever one person has
in the way of basic necessities, they all have, and (3) obligation to work –
every member of the family, and every guest who shares in the right to eat
and have shelter, taking for granted the duty to join in whatever work needs
to be done (Hyden, 1982). The devastating economic and social effects of
‘ujamaa’, when it became a formal government policy, and the retreat from
‘ujamaa’ are part of Tanzania’s recent history. The attachment to communal
ways of thinking and acting, however, lingers on even today.
5.4 Religion
European (E) transition in the institutional sphere of religion, and checks
these against the state of relevant aspects in contemporary Tanzania (T).
The prerequisites in the institutional sphere of religion appear to be specific
for the European transition from the agrarian to industrial society, and do
not apply to contemporary Tanzania. From a historical point of view, they
are probably coincidental.
5.5 Education
The influence of a high level of education and of science on the European
transition has often been overestimated. In the first phase of the transition,
Table 12.5 Overview of identified prerequisites in the sphere of religion
Prerequisites E T
Structural
Religious competition (contributing to the establishment of
independent nation states) + –
Cultural
Zeal to convert + –
‘Calvinist’ labour ethics + –
it was the artisans that led the way, and science did not play an important
role as yet. In the institutional sphere of education, the lack of skilled arti-
sans in key areas is, therefore, the only prerequisite not met.
5.6 Health care

A healthy labour force enhances productivity and is, therefore, important
in a competitive industrial environment. Over the past two decades, health
care in Tanzania deteriorated (Schulte, 1996), and this must have affected
the productivity of the labour force in a negative way. The question,
however, is whether a healthy labour force is a prerequisite for the transi-
tion from an agricultural to an industrial society. European history has
shown that it is not (see Section 4.6).
6 Can Tanzania learn from European history?
In absolute terms this question cannot, of course, be answered. Europe

around 1800 is not the Tanzania of today, and the social and physical envi-
ronments differ. However, the lesson that can be learned is that all coun-
tries that were among the first to make the transition from dominantly
agrarian to industrial society shared, right from the start, a number of
important innovative aspects in the institutional spheres of politics,
economy, kinship and religion. Innovative aspects in the institutional
spheres of education and health care apparently played a lesser role. These
shared aspects – and particularly the balance and interaction between them
– to a significant extent determined the possibilities for success of an early
take-off.
Institutional development in Tanzania appears, compared to the early
European one, to suffer from an imbalance. Important aspects in the politi-
cal and kinship spheres are still those of a dominantly agrarian society, and
they insufficiently support the further development of the institutional
sphere of the economy towards a dominantly industrial one.
In the sphere of politics, Tanzania finds it difficult to give members of
the new economic elite a say in politics and to let them form strategic
alliances. Even the present multi-party state is still dominated by the old
elite. The state, for a variety of reasons, is not able to exploit the existing
taxation system efficiently and effectively. This reduces the possibilities to
manipulate tax revenues in favour of trade, commerce and industry, and to
provide adequate collective provisions in the field of transport, communi-
cation, trade and education infrastructure. Additional political reform
would appear to be necessary for further economic development.
In the sphere of kinship, Tanzania’s development is hampered by rela-
tively low population densities and an unequal distribution of people over
the territory. This makes the development and maintenance of an adequate
infrastructure in almost all spheres prohibitively expensive. In view of the
low income per capita, consumption-oriented families are few, and the
local market for manufactured goods remains relatively small. Extended
family obligations weigh heavily on individual households, reduce their
saving capacity and require significant investments in social time, all
aspects that do not really enhance further economic development.
Communal ways of thinking and acting – having their roots in the tradi-
tional East African subsistence peasant societies and re-emphasized in the
philosophy of ‘traditional ujamaa living’ in the 1950s and 1960s – still
linger on and, by and large, have not been replaced by more individualis-
tic, utilitarian and meritocratic ways of thinking and acting. Cultural
changes of this type are hard to enforce by political means, and a further
change will take considerable time.
In the sphere of the economy, Tanzania lacks a highly productive agri-
cultural system and staple food prices that reflect the benefits of such a
system. Further economic development is hampered by a large class of sub-
sistence peasants who have great difficulties producing more than their
basic family requirements. Reviewing the situation of this sizeable part of
the population, and taking appropriate measures, would appear to be an
important political priority. In addition, credit facilities are limited, essen-
tial raw materials are not (yet) locally available, energy production and dis-
tribution are inadequate, and a good trade and communication
infrastructure is lacking. Various extended-family characteristics and the
concurrent communal ways of thinking and acting do not facilitate further
economic development, and will be difficult to change in the short term.
Notes
13
Macroeconomic Policy and
Performance of the Manufacturing
Sector in Tanzania: Has
Liberalization Helped? An
Econometric Approach
A. V. Y. Mbelle*
1 Introduction
After a decade of ‘de-industrialization’ during the 1980s, there are new

hopes for the Tanzanian manufacturing sector for the 1990s and beyond.
The Sustainable Industrial Development Policy for Tanzania, 1996–2020 (URT,
1996a), spells out the mission of the sector as making contributions to the
achievement of overall long-term development goals and enhancing the
sustainable development of the sector itself. The Development Vision 2025
for Tanzania (8th version, URT, 1998) sets more ambitious targets. In 2025
Tanzania should be a diversified, semi-industrialized economy with the
industrial sector (manufacturing, mining, construction and utilities)
accounting for 40 per cent of GDP (against 16.5 per cent in 1998).1 In addi-
tion to this, the sector is expected to form the basis for broad-based growth,
income generation, employment creation and foreign exchange earnings.
The sector is also seen as a vehicle for eradicating poverty in Tanzania
through informal industrial activities, rural industrialization and the emer-
gence of processing establishments in villages (National Strategy for
Overseeing the Implementation of Poverty Eradication in Tanzania, URT,
1997).2 The 1998/99 and 1999/2000 budgets for Tanzania promise a bright
future for the sector.
The high hopes for the manufacturing sector beyond the 1990s pose a
number of questions. Is the sector prepared for this mission? Is the sector
receiving sufficient attention (in substantive terms) to strengthen it for the
mission? What is the news that warrants such hopes?
300
Macroeconomic Policy and Manufacturing Performance 301
This article attempts an assessment of the performance of the manufac-

turing sector in Tanzania in the context of broad (economy-wide or macro-
economic) policies with a view to gauging the impact of such policies on
the sector. The article is organised into five sections. Section 2 highlights
recent developments in the manufacturing sector in Tanzania. Section 3
presents a survey of the deployment of macroeconomic policies for indus-
trial development to cover both the theoretical underpinning and evidence
on the ground. The empirical results are reported in Section 4, with the last
section being devoted to concluding remarks.
2 Recent developments in the manufacturing sector in

Tanzania
2.1 Performance
Table 13.1 below presents data on the performance of the sector in the
1990s.
2.1.1 Growth and share in GDP

In terms of growth ‘sustained’ positive rates were realised after 1995, rising
from 1.6 per cent in 1995 to 8.1 per cent in 1998, picking up from the
negative growth rates experienced between 1992 and 1994. The share of
manufacturing in GDP varies between 8.8 per cent, recorded in 1990, and
8.4 per cent, recorded in 1998, with a low of 7.9 per cent in 1995.
2.1.2 Export performance

The growth of manufacturing exports increased dramatically from –19 per
cent in 1993 to 48.1 per cent in 1994. This can partly be explained by
Table 13.1 Performance of the manufacturing sector in Tanzania, 1990–98

(1992 prices)a
1990 1991 1992 1993 1994 1995 1996 1997 1998
Real growth (%) 4.1 1.9 –4.1 –0.6 –0.2 1.6 4.8 5.7 8.1
Real mfg export 13.3 –27.7 –8.7 –19.0 48.1 41.9 1.4 –5.7 –42.6
growth (%)
Share in GDP 8.8 8.7 8.2 8.2 8.1 7.9 8.0 8.1 8.4
Share in total exports 23.8 19.6 16.2 11.8 14.8 16.0 14.6 14.5 8.8
Average capacity 50 50 38 50 46 46 48 60 55
utilization
Share in NT exportsb 48.9 43.7 38.8 28.4 42.2 36.5 35.3 29.2 18.2
Employment (000) 123 121 126 126 130 131 129 132 137
Sources: URT (1999a; 1999c).

Notes:
a
For establishments with 10 or more employees.
b
Non-traditional exports.
changes in the management of the exchange rate, though exports ran out
of steam by 1996. The share of manufactured exports in total exports
declined from 23.8 per cent in 1990 to a mere 8.8 per cent in 1998. The
dominance of manufacturing exports in non-traditional exports was greatly
eroded by 1998. The factors accounting for this drastic decline include
trade sanctions against Burundi, unrest in the Great Lakes region, and dis-
ruption of transport network caused by El Niño rains, thus hindering deliv-
ery of industrial exports to neighbouring countries, as well as leading to
reduced inputs for agro-industries.
2.1.3 Capacity utilization

There was a marked stability in the average capacity utilisation rate
between 1990 and 1998. This can mainly be attributed to a few industries
which have shored up the average, following successful joint ventures with
foreign investors in the ongoing privatisation exercise. Such industries
include beer brewing (especially Tanzania Breweries Ltd), cigarette manu-
facturing (a single-firm industry, the Tanzania Cigarette Company), the
cement industry, oxygen production (Tanzania Oxygen Ltd) and the soft
drinks industry. These industries have benefited from injections of foreign
capital, rehabilitation of plant and machinery, and improved management.
2.1.4 Employment
Trends in manufacturing employment were on the increase between 1991
and 1998. When compared to the sector’s real growth, the increase in
employment was more rapid than is warranted by performance of the
sector. Further analysis of the employment data indicates most employ-
ment is found in food and beverages (ISIC 31) and textiles and leather (ISIC
32). For example, in 1998, these two categories accounted for 67.1 per cent
of total manufacturing employment: ISIC 31, 37.3 per cent, and ISIC 32,
29.8 per cent.3 For the food and beverages group, this reflects increased
investments, especially after 1995, with an average of 31 per cent of all new
industrial licences issued between 1995 and 1998.4
2.2 Other signs of recovery

One of the notable new developments is the participation of the manufac-
turing sector in the stock market. After the establishment of the Dar es
Salaam Stock Exchange (DSE) on 15 April 1998, the first ‘firm’ which listed
itself was Tanzania Oxygen Ltd, selling shares to the public (at about
US$0.9 per share). Tanzania Breweries Ltd started selling its shares (10 per
cent) to the public towards the end of May 1998, at about US$0.9 per share.
Other industries (or firms) are in the preparatory stage for joining the stock
market. These include the Tanzania Cigarette Company, Tanga Cement
Company Ltd and Portland Cement (Wazo) Ltd. Again one notices that
these are the industries (or firms) which have concluded successful joint-
venture agreements with foreign investors, thus attracting foreign capital

investments.
2.3 Manufacturing sector dynamics in Tanzania

2.3.1 Export competitiveness
Dynamism in the export sector refers to the ability to penetrate ‘sophisti-
cated’ markets where quality, delivery schedules and so on must be strictly
adhered to. Such markets will mainly be the markets in developed coun-
tries (OECD, USA, and so on). There are different measures of such
dynamism (see Rodriguez, 1998, for an extensive discussion). Rodriguez cal-
culated export competitiveness indices for ten Sub-Saharan African coun-
tries, including Tanzania, and concluded that there were substantial
declines in global market shares, and that market participation in advanced
country markets declined even more. Compared to, for example, Latin
American countries, sub-Saharan African countries failed to adapt to exter-
nal markets between 1985 and 1995. These results point to issues of price
competitiveness, quality and aggressiveness in marketing. It is not surpris-
ing that an increasing proportion of African exports is channelled to
regional markets, partly because of the success of regional arrangements
but mainly as a result of loss of competitiveness in advanced markets.
2.3.2 Technological dynamism

Studies by Mbelle and Bagachwa (1995) and Wangwe, Semboja and Kweka
(1997) show some evidence of dynamism in manufacturing for a few firms.
‘Offensive Strategies’5 have been adopted with encouraging results. This is
true even for firms which have not concluded successful joint ventures
with foreign investors (see also Alange, 1987).
3 Macroeconomic policies for industrial development in

Tanzania
In this section we review the deployment of macroeconomic policies in

Tanzania and discuss existing evidence on the effect of such policies on the
manufacturing sector in Tanzania. Policies are dealt with in Section 3.1,
while the evidence concerning the impact of policies is presented in
Section 3.2.
3.1 Macroeconomic policies for industrial development in Tanzania

Macroeconomic policies are those that have an influence on aggregates in
the economy. Typical policies will deal with, for example, aggregate expen-
diture and consumption, government deficits, money supply, investments,
exchange rate policies, interest rate policies, and so forth. Results are
reflected in GDP growth rates, inflation and so forth. Although generally
sound macroeconomic policies are good for economic performance, their
impact on industrialisation may not be directly beneficial. This depends to a

considerable extent on the characteristics of the industry (whether it is a net
exporter or a net importer), its financial standing and so forth. For example,
a tight monetary policy, while credited with checking inflation in many
developing countries, may lead to decreased credit in the economy, which
may impact negatively on industrialisation. A liberalised exchange rate
regime, while providing incentives to exporters, may harm high-import
content manufactures, as they spend more in local currency terms for the
same volume of imports. Pieper (1998) demonstrates extensively how poli-
cies, especially exchange-rate-based policies, favour industrialization.
3.1.1 Recent trends in selected macroeconomic aggregates

Table 13.2 shows selected macroeconomic variables in Tanzania for the
period 1990–99. The picture of macroeconomic performance emerging
from the table is quite impressive. For some aggregates, however, the
targets set for the 1990s were achieved late or only partially, for instance
single-digit inflation, a real annual GDP growth rate of 5 per cent and a
growth of money between 8 and 10 per cent.
With regard to the exchange rate, the unification in August 1993 seems to
have had an impact in reducing the foreign exchange premium to single-
digit figures in 1993 and to less than 1 after 1995. On the other hand, the
liberalisation of interest rates in 1993 seems to have led to one queer result:
widening the spread after 1993 (for example, 1990: 5 per cent; 1993: 15 per
cent; 1997: 24.7 per cent; 1999: 15 per cent). This has the potential impact
of discouraging investments in the real sector of the economy by making
credit too expensive compared to interest earned on deposits.
Table 13.2 Selected macroeconomic indicators in Tanzania 1990–98
Variable 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999d
Growth in money
supply (M2) 45.4 26.1 38.5 28.8 32.5 26.2 11.6 11 11.1 7.9
Inflationb 32 34 24 24 27 34 15.4 15.4 15.2 7.8
Forex Premiumc 66.7 43.8 36.0 5.3 1.9 <1 <1 <1 <1 <1
Savings rate 1 yr av. 26 26 26 24 24 15 11 7.3 8.7 10
Lending rated 31 31 31 39 39 40 38 32 24.3 25
GDP growth (real) 5 6 4 4 3 4 4.2 3.3 4.0 n.a.
Notes:
a
As of 30 June, 1999.
b
Headline inflation.
c
Defined as the difference between the parallel exchange rate and official exchange rate
expressed as proportion of the official rate.
d
Maximum, medium and long term.
Sources: Bank of Tanzania, various issues; Mbelle (1998).
3.1.2 Discussion
Prudent macroeconomic policies are supposed to provide an overall
environment which is benign for economic agents to take advantage of.
Tanzania has gone a long way towards improving the macroeconomic
environment (for example, liberalizing the exchange rate towards
a market-determined rate). However, the extent of growth of GDP
does not reflect the intensity and extensiveness of the policies being
pursued.
3.1.3 Future prospects for the sector

Since 1998/99, the government has introduced a number of measures
which have potential benefits for the manufacturing sector. Such measures
include:
(a) Introduction of value added tax (VAT) on 1 July 1998, replacing sales
tax and stamp duties for firms registered in the VAT system. This will
simplify taxation procedures (depending on co-operation of agents and
good record-keeping).
(b) Depositing duty draw-back funds in a special account instead of in the
general account of the government where the money had been used for
other purposes, thus discouraging exporters.
(c) Further liberal tariff reforms with only four bands: a low tariff rate of
5 per cent to be imposed on raw materials, replacement parts and
investment goods; a 10 per cent tariff for semi-processed inputs; a 20
per cent tariff for fully processed inputs and a top rate of 25 per cent for
consumer goods. A zero rate for strategic and leading investment
sectors continues to apply as well.
(d) Pre-shipment inspection to be introduced for textile materials, in order
to curb illegal imports.
(e) Reduction of duties for COMESA (Common Market for Eastern and
Southern Africa) goods to 20 per cent.6 By June 1999 the rate averaged
7 per cent.
An informed assessment of the benefits is not easy, given the short time
span. However, casual observation of economic performance in 1998 would
point to a positive impact of these measures.
3.2 Evidence of the impact of macroeconomic policies on the

Tanzanian manufacturing sector
There are several studies in Tanzania that have attempted an assessment of
the impact of these policies right to the level of firms – assessing the
impact of structural adjustment policies, the break point being the adop-
tion of the first Economic Recovery Programme (ERP) in 1986. ERP had
‘promised’ to revive capacity utilization rates in manufacturing from the

obtaining levels of 20–30 per cent to levels of 60–70 per cent, through
rehabilitation of plants and machinery, and increased resources (URT,
1986a). It is only in 1997 that 60 per cent capacity utilization was
achieved (URT, 1999a).
Studies which have reflected on some adverse impacts include Mbelle
(1988), RPED (1994), Wangwe, Semboja and Kweka (1997), Mbelle et al.
(1998a,b), and Engblom and Veres (1998) for the worst scenario in the
textile industry.
Areas of concern include:
(a) expensive credit, implying that the liberalisation of the banking sector
has not helped; in the survey by the Regional Programme on Enterprise
Development (RPED) 64 per cent of the industries ranked this as the
most severe constraint
(b) ‘expensive’ foreign exchange – accessibility is mainly guaranteed to
those with connections to foreign investors
(c) ‘excessively’ restricted circulation of money in the economy
(d) markets swamped by cheap imports from abroad, which have, for
example, adversely affected the textile industry
(e) increases in general unemployment brought about by retrenchments in
the civil service and in the manufacturing sector (though for this sector
it doesn’t show so much), as privatisation of public enterprises is taking
its course (market forces)
(f) still too many taxes and too high levels of taxation.
3.2.1 Discussion
Looking at Tanzania’s history since independence, one sees differences
in the reliance on macroeconomic policies for improved production
(see the overview in Chapter 1). While in the 1970s sectoral policies
were given equal weight (for example, in agriculture and industry in
the Basic Industry Strategy see Skarstein and Wangwe, 1986; Musonda,
1992; Ndulu and Semboja 1994, 1996; Parker et al., 1995), the 1980s
witnessed an absence of such meso-policies following up on macro-
policies with concrete targets. The SIDP for Tanzania points out this fact.
In the introductory remarks it is acknowledged that one of the reasons
for launching SIDP 1996–2020 is that the Basic Industry Strategy
(1975–95) had expired (in 1995). In fact it had already expired in late
1970s, when the economic crisis unfolded. It is only now that a vision
for the sector, and one for the country as a whole, is being formulated
again.
Against this background, the following section presents an econometric
investigation of the impact of macroeconomic policies on the manufactur-
ing sector in Tanzania.
4 Empirical analysis of the effects of macroeconomic policy

on manufacturing in Tanzania
In this section we report the results of the estimation of macroeconomic

policy impacts on manufacturing. The section covers methodological
issues, modelling, results and comparison of the results with the results of
other studies.
4.1 Methodological issues

4.1.1 The models and their justification
Following the discussions in Section 3, two macroeconomic instruments
were selected on the basis of their important potential impacts. These are
the exchange rate (for the external sector) and the domestic interest rate
(for the domestic sector). We did not estimate the determinants of manu-
facturing output growth since this has been tackled in many studies, for
example, Mumburi (1998).
(a) Effect of the exchange rate on manufacturing exports

Here we estimate the responsiveness of manufacturing exports to exchange
rate incentives. Such incentives include streamlined administrative proce-
dures for acquiring foreign exchange, leaving the market to determine the
exchange rate, and freedom of retention of foreign exchange. It is expected
that exporters of manufactures will increase exportation, also to take
advantage of increased earnings in local currency due to more favourable
exchange rates.
(b) The effect of interest rates on manufacturing investments

In economic theory high lending rates discourage borrowing for invest-
ment. Thus investments in the real sector, for example, manufacturing, will
be discouraged. The liberalisation of interest rates in 1993 is expected to
have a negative effect on investments in manufacturing.
4.1.2 The models and expected signs (hypotheses)

(a) The determinants of manufacturing exports
log CMAE = + 1 log CMAO + 2 log CRERM + 3 log TOT + 4 DU84 +

5 DU86 + 6 DU93 + e
where:
CMAE = changes in real manufactured exports

CMAO = changes in manufacturing output (supply capacity variable) (+);
as manufacturing output increases, we expect manufactured
exports to increase also
CRERM = changes in real exchange-rate misalignment, as a measure of real

overvaluation; this is done so as to assess the objective of main-
taining an equilibrium exchange rate (eliminating overvaluation)
in Tanzania (–)7
TOT = net barter terms of trade (–), with 1977 as the base year; the
better the terms of trade, the more expensive exports become
compared to imports – hence a hypothesised fall in the volume
in exports
DU84 = dummy to capture the influence of the partial liberalisation mea-
sures instituted in 1984 with the introduction of the own funds
window (1 after 1984, zero before 1984) (+)
DU86 = dummy for the ‘true’ liberalisation carried out in 1986 (1 after
1986) (+)
DU93 = dummy for the unification of the exchange rate in 1993 (1 after
1993) (+)
e = error term
, = parameters.
(b) Manufacturing investment
The objectives of liberalising the financial market include deepening the

market. This can be achieved, for example, through interest rate deregula-
tion. However, existing evidence is conflicting on whether interest deregu-
lation leads to increased savings, and hence an increased supply of loanable
funds, especially in countries where credit markets are undeveloped, and
inflation rates are high. One of the key aspects of liberalisation of the
financial sector thus becomes the relationship between capital and domes-
tic investment.
In countries such as Tanzania, where historically the government has
led industrialization, it becomes a question of policy priorities – how
much of capital is invested in different activities. This then becomes a
macroeconomic issue. Tracing changes in the absolute values of the
investments themselves will miss the policy-priority aspect where a single
investor is dominant. Below we therefore hypothesize that it is the propor-
tion of total investible funds devoted to manufacturing that counts and
reflects macroeconomic policy. The use of ratios, for example investments
to GDP, is commonly resorted to in many econometric works in such
cases.
In economic theory investment is mainly determined by the interest
rate. The government in this case is then seen as the saver and investor
(for example, through parastatals which acquire investible money in
normal ways and thus face the lending rate). Our assessment of liberaliz-
ation pits both government and private-sector agents against each other
in the financial market. Here the lending rate is supplemented by general
liquidity (money supply) and effectiveness of banking facilities (how the

numerous commercial banks under liberalisation can influence public
holding of money). If money in public hands increases, then the banking
facilities are less effective. Hence marshalling of investible funds becomes
less and the potential for borrowing from the banks for investing (for
instance in manufacturing) becomes less. This will be true for other sectors
as well.
Lynch (1995) uses the same relationships (though with more variables) to
estimate investments. We thus formulate our investment equation as
follows (with expected signs between brackets):
log CMAI = + 1 log CARE + 2 log CAMS + 3 log EFB + 4 DU86 +

5 DUP93 + e
where:
CMAI = changes in investment in manufacturing as a proportion of total

investment
CARE = changes in real lending rate (–); the higher the lending rate, the
less borrowing there will be for investment purposes
CAMS = changes in money supply (broad money, M2) (+); this has the
potential of increasing the supply of loanable funds (for invest-
ment)
EFB = the ratio of narrow money to broad money (M1/M2) (–); this mea-
sures the effectiveness of banking facilities, that is, the extent to
which money is held in public hands (circulation outside the
banks); in the estimation procedure, we deal with changes in the
variable.
DU86 = liberalisation in 1986 (1 for period 1986–96, else 0) (–); this is
intended to capture the expansion of activities in the economy
and whether manufacturing was still an attractive avenue for
investment; we hypothesize that manufacturing will compete
with emerging activities for loanable funds, hence manufacturing
investments will suffer
DU93 = liberalisation of the interest rate in 1993 (1 for 1993–96 or else 0)
(–); we expect interest rates to rise with deregulation, hence dis-
couraging manufacturing investments
e = error term.
4.1.3 Estimation techniques

The ordinary least squares method (OLS) was used for its simplicity and
wide use. The logs were preferred so that the regression coefficients can
be interpreted directly as elasticities. The equations are (log) linear and
additive.
4.1.4 Data sources

Official published data sources were consulted mainly from the (Central)
Bank of Tanzania.8 The periods covered are 1964–96 for the export equa-
tion and 1966–96 for the investment equation. For the latter, 1966 cut-off
was used, as this was the year in which Tanzania established its own
Central Bank following the break-up of the East African Currency Board,
which issued the East African shilling. Thus actions by Tanzania’s Central
Bank, for instance in money supply, are traced. There were no data gaps,
hence we are confident of the information. Supplementary sources
included Economic Surveys. Real exchange-rate misalignment data were
obtained from calculations in Elbadawi’s study (1998).9
4.1.5 Time series characteristics of the data

It is becoming common practice in econometric estimations to analyse the
time series characteristics of the data when trends are used (time series
data). This will be done in the following section.
4.2 Time series analysis

(a) The determinants of manufacturing exports
Since most macro-variables are likely to be non-stationary, we begin by
testing for the order of integration. This means that the variables may have
a mean that changes with time and a non-constant variance. Working with
such variables in their levels will give a high likelihood for spurious regres-
sion results. Furthermore, no inference can be done since the standard sta-
tistical tests such as the F-distribution and the t-distribution are invalid.
Hence the first step is to test the level of integration through the unit
root process by using standard testing procedures, such as those developed
by Dickey and Fuller and Sargan and Bharghava. The results of these tests
are presented in Table 13.3.
After establishing the order of integration, the next step is to test for
cointegration. At this step, we seek to establish whether a linear combina-
tion of a set of non-stationary variables is stationary. As can be seen from
Table 13.3 Time series characteristics of the dataa
Variable SBDW DF ADF Order of integration
CMAE 1.661 –4.508 –3.026 I(1)

CMAO 1.168 –3.806 –3.475 I(1)
RERM 0.1195 –0.8245 –1.713 I(1)
TOT 0.3073 –2.572 –2.839 I(1)
Notes: a The critical values are –3.556, –3.575 respectively for the, DF and ADF. All
variables are in log levels.
Table 13.4, we cannot reject the null of cointegration in all three cases. We
then proceed to estimate a static regression involving all the variables and
the residual generated therefrom. One period lag is included in the general
autoregressive distributed lag (ADL) model as the error correction term.
After estimating the general model, 10 we went through a series of
simplification procedures. We proceeded principally by eliminating those
variables with t-values of less than one in absolute terms, the efficiency of
which was assessed at each stage using the F-statistic, the information crite-
rion and the model diagnostic tests, until we arrived at a preferred and a
certainly more parsimonious characterization of the data.
By reducing the model, the F-statistic increased from 3.1555 to 5.4806 in
the general and preferred models respectively. The standard error of the
regression declined from 0.965761 in the general model to 0.903656 in the
preferred model. The information criterion11 also showed a decline, starting
with the SC from 0.903278 to 0.480687, HQ from 0.44081 to 0.183386,
and FPE from 1.39904 to 1.07907 in the general to the preferred model
respectively.
The model diagnostic tests are insignificant at the conventional test
levels starting, with the AR test for higher order autocorrelation. The ARCH
test for autoregressive conditional heteroscedasticity, the Bera–Jaque test
for residual normality, Whiter test for heteroscedasticity and the REST test
for functional form misspecification were carried out. There is nothing to
suggest that our model is misspecified in any way.
The error correction term (Res4_1) has the expected sign and is
significant at the 1 per cent test level. This is a further confirmation that
the variables CMAE, CMAO, TOT, and RERM indeed co-integrate.
In addition, we ran the estimation excluding dummy variables to test for
parameter constancy. The insignificant Chow test value and the forecast
test value all point to parameter constancy over the period 1988 to 1996.
The model can thus be used for forecasting.
The preferred model results are presented in Table 13.5.
4.2.1 Interpretation of the results

With respect to the signs output of manufactures, real exchange-rate mis-
alignment and the liberalisation measures of 1984 and 1993 behaved as
Table 13.4 Testing for Co-Integration
Variables SBDW DF ADF Co-integration accept/reject
CMAE v CMAO 1.829 –3.885 –3.711 Accept

CMAE v TOT 1.681 –3.702 –3.562 Accept
CMAE v RERM 1.654 –3.690 –3.625 Accept
Note: Tests are applied on residuals.

Table 13.5 Results: the determinants of manufactured exports in Tanzania
Variable Coefficient Std error t-value
Constant 0.19995 0.28903 0.692

DLCMAO2 0.35357 0.19950 1.772
DLRERM –3.8352 1.3399 –2.862
DLRERM_2 –2.9556 1.2872 –2.296
DLTOT 3.1427 1.5071 2.085
Res4_1 –1.1777 0.20160 –5.842
s1984 1.1233 0.71479 1.572
s1986 –2.8081 1.0382 –2.705
s1993 1.6063 0.74558 2.154
Note: R2 = 0.697666; F(8,19) = 5.4806 [0.0011]; s = 0.903656; DW = 2.25; RSS = 15.51529265 for
9 variables and 28 observations.
postulated. This suggests that increased manufacturing output leads to

increased export of manufactures. The exchange rate variable confirms the
hypothesis that overvalued exchange rates act as a disincentive to
exporters. This is the macroeconomic policy instrument we were interested
in. The variable is significant at the 10 per cent level.
The terms of trade variable came with an unexpected sign. This can be
explained by the fact that in Tanzania’s history agricultural exports have
been the main source of foreign earnings, while manufactures have been
relatively unimportant. Thus manufactured exports are not likely to
respond significantly. However, the variable is insignificant at the conven-
tional test levels.
With regard to liberalisation measures the behaviour of the 1986 dummy
is as is to be expected. While the 1984 and 1993 measures were more tar-
geted to foreign exchange holdings and dealings respectively, the general,
economy-wide reforms of 1986 led to a greater inflow of imports, which in
the final analysis probably led to a reduced output of manufactures and
thus less ability to invest in exporting.
4.2.2 Comparison with other results

The positive responsiveness of manufacturing exports to foreign exchange
incentives corroborates the results of many such studies in Tanzania, for
example, Ndulu and Semboja (1996) and Rutasitara in 1996 (cited in
Mbelle et al., 1998 a,b) as well as the results of Pieper (1998):
(b) The determinants of manufacturing investments
While reporting these results we will not repeat the procedures outlined in
the estimation of the first model. 12 Only a few tests are given (Tables 13.6
and 13.7). The rejections in Table 13.7 may point to problems with the
Table 13.6 Test for unit roots: the manufacturing investment equation
Variable SBDWa DFa ADFa,b Order of integration
LAMS 1.432 –4.039 –2.857 I(1)

LARE 0.7079 –2.165 –1.364 I(1)
LMAI 0.3202 –2.331 –1.355 I(1)
LEFB 0.5424 –4.206 –1.788 I(1)
Notes:
a
Critical values at 5% level are: DF –3.567; ADF –3.587; SBDW 1.72.
b
Lags were used in the auxiliary equation for ADF.
Table 13.7 Testing for co-integration
Variables SBDW DF ADF Co-integration accept/reject
LMAI v LEFB 0.5098 –2.261 –0.1391 Reject

LMAI v LARE 0.5136 –1.91 –0.8599 Reject
LMAI v LAMS 0.3747 –1.553 –0.436 Reject
RES5 0.8633 –3.238 –0.9949 Reject
Note: Tests are applied on residuals.
testing statistic when the error correction term is also rejected. With this
assumption we then proceeded to include the error correction term in the
general model and re-estimated the equation. If the ECT turns out to be
significant and bearing a negative sign, then our assumption is correct. The
results of the general model confirmed the correctness of our assumption. 13
In Table 13.8 we present the final results.
4.2.3 Interpretation of the results

The real lending-rate variable came out with an unexpected sign, though
the coefficient was insignificant. This may reflect the fact that the rate has
been negative for quite a considerable period of time. This is documented
in Appendix Table A.2. Thus changes in real lending rate (DLARE), even
when lagged, did not deter manufacturing investments. Partly, this is also
due to the fact that investment in manufacturing was government-led.
Changes in the money supply (DLAMS), though not significant, did not
favour manufacturing investment (counter to the hypothesized positive
relationship). The same conclusion holds for the effectiveness of banking
facilities. Increased effectiveness of the banking system was not positively
related to manufacturing investment. The influence is however
insignificant.
As expected the liberalisation measures instituted in 1986 resulted (ini-
tially) in a declining share of industrial investment. This variable is
significant at the 1 per cent level.
Table 13.8 Results: the determinants of manufacturing investments in

Tanzania, preferred model (sample 1969 to 1996)
Variable Coefficient Std error t-value
Constant 0.32713 0.14338 2.282

DLARE 0.39923 0.23991 1.664
DLARE_1 0.64201 0.27986 2.294
DLARE_2 0.57344 0.21624 2.652
DLAMS –0.24067 0.17319 –1.390
Res5_1 –0.58199 0.18704 –3.112
s1986 –0.96219 0.29241 –3.291
DELFB_1 –53.804 41.971 –1.282
Note: R2 = 0.534871; F(7,20) = 3.2855 [0.0172]; s = 0.414869; DW = 1.74; RSS = 3.442327352 for
8 variables and 28 observations.
4.2.4 Comparison with other results

Though, to the best of our knowledge, no similar analyses exist for
Tanzania, our finding on the significant negative impact of liberalization
corroborates the results of many non-econometric studies, for example
those reported in Section 3, and the general impression that investible
funds tend to be channelled more into the non-real part of the economy,
for instance, services.
We set out to estimate the impact of macroeconomic policies on manufac-

turing performance in Tanzania, with special attention to liberalization
measures (a policy regime shift). Our main interest in carrying out this
exercise was to determine the influence of liberalization (a regime shift
away from a command and control economy). As our analysis has shown,
the impact of liberalization on manufacturing performance is not unam-
biguous. Some policies are counterproductive. This leads to the conclusion
that in order to ameliorate some of the negative sectoral impacts of macro-
policies, sector-specific policies have to be designed and targeted, and sub-
sequently implemented.
Appendix
Table A.1 Data for manufacturing exports equation
MAE MAO RERM TOT
1964 176 1 186 29.1 69

1965 200 1 203 29.0 68
1966 271 1 289 29.0 69
1967 390 1 315 29.0 69
1968 336 1 269 29.1 66
1969 389 1 523 34.9 82
1970 378 1 744 34.2 80
1971 459 2 113 35.0 70
1972 416 2 516 39.3 68
1973 439 3 048 34.9 84
1974 498 3 888 36.3 77
1975 566 4 546 43.2 66
1976 667 5 517 55.3 90
1977 673 6 524 53.2 100
1978 625 9 332 42.6 82
1979 1 003 9 832 49.3 82
1980 1 229 10 377 58.8 71
1981 802 10 828 74.8 61
1982 708 11 038 87.3 63
1983 791 12 768 94.5 65
1984 1 217 16 944 92.0 68
1985 1 425 20 562 99.7 64
1986 1 641 28 171 72.3 74
1987 4 036 43 521 34.1 64
1988 7 938 51 759 25.5 62
1989 11 430 95 825 23.0 58
1990 11 282 124 572 18.9 50
1991 7 672 135 783 19.0 50
1992 7 012 168 371 16.4 45
1993 5 680 207 096 16.7 45
1994 8 406 238 515 15.6 46
1995 11 852 267 137 15.6 50
1996 12 208 278 081 15.6 54
Sources: Bank of Tanzania, various; URT, Economic Surveys, various issues.

Table A.2 Data for manufacturing investment equation
MAI ARE AMS EFB
1966 14.8 4.3 3.0 10.22

1967 16.2 4.3 2.0 10.38
1968 13.2 4.3 –3.0 10.29
1969 17.0 4.4 23.0 10.27
1970 13.4 4.5 20.0 10.28
1971 11.0 4.5 16.0 10.30
1972 11.8 –0.5 16.0 10.29
1973 15.6 –1.5 14.0 10.29
1974 15.6 –11.5 21.0 10.28
1975 17.9 –17.5 22.0 10.26
1976 26.6 1.5 21.0 10.26
1977 37.2 –3.5 18.0 10.27
1978 36.7 –4.5 12.0 10.26
1979 36.9 5.5 38.0 10.25
1980 33.8 –18.0 24.0 10.28
1981 21.7 –14.0 17.0 10.28
1982 21.3 –15.5 18.0 10.30
1983 15.1 –18.5 16.0 10.27
1984 13.2 –25.5 4.0 10.27
1985 12.2 –17.0 25.0 10.33
1986 2.6 –16.0 26.0 10.34
1987 11.8 –1.0 28.0 10.35
1988 6.7 –2.0 30.0 10.35
1989 2.4 1.0 26.0 10.35
1990 3.0 –4.9 36.0 10.33
1991 4.8 2.2 26.1 10.29
1992 4.6 9.1 38.5 10.31
1993 4.0 13.8 28.8 10.28
1994 2.4 4.9 32.5 10.27
1995 2.5 10.7 26.2 10.31
1996 2.7 22.6 11.6 10.29
Sources: Bank of Tanzania, various; URT, Economic Surveys, various issues.
Notes
* Department of Economics, University of Dar es Salaam
1. The shares for manufacturing, mining, construction and utilities being 8.4%,
2.0%, 4.4% and 1.7% respectively (URT, 1999a).
2. ‘Mkakati wa Taifa wa Kusimamia Utekelezaji wa Sera za Kuondoa Umaskini’
(National Strategy for Poverty Alleviation) (URT, 1997).
3. Computed from URT (1999a), table 60, p. 150.
4. Computed from URT (1999a), table 64, p. 154.
5. Introduction of new products, higher expenditure in R&D and product
development.
6. Tanzania recently (July 1999) announced it was pulling out of COMESA.

7. Real exchange rate is defined as the ratio of the consumer price index (CPI) to
the nominal effective exchange-rate index multiplied by the weighted price
indices of major trading partners. A misalignment thus shows the extent of
departure from the equilibrium (of a chosen base year) for which the economy is
considered to be in equilibrium. For extensive treatment see for example
Elbadawi (1998).
8. This is the most reliable source in Tanzania (accepted by all).
9. ‘Real Exchange Rate Policy and Export Competitiveness in Sub-Saharan Africa’,
paper presented at the UNU/WIDER Conference, Addis Ababa, 16–20 March
1998.
10. This model can be obtained from the author.
11. A decline in the value of these statistics is an indication of increased model
parsimony.
12. These can be obtained from the author.
13. This can be obtained from the author.
14
The Urban Informal Manufacturing
Sector in Tanzania: Neglected
Opportunities for Socioeconomic
Development
Herman Gaillard and Amber Beernink*
1 Introduction
This chapter discusses the potential contributions of the urban informal

sector to socioeconomic development. It presents a discussion of studies of
the Tanzanian informal sector in the context of a review of comparative
data on the informal sector in developing countries.
First, in Section 2 we explain why there is increasing interest in the urban
informal sector during the past decades, both in Tanzania and elsewhere.
Then, the state of theoretical and empirical research is addressed in Section 3.
Methodological problems and shortcomings in research on the informal
sector are identified in Section 4. In Section 5 empirical evidence is pre-
sented with regard to the following research questions:
• the degree of empirical support for the formal–informal dichotomy (5.1)

• the nature of the relations between the informal sector and the formal
sector (5.2)
• the potential contribution of the Tanzanian urban informal sector to
socioeconomic development (5.3)
• constraints for the future growth of the urban informal sector (5.4).
Finally, Section 6 contains policy recommendations with regard to the

urban informal sector in Tanzania.
Studies of the urban informal sector in Tanzania are rather scarce (for an
overview see References, ‘Relevant informal sector studies in Tanzania’).
Furthermore, most surveys are either very global or very restricted in scope
and content. This is not typical of Tanzania, but applies to most informal-
sector studies in developing countries all over the world. The still scanty
318
Urban Informal Manufacturing Sector 319
detailed information concerning the urban informal sector of Tanzania will

become more meaningful, when compared – whenever possible – with data
from other case studies and surveys (see References, ‘List of case studies’, for
material included in this survey). It should be stressed, however, that all
the data about urban informal sectors in developing countries should be
interpreted with caution. Most of them are estimates based on limited
samples.
The data in the Appendix indicate that urban informal sector studies in
Tanzania are basically quite typical of studies performed elsewhere. The
scores for Tanzanian studies are either similar to the average scores of case
studies and surveys in a wide range of countries, or they are within the
ranges observed in these other studies. Thus, by and large, the same charac-
teristics can be observed in Tanzania as in the majority of urban informal
sector studies carried out elsewhere.1
2 Interest in the urban informal sector
In most developing countries in the post-war period, increasing population

pressure and insufficient job creation in the formal sector, together with
rural–urban migration due to a variety of push and pull factors, have
created a spillover of employment into the urban informal sector.
According to some observers, the activities in this sector are illegal, mar-
ginal, parasitic or even criminal; according to others this sector represents
productive, small-scale ‘self-employment’ and therefore should be consid-
ered as a potential for stimulating growth, employment and achieving a
less skewed income distribution.
The interest in the meaning of development and how a developing
country may reach higher levels of welfare started after World War II.
Nowadays, there is general agreement on the fact that the first priority in
development is to combat poverty and its causes. Opinions on how to
realize this, however, still diverge widely.
In the 1950s the opinion dominated that development was only a matter
of more investment in capital, as was expressed by the use of the well-
known capital–output ratio. By the beginning of the 1960s it was realised
that this magic formula, so successfully applied in West European countries
after World War II, would not have the same effects in developing coun-
tries, because the necessary knowledge and know-how were lacking there.
More schooling (‘investment in human capital’) seemed to be the answer,
but soon it was demonstrated – among others by Galtung (1968) – that the
effects of improved schooling can only be modest in a global situation
characterised by an ossified international division of labour between rich
and poor countries. Nevertheless, until the end of the 1960s people were
still convinced that there was only one road to development. All invest-
ments should be directed towards the growth of the modern industrial
sector. Only this sector was supposed to be capable of achieving real

progress. This one-sided, implicitly unilinear way of thinking is most
clearly recognizable in the ideas of Rostow (1960) concerning the stages of
economic growth through which every country has to pass. Thus, in devel-
opment thinking attention was mainly directed to large-scale, capital-
intensive industries concentrated in urban areas.
This strategy was doomed to fail. In terms of employment the modern
industrial sector concerns a rather small sector, which often grew only
slowly. Employment creation in this sector lagged behind the growth of
production because of rising productivity. In the meantime, the population
was increasing rapidly along with rural–urban migration, leading to a
growth of the urban population by about ten per cent per year.
Furthermore, the modern industrial sector remained an enclave: forward
and backward linkages were limited. Also, there was attention for the infra-
structural development of only the largest cities – frequently only the
capital city – in order to attract industry and tourism. Smaller cities and
rural areas were neglected. At the end of the 1960s, it was more and more
clear that this strategy had resulted in a widening not only of the gap
between the rich and the poor countries, but also of the gap between the
poor and the rich within each developing country separately.
It is only then that international organisations – and to a much lesser
extent governments – began to show some interest in agriculture as a poss-
ible engine for development. When the results of this strategy also turned
out to be disappointing, to say the least, interest was directed to a number
of target groups, such as the working poor, small farmers, the landless, the
unemployed, slum-dwellers and the urban informal sector. The interest in
the urban informal sector can therefore largely be explained by the failure
of other strategies. As Bromley (1978) puts it rather cynically: ‘the informal
sector concept was adapted because it arose through effective communica-
tion channels at a convenient moment, and because it embodied policy
implications which were convenient for international organisations and
governments’.
3 Current theoretical points of view and suggested major

research topics
We would rightly be accused of being overly pessimistic if we were to state

that no progress can be observed during the past years, whether in theory
or in research, with regard to the phenomenon of the urban informal
sector. However, we would – also rightly – be accused of gross exaggeration,
if we stated that recently brilliant progress has been made with regard to
theoretical issues, in the field, or with regard to policy formulation.
The international theoretical debate is still largely concentrated on a
rather limited number of different approaches. Hardly any progress can be
noted in the debate. The advocates of the so-called sectoral approach (among
others, Souza and Tokman, 1976; van Dijk, 1980; Sethuraman, 1981;
Liedholm and Mead, 1987; De Soto, 1989) still demonstrate little consen-
sus, not only with regard to the number of sectors that should be distin-
guished (two or more), but also with regard to the criteria on the basis of
which such distinctions should be made. Neo-Marxist authors (for example,
Lebrun and Gerry, 1975; Davies, 1979; Guerguil, 1988) make a distinction
between different modes of production within one capitalist system, and
try to prove that ‘petty production’ is subordinated to – but also integrated
in – the capitalist mode of production. This approach is not very convinc-
ing in the light of the (scanty) empirical evidence. The third theoretical
approach, the so-called continuum approach (see, for example, Bromley and
Birbeck, 1984; Tezler and Molenaar, 1989), seems more promising. In this
literature, the assumption is made that it makes sense to distinguish two
sectors (formal and informal). However, instead of supposing that the two
sectors operate essentially independently of each other, it seems more prob-
able and realistic according to these authors that the two sectors interact
with each other in a continuous fluctuating manner, and that parts of the
informal sector can be dominated or even created by parts of the formal
sector.
Since the introduction of the formal–informal dichotomy (Hart, 1971)
some progress has certainly been made in empirical research. For example,
empirical research has demonstrated quite convincingly that it is incorrect
to declare the informal sector synonymous with the urban poor. In most
cases, paid workers in the formal sector earn considerably less than entre-
preneurs in the informal sector (see Table A.27). Nevertheless, one may
legitimately conclude that still relatively little is known about the urban
informal sector in developing countries.
The following major and often related research topics are frequently
mentioned in the literature:
• What exactly belongs to the urban informal sector? What are the
specific characteristics of this sector? To what degree do these character-
istics differ from those of the formal sector?
• What is the effect of the informal sector on income distribution and
poverty alleviation? To what degree and how does absorption of labour
take place? What are the advantages of the informal sector for the urban
economy or, put in more general terms, what is the relevance of the
urban informal sector for socioeconomic development?
• What kinds of relationships exist between the formal and the informal
sector? Do the two sectors function in relative isolation from each
other? Do they compete with each other or are they complementary to
each other? Should one speak of exploitation or mutual interest?
• What are the constraints on further growth of the informal sector?
4 Methodological problems
Our answers to the above questions can only be approximate and prelimi-
nary, mainly because of the methodological problems facing urban infor-
mal sector research. We will briefly discuss the main methodological
problems with regard to urban informal sector research as observed in most
of the case studies collected (see References). The scientific and the social
value of research results is mainly determined by their degree of comparabil-
ity, generalization and replicability.
The first issue is concerned with conceptual definitions, units of research
and theoretical frameworks. A large majority of case studies typically does
not give any definition at all of the concept of the urban informal sector or
provides only a vague description. Furthermore, different units of research
are used: sometimes individuals, sometimes households, and sometimes
production units. A theoretical framework, indicating what exactly is to be
investigated and how these elements are expected to be interrelated to each
other, is hardly ever given.
The second issue is mainly concerned with sampling. The large majority
of the case studies typically does not give any information at all about the
applied sampling method. Furthermore, samples are usually very small. The
population distribution according to economic activity is very seldom
taken into account, resulting in substantial bias that is mentioned only
very rarely by the researcher.
The third issue is especially related to the operational definitions. Hardly
any of the case studies provide the research instrument used, indicating
how different concepts actually have been measured.
In conclusion, much of the research on the urban informal sector in
developing countries is still of a predominantly pre-scientific nature. As a
result – as stated above – many of the relevant major research questions
can so far only be answered partially and in a preliminary fashion.
Although quite a few case studies have been executed, these studies are
largely of a pre-scientific nature. The lack of generalisation power and the
scant possibilities for comparison and replication seriously hamper theory
formation.
5 Empirical evidence
Nevertheless, an attempt will be made to give (preliminary) answers to the

major research questions mentioned in Section 3. These answers are based
on the results found in the case studies, summarised in the References and
the tables in the Appendix, grouping these questions under four different
headings. In the tables unweighted averages of the case study results are
presented for three continents. Results for Tanzania are reproduced in a
separate column.
5.1 Empirical support for the formal–informal dichotomy

Appendix A.1 indicates that many basic characteristics of informal-sector
organisations seem to be quite different from those of formal-sector organi-
sations. Observed differences concerned location, type, size, age, owner-
ship, mode of production, product mix and product variation.
Typically, informal-sector organisations:
• have a variable location or a fixed location with a temporary structure

(Table A.2)
• are predominantly one-man firms (Table A.3)
• have recently been established (Table A.5)
• have only one or two employees, frequently apprentices (Tables A.3 and
A.4)
• have machines and tools among their own property, but not the work-
shop (Table A.6)
• produce mainly to order (Table A.7)
• demonstrate a very limited product mix coupled with large product vari-
ation (Tables A.8 and A.9).
On the basis of these findings, one may cautiously conclude that the infor-
mal-sector organisation does have specific characteristics, diverging sub-
stantially from those of the formal-sector organisation and that, therefore,
the distinction formal–informal is a meaningful one, as stated by protago-
nists of the sectoral approach (see Section 3).
5.2 The relations between the informal sector and the formal
sector
Various different kinds of relationships have been investigated in the case

studies: output and input relations, sources of capital, competition, type of
output market, actual and desired types and degrees of government inter-
ference and formal education received (see Appendix A.2).
Looking at the data one can hardly maintain that in general the informal
sector is overwhelmingly exploited by the formal sector. To a substantial
degree ‘exploitation’ seems to occur by formal sector households, not
however by formal sector organisations (Table A.10). One could state that the
informal sector to a large extent fulfils the basic needs not only of informal-
sector households but also of many formal-sector households. Inputs origi-
nate primarily from other informal-sector organisations (Table A.11); formal
sources of capital are hardly available and/or used (Table A.12); government
intervention is only marginal (Table A.16), although specific kinds of gov-
ernment intervention are frequently desired (Table A.17).
All in all, it seems that relationships between the formal and the infor-
mal sector are rather limited. Only the following relations have been
observed. Formal education is frequently obtained by members of the infor-

mal sector (Table A.15). Much competition from formal-sector organisa-
tions is experienced by a large majority of informal-sector organisations
(Table A.13). However, it should be noted that this competition is not felt
to be more fierce than competition between informal-sector organisations,
and it occurs in the context of still expanding markets (Table A.14). One
may conclude that these rather limited relationships are consistent with
the view expressed in Section 5.1, namely, that it makes sense to distin-
guish between formal and informal sectors. One could even speak of two
societies in one country, each developing rather autonomously.
5.3 Relevance for socioeconomic development

Various aspects of the informal sector which are of relevance for socioeco-
nomic development have been assessed in the case studies. They include:
the share of the informal sector in the urban working population, the share
in urban income generation, the growth of employment, the type of
employment created, the use of capital, sources of training, costs and dura-
tion of informal-sector training, income distribution and income generation
(see Appendix A.3). There can no doubt that both with regard to urban
income and especially with regard to urban employment creation the con-
tribution of the informal sector is considerable, notably in African countries.
In Tanzania 56 per cent of the urban active population is working in the
informal sector, generating 32 per cent of total urban income (Table A.18).
It is also well known – and supported by the data – that the continuous
intra-regional and inter-regional migration (Table A.24) is largely absorbed by
the informal sector, as indicated by its average annual growth of employment
of about 15 per cent. The figure for Tanzania is 16 per cent (Table A.19).
It also seems clear that employment creation not only concerns self-
employment but also wage-workers, apprentices and (unpaid) family
workers. In Tanzania the latter category strongly predominates: 52 per cent
of informal-sector workers consist of unpaid family workers (Table A.20).
These large amounts of different types of employment are created with
only a minimum of capital outlay. Initial investments per person are aston-
ishingly low and only a fraction of those in the formal sector, though there
are differences between (sub)sectors. Data for Tanzania, as far as available,
do not deviate substantially from the observed patterns for other countries
(Table A.21).
Although very limited use is made of capital, urban informal-sector
entrepreneurs manage to create an income for themselves that on average
is well above the legal monthly minimum wage in the formal sector.
Differences between the various (sub)sectors of economic activity are sub-
stantial, but in all (sub)sectors – on average – the income of the entrepre-
neur in the informal sector is above the legal monthly minimum wage in
the formal sector (Table A.27).
Although employees and apprentices on average earn less than the legal
monthly minimum wage in the formal sector, it should be noted that the
income distribution within the informal sector is much more equal than in
the formal sector of developing countries. In Tanzania on average employ-
ees in the informal sector even seem to earn more than the legal monthly
minimum wage (Table A.25).
Furthermore, the training of informal-sector entrepreneurs, workers and
apprentices is overwhelmingly done within the urban informal sector at
ridiculously low costs, again in comparison with the formal sector. Thus, in
Tanzania 91 per cent of the informal-sector entrepreneurs received their
training in the informal sector, either ‘on the job’ (35 per cent) or via an
apprentice system (57 per cent). The latter system seems predominant in
Africa (Table A.22).
Summarising, the available data clearly demonstrate the relevance of the
urban informal sector for socioeconomic development. The urban informal
sector generates substantial amounts of employment, income and training,
absorbs migration and provides a relatively more equal income distribu-
tion, all this at only minimal capital costs. Thus, on the basis of our survey
of the available empirical evidence, we may conclude that the urban infor-
mal sector – much more than its formal counterpart – can create a substan-
tial middle class, ensuring socioeconomic development, not only for
the happy few in the formal sector, but for larger segments of the whole
population.
5.4 Constraints for future growth of the Tanzanian urban informal

sector
In Section 5.3 we emphasised the relevance of the urban informal sector for
development. If our conclusion is correct, the logical consequence should be
to support this sector. In order to arrive at meaningful policies, the first step
should be to identify the major constraints to the further development of the
informal sector. Next, stimulating policies can be formulated accordingly.
In the Appendix some indications have been gathered concerning the
major constraints experienced in the Tanzanian urban informal sector.
Unavailability of capital or credit is reported as the most important problem
(mentioned by 41 per cent of the respondents) for starting informal-sector
enterprises (Table A.28). Potential starters have difficulty in starting busi-
nesses because they have little capital or credit to purchase inputs and raw
materials, and have limited access to credit. Credit institutions often set
difficult conditions which have to be fulfilled in order to get credit, but
which rarely can be met. These include possession of fixed assets or leasehold
land title deeds to act as securities, or the preparation of detailed feasibility
studies for the proposed projects. Lack of equipment and spare parts is the
second most important problem for starters: it is mentioned by 16 per cent of
the respondents. The other problems mentioned seem far less important.
Unavailability of credit is again mentioned (by 32 per cent of the respon-

dents) as the major problem which hinders the day-to-day operation of infor-
mal-sector enterprises, followed by non-payment of debts (mentioned by
20 per cent) and, again, unavailability of capital equipment (mentioned by
15 per cent) (Table A.29). Non-payment of debts means that net profit is
eroded by customers who do not pay their debts promptly, and this reduces
the amount of capital for further investment.
It is not surprising then that by far the most important form of assistance
needed – stated by 61 per cent of the respondents – is access to loans
(Table A.30). As mentioned earlier, the present conditions for loans cannot
be met by most informal entrepreneurs. What seems to be required at the
least, therefore, is to simplify the procedures and to soften the credit terms
in order to enable the entrepreneurs to secure the loans within a short time.
Permanent sites were the second most important form of assistance
needed, stated by 16 per cent of all respondents. It is therefore also
necessary that municipal and city councils reserve a certain number of
plots for setting up informal-sector enterprises.
It should be noted that training is the least important form of assistance
needed, mentioned by only 2 per cent of the respondents. This finding
seems congruent with the conclusion in Section 5.3 that the informal
sector is self-supporting with regard to training. Thus, most entrepreneurs
do not recognise it (yet) as a needed form of assistance. However, it does
not seem wise to conclude from this that training for upgrading should not
be part of any assistance programme. The few respondents that did indicate
that training was a form of assistance needed overwhelmingly mentioned
technical training, both on the job (46 per cent) and formal (45 per cent).
6 Policy formulations for the urban informal manufacturing

sector in Tanzania
As stated in the introduction, urban informal sector studies – both surveys

and case studies – are either very global or very limited in scope and
content, both in Tanzania and elsewhere. As a consequence, policy recom-
mendations for the urban informal sector can only be very global, since
specifications cannot be made, or very limited, since the content of the
research has been very limited.
In 1972 the Kenya mission of the ILO (1972) already suggested a number
of global measures to stimulate the development of the urban informal
sector. Since then similar ideas have been put forward, in a multitude of
publications, albeit in different combinations. One can safely conclude
that, generally speaking, the attempts to formulate a policy for the urban
informal sector have been either non-existent or too global. For example,
in the Tanzanian informal sector studies it also becomes clear that access to
credit facilities is of vital importance (see Section 5.4). However, the most
important questions have not been asked: How much credit? All at once or
in phases? How much in kind and how much in money? For which activi-
ties? For which enterprises? Credit for what (which equipment, which raw
materials, which semi-manufactured products, which tools, which
premises)? How must credit be supervised and be repaid?
In short, specification of the measures according to activity and
phasing in time is lacking. Furthermore, the mutual relationships
between the measures are often obscure, and alternatives are seldom for-
mulated. It is even more rare for alternatives to be compared with each
other on the basis of explicit criteria, as is necessary for deciding on
policy priorities.
When a government has only limited resources – a rather typical situa-
tion not only for most governments in developing countries – it seems
obvious that it should distribute these scarce resources as optimally as pos-
sible, in the sense that the expected contribution to future growth and
development is as high as possible. However, this will only be possible with
the help of a detailed selection procedure based on questions as outlined
above, resulting in alternative congruent packages of mutually supporting
measures for specific sectors and subsectors. As long as such efforts are not
made, only global policy formulations will be possible, leading to a subop-
timal allocation of scarce resources.
Appendix
Table A.1 Classification of urban informal sector enterprises according to ISIC
Tanzania Case studies
X R
Agriculture 8 5 10–3
Manufacturing 24 15 28–10
Food 2 5 8–2
Clothing 5 6 10–3
Wood 6 1 7–<1
Non-metal 2 1 3–<1
Metal 1 1 4–1
Other 8 1 10–0
Construction 7 6 11–4
Trade 51 53 68–46
Transport 2 6 8–1
Services 6 15 20–6
Sources: Tanzania: References 1, 2, 3, 6, 7; (case studies) 19–21, 26–28, 32, 34–36, 39, 44–48,
53–54, 57–58, 63–64, 67–68, 75.
Notes: (Tables A.1–A.27); x = average of the case studies; (x) = variable included only when or
more case studies provide data; R = range of the case studies.
Table A.2 Location according to type
% of organisations with Tanzania Africa (x) Asia (x) LA (x) R (x)
Variable location 24 24 15 41 75–7

Fixed location, temporary structure 32 46 41 11 70–7
Fixed location, permanent structure 39 30 39 48 88–10
Sources: Tanzania: References 2, 4, 5; Africa: References, 20, 25, 34, 36, 53, Asia: References
44–45, 48, 54, 57–58, 63, LA: References 19, 21, 27, 32, 47, 51, 67, 75.
Table A.3 Type of organisation
% of organisations Tanzania Africa (x) Asia (x) LA (x) R
One man firms 74 51 78 63 90–15

Partnerships 1 6 5 10 15–1
Master and apprentices/employees 26 44 17 27 82–10
Sources: Tanzania: as Table A.2; Africa: as table A and 37, 39 50, 66, 70; Asia: References 42, 48,
54, 57, 58, 63–64; LA: as Table A.2, and 25, 55, 58, 62, 65.
Table A.4 Size of organisation, entrepreneur included
Africa (x) Asia (x) LA (x) R
No. of persons per unit 3.3 1.8 2.0 5.9–1.1
Sources: Africa: as Table A.3, Asia: References 24, 44–45, 48, 54, 57, 58, 53, 64, 72; LA: as
Table A.3.
Table A.5 Age of organisations
% of organisation Tanzania Africa (x) Asia (x) LA (x) R
<1 year 16 17 10 9 28–7

1–2 years 23 16 12 10 25–7
Sources: Tanzania: References 2, 4; Africa: 20, 26, 33, 34, 35, 36, 46, 49, 50, 54; Asia: References
44, 45, 48, 58, 54, 57, 75; LA: References 24, 27, 38, 47, 55, 67.
Table A.6 Ownership
% of entrepreneurs with Africa (x) Asia (x) LA (x) R
Machines/tools in own property 87 80 95 100–74

Workshop in own property 15 28 43 65–4
Hired workshop 61 50 38 81–20
Free use of workshop 24 22 19 35–5
Sources: Africa: References, 20, 34, 35, 36, 37, 40, 49, 50, 53; Asia: References 44, 45, 48, 54, 57,
63; LA: References 19, 21, 24, 51, 62, 65.
Table A.7 Mode of production
% of organisations that Africa (x) Asia (x) LA (x) R
Produce only or mainly to order 67 62 64 80–62

Produce only or mainly in stock 33 38 36 60–20
Source: Africa: References 2: 20, 34, 36, 49, 50, 53; Asia: References 2: 48, 58, 54, 59, 63;
LA: References 2: 17, 18, 21, 27, 41, 67, 71.
Table A.8 Product mix
% of organisations with Africa (x) Asia (x) LA (x) R
1–2 kinds of output 85 76 72 92–64

>2 kinds 15 24 28 36–8
Source: Africa: References 20, 34, 36, 39, 50, 53; Asia: References 48, 58, 54, 59, 63;
LA: References 17, 18, 21, 27, 41, 67, 71.
Table A.9 Product variation
% of organisations with Africa (x) Asia (x) LA (x) R
Only or mainly standard output 22 26 28 36–12

Only or mainly output to customer specification 78 74 72 95–64
LA: References 17, 18, 21, 27, 41, 67, 71.
Table A.10 Type of customers
% of output going to Tanzania Africa (x) Asia (x) LA (x) R
Individual households 93 75 85 81 100–20

in informal sector 35 35 32 60–15
in formal sector 40 50 49 72–17
Industrial organisations 7 25 15 19 80–0
in informal sector 5 17 8 13 32–4
in formal sector 2 8 7 6 49–0
Source: Tanzania: References 16, 18, 19; Africa: References 20, 25, 34, 35, 36, 37, 49, 50, 53, 70;
Asia: References 26, 38, 48, 58, 54, 57, 59, 63; LA: References 15, 16, 17, 18, 19, 21, 32, 57, 67,
71, 75.
Table A.11 Type of suppliers
% of inputs coming from Tanzania Africa (x) Asia (x) LA (x) R
Informal sector organisations 75 66 79 66 90–22

Formal sector organisations 25 33 21 34 78–10
Importation 0 1 0 0 2–0
Source: Tanzania: References 16, 18; Africa: References 20, 25, 34, 35, 36, 50, 53, 70;
Asia: References 26, 38, 43, 48, 58, 54, 57, 59, 63; LA: References 15, 16, 17, 18, 19, 21, 32, 57,
67, 71, 75.
Table A.12 Main sources of capital
Tanzania Africa (x) Asia (x) LA (x) R
Income from sales 47 74 79 78 93–51

Money lenders 0 1 2 1 3–1
Personal networks (family, friends) 52 17 12 11 29–2
Advance from customers 0 2 1 1 5–1
Credit from supplies 0 4 2 1 15–1
Banks, government 1 2 4 8 8–1
Source: Tanzania: References 16, 17, 18; Africa: References 20, 22, 25, 28, 34, 35, 36, 39, 49, 66,
70; Asia: References 26, 38, 42, 43, 48, 58, 54, 57, 59, 63; LA: References 15, 16, 17, 18, 19, 21,
24, 27, 32, 38, 55, 71, 75.
Table A.13 Competition
% of entrepreneurs that experience much competition Africa (x) Asia (x) LA (x) R
with
Informal sector organisations 72 65 58 84–50

Formal sector organisations 67 75 60 94–40
LA: References 15, 16, 17, 18, 19, 21, 24, 27.
Table A.14 Type of output market
% of entrepreneurs that mention Africa (x) Asia (x) LA (x) R
A decrease in demand 17 22 23 40–4

A constant demand 17 38 35 50–10
An increase in demand 67 40 42 86–40
Source: Africa: References 20, 25, 28, 34, 35, 39, 40, 46, 53, 68; Asia: References 48, 58, 54, 57,
59, 63; LA: References 17, 18, 19, 21, 38, 71, 75.
Table A.15 Formal education received
Tanzania Africa (x) Asia (x) LA (x) R
No. of school years – 4.3 5.2 5.5 9–3

% with no formal education 17 54 24 39 74–3
% with primary education 77 29 65 46 97–12
% with secondary education 2 16 11 15 60–0
% with vocational education 2 4 9–0
Source: Tanzania: References 16, 17, 18; Africa: References 20, 22, 28, 34, 35, 36, 46, 49, 54;
Asia: References 26, 43, 44, 45, 48, 54, 57, 59, 63, 64; LA: References 17, 18, 21, 57, 65, 75, 73.
Table A.16 Actual government interference
% of enterprises Africa(x) Asia(x) LA(x) R
With a government licence 20 25 12 25–12

With inspection by government 14 21 12 21–12
That pay taxes 15 20 20–15
With government as a customer 4 20–0
Source: Africa: References 20, 28, 34, 35, 39, 40, 49, 68; Asia: References 26, 48, 54, 57, 59, 64;
LA: References 17, 18, 21, 32, 67, 71, 75.
Table A.17 Desired government interference
% of enterprises that desire on the part of the government Africa(x) Asia(x) R
Professional training courses 34 58 90–0

Credit facilities 62 76 80–55
Infrastructural provisions 46 59 66–40
Measures to stimulate demand 34 23 55–10
Source: Africa: References 20, 28, 34, 35, 52, 53, 68; Asia: References 34, 48, 54, 57, 59, 64.
Table A.18 Size of urban informal sector
Tanzania Africa Asia LA
(x) R (x) R (x) R
% of urban active population

working in informal sector 56 50 73–31 40 69–19 42 60–24
% of urban income originating
from informal sector 32 40 55–28 29 45–16 29 40–14
Source: Tanzania: References 15, 16, 18, 19, 20–23; Africa: References 20, 28, 33, 34, 35, 36, 37,
39, 40, 46, 49, 52, 53, 60, 61, 68, 70; Asia: References 34, 38, 42, 44, 45, 48, 54, 57, 63, 64, 72,
74; LA: References 15, 16, 17, 18, 19, 21, 32, 57, 67, 71, 75.
Table A.19 Growth of employment, average annual
% growth, average annual Tanzania Africa(x) Asia(x) LA(x) R
In formal urban sector 5

In informal urban sector 16 15 9 9 28–6
53, 68; Asia: References 26, 44, 45, 48, 54, 57, 63, 64; LA: References 15, 16, 17, 18, 19, 21, 32,
57, 67, 71, 75.
Table A.20 Type of employment
% of workers that are Tanzania Africa(x) Asia(x) LA(x) R
Wage workers 20 40 60 77 83–7

Unpaid (family) 52 10 35 16 50–7
Appentices 28 50 5 7 86–0
68, 70; Asia: References 38, 43, 44, 45, 48, 58, 54, 57, 59, 63, 74; LA: References 15, 16, 17, 18,
19, 21, 67, 71, 75, 73.
Table A.21 Use of capital (initial investments per person/legal monthly

minimum wage, formal sector)
Tanzania Africa
Informal sector Formal sector FS/IS

(x) R (x)
All activities 1.8 10.0–0.5 25.0 13.9

Manufacturing 1.4 2.5 8.5–0.9
Food, beverages 1.4 1.5–1.2 50.0 35.7
Textiles 1.8 3.4–0.5 20.0 11.1
Apparel 2.2
Leather 1.2 1.7–0.9 64.0 53.3
Wood 2.6 5.4–1.0
Paper 7.9
Metal 4.7 6.0–2.3 20.0 4.3
Other 5.0 8.5–2.0
Construction 5.4 10.0–0.7
Trade 1.4 1.3 2.0–0.5
Transport 4.0 6.4–1.6
Services 2.3 3.0–1.5
Repair/maintenance 0.7 0.8–0.5
Source: Tanzania: References 16, 18; Africa: References 20, 28, 34, 35, 39, 40, 53.
Table A.22 Training, sources of training for informal-sector entrepreneurs
% of entrepreneurs that received their training Tanzania Africa(x) LA (x) R
In the informal sector 91 88 85 94–82

‘On the job’ 35 23 70 70–0
By the apprentice system 57 65 15 92–15
In the formal sector 9 12 15 18–6
Government 2 3 4 4–3
Private enterprises 3 6 8 10–2
Training institutions 4 3 3 8–0
68; LA: References 15, 16, 17, 18, 19, 57, 71.
Table A.23 Training, costs and duration of informal-sector training
Africa(x) R
Total costs (US dollars) 78 125–50

Duration (no. of years) 2.7 4.5–0.4
Source: Africa: References 20, 28, 36, 37, 53, 68.

Table A.24 Migration, overall (% of enterpreneurs not born in town of actual

residence) and according to type
Africa(x) Asia(x) LA(x) R
Overall 76 37 64 98–14
International 7 0 3 13–0
Inter-regional 32 10 11 42–6
Intra-regional 36 27 50 60–4
Source: Africa: References 20, 36, 37, 53; Asia: References 48, 54, 57, 63, 64; LA: References 17,
18, 32, 57, 75.
Table A.25 Income-distribution, wage structure in the informal sector
Tanzania Africa(x) Asia(x) LA(x)
Payment, apprentice 0.14

Wage, employee 1.22 0.71 0.33 0.50
Income, entrepreneur 2.60 2.14 1.33 1.50
Legal minimum wage in formal sector 1.00 1.00 1.00 1.00
Source: Tanzania: References 16, 18, 19; Africa: References 20, 34, 35, 36, 37, 53;
Asia: References 43, 48, 54, 57, 49, 63; LA: References 15, 16, 17, 18, 19, 21, 55, 71, 73.
Table A.26 Income-Distribution, wage structure of employees according to

subsector
Africa(x) Asia(x)
Trade 1 1
Services 1.3 1.8
Construction 4.9
Manufacturing 2.4
Source: Africa: References 20, 34, 35, 36, 37, 53; Asia: References 43, 48, 54, 59, 63.
Table A.27 Entrepreneur income informal-sector (monthly income

owner/legal monthly minimum wage, formal sector)
Tanzania Africa(x) Asia(x) LA(x)
All activities 2.6 2.1 1.4 1.5

Manufacturing 2.7
Food 3.5
Textiles 3.8
Apparel 2.2
Leather 1.3
Wood 3.4
Metal 2.4
Other 4.0
Construction 7.0
Trade 1.6
Transport 3.8
Services 2.5
Repair/maintenance 1.3
Source: Tanzania: References 16, 18; Africa: References 20, 34, 35, 36, 37, 53; Asia: References 43,
48, 53, 54, 57, 59, 63; LA: References 15, 16, 17, 18, 19, 21, 55, 71, 73.
Table A.28 Most important problem faced when establishing enterprises
Tanzania (%)
1. Capital or credit 41
2. Lack at equipment/spare parts 16
3. Finding premises 12
4. Market/customers 11
5. Government regulations 6
6. Raw materials 6
7. Transport 3
8. Skilled workers 1
9. Service, e.g. water 1
10. Other 3
Source: (Tables A.28–A.31) Planning Commission, Ministry of Labour and Youth Development
(1991).
Table A.29 Most important problem encountered during day-to-day operation
Tanzania (%)
1. Credits 32
2. Non-payment debts 20
3. Unavailability of capital equipment 15
4. Raw materials 7
5. Transport 5
6. Space 4
7. Management skill 3
8. Regulations/law 2
9. Taxes/licence fees 2
10. Thieves 2
11. Spare parts 1
12. Skilled personnel 0
13. Other 8
Table A.30 Most important form of assistance needed
Tanzania (%)
1. Loans 61
2. Permanent site 16
3. Raw materials 6
4. Transport 5
5. Marketing 4
6. Modern technology 4
7. Government regulations 2
8. Training, self 2
9. Training, workers 0
10. Other 0
Table A.31 Training needed by type
Tanzania (%)
1. Technical skills – on the job 46

2. Technical skills – formal 45
3. Commercial/financial 6
4. Literacy 3
5. Managerial 0
Notes
1. The list of Tanzanian studies in the References (1–14) is complete and updated.
The list of case studies in the References (15–75) contains only two references
after 1990. Thus the comparative database refers to studies from the period
1975–90.
References
Relevant informal sector studies in Tanzania

Informal sector: surveys
1. Aboagye, A. A. (1985), An Analysis of Dar es Salaam’s Informal Sector Survey, Addis
Ababa: Jobs and Skills Programme for Africa.
2. Planning Commission, Ministry of Labour and Youth Development (1991), The
National Informal Sector Survey, Dar es Salaam.
3. Planning Commission, Ministry of Labour and Youth Development (1995), Dar
es Salaam Informal Sector Survey, 1994/5, Dar es Salaam.
4. Tegelaars, M. (1994), ‘Performance Upgrading of Informal Building Contractors
in Dar es Salaam (Tanzania)’, Eindhoven: MSc Thesis, Technology and
5. Treffers, M. (1996), ‘The Informal Building Process in Dar es Salaam, Tanzania. A
Contribution to the Upgrading of the Informal Construction Sector and the
Formulation of a Management Plan for Hazard Land’, Eindhoven: MSc Thesis,
Technology and Development Studies, Eindhoven University of Technology.
Informal sector: estimates

6. Bagachwa, M.S.D. (1993), ‘Estimates of Informal Parallel and Black Market Activities
in Tanzania’, paper for the International Conference on Development
Challenges and Strategies for Tanzania: An Agenda for the 21st Century, Dar es
Salaam: Economic Research, October 25–1993.
7. Bagachwa, M.S.D. and A. Naho (1995), ‘Estimating the Second Economy in
Tanzania’, World Development, 23, 8, 1387–99.
8. Maliyamkono, T.L. and M.S.D. Bagachwa (1990), The Second Economy in
Tanzania, London: Eastern African Studies.
9. Mtatifikolo, F.P. (1995), ‘The Subterranean Economy in Tanzania. Its Structure
and Annual Aggregates mid-1970s to mid-1980s’, unpublished paper.
Informal sector: policy

10. Bagachwa, M.S.D. (1983), ‘Structure and Policy Problems of the Informal
Manufacturing Sector in Tanzania’, ERB Paper 83.1, Dar es Salaam: Economic
Research Bureau.
11. Bagachwa, M.S.D. (1995), ‘The Informal Sector Adjustment in Tanzania’, in: L.A.
Msambichaka, A.A.L. Kilindo and G.D. Mjema (eds), Beyond Structural Adjustment
Program in Tanzania, Successes, Failures and New Perspectives, Dar es Salaam:
Economic Research Bureau.
12. Bienefield, M. (1975), ‘The Informal Sector and Peripheral Capitalism: The Case
of Tanzania’, Institute of Development Studies Bulletin, 6, 3, 33–73.
13. ILO (1993), The State and the Informal Sector in Tanzania: Report of a Study on the
Development Policies and Institutional Environment for the Informal Sector in
Tanzania, Addis Ababa: Jobs and Skills Programme for Africa.
14. Vitta, P.B. (1985), ‘The Informal Sector in Eastern Africa: Selected Policy-Related
Issues’, Africa Development, 10, 4, 59–71.
List of case studies

15. Alba Vega, C. and D. Kruyt (1988), Los Empresarios y la Industria de Guadalajara,
Guadalajara: El Colegio de Jalisco, Mexico.
16. Alonso, J. (1979), ‘The Domestic Seamstresses in Nezahualcoyad: A Case Study
on Feminine Over-Exploitation in a Marginal Urban Area’, New York: PhD
Thesis, New York University.
17 Alonso, J. (1983), ‘Domestic Clothing Workers in the Mexican Metropolis and
their Relationship’, International Labour Review, 131, 3, 161–172.
18. Arias, P. (1985), Guadalajara, la Gran Ciudad de la Pequena Industria, Michoacan:
El Colegio de Michoacan.
19. Arizpe, L. (1977), ‘Women in the Informal Labour Sector: The Case of Mexico
City’, Signs, 3, 1, 25–37.
20. Aryee, G. (1981), ‘The Informal Manufacturing Sector in Kumasi’ in S.
Sethuraman (ed.) (Reference 69).
21. Arocena, M. (1982), ‘Forms of Production in the Shoe Industry in Léon,
Guanajuanto’, Austin: PhD Thesis, University of Feras.
22. Baron, J. and W. Bielby (1984), ‘The Organization of Work in a Segmented
Economy’, American Sociological Review, 49, 4, 454–73.
23. Beneria, L. and R. Roldan (1987), The Crossroads of Class and Gender: Industrial
Homework Subcontracting and Household Dynamics in Mexico City, Chicago:
University of Chicago Press.
24. Birkbeck, C. (1978), ‘Self Employed Proletarians in an Informal Factory: The Case
of Cali’s Garbage Dump’, World Development, 6, 9–10.
25. Bosch, E. (1983), Marktvrouwen van Bobo, het Leven en Werken van Handelaarsters
in de Stad Bobo-Dioulassa in Boven Volta, Leiden: Instituut voor Culturele
Antropologie.
26. Bose, A. (1974), The Informal Sector in the Calcutta Metropolitan Economy, ILO,
WEP 2-19, WP 5.
27. Bromley, R.J. (1978), ‘Organization, Regulation and Exploitation in the So-Called
Urban Informal Sector: The Street Traders of Cali Columbia’, World Development,
Sept.-Oct., 1161–86.
28. Bromley, R. and C. Gerry (ed.) (1979), Casual Work and Poverty in Third World
Cities, Chichester: Wiley.
29. Bunster, X. and E. Chaney (1985), Sellers and Servants, Working Women in Lima,
Peru, New York: Praeger.
30. Calleja Pinedo, M. et al. (1980), Unidad Doméstica y Organización del Trabajo de la
Industria Calzado en León, Guanajuato.
31. Conradi, P. (1988), Los Campamentos: Huishoudens en Overlevingsstrategieën; een
Exploratief Onderzoek naar het Functioneren van Overlevingsstrategieën in een Lage
Inkomenswijk in San Louis Potosi (Mexico), Geografisch Instituut, Rijksuniversiteit
Utrecht.
32. Defranco, S.D. (1980), Employment and the Urban Informal Sector: The Case of
Managua, Ann Arbor, Michigan: Xerox Univ. Microfilms.
33. Diemer, G. and W. van der Laan (1981), ‘The Informal Sector in Historical
Perspective: The Case of Tunis’, Cultures et Développement, 13, 1–2.
34. Dijk, M.P. van (1980), ‘De Informele Sector van Ouagadougou en Dakar’,
Amsterdam: PhD Thesis, Vrije Universiteit.
35. Effron, L. (1980), The Informal Sector in Abidjan (Ivory Coast) in: S. Sethuraman
(ed.) (Reference 69).
36. Fapohunda, O.J. (1981), ‘Human Resources and the Lagos Informal Sector’ in
S. Sethuraman (ed.) (Reference 69).
37. Fowler, D.A. (1981), ‘The Informal Sector in Freetown: Opportunities for Self-
Employment’, in S. Sethuraman (ed.) (Reference 69).
38. Gelder, P. van (1985), Werken onder de Boom: Dynamiek en Informele Sector: de
Situatie in Groot-Paramaribo, Dordrecht: Foris Publications.
39. Gerry, C. (1974), ‘Petty Producers and the Urban Economy: A Case Study of
Dakar’, Working Paper, World Employment Programme Research, no. 8, Geneva:
International Labour Office.
40. Gerry, C. (1978), ‘Petty Production and Capitalist Production in Dakar: The
Crisis of the Self-Employed’, World Development, 6, 9/10, 1127–60.
41. Gonzalez de la Rocha, M. (1984), ‘Domestic Organisation and Reproduction of
Low-Income Households, the Case of Guadalajara, Mexico’, Manchester: PhD
Thesis, University of Manchester.
42. Guisinger, S. and M. Irfan (1978), ‘Pakistan’s Informal Sector: The Case of
Rawalpindi’, Journal of Development Studies, 412–426.
43. Harris, J., ‘Character of an Urban Economy; Small-Scale Production and Labour
Markets in Coimbatore’, Economic and Political Weekly, 17, 23, 945–54, 24,
993–1002.
44. ILO (1974), Calcutta, see A. Bose (Reference 26).
45. ILO (1975), Djakarta.
46. ILO (1975), Abidjan.
47. ILO (1976), Sao Paulo.
48. Jurado, G.M. et al. (1981), ‘The Manilla Informal Sector: In Transition’, in
S. Sethuraman (Reference 69).
49. Kennedy, P. (1981), ‘The Role and Position of Petty Producers in a West African
City (Accra, Ghana)’, Journal of Modern African Studies, 565–94.
50. King, K.J. (1974), ‘Kenya’s Informal Machine-Makers: A Study of Small Scale
Industry in Kenya’s Emergent Artisan Society’, World Development, 2, 4, 9–28.
51. Laenen, A. (1988), Dinamica y Transformacion de la Pequena Industria en
Nicaragua, Dordrecht: Foris Publications, series: CEDLA Latin American Studies,
42.
52. Livingstone, J. (1991), ‘A Reassessment of Kenya’s Rural and Urban Informal
Sector’, World Development, 19, 6.
53. Mabogunje, A.L. and M.O. Filani (1981), ‘The Informal Sector in a Small City:
The Case of Kano’, in S. Sethuraman (Reference 69).
54. Marga Institute (1981), ‘Informal Sector without Migration: The Case of
Colombo’, in S. Sethuraman (Reference 69).
55. Merrick, T.W. (1976), ‘Employment and Earnings in the Informal Sector in
Brazil: The Case of Belo Horizonte’, Journal of Developing Areas, 10.
56. Middleton, A. (1981), ‘Petty Manufacturing, Capitalist Enterprises and the Proces
of Accumulation in Ecuador’, Development and Change, 12, 4, 505–24.
57. Moir, H. (1981), ‘Occupational Mobility and the Informal Sector in Jakarta’, in
58. Mosselman, L. (1992), Kleine Ondernemers in Kerala (India), Rotterdam: Erasmus
Universiteit Rotterdam.
59. Naponen, H. (1987), ‘Organizing Women Petty Traders and Home-Based
Producers: A Case Study, India’, in Singh A. et al. (eds), Invisible Hands, Women
and Home-Based Production, New Delhi: Sage Publications.
60. Nihan, G. and R. Jourdan (1978), ‘The Modern Informal Sector in Nouakchott’,
International Labour Review, 117, 6, 709–719.
61. Nihan, G., et al. (1979), ‘The Modern Informal Sector in Lomé’, International
Labour Review, 118, 5, 631–44.
62. Pansters, W.G. (1985), De Dialektiek van Mikro en Makro, een Onderzoek naar het
Funktioneren van Kleinschalige Bedrijvigheid in Ciudad Juarez, Mexico,
Rijksuniversiteit Utrecht: Geografisch Instituut, Diskussiestukken van de vak-
groep Sociale Geografie van Ontwikkelingslanden, nr. 30.
63. Papola, T.S. (1978), The Informal Sector in an Urban Economy: A Study in
Ahmedabad, New York: PhD Thesis, New York University.
64. Papola, T. (1981), Urban Informal Sector in a Developing Economy, New Delhi:
Vikas.
65. Pfeiffer, P. (1987), Politische, Sozio ökonomische und Urbanistische Aspecte der
Favelas und ihre Sociale Organisation in Rio de Janeiro, Stuttgard: PhD Thesis,
University of Stuttgard.
66. Regt, M. de (1990), Tapijtknoopsters in Rabat-Salé (Marokko), Nijmegen: PhD
Thesis, University of Nijmegen.
67. Sanchez, C.B. et al. (1981), The Informal and Quasi-Formal Sectors in Cordoba in
68. Sethuraman, S.V. (1977), ‘The Urban Informal Sector in Africa’, International
Labour Review, 116, 3, 343–352.
69. Sethuraman, S. (ed.) (1981), The Urban Informal Sector in Developing Countries,
Employment, Poverty and Environment, Geneva: ILO.
70. Simon, D. (1984), ‘Urban Poverty, Informal Sector Activity and Intersectoral
Linkages: Evidence from Windhoek (Namibia)’, Development and Change, 15, 4.
71. Smith, S. (1988), Een Studie van de Ontwikkelingsmogelijkheden van de Informele
Kleding- en textielnijverheid in Aguascalientes (Mexico), Nijmegen: Katholieke
Universiteit Nijmegen.
72. Sit, V.F.S. et al. (1980), ‘Ambulatory Labour in Hong Kong’, International Labour
Review, 119, 4, 505–514.
73. Suarez, R. (1988), Informal Sector, Labour Markets and Returns to Education in Peru,
Washington D.C.: World Bank.
74. Tom, J. (1987), De Inschakeling van Vrouwenarbeid in de Zijde-Industrie in Zuid-
India: de Dynamiek van de Informele Sector Belicht, Amsterdam: Vrije Universiteit:
Geografische en Planologische Notities, no. 51.
75. Tosta Berlinck, M. et al. (1981), ‘The Urban Informal Sector and Industrial
Development in a Small City: The Case of Campinas’, in S. Sethuraman
(Reference 69).
15
The Impact of Reforms in Tanzania:
The Case of Privatized
Manufacturing Industries
Humphrey P.B. Moshi*
1 Introduction
The main objectives of public-sector reforms, including the restructuring of

public expenditures and reforming the framework for public enterprise
management, ownership and operation, are: to increase the efficiency of
the productive sector and raise economic growth; to reduce government
budgetary costs and increase revenue, thereby improving the fiscal balance;
to broaden direct private ownership and control of productive assets; and
to reduce and reorient the role of government to that of providing social
and economic infrastructure and implementation of economic policy,
rather than to being involved in direct production.
This article focuses on privatised manufacturing industries with a view to
analyse the trends in capacity utilization, productivity and technological
acquisition with reference to a selected sample of industries.
2 Conceptual framework
Privatisation is a term used to describe a wide range of policies designed to

effect a transfer of activity from the public sector to private sector.
However, privatisation does not necessarily require total divestiture.
Franchising, contracting out and leasing, whereby public-sector activities
are undertaken by the private sector, are widespread forms of privatization
(Hemming and Miranda, 1991). For the purpose of this study privatization
will embrace cases of partial and total divestiture. Partial divestiture refers
to the creation of joint ventures between public and private sector organi-
zations, with the private-sector organization holding majority shares and
assuming management responsibilities.
The principal rationale for privatization is the assumption that the
private sector will be able to increase the efficiency of production.
341
Therefore, the sale of state enterprises is designed to prevent the private

sector from being squeezed out by the public sector. It can also reduce the
channels and therefore possibilities for public-sector corruption (Biersteker,
1991). In the above context, it can be argued that reform of the state enter-
prise sector and privatization are mutually supporting strategies aimed at
achieving wider objectives of creating a more efficient, market-oriented and
more productive economy.
Broadly speaking, two sets of conclusions emerge from the literature on
privatization. The first finds private enterprises clearly superior.
• There is robust evidence that state enterprises and mixed enterprises are
less profitable and less efficient than private corporations (Boardman
and Vining 1989, p. 17).
• There is virtually universal consensus that privatisation improves
efficiency (Boyco, et al., 1993, p. 1).
• Without exception, the empirical findings indicate that the same level
of output could be provided at substantially lower costs if output were
produced by the private rather the public sector (Bennett and Johnson,
1979, p. 59).
A second body of literature draws rather different conclusions. The earliest

(to our knowledge) and most influential formulation of this position is that
of Caves and Christensen (1980, p. 974): ‘Contrary to what is predicted in
the property rights literature, we find no evidence of inferior efficiency per-
formance by government owned railroads … public ownership is not inher-
ently less efficient than private ownership … the often-noted inefficiency
of government enterprises stems from their isolation from effective compe-
tition rather then their public ownership per se.’
The first set of theories links privatization to increased productive and
allocative efficiency. Privatization is expected to reduce the productive
inefficiency arising from public ownership. However, representatives of the
second perspective caution that improvements in allocative efficiency will
generally only be achieved when privatization goes hand in hand with
increased exposure of firms to competition (Hemming and Miranda, 1991).
3 Reforms and industrial performance
3.1 The reforms

The drive to address the socioeconomic crisis began in 1981 and focused on
four main programmes (see also Chapter 1): the national economic survival
programme (NESP, 1981/82); the structural adjustment programme
(SAP, 1982/83–1984/85); the Economic Recovery Programme (ERP,
1986/87– 1988/89); and the economic and social action programme
Impact of Reforms in Tanzania 343
(ESAP, 1989/90– 1991/92). The overall performance of NESP and SAP had been
very poor because these programmes failed to improve the domestic produc-
tion capacity of the economy. The critical weakness of the pre-1986 reforms
was their failure to address fundamental constraints inherent in a highly con-
trolled and state dominated socioeconomic system. In the subsequent reform
efforts, however, these impediments did receive adequate attention.
In an attempt to improve the performance of the industrial sector two
policies were adopted. The first policy shift was in favour of financing
improvements in utilization of existing capacities rather than in creation of
new capacities. This shift was informed by the realization that the speed at
which capacity was being created had outstripped the ability to finance
recurrent inputs, especially imported inputs (Wangwe, 1992b). Therefore,
resources were directed towards rehabilitation of productive sectors as well
as of the infrastructure. To that effect donor initiatives took the form of
balance of payment support, import support and commodity import
support. The overall impact of this policy shift is reflected in fairly modest
real growth rates of investment.
The second policy shift, which was adopted in the early 1990s, was that
of privatisation of public enterprises as a way of turning around the perfor-
mance of the economy in general. At this juncture it was realized that there
was a need to redefine the roles of the state and the private sector in the
national economy in order to achieve an appropriate balance, supportive of
growth and efficiency.
The response of the private sector to the new policy environment has
been quite positive as far as investment is concerned. The share of private
investment in total investment increased from about 50 per cent in the
mid-1970s to 60 per cent in 1986, and further to 70 per cent in the early
1990s. This partly reflects the slow-down of investment in the parastatal
sector and partly the private-sector response to economic liberalization and
an improved government attitude towards private enterprise.
An analysis of the investment behaviour of both the private and public
sector for the 1967–95 period suggests that public investment reduced its
contribution to GDP to 8.1 per cent during 1981–85 and rose slightly to
10.3 per cent in the reform period. In the case of private investment, a dif-
ferent pattern is observed. Its share of GDP depicts a rising trend, rising
from a low 11.3 per cent in 1967–80 to 17.6 per cent during the reform
period (Likwelile, 1997).
The public-sector trend observed with respect to GDP is also reflected in
the case of gross fixed capital formation. Having contributed significantly
to GFCF during the pre-reform period (54.1 per cent), the public sector
share fell to 32.8 per cent (1981–85) and then slightly rose again to 37.6 per
cent (1986–95), the average for the entire period being 41.5 per cent. The
contribution of the private sector to GFCF rose from 45.9 per cent
(1967–80) to 62.4 per cent during the reform period, with an average of
53 per cent for the whole period. The decrease in public investment should
be celebrated with some caution, however. Whether or not public invest-
ments crowd out private investment depends very much on the nature of
those public investments. For instance, investment in infrastructure tends
to complement private investment (Moshi and Kilindo, 1995). This points
to the need for public investment to concentrate on those projects which
demonstrate positive relationships between private and public investment.
On the privatization front, evidence shows that the privatization exercise
has not been very successful in terms of speed and numbers of enterprises
involved. So far only about 68 per cent (270 out of 395) of the firms target-
ted for divestiture have actually been divested. Further, the privatization
exercise has not been able to attract the envisioned massive amounts of
foreign direct investment (FDI).
The performance of the industrial sector shows some improvement.
Overall manufacturing output expanded as access to foreign exchange
improved, and trade and exchange-rate reforms reduced the prevailing
anti-export bias. Capacity utilisation has increased to an average of 50 per
cent. Nevertheless, the share of manufacturing GFCF has almost remained
the same, at 22.2 per cent of GDP for the last five years. Further, in the
recent past (1991–95) the growth of industrial output declined to around
two per cent (Wangwe, Musonda and Kweka, 1997).
3.2 Privatised firms: some insights

In this section we analyse the performance of two firms (one consumer
goods and one capital goods producer) which were hitherto under govern-
ment ownership. The analysis should help us assess whether or not priva-
tised firms are ‘bearers’ of hope for industrial revival and competitiveness.
3.2.1 Tanzania Breweries Limited (TBL)

Background. Prior to its privatisation in 1993 TBL was 100 per cent owned
by the Tanzanian government. At present TBL is a joint venture of the
government (40.4%), INDOLL International B.V. (45.9%), International
Finance Co. (9.8%), PROPARCO (1.8%), Dresdener Bank Luxembourg S.A.
(1.6%) and the Dutch Financing Corporation for Development (FMO)
(0.8%). Recently, the government has shown interest in selling 20 per cent
of its shares to the public.
TBL’s main line of business is that of the production, distribution and
sale of malt beer in Tanzania. It operates breweries in Dar es Salaam, Arusha
and Mwanza. The management of the firm is under contracted manage-
ment with INDOLL, a South African company
Performance. The performance of the firm prior to privatization in 1993

shows negative growth. In 1992 output was six percentage points below
the 4.2 million crates produced in 1988. The post-privatisation period

depicts an average annual growth, in terms of production, of 34.1 per cent.
Production increased threefold from 4.2 million cases in 1993 to
13.6 million crates in 1997.1
During the period of total government ownership, capacity utilization at
TBL was very low. It averaged 20 per cent for the Dar es Salaam and Arusha
plants and 0 per cent in the case of the Mwanza plant. After privatisation
the utilization rate rose to 65 per cent (over 90 per cent in 1997) for the
former plant and 85 per cent for the latter two plants.
Productivity and efficiency levels have improved when compared to the
situation before divestiture. In this context the number of productive man-
hours increased from 30 per cent (1993) to 40 per cent (1994) and further
to 42.2 per cent (1995). Likewise, the productivity of labour increased from
177 hectolitres per employee (1993) to 592 hectolitres per employee (1995).
The target is to produce 1000 hectolitres per employee. Further, efficiency
in terms of utilisation of raw materials and consumption of water and elec-
tricity has been enhanced, and beer losses during brewing, packing and dis-
tribution have been significantly minimized.
The financial performance of TBL has been positive. In 1993 the firm
earned a profit of US$15 million, and all subsequent years have witnessed
increased profits. Thus, the firm contributed positively to the government
budget both through taxes and dividends. The financial soundness of the
firm has made it possible for the management to offer an attractive
financial package to its workers as an incentive for further enhancement of
productivity. Thus, a virtuous circle of higher profitability, enhanced pro-
ductivity and higher wages emerges.
3.2.2 ABB TANELEC

Background. The Tanzanian Electrical Goods Manufacturing Company
(TANELEC) started operations in 1981. It produces transformers, switch
gears, cookers and light fittings. It was a joint venture owned by the
National Development Corporation (60%), Tanzania Electricity Supply
Company (TANESCO, 20%) and the National Industry of Norway (20%).
The firm was divested in 1995 and the new ownership structure is as
follows: ABB sub-Saharan Africa (50%), ABB Norway (20%), TANESCO
(20%) and NDC (10%). The management contract is with ABB Norway.
Performance. The production performance of transformers, switch gears,

cookers and light fittings for the 1983–91 period was quite erratic, but gen-
erally showed a declining trend. For example, the production of transform-
ers declined from 217 units to 201 between 1983 and 1984. It later rose to
456 units in 1985 and further to 638 units in 1990. The production of
switch gears increased from 86 units in 1983 to 114 in 1985, dropping
drastically to only 7 units in 1990. The production of cookers declined

from 9393 units in 1985 to 6646 units in 1990.
There are no data on trends of production per product line after privati-
zation. However, management claims that capacity utilization has
increased from below 50 per cent (pre-divestiture period) to 100 and 80 per
cent in the case of transformers and switch gear, respectively. The increase
in capacity utilization should also mirror improved efficiency in terms of
productivity as well as higher profitability rates.
On the export front the firm has performed quite well. The firm exports
over 70 per cent of its products compared to the pre-divestiture period
when it used to export about 30 per cent of its output. The enhanced com-
petitiveness of the firm underscores the positive effects of the privatization
exercise.
4 Factors accounting for the turnaround
The two case studies of successful privatisation efforts discussed above

should not be seen as isolated cases. Evidence (Due and Temu, 1996;
Moshi, 1996) shows that the majority of the privatised industrial firms
have turned around. These include the Tanzania Cigarette Company
(TCC), Bora Shoe Company, Morogoro Shoe Company, and Tanzania
Portland Cement. There are also examples where turnaround was initially
not successful, but is expected to take place over time. These include
African Tanneries Ltd (AT), Mwanza. Indeed, one needs to appreciate the
fact that the pace and the magnitude of the turnaround of the divested
enterprise will vary depending on the nature of the sector or subsector in
which an enterprise operates, and on the initial condition of an enterprise
at the time of divestiture.
A number of factors have accounted for the successful turnaround of the
divested enterprises. These are the following:
• Improved access to finance via foreign direct investment for the compa-
nies which were sold to multinationals. This accounts for about a dozen
cases. The establishment of private-sector financial institutions in
Tanzania has further improved the accessibility of finance.
• Acquisition of modern technology, which improves both labour and
capital productivity and the quality of product(s). Massive investment
by the new owners of breweries (US$ 48.5 million) and of TANELEC
(US$ 2.5 million) facilitated the acquisition of modern machinery and
equipment.
• Competent and accountable management that is cost conscious and
flexible, and adheres to a ‘management by objectives’ approach. The
post-privatisation retrenchments are indicative of management’s quest
for profits through cost-reduction initiatives.
• Improved workers’ welfare. This includes better pay packages, which boost
the morale of the labour force and consequently improve productivity.
• Targeted training of employees as a way of enhancing their capabilities by
exposing them to modern organisational and production technologies.
• An improved policy environment that is more friendly to the private
sector, coupled with wide-ranging complementary reforms. That is, a
market-oriented environment.
• Near monopoly positions of the firms studied. For example, TANELEC
controls 90 per cent of the domestic market as far as transformers are
concerned. In the case of switch gears it has a total (100%) monopoly.
Likewise, TBL is a giant in the beer industry, having a market share of
about 85 per cent in 1998. Though there are small competing breweries,
a monopoly exists at the level of brands. They belong exclusively to a
particular brewery. In the case of TBL the brands are Safari, Ndovu,
Kilimanjaro and Bingwa. The Associated Breweries has a monopoly in
the production of Serengeti and kick brands.
In spite of the performance-enhancing factors listed above, there are also

constraints which have either increased the transaction costs of doing busi-
ness or have hindered the privatised firms from realising performance
targets. These constraints include the following:
• Inadequate and unreliable supply of infrastructural services especially

transport, telecommunication, water and energy.
• Lack of credible public supporting institutions, arising from unclear
definition of the roles of the Tanzanian Investment Centre (TIC), the
Ministry of Trade and Industries (MIT) and the Treasury.
• Competition from cheap but poor-quality imports, which are either not
taxed at all or lowly taxed.
• High interest rates, which tend to discourage borrowing and thus hinder
capacity expansion or the acquisition of modern technology.
• Lack of an effective legal and institutional framework for prompt recov-
ery of debts, especially from government ministries and parastatals.
• A weak private sector and the concomitant narrow and shallow charac-
ter of interfirm business linkages.
• A civil service that is inefficient, poorly paid, unmotivated and corrup-
tive, failing to deliver services adequately and efficiently. The mindset of
the civil servants is far from being in tune with an economy which is
private-sector led.
It is evident from the two cited case studies that the privatisation of public
enterprises in the context of ongoing macroeconomic policy and
institutional reforms has impacted positively on the manufacturing sector.

This has resulted in higher output, better utilization of capacity, higher
productivity, improved product quality and increased competitiveness.
Thus, it needs to be underscored that privatization has made it possible for
enterprise adjustment to take place. This has taken the form of changes in
the product mix, switching to new markets, restructuring of costs, estab-
lishing more efficient operations, shedding of labour, and internal reorga-
nization of enterprises and their management. Nevertheless, the analysis
shows that the environment in which the divested enterprises are operat-
ing is still far from enabling. Consequently, the full realization of potential
benefits of privatization has been inhibited. It needs to be emphasized that
if an enabling environment were in place, the positive performance aggre-
gates would have been much more pronounced. To this effect the follow-
ing policy recommendations can be formulated.
First, there is a need to continue to reform the socioeconomic and insti-
tutional environment, with an emphasis on the selection of appropriate
policies, the proper sequencing of policy reforms, and the effective imple-
mentation of such policies. In the second place, the capabilities for manag-
ing the privatization process have to be enhanced, and the institutional
arrangements such as the Parastatal Sector Reform Commission (PSRC), the
Tanzania Investment Centre (TIC), the Ministry of Industry and Trade, the
Treasury and the Planning Commission have to be well targeted in terms of
clarity of mission, and effectively coordinated. These institutions should
strive to see that the divested enterprises are a success. Third, privatization
is most effective when it is part of a broader programme of economic
reform and private-sector development. Therefore, there is an urgent need
to develop and strengthen private-sector entrepreneurs through an ad-
equate provision of the necessary business services and infrastructure. Such
provision will in turn elicit private-sector participation in the economy.
Finally, the ongoing reforms in the civil service sector should also aim at
creating a market-oriented bureaucracy that is not only competent in the
articulation and implementation of private-sector-friendly policies, but also
capable of identifying and redressing the economic, legal and institutional
constraints that inhibit private sector development.
Notes
* Economic Research Bureau, University of Dar es Salaam.
1. Source: TBL data files.
16
Economic Reforms, Industrialization
and Technological Capabilities in
Tanzanian Industry
Samuel M. Wangwe*
1 Introduction
The post-independence period was perceived by African countries as an

opportunity to develop their economies. Impressed by the development
experiences of advanced economies, where the patterns of development had
demonstrated the increasing importance of industry and manufacturing, in
particular as income per capita rises, industrialization came to be perceived
as an integral part of the development agenda. The desire to replicate devel-
opment patterns in the advanced countries was supported by dual economy
models portraying the desire to replace traditional sectors of the economy
by modern sectors, which were supposed to possess characteristics, institu-
tions and values similar to those found in developed economies.
Industrialization was expected to facilitate the structural transformation
from predominantly agricultural economies to modern industrial
economies, and to enhance realization of high labour productivity and its
continued growth. In addition, industrialization was perceived to have the
ability to realise technological development. In the external sector, it was
envisaged that industrialization would bring about a restructuring of the
predominantly primary export sector to a more diversified export sector in
which exports of industrial products would increasingly play an important
role. Industrial exports were associated with dynamic products in which
specialisation and technological learning were attained.
Tanzania has been no exception to the general desire to industrialize.
Industrial development was expected to lead the process of modernization
and transformation to high productivity and high growth. All development
plans since independence, therefore, envisaged a substantial increase of the
share of industry in GDP.
Policy choices in the post-independence period were influenced by the
desire to replicate development patterns in the advanced countries. At the
349
same time, it was perceived that circumstances in which development was

to take place had changed. Factors such as the rejection of colonial policy
tools, the legacy of the Great Depression, the legacy of the Russian
Revolution and economic thought favouring planning, government owner-
ship, intervention and control featured quite conspicuously at that time.
The Great Depression had exhibited pervasive market failures, a situation
which led to considerable mistrust of markets. That experience undermined
the core of traditional economic theory based on efficient allocation of
resources and comparative advantages (Krueger, 1992). When most sub-
Saharan African countries were getting their independence in the 1950s
and 1960s, conventional wisdom then guided them to mistrust markets
and adopt the conventional idea that failures of markets contributed to
their underdevelopment. They also believed that their economic structures
were biased towards primary products. This reflected weakness in domestic
activities and an inability to compete in international markets, the latter
reinforced by elasticity pessimism (Krueger, 1992). These considerations
influenced the choice of development policies and industrialization strate-
gies, and the associated forms of state intervention in development.
These experiences and ideas led sub-Saharan African countries to adopt
planning using input–output based models and other planning models to
address the problems of a dual economy, focusing on the elasticity pes-
simism and the need to industrialize through state-led import substitution.
The policies which were adopted largely discounted the value of traditional
commodities, gave a large responsibility to government and adopted
import substitution industrialization with protection as a response to weak
domestic activities, lop-sided economic structures and an inability to
compete. The logic behind ISI is that developing countries such as
Tanzania needed to alter the structure of their economies towards
industrialization.
ISI was adopted during the 1960s at a time when development thinking
was sceptical about the functioning of markets, when capital accumulation
occupied a central position consistent with the Harrod-Domar model and
development was largely perceived as replicating development patterns of
the advanced countries. The key role of capital accumulation meant that
foreign aid was a logical option to facilitate development in countries with
low saving rates.
To industrialize in an environment where more industrialized economies
already existed would require protection from imports from these
economies. Hence, the implementation of ISI came to be buttressed by pro-
tection of all kinds. Approaches to implementation of ISI gravitated around
planning, protection (using tariffs, quantitative restrictions and exchange
rates) and efforts to increase savings and investments. Further policy distor-
tions occurred in capital markets (with subsidised interest rates), labour
markets (with fixed minimum wages) and goods markets (with widespread
Economic Reforms and Technological Capabilities 351
price controls). The share of agriculture in GDP was expected to decline

with the level of development. In practice, surplus was squeezed from agri-
culture for investment in other sectors. This was done through price con-
trols, exchange rate overvaluation and taxation.
The process of import substitution industrialization was given further
impetus in the 1970s by enhancing the role of the state, in what came to be
known as the basic industry strategy (BIS). The BIS approach put greater
emphasis on structural change and self-reliance. The basic industry strategy
was formulated in the early 1970s and was supposed to guide the industri-
alization process in Tanzania for two decades (1975–95). This strategy
emphasises industrial activities which transform the economy so that
domestic demand would be met primarily by domestic resources (Wangwe,
1998). The basic industry strategy therefore emphasized two sets of indus-
trial activities. The first set consists of industries which meet the basic
needs of the people. Important components of the selected industrial activ-
ities in this set include food processing, textiles, clothing, footwear, build-
ing materials, and materials and facilities to meet the requirements of
education, health services, transportation and water supply. The second set
of industries consists of activities which can use domestic resources to
produce and supply intermediate inputs and capital goods to industries in
the first set. In doing so, one can identify a core group of industries whose
products are consumed by most other industries as suggested by Thomas
(1974). This core group, which constitutes the base of industrial produc-
tion, consists of industries such as iron and steel, metalworking and engi-
neering, industrial chemicals, paper, textiles, leather, construction
materials and electricity. In this context, for example, the metalworking
and engineering industry receives considerable emphasis because of its
capacity to supply machinery and equipment, while iron and steel is a pri-
ority industry in its capacity as a supplier of inputs into metalworking and
engineering industries. The logic of the BIS is fundamentally different from
that of ISI. In practice, however, its implementation did not show any
significant departure from the practice of ISI (Barker, et al., 1986; Skarstein
and Wangwe, 1986; Wangwe, 1990, 1993).
2 Performance of industry
At independence in 1961, there were fewer than 400 manufacturing

establishments (with 10 or more workers). Consumer goods industries
accounted for 70 per cent of total manufacturing sector employment,
while capital and intermediate goods accounted for the rest. The GDP
share was 4 per cent, meeting less than 15 per cent of the national
demand (Ministry of Industries and Trade, 1996). It is important to
address industrial performance in the perspective of an activity that
started from a very low initial base.
The number of manufacturing enterprises employing more than 10 workers

increased from 101 in 1946 to 380 in 1964, employing 30 000 people, and in
other sectors mechanization began to have effect in this period. Productivity
increased. Asian capital continued to expand quite rapidly in the post-
independence period, even though it received little government support
(Barker et al., 1986). A major stimulant of this growth was the Kampala agree-
ment of 1964, reinforced by measures towards protection. The year 1964
marks the beginning of a period of expansion, with manufacturing value
added growing over 10 per cent per year between 1964 and 1972, and the
share of industry growing from 7.1 per cent to 11 per cent of GDP. In terms of
structure, 74 per cent of manufacturing value added comprised consumer
goods, mainly food and beverages, whereas intermediate goods and simple
capital goods accounted for 23 per cent and 3 per cent respectively.
Processing of agricultural products consisted mainly of sisal decorticating and
cotton ginning industries (Ministry of Industry and Trade, 1996).
Despite the adoption of the basic industrial strategy in the early seven-
ties, which aimed at raising the share of producer goods, the sector contin-
ues to be dominated by simple consumer goods in the food, beverages,
tobacco, textile and leather subsectors (Wangwe and Semboja, 1997). The
manufacturing sector remains small (about 8 per cent of gross domestic
product), and has remained largely untransformed and undynamic, still
dominated by primary processing and simple consumer goods industries
and some simple intermediate goods. In the reconstructed series of data,
included in this volume, the degree of structural change between 1966 and
1989 is found to be even less pronounced than the unadjusted data would
suggest (Prins and Szirmai, 1998).
Towards the end of the 1970s, with variations across sectors, there were
indications that the high expectations from industrial development were
under threat. The main indications for this were low returns, inefficiency,
low capacity utilisation and low labour productivity. Infant industries were
not growing up as envisaged. As it turned out, the manufacturing sector in
Tanzania has performed poorly in terms of output, employment, structural
change and competitiveness.
The accuracy of statistics on manufacturing has been known to be very
fragile. Data on growth of manufacturing and its contribution to the GDP
differ from one source to another. The best work that has been done to
date in addressing these weaknesses and reconstructing the data is by Prins
and Szirmai (1998; see also Chapter 3 of this volume). They have suggested
adjustments to existing data taking into account new information on cov-
erage, non-response and conceptual issues, calculated new indices for six
major branches of industry, and reconstructed employment series that are
consistent over time with data on manufacturing value added. These recon-
structed series are a major improvement on the previous rather incomplete
and inconsistent data.
According to the series generated by Prins and Szirmai (1998), manufac-

turing output peaked in 1978, having risen about five-fold from the 1961
level. The industrial production index fell from 100 in 1976 to only 72 in
1985, picked up to 84 in 1990 and 1991, after which this recovery seems to
have petered out. This observation is consistent with the finding that
exports of manufactures recovered after 1986, but that by 1990 this recov-
ery had run out of steam. The overall picture of growth of manufacturing
shows that it continued fairly steadily until 1978 when it faced a decline
until 1985, after which the declining trend was reversed. This recovery has
not been sufficiently strong to date, because it has largely been based on
increased levels of utilisation of existing capacities, rather than investments
in new capacities based on improved technology and training, which are
essential for attaining international competitiveness (Bagachwa and
Mbelle, 1995; Wangwe, 1995). With the coming on-stream of new invest-
ments in the recently privatised industries, such as Tanzania Breweries, and
new investments, such as those in soft drinks, especially Coca-Cola and
those in the soap industry, there are indications that the growth of manu-
facturing may be picking up (Ministry of Industries and Trade, 1999).
According to the Minister’s speech to Parliament in June 1999, growth of
manufacturing increased by 5.7 per cent in 1997 and 8.1 per cent in 1998.
Productivity in manufacturing increased steadily up to 1973, after which
it declined in spite of continued capital investments. Thus labour produc-
tivity (in terms of manufacturing value added at 1989 prices per person
engaged) is shown to have declined from a peak of TSh 520 000 in 1973 to
a low level of TSh 250 000 in 1987 (Prins and Szirmai, 1998). By 1990
labour productivity in manufacturing was 21 per cent lower than in 1965.
The major contributor to the decline in labour productivity was the paras-
tatal sector that dominated the manufacturing sector. As many studies have
demonstrated, productivity in the parastatal sector faltered. For example,
Coulson (1979) examined several parastatal projects (for example, fertilizer
and bakery projects) and found their performance weak and riddled with
bureaucracy and inefficiency. Productivity in parastatals was found to be
low (Clark 1978; Shao, 1978; Barker et al. 1986). In particular, Clark (1978)
found that large parastatals had low profit rates, standing at only a quarter
of smaller companies in 1971. Shao (1978) found that output per employee
was higher in Asian-run enterprises (for example, textiles, grain milling and
sawmills) than in parastatals. Jedruszek (1979) found that productivity in
industry declined by 14 per cent between 1967 and 1977. Manufacturing
productivity was below average over the period in spite of large capital
investments. Jedruszek suggested that productivity was low because man-
agers had no incentive to control production costs and to promote
efficiency and productivity, while workers exercised their rights and not
their responsibilities. Overall productivity in the manufacturing sector was
very low to the extent of making it difficult for the sector to compete with
imports. Poor competitiveness is also reflected in labour productivity. In

1990 labour productivity in spinning measured in terms of number of
man-hours required to produce 100 lb or 45.4 kg of yarn averaged
66.2 hours in Tanzania. This is about two to three times that obtained in
Kenya, Turkey and India. Labour productivity in weaving measured in
man-hours required to produce 42 m 2 of cloth was 18.2 in Tanzania. This
again compares unfavourably with 4.0 in Kenya, 3.5 in India or 1.6 in
Turkey (Komba, 1999).
The comparative analysis of productivity in Tanzania and the USA in
1989 revealed that aggregate labour productivity in Tanzanian manufactur-
ing was as low as 3.7 per cent of that in US manufacturing (Szirmai and
Schulte, 1998, and Chapter 3). This is an indication of a large technology
gap and wide differences in the level of competitiveness between industrial
operations in Tanzania and the USA.
The above indicators of performance underscore the emerging problems
with ISI. These problems surfaced more conspicuously in the early 1980s in
terms of slow growth of employment, low productivity growth (or even
decline after 1973) and balance of payments problems as imports were
rising faster than exports. Transfer of technology and knowledge (adminis-
trative and marketing) proved to be more complex than had been
expected. Indigenous learning processes were not forthcoming in ISI.
Structural change was not leading to an automatic change in the capacity
to learn and to accumulate knowledge. These problems and disappoint-
ments with the performance of ISI influenced the decision to undertake
policy reforms in 1986, in response to the deterioration of the economy in
general and of industrial performance in particular.
3 Restructuring the industrial sector
3.1 Rationale of economic reforms

One implicit assumption of economic reforms and industrial restructuring is
that enterprise level inefficiencies are a reflection of distorted or inappropri-
ate macroeconomic policies. It is suggested that if appropriate adjustments
could be put in place at the macro-level, enterprises would receive the right
signals through the market. In response to these signals, enterprises would
restructure appropriately. According to this approach, appropriate changes
in policies (for example, with respect to market prices, realistic exchange
rates, interest rates, competition) are expected to induce restructuring by
favouring the expansion of efficient enterprises and contraction of
inefficient ones. This approach has been associated with the World Bank,
especially as reflected in its earlier publications (World Bank, 1981, 1989).
According to this approach, reform and restructuring of industry are essen-
tially a macroeconomic issue. Policy aims are restructuring the supply side
by putting in place appropriate macroeconomic and sectoral policies.
This approach has been adopted by the ‘adjusting countries’ in Africa,

and the outcome is still debatable (see also Chapter 13). During the 1970s,
10 countries in Africa experienced de-industrialization and 11 others in the
early 1980s (World Bank, 1989). According to the World Bank (1989),
adjustment programmes have both positive and negative effects on indus-
try, with the net effect varying widely by product and firm. The World
Bank’s assessment is that the overall effect has been positive (World Bank,
1994c).
3.2 Recovery of industry?

The policy reforms of 1986 were accompanied by a release of donor funds
that had been withheld pending agreement with the international financial
institutions (IFIs). In the case of Tanzania, at the level of the manufacturing
sector as a whole, the shift of resources towards capacity utilization coupled
with a greater inflow of foreign resources seems to have had a positive
influence on output. This influence occurred within the same industrial
structure under which the decline of industry had occurred in the first
place. The question yet to be answered is whether this recovery is sustain-
able. It would appear that it is more of a capacity utilisation boom, the sus-
tainability of which will largely depend on the capacity of the economy to
shift away from dependence on donor financing towards greater reliance
on export earnings. Even more important, sustainability of industrial recov-
ery will require that industrial restructuring be effected in the direction of
efficiency and competitiveness. This situation cannot be attained unless
macroeconomic reforms are accompanied by appropriate action at industry
level and at firm level (Wangwe, 1995, 1998).
3.3 Restructuring entails more than macroeconomic reform

Various industrial studies have revealed that restructuring the industrial
sector entails much more than macroeconomic management. For instance,
findings of various enterprise-level and sectoral studies in Tanzania in the
1980s lend support to this observation. Findings made by the Tanzania
Industrial Studies and Consulting Organisation (TISCO) of the techno-
economic review of eight Tanzania Investment Bank (TIB) client companies
in the mid-1980s revealed many operational problems other than those
calling for mechanical rehabilitation in the engineering sense. Specifically,
the review brought out problems of the exchange rate, government priori-
ties, project planning and programming, linkages, financing, organization
and rationalization, quality of management, incentive structures, and infra-
structural requirements. These broader concerns are also quite explicit in
the findings of six subsectoral studies on the industrial sector in Tanzania.
Studies of the industrial sectors in several African countries (UNIDO,
1988b, 1990) confirm that the problems of industry cover all levels (macro-
sectoral and plant level) and involve all aspects of performance, including
economic, financial, managerial, technical and marketing. In this context,

three challenges are recognized: to identify enterprises suited for rehabilita-
tion; to combine plant rehabilitation with a restructuring programme of
the industrial sector to ensure growth and dynamism (this would entail
mainly new investments); and to adjust the policy and administrative
framework. In his study of the performance of Tanzania’s textile sector,
Peter de Valk (1992) has demonstrated that performance of enterprises is
determined not only by macroeconomic factors but also by international
factors, sectoral policies and characteristics at firm level, and that all these
interact in a rather complex way (de Valk, 1992). These findings indicate
that macroeconomic policy may be a necessary component of industrial
restructuring but it is by no means sufficient. It has to be complemented by
actions at sectoral and enterprise level, with due consideration of develop-
ments at the international level.
3.4 Restructuring and technological capabilities

One major determinant of international competitiveness is investment in
technology. This implies that technological learning and acquisition of
technological capabilities cannot be ignored in the process of industrial
deepening and structural change. The industrialization strategy adopted
should provide an impetus to the demand side as an inducement mech-
anism for technological change. The particular type of industrialization
which has so far been implemented in Tanzania has not provided a strong
inducement mechanism for technological change on the demand side.
During the import liberalization phase there has even been a tendency to
establish more and more of the finishing types of industries. Industrial
deepening and technological change have not received any notable priority
(Wangwe, 1993).
A study of the early phase of industrial restructuring (1985–90) showed
that technological capabilities remained low. Four indicators of technolog-
ical learning were used in that study (Wangwe, 1992a): the degree of local
participation in the identification and implementation of the rehabilitation
programmes; the balance between output and learning objectives; the
extent of upgrading of technical and managerial skills through training;
and the implications for the domestic capability to manufacture spare parts
and components. On all four counts the rehabilitation programmes were
found to have paid little attention to the issue of raising the level of local
technological capabilities. These findings point to the possibility that reha-
bilitation under the current economic reforms may not address the prob-
lems which have inhibited technological learning in the manufacturing
sector (Wangwe, 1993).
These findings suggest that although changes in the macroeconomic
environment have introduced competitive pressures among manufacturers
(for example, injecting a sense of quality consideration), many of the earlier
weaknesses in technological development have not been resolved. Changes

in macroeconomic policy and enterprise-level rehabilitation programmes
alone are not sufficient for industrial restructuring to be realised. Action at
the intermediate level (below macro-level and above micro-level) is still
insufficient (Wangwe, 1993). Some sectors (for example textiles and gar-
ments) have even suffered considerable de-industrialization (Semboja and
Kweka, 1998). During the trade liberalisation phase there are signs that even
the modest degree of product complexity has been further eroded by imports
(Lall et. al., 1997; Semboja and Kweka, 1997; Wangwe and Semboja, 1997).
This is a reflection of the persistence of weaknesses in the institutional
mechanisms for the support of industry and in the incentive framework
relating to the building of technological capabilities. The importance of
action at this level is well demonstrated in the case of the textile sector
(de Valk, 1992). More recent studies of the industrial sector have confirmed
this point in the textile and clothing and engineering sectors (Bagachwa
and Mbelle, 1995; Lall et al., 1997; Semboja and Kweka, 1997). Such action
needs to be generalised to the whole industrial sector in the framework of
an industrial strategy.
4 Why a competitive industry has not been realized
The envisaged industrialization process has not been realised on account of

policy failures, inadequate institutional capacity, poor infrastructure,
insufficient commitment to make the necessary investments to cope with
technological change, and failure to invest in the required human
resources.
4.1 Policy failures

The state-led industrialization programme was implemented in a policy
environment which was riddled with macroeconomic imbalances. Inward-
looking import substitution policies shielded firms from competitive pres-
sures. The incentive structure was not conducive to developing a
competitive industrial structure. Investors were securing niches and pre-
venting rivals from entering the domestic or regional market, characterised
by monopolistic relationships. In addition, first-comer investors obtained
guarantees from the state that no other company would be licensed to
manufacture or import the same final product for several years. Profitability
was assured irrespective of the level of efficiency, because of price controls
whereby output prices were set on a cost-plus basis supported by high
tariffs. The exchange rate was overvalued, rendering imports cheaper than
under free market conditions and making exporting unattractive. Interest
rates were set at below inflation, and credit was rationed in rather unpre-
dictable ways. The actual allocations of credit have been shown to have
favoured large-scale parastatal firms (Bagachwa and Mbelle, 1995).
Resource allocation favoured expansion rather than efficient utilisation

of the capacities which had already been created (Wangwe, 1979). The
resulting imbalances exerted strains on the resources for implementing the
BIS. One aspect of the macroeconomic environment in the last two decades
has been that of pressure on the limited investment and foreign exchange
resources. These became major factors in determining the pace and pattern
of project implementation during this period. At enterprise level this pres-
sure was manifested in project implementation delays and cost overruns
(Wangwe, 1993).
The foreign finances available for industrialization were channelled into
industrial expansion in the form of ill-designed projects and choice of tech-
nology that required high levels of foreign exchange to sustain it. The crisis
was deemed to be temporary. Longer-term decisions in investment did not
take the impinging balance of payments crisis and future implications of
the burden of debt servicing on the balance of payments into considera-
tion. The phenomena of pressure on investment resources and a large pro-
portion of foreign finance in industrial financing are not necessarily bound
to cause problems. However, the particular form that foreign finance took
in Tanzania had far-reaching implications for the pace and pattern of
project implementation.
4.2 Inadequate institutional capacity

The institutional capacity to programme, monitor and provide supportive
services to industry has been a constraining factor in industrial develop-
ment. In particular, the institutions which had been created to provide
technological support services to industry (for example, the Tanzania
Industrial Research and Development Organization, Tanzania Bureau of
Standards) were not allocated the resources needed to provide the requisite
support to industrial enterprises. The allocation of resources to supportive
services became even more difficult with the budget cuts during the period
of structural adjustment. The restructuring of the providers of industrial
support services has not been completed as yet. This delay has continued to
afflict the process of industrial recovery.
4.3 Weak linkages between industry and other sectors and support
institutions
Linkages between industry and related sectors were not articulated, and
complementary investments were not made to ensure integrated develop-
ment of major sectors. For example, investments in mining, agriculture
and infrastructure were not articulated in line with the BIS (Wangwe,
1993).
Although technology institutes were established in the late 1970s and
early 1980s, they did not function in ways which served the needs of
industry. For example, in their study of TIRDO, Bongenaar and Szirmai
(1999) found that out of twelve projects, ten were initiated by TIRDO
officials themselves or by one of its governing bodies, rather than by enter-
prises, which are the main clients. Entrepreneurs were not involved in the
project selection process, and insufficient attention was paid to the transfer
of technologies to the users. For these reasons TIRDO could not pay ad-
equate attention to the appropriateness of technologies to the Tanzanian
environment. In general industrial linkages between TIRDO and domestic
industry were found to be weak (Bongenaar and Szirmai, 1999, and
Chapter 7 of this volume).
4.4 Inadequate infrastructure

The level and cost of infrastructure become obstacles to efficient industrial
development. The roadmap study comparing Tanzania with Uganda,
Namibia and Ghana showed that Tanzania was in a disadvantaged position
relative to other countries (TSG, 1996). Subsequent studies have shown
some improvement, and prospects for further improvement are bright, but
the situation is still quite bad (HIID, 1998). In particular the inadequate
availability and high cost of energy, transport and communication and
institutional infrastructural support services (technical, marketing, infor-
mations flows, finance) have inhibited the development of an efficient
industrial structure. The poor state of infrastructure has contributed to the
high cost of doing business.
4.5 Low level of human resource development

The positive role of human resources in development of a competitive
industry has been recognized. However, human development is one area
where Tanzania has lagged behind many countries. The level of tertiary
education in Tanzania is one of the lowest in Africa. The wrong mix of dif-
ferent levels of technical skills aggravates the low quantity of technical edu-
cation that is offered. While the normal ratio of engineers to technicians to
craft workers is 1:5:25 in advanced countries, it was found to be 1:5:13 in a
sample cross-section of modern enterprises in Tanzania. The same ratio was
1:2:5 in the actual output of formal education in Tanzania in the late 1970s
(Barker et al., 1986). There are no indications that this ratio had changed
significantly by the late 1990s. There is no coordinated effort to link train-
ing to imported machinery. In general, the level of investment in human
resources for industrialization has been unable to cope with the challenges
of technological development of a competitive industry. The level of ter-
tiary education in general, and technological and engineering personnel for
industrial development, has remained low.
The quality and relevance of technical education that is offered in local
vocational training centres and technical schools and colleges is not ade-
quate (Chapter 9). Industry-specific training institutes were established, but
even these were geared to basic and low-level skills oriented towards
operation of existing plant, rather than to adaptations, innovations and

improvements of imported technology.
With regard to training, it seems that Tanzanian firms do not attach due
importance to training, especially formal and long-term training for the
low cadre employees. A study by ESRF (1997) in the metal subsector indi-
cated the need for upgrading of production skills for all levels of employ-
ment if firms are to install new production technology. This is despite the
fact that about 90 per cent of the industries in the sample of 74 firms
conduct informal in-house training (less than 20 per cent of the firms were
found to conduct long-term and formal training) (Wangwe et al., 1997b).
Other studies, such as that of Masuha (ESRF, 1997), indicated that in
many metal engineering companies little thought is given to designing
training programmes for employees in activities such as quality control and
maintenance. Additionally, according to the competitiveness strategies of
nine industries studied by Bagachwa and Mbelle (1995), only one company
(Matsushita Electric Co., which is a subsidiary of Matsushita Electric, Japan)
adopted all forms of training at all levels as one of their very important
competitive strategies, which is typical of Japanese companies
Despite the diversity of final products generated in manufacturing, it
has been found that most of the jobs undertaken require skills of a very
low order usually associated with attending machines and doing repetitive
part work (Barker et al., 1986). However, Barker et al. (1986) identified a
few enterprises that demonstrated that development of worker skills is
only partly determined by the type of machinery and its technological
level. An important role is also played by the way the labour process itself
is organised. For example, enterprises that organised their workers in coor-
dinated teams of workers, engineers and managers in performing a set of
operations seem to have achieved higher levels of skills. Policies towards
management of the labour process are an important dimension of build-
ing up technological skills. The low level of skill development in
Tanzanian industry is partly a function of the technology employed and
partly a function of management of the labour process. In this respect it
has been shown that in Tanzanian industries most of the workers are
excluded from planning, designing and control tasks. They are largely
confined to activities relating to repetitive tasks, quality control, and
repair and maintenance work. Engineering functions related to design
work are performed in the countries from which equipment is imported.
(Barker et al., 1986).
The manner in which the labour process is organised is not particularly
conducive to learning and the development of human resources through
learning by doing. It has been found that technicians in Tanzania have a
tendency to spend most of their time in offices at a distance from the shop
floor, instead of liaising between engineering design and planning on the
one hand and implementation by shop-floor workers on the other. This
white-collar mentality is an influence from an education system that does

not provide sufficient practical training (Barker et al., 1986). The manager-
ial strata were separated from the shop-floor workers by huge gaps in edu-
cation, training, salaries, privileges and power.
During the time the parastatals were dominant, the learning process of
Tanzanian managerial and technical personnel was impaired by the pre-
ponderance of foreign personnel in key positions in top management. Top
management in parastatals largely consisted of expatriates who outnum-
bered local counterparts by a factor of two to three (Barker et al., 1986).
Most of the Tanzanian managers were trained in fields other than engineer-
ing or science and technology. Most of them were trained in finance and
administration. The industrial skilled, semi-skilled and unskilled workers
were mostly first-generation industrial employees. They required concerted
efforts to mould them into an industrial culture before they could con-
tribute to technological development activities. A sample of employment
in modern industrial enterprises in the late 1970s showed that employees
were largely unskilled (20.6 per cent) and semi-skilled (57.5 per cent)
workers, with only 10.5 per cent being skilled workers; 4.7 per cent were
supervisory personnel, 4.6 per cent administrative personnel, 1.1 per cent
middle-level management and 1 per cent top management. Most of the
unskilled and casual workers had no school education, so their tasks were
limited to unskilled physical work, such as moving, sorting, cleaning and
packing. No provision was made for their training to enable them to learn
skills. Most of the supervisory level and skilled and semi-skilled workers
had only primary education. This educational composition of the labour
force demonstrates that the level of education, and in particular secondary
and higher-level education, could have been a limiting factor in the
processes of technological learning.
Bureaucratisation of industrial management was institutionalised in the
parastatal sector, whereby management lacked initiative towards the design
of new products, processes and maintenance systems. In some cases there
were even instances of obstructing useful innovations, such as local manu-
facture of enzymes for the breweries. The bureaucratised management in
the parastatals shied away from addressing critical production problems.
Instead it either called in experts from abroad, or top management trav-
elled abroad supposedly to learn how similar problems were solved
elsewhere.
The environment in which industry was operating was not putting com-
petitive pressure on firms to train workers with a view to attaining compet-
itiveness. As Pack and Paxson (Chapter 2) have indicated, the simple
existence of trained and experienced managers and workers does not guar-
antee high firm productivity. They need to be complemented by competi-
tive pressures to induce firms to innovate and by inflows of new
technologies.
4.6 Low investment in technology and technological learning

Investments in industry favoured initial capital investments, but invest-
ments in technology were not made on a continuous basis with a view to
keeping pace with the challenges of building a competitive industrial
structure.
The parastatal holding corporations which were responsible for coordi-
nating several public enterprises in specific industrial branches, and for the
promotion of new projects in those branches, became increasingly engaged
in the mobilization of investment resources. In particular they gave high
priority to mobilization of foreign finance. This tendency has not been
conducive to promoting the utilization of local capabilities and technolog-
ical learning, and building of indigenous technological capabilities through
learning by doing. Instead, it has pre-empted the mobilization and utiliza-
tion of local resources and capabilities. One consequence of this is that the
question of choice of technology was accorded low priority. The main
public-enterprise development institution, NDC, usually left the choice of
technology for its projects entirely to the foreign partner (Barker et al.,
1986). The foreign partner had considerable flexibility in the way manage-
ment agreements were framed. They usually stated rather vaguely that the
foreign partner would select modern processes, taking into account the
objective conditions of Tanzania. Feasibility studies were preoccupied with
the range of products to be produced and their market and financial viabil-
ity, rather than with implications of technology development issues.
Several factors explain this laxity in choice of technology:
• Foreign partners were generally unwilling to participate in a project

unless they were given leeway in the choice of technology. This was
done under the pretext that they could otherwise not be held responsi-
ble for performance of the project. However, even when they were given
freedom to choose the technology, there was no mechanism put in
place to hold them responsible for performance. Most of the time the
management fees were not related to any performance indicators.
• Tanzania did not have qualified and experienced engineers and other
professionals who could unpackage the technology and negotiate terms
and conditions under which different parts could be supplied.
• The Tanzanian state had little access to global information on technol-
ogy sources and costs, and did not put in place policies and mechanisms
to build the capacity to acquire the necessary information. The incentive
structure encouraged the top bureaucracy in the parastatals to give high
priority to setting up as many enterprises as possible in the shortest
time. The number of projects implemented was equated with a high rate
of industrialization. The high propensity to speed up implementation of
projects has tended to overshadow more important concerns with regard
to technological learning. Concerns with speed of project implementation
have been overriding, and the engagement of foreign finance packaged

with foreign personnel has been deemed more appropriate for the
attainment of this objective. James (Chapter 5) has invoked the public-
choice approach to bureaucracy to make two main arguments related to
this observation. First, bureaucrats had preferences defined over projects
rather than technologies. Second, managers of parastatals sought to ini-
tiate as many projects as possible, mainly on the basis of foreign aid.
Nationalization and the use of foreign partners did not lead to successful
restructuring and improved technological performance of the firms
inherited from local private entrepreneurs.
• Nationalization alienated the former owners and other actors who pos-
sessed detailed knowledge of the history and developments in the facto-
ries. The new group, which took over, inherited technologies they did
not choose and lacked the experience and background information
about the plants. A situation of discontinuity was forced onto the opera-
tions and technological learning. In this connection, it is observed that
the experiences of parastatals can be contrasted with those of the smaller
private-sector enterprises. For example, in the case of firms owned by
domestic and East African private capital, the owners kept a close watch
on the choice of technology by participating directly or engaging experi-
enced expatriates for technical help (Barker et al., 1986).
5 Towards developing a competitive industry
Industrial development in the 21st century needs to accommodate the

changing external and internal environments. External conditions have
changed towards openness and competitiveness. The industrial drive needs
emphasis on competitiveness. Theories of long-term development have
shifted from stressing the problems of capital accumulation and are now
singling out technological development and growth of productivity
(Pasinetti, 1981; Romer, 1986; Pasinetti and Solow, 1994; Ruttan, 1998). It
has become increasingly recognised that long-run growth is primarily
driven by the accumulation of knowledge. Internally, there are concerns
about creating a dynamic economy which not only grows fast but places
high human development high on the agenda. To the extent that different
ways of generating growth of real income may have different effects on the
possibility of conversion of real income growth into achievements that are
highly valued for human development, there is a case for worrying about
other aspects of human development. For example, the nature of employ-
ment orientation in GDP growth and the role of education and learning
have a clear bearing on the promotion of good living conditions (Sen,
1994). The kind of industrial development that is desirable should be able
to generate broad-based employment and incomes directly or indirectly
through strong linkages with sectors which engage the bulk of low-income
earners. In the case of Tanzania, 92 per cent of the poor live in rural areas
where the majority of them depend on agriculture for a living. Industrial
development will need to have a positive influence on agricultural
incomes, if poverty is to be reduced considerably.
5.1 Developing a dynamic and competitive industrial structure

The following industries have achieved higher productivity and quality and
can therefore effectively compete both locally and in export markets: bev-
erages, such as beer, soft drinks and bottled water, which take more than
80 per cent of the local market, and cigarettes and food, which take 50 per
cent and 35 per cent respectively (URT-budget speech, 1999c). This is partly
due to reorganisation in capital plants and management and also due to
local availability of a significant part of production inputs. On the other
hand, industries which seem the least competitive are those in the textile
sector. The fact that 35 out of 37 establishments were closed right after lib-
eralisation justifies this contention. Local garment firms have been unable
to compete with the vast second-hand trade (mitumba).
In general, the available evidence suggests that in Tanzania manufac-
tured exports constitute a dismal share of total exports and gross manufac-
turing output (Komba, 1999). According to Wangwe et al. (1997b), the
share of exports in the total output of the sampled manufacturing firms in
Tanzania averaged no more than 5–15 per cent.
There is a need to put in place a dynamic industrialization programme
based on the technological revolution in industrial production and organi-
sation, to ensure faster growth of high technology manufacturing.
Industrial policies in the 21st century will have to be made in full consider-
ation of the emerging technological setting driven by globalization and
rapid technological progress (Wangwe, 1998). Industrial policy will be
expected to cope with the new paradigm of new technologies, new product
ranges, new management and organisational techniques, and linkages and
networks. In order to tap the opportunities arising from this process, it will
be necessary to break the constraints relating to institutional weaknesses,
inadequate infrastructure and low levels of human resource development.
5.2 Focus on technological learning and building competitiveness

The accent will need to shift towards technological learning and building
of capabilities for attaining competitiveness in the domestic market and in
exports to the region and to global markets. There is a need to acquire and
maintain competitiveness in local-resource-based industries, agro-industry,
industries producing agricultural inputs, mining and tourism, which are
capable of generating employment and incomes broadly in the economy,
while developing the capability to seize existing and new opportunities in
export markets. Exploiting initial comparative advantages will provide
investible surpluses and capacities to develop into activities with dynamic
comparative advantages. The strategy will promote regional industrial part-

nerships with a view to tapping economies of scale and attaining competi-
tiveness in regional and international markets.
5.3 Priority of agro-based small and medium-sized enterprises (SMEs)

The industrial strategy should focus on agro-based small and medium
industries capitalizing on the dynamic synergies between industrial and
agricultural development. This approach should effectively address poverty,
unemployment or underemployment and food insecurity. It should
provide broad benefits to society.
Factors that have proved to explain the dynamism of SMEs in recent
years need to be made the driving forces of this development. Three key
factors are identified as driving forces: the internal configuration of enter-
prises must take advantage of efficient organisational and production tech-
nology; the economic and social context in which the SMEs will function
must reflect a competitive environment; and linkages to other enterprises
must attract deliberate focus.
Support should be provided to rural industry as a means of generating
rural non-farm employment and income. Rural non-farm activities should
be able to generate further farm income through various linkages. The
potential of rural industries for employment and income generation and
export growth lies in the production of labour-intensive and local-resource-
using products, and a continuous upgrading and improvement of organisa-
tional and production technologies.
5.4 Enhance the role and capacity of government support

The experience of Tanzania in the reform period clearly demonstrates that
while opening up the economy to competition from imports may be a nec-
essary condition for developing a dynamic and competitive industrial
sector, it is in itself far from sufficient. There is need to complement the
process of opening up with government intervention through supportive
policies aimed at helping enterprises upgrade their technologies and com-
petitiveness, guided by a clear strategy. The pace of opening up should be
geared to the learning needs of different activities in the economy. The role
of government will need to be enhanced to support and promote the
restructuring of subsectors and the potential for growth and competitive-
ness, and to support human resource development and technology pro-
grammes. Government intervention will need to be directed towards
promoting the building of institutional infrastructure to meet requirements
for finance (especially development finance), training, development of
quality standards, generation of the necessary information, promotion of
investments and technology flows, and the development of applied
research in appropriate fields for industrial development. Government
intervention will be necessary in the following areas to:
• identify and facilitate the pursuit of dynamic comparative advantages

with a view to attaining an optimal industrial structure which is com-
petitive
• ensure that the balance of payments is sustainable
• Ensure long-term growth of domestic demand: a large domestic market,
assisted by some degree of protection, could facilitate generation of high
profits, which could provide conditions for high rates of investments,
learning by doing and improving product quality
• facilitate the acquisition of foreign technology through facilitating FDI,
licensing and or other means of channelling foreign technology to
promote competitiveness in economic activities in the country
• facilitate exporting.
6 Conclusion
The rapid changes in market and technological conditions pose a great

challenge for industrial transformation. The implementation strategy must
place top priority on building the capacity for industrial competitiveness.
The main components of this strategy are:
• industrial policy based on economic analyses of industrial competitive-

ness and identification of dynamic comparative advantages on a contin-
uous basis
• nurturing and promoting public–private sector consultation mecha-
nisms based on industrial subsector surveys and analyses, as well as
industrial development experiences in the country and in other coun-
tries
• human resource development through restructuring education, training
and entrepreneurship for micro-, and small- and medium-sized indus-
tries
• development of industrial institutions and services to support enterprise-
level efforts towards attaining competitiveness in regional and interna-
tional markets
• linking industry and agriculture to enhance productivity and competi-
tiveness in agro-industries.
Notes
* Executive Director, Economic and Social Research Foundation, Dar es Salaam.
17
Highlights of the Sustainable
Industrial Development Policy in
Tanzania, 1996–2020
A.K. Maziku*
1 Introduction
In 1996 Tanzania launched the Sustainable Industrial Development Policy

(SIDP), 1996–2020, to succeed the basic industrial strategy (BIS), 1975–95.
The BIS was Tanzania’s first long-term industrial development strategy
since independence in 1961. However, the government had already began
to pay serious attention to industrial development during the second five-
year development plan (1969–74), when it designated ten towns as urban
‘growth centres’ to which ‘suitable new factories and employment opportu-
nities will be diverted’ (URT, 1969). The expiry of the BIS in 1995, and the
economic crisis facing the country, necessitated the formulation of a new
policy that would take cognizance of the emerging macroeconomic envi-
ronment. This paper presents the salient features of the new policy, paying
special attention to the implementation mechanism, which is a major
departure from the past.
2 The Sustainable Industrial Development Policy (SIDP)
In his foreword to the new policy document, the Minister for Industry and
Trade writes: ‘Two major factors accounted for the need to prepare a new
policy for the industrial sector. One is the expiry of the Basic Industry
Strategy in 1995; the other is the government’s decision to phase itself out
of investing directly in productive activities and let the private sector be
the principle vehicle to the role’ (URT, 1996a).
The second factor described by the Minister is in fact one of the implica-
tions of the wide-ranging reform programme that the country has been
implementing since the mid-eighties. So far, Tanzania has successfully
transformed its political system from a one-party system to a multiparty
system. It has opened up its financial sector and implemented other
367
financial and monetary reforms, and the government is in the process of

hiving off non-traditional functions of the state. The emerging sociopoliti-
cal and macroeconomic environment is different from that existing in the
country when the BIS was designed in 1975. Not only did that strategy
expire in 1995 but, as described in Chapter 1, its basic assumptions were no
longer valid as early as the beginning of the eighties.
Another major development worth noting is the formulation of a
national vision, Development Vision 2025, that will generally guide socioe-
conomic development up to the year 2025. In Development Vision 2025,
Tanzania has set itself the following socioeconomic targets:
• an increase in per capita income from the current US$200 to US$1200

• a decline in infant mortality from the current 99 per 1000 inhabitants to
20 per 1000 inhabitants
• an increase in life expectancy from the current 52 years to 70 years
• an increase in the share of industry in GDP from 17 per cent to 40 per
cent
• a decline of the population growth rate from 2.8 per cent per year to
2.0 per cent.
2.1 The industrial sector mission and objectives

The mission of the industrial sector in the 1996–2020 period is two-fold:
(i) to contribute towards the achievement of the overall national long-

term development goals as contained in the Development Vision 2025
(ii) to enhance sustainability in the development of the sector.
The objectives of the new industrial policy include: overall contributions

towards human development and the creation of employment opportuni-
ties; economic transformation in order to achieve sustainable economic
growth; external balance of payments; environmental sustainability; and
equitable development. All of these objectives are consistent with the
National Vision.
The industrial sector’s role with regard to human development is two-
pronged. On the one hand, it is expected to produce goods fulfilling basic
needs for the majority of the people. On the other hand, it is expected to
create employment opportunities, directly and indirectly, through the
expansion of the industrial sector and by stimulating informal sector devel-
opment. The goal of economic transformation will benefit from the pro-
jected increasing share of industry in GDP. The development of
intermediate and capital goods industries will enhance sustainable produc-
tivity increases, technological progress, structural change and inter-sectoral
linkages.
Sustainable Industrial Development Policy 369
One of the major constraints to the implementation of the BIS was the
lack of foreign exchange to finance imports of required machinery, spares
and raw materials. Therefore, the SIDP calls for a balanced approach
between import-substitution and export-oriented industrialization.
Regional and rural–urban disparities will also be targeted so as to bring
about equitable development. The current drive to develop physical and
economic infrastructure across the country, the provision of incentive
packages to large-scale industrial investors in less developed areas and the
promotion of small and micro-enterprises are expected to contribute to
such an equitable development.
2.2 Implementation strategy

The 25-year period to which the SIDP refers has been divided into three
phases to enable specific quantitative targets to be set for indicators such as
the rate of industrial growth, the structure of value added, exports and
employment generation. For each phase, priority activities are identified.
The three phases are short term (1996–2000) medium term (2000–2010)
and long term (2010–20). Identification criteria in selecting specific priority
activities are profitability and competitive advantage.
The private sector will play a central role in the realisation of the new
policy’s objectives and goals. Apart from creating and sustaining a con-
ducive environment for the growth of the private sector, the government
will have to pursue a vigorous vocational and technical educational train-
ing programme that will include entrepreneurship development. At the
same time it will also have to provide assistance to informal-sector activi-
ties – a potential source of entrepreneurship – so that informal-sector entre-
preneurs can progressively enter the formal sector. The priorities in the
short, medium and long term are the following:
Short-term priority programme (1996–2000)
(i) Rehabilitation and consolidation of existing industrial capacities

through financial, capital and management restructuring
(ii) creating and sustaining an enabling environment
(iii) development of agro-processing activities in which Tanzania has a
comparative advantage.
Medium-term priority programme (2000–2010)
(i) Creation of new capacities in activities with clear competitive advan-

tages for export
(ii) promotion of intermediate goods industries, light capital goods and
machine making
(iii) carrying out techno-economic preparations for exploitation of iron-ore

deposits.
Long-term priority programme (2010–20)

Promoting full-fledged investments in basic capital goods industries.
3 The partnership between the government and the private

sector
In the SIDP the government role will change from that of being a direct
and indirect investor, through the parastatal enterprises, to a ‘facilitating
and supportive’ role by way of putting in place an enabling environment in
the form of appropriate macroeconomic policies, legal frameworks and
efficient infrastructures. On the other hand, the private sector will be
expected to invest in a manner that will contribute to the realization of the
objectives of the SIDP. This requires a very close working relationship or
partnership between the government and the private sector, based on a
convergence of purpose, interest, trust and transparency, all of which ulti-
mately depend on each side playing its role effectively. On the one hand,
government policies must be transparent, realistic, practical and predictable
and sustainable, while on the other, the private sector must be capable of
meeting the new challenges. There are three essential preconditions for this
to happen:
(i) the development of the private sector

(ii) the creation and sustainability of an enabling environment
(iii) the setting up of an effective consultative mechanism.
3.1 The development of the private sector
The emerging economic and business environment in Tanzania requires,

among other things, greater dependence on market forces and private
initiatives, hence the growing need for a stronger private sector. It also
requires good governance, and transparency and accountability in
administrative processes, which inspire current and future business
confidence.
The private sector in Tanzania is currently very small, and its institu-
tional infrastructure is largely undeveloped, weak and uncoordinated.
Efforts are under way to form a National Business Council. It is hoped
this will function as a central coordinating, apex business body that will
be an effective representative of private-sector interests. It is also expected
to engage in an effective dialogue with the government and assume some
of the responsibilities of the government in business and investment
promotion.
Private-sector development requires increased emphasis on vocational

and technical training as a source of entrepreneurial skills. Informal-sector
activities need to be assisted so that informal-sector firms can progressively
enter the formal sector, and micro-enterprises can grow into larger enter-
prises. The introduction of the Dar es Salaam stock exchange and continu-
ing financial sector reforms are expected to contribute to the creation and
development of a much needed capital market.
3.2 Creation and sustainability of an enabling environment

The creation and sustainability of an enabling environment refer to an
ever-changing economic situation. Therefore ongoing changes and adjust-
ments in fiscal, monetary and trade policies are required to address specific
situations as they arise.
The following measures will be implemented within the SIDP framework.
3.2.1 Trade
(i) A new trade policy will be formulated to streamline export develop-
ment, cross-border trade, import management and business licensing
(ii) further measures will be undertaken to implement the 1994 Fair Trade
Practices Act
(iii) special export incentives will be provided under the export promotion
strategy, which includes the establishment of export processing zones
(EPZs).
The new trade policy will chart Tanzania’s strategy in subregional eco-
nomic groupings, and other international and global programmes in the
context of the post-Uruguay Round agreements under the WTO. Measures
geared towards opening up the economy to external competition will con-
tinue to be taken, but selective protection within the framework of WTO
agreements will be applied.
3.2.2 Fiscal policies

Adjustments in the tax structure will continue to be made with the objec-
tive of ‘levelling the playing field’, improving tax compliance and reducing
the inflow of inferior goods.
3.2.3 Monetary policy

Interest rates associated with the limited existing credit facilities are too
high and unstable, and do not differentiate between trading and industrial
operations. Further financial sector reforms are needed to ensure interest
rate stability, encourage domestic savings and determine appropriate inter-
est rates that will distinguish between purely trading and industrial opera-
tions. Long-term financial and working capital needs for industry require
the development of a domestic capital and securities market.
3.2.4 Investment promotion

The old Investment Promotion Centre (IPC) was based on the Foreign
Investment Protection Act, which has been repealed and replaced by a new
law that provides for the establishment of a one-stop-centre, the Tanzania
Investment Centre (TIC), which will design appropriate measures for both
local and foreign investors.
3.2.5 Infrastructural development

The government will continue with efforts to expand and improve the pro-
vision of water, power, communication facilities and transport networks,
which are essential for investment and trading.
3.2.6 Action on other related issues

Other measures include: a new land policy that, among other things, will
give value to land; human resource development to improve skills and
managerial capabilities; a review of labour laws; promoting product stan-
dards and quality assurance; development, consolidation and strengthen-
ing of technology and R&D institutions and activities; and the
development of key complementary sectors of agriculture, livestock,
mining and tourism.
4 The institutional framework and the consultative

mechanism for policy implementation
There are four key players in implementing the SIDP. First, the government
will continue with the consolidation of the democratization process, the cre-
ation of a smaller, more streamlined and more efficient bureaucracy, and the
protection of individual property rights. Second, the ministry responsible for
industry will be in overall charge of the operationalization, monitoring and
review of the SIDP at the national level, in collaboration with the ministries
responsible for finance and planning. Third, the industrial support organiza-
tions, which offer essential services in the field of standardization, quality
assurance and certification, research and development, and environmental
management, will continue to play that role jointly with the government.
The fourth and last key player is the private sector. In order to be able to play
its role effectively, it will not only need to be developed (especially the local
private sector), but will need to reorganize itself so that it develops a viable
institutional network which can enter into dialogue with the government
and engage in productive joint planning on a regular basis.
The most important departure from the old plan implementation
modality will be the partnership approach adopted in the SIDP. This
approach is required for the full involvement of the private sector in the
realisation of the objectives of industrial policy. Based on the fundamental
assumption of a commitment by the partners to industrialization in
particular and economic development in general, a legally binding institu-

tional mechanism will be put in place to be known as the national consul-
tative process for industrialization (NCPI). The framework will require the
establishment of an apex umbrella organisation, which will bring together
capable people conversant with the industrial sector, from both govern-
ment and the private sector.
The main functions of the apex body will be:
(i) to recommend practical programmes of action to the government that

will sustain developments in each industrial subsystem
(ii) to provide strategic information to enterprises on technology, markets
and competitive conditions
(iii) to monitor tendencies towards excessive concentration of production
and the emergence of other forms of monopolistic practices
(iv) to participate in reviewing and monitoring the implementation of poli-
cies
(v) to act as an intermediary between enterprises, working consultative
groups and the government.
At the apex level, the organisation will be assisted by a technical support

group, which will conduct technical studies for working consultative
groups (WCGs), one for each identified industrial subsystem. The term
industrial subsystem (ISS), as used here, refers to a segment or branch of the
industrial sector comprising core enterprises organised around a product or
group of products, operating in the same markets for raw materials and
consumers, and using more or less the same type of technology and skills.
The critical requirement is for the enterprises to have a minimum level of
common interests that will make them work together, collaborate to for-
mulate development strategies, and design action plans and programmes to
improve their operating environment for their mutual benefit.
The bulk of the analytical work in terms of the problems, constraints,
strengths, weaknesses, opportunities and threats facing each industrial sub-
system will be done jointly by the technical support group and the working
consultative groups. Another unit that will assist the apex organisation will
be a management support group (MSG) which, apart from acting as a secre-
tariat to the national body, will handle financial and administrative
matters relating to the national consultative process.
The working consultative group for each subsystem will carry out the fol-
lowing main functions:
(i) to deliberate on the strengths, weaknesses, opportunities and threats in

the subsystem (SWOT)
(ii) to formulate strategies and policies in order to improve the subsystem’s
overall operating environment
(iii) to formulate action plans and programmes

(iv) to formulate actions to be taken at the level of individual enterprises.
The implementation of the national consultative process for industrializa-

tion involves the following phases:
(i) A preparatory phase that will involve putting in place a legal frame-
work within which the NCPI will operate; setting up the national co-
ordinating body; organising and staffing the technical and
management support groups; identifying industrial subsystems and
setting up their corresponding WCGs.
(ii) A study phase to establish base line data that will enable the new
approach to take off and provide the basis for detailed future studies,
including:
• diagnostic studies of the identified industrial subsystems
• analysis of the industrial structure with a view to establishing critical
gaps that need to be filled
• situation reports on the informal sector: micro- and small-scale
enterprises
• assessment of natural resource endowments for industrialization
• assessment of the performance of industrial support organizations
and the services they provide.
(iii) A consultative phase that will include initiating the consultative

process in the working groups of the industrial subsystems; the devel-
opment of strategic action plans and programmes; and the formaliza-
tion of the action plans and programmes.
(iv) A take-off phase, during which programmes and plans are imple-
mented. The implementation, monitoring, feedback and review of
strategies, programmes and action plans are a continuous process
during the whole period of the SIDP.
The implementation of the twenty-year (1975–95) basic industry strategy

formally ended in 1995. A new 25-year policy, the Sustainable Industrial
Development Policy (1996–2020), was designed and launched in 1996. The
SIDP is expected to achieve the following: improvement of the quality of
human life, creation of employment opportunities, economic transforma-
tion for achieving sustainable economic growth, external balance, environ-
mental sustainability and equitable development. These objectives are
consistent with the National Vision 2025. The success of the Sustainable
Industrial Development Policy depends on the success of the economic
reform programme in creating better prospects for the effective reversal of

the declining performance of the industrial sector.
The major departure of the new policy from the previous ones is with
respect to the implementation modality. Under the liberalized economic
system the private sector and market forces are increasingly replacing the
government and its organs. The government is initiating a national consul-
tative process for industrialization that will constitute an apex body, and
working consultative groups for each main branch of industry to supervise
and monitor policy implementation. For this to succeed, a government–
private-sector partnership must be established and supported by a struc-
tured dialogue and continuous consultative process.
Note
* Ministry of Industry and Trade.
Bibliography
Adam, C. (1994), ‘Privatization and Structural Adjustment in Africa’, in: W. van der
Geest (ed.), Negotiating Structural Adjustment in Africa, London: James Curry.
Agrawal, N., Z. Ahmed, M. Mered and R. Nord (1993), Structural Adjustment, Economic
Performance, and Aid Dependency in Tanzania, Policy Research Working Paper
WPS1204, Washington: World Bank.
Alange, S. (1987), Acquisition of Capabilities through International Technology Transfer:
The Case of Small Scale Industrialization in Tanzania, published PhD thesis, Sweden:
Chalmers University of Technology.
Alcorta, L. (1995), The Impact of Industrial Automation on Industrial Organisation:
Implications for Developing Countries’ Competitiveness, submitted as Background
Paper no. 7 to the UNIDO Global Forum on Industry.
Appelbaum, R.P. and W.J. Chambliss (1995), Sociology, New York: Harper Collins.
Ark, B. van (1993), ‘The ICOP Approach – Its Implications and Applicability’, in: A.
Szirmai, D. Pilat and B. van Ark (eds), Explaining Economic Growth: Essays in Honour
of Angus Maddison, North Holland: Elsevier, 375–98.
Arkadic, B. van (1986) (ed.), External Finance and Policy Adjustment in Africa, London,
Sage.
Arrow, K. (1962), ‘Economic Welfare and the Allocation of Resources for Inventions’,
in: R.R. Nelson (ed.), The Rate and Direction of Inventive Activity, Princeton:
Princeton University Press.
Bagachwa, M.S.D. (1992), ‘Choice of Technology in Small and Large Firms: Grain
Milling in Tanzania’, World Development, 20 (1), 97–107.
Bagachwa, M.S.D. (1993a), ‘Impact of Structural Adjustment on the Small Enterprise
Sector in Tanzania’, in: A.H.J. Helmsing and T. Kolstee (eds), Small Enterprises and
Changing Policies, London: Intermediate Technology Publications, 91–113.
Bagachwa, M.S.D. (1993b), Export Competitiveness and Firm Level Capabilities in
Selected Tanzanian Industries, Economic Research Bureau, University of Dar es
Salaam.
Bagachwa, M.S.D. (1994), Export Competitiveness and Firm Level Capabilities in Selected
Tanzanian Industries, Research Paper, ERB, University of Dar es Salaam.
Bagachwa, M.S.D. and A.V.Y. Mbelle (eds) (1993), Economic Policy under a Multiparty
System in Tanzania, University of Dar es Salaam, DUP.
Bagachwa, M.S.D. and A.V.Y. Mbelle (1995), ‘Tanzania: Export Manufacturing’, in:
S.M. Wangwe (ed.), Exporting Africa: Technology, Trade and Industrialization in sub-
Saharan Africa, London Routledge.
Bank of Tanzania (1970 [1973, 1977, 1978, 1981, 1982]), Economic and Operations
Report June 1982 [1973, 1977, 1978, 1981, 1982], Dar es Salaam: Bank of Tanzania.
Bank of Tanzania (1975), Economic Bulletin, 7, 1.
Bank of Tanzania (1981), Tanzania: Twenty Years of Independence (1961–1981), A
Review of Political and Economic Performance, Dar es Salaam: Bank of Tanzania.
Bank of Tanzania (1984), Tanzania: Twenty Years of Independence (1961–1981): A
Review of Political and Economic Performance, Dar es Salaam.
Bank of Tanzania (1988), Economic and Operations Report 1988, Dar es Salaam: Bank of
Tanzania.
376
Bibliography 377
Bank of Tanzania (1989), Economic and Operations Report for the Year ended 30th June
1989, Dar es Salaam: Bank of Tanzania.
Tanzania.
Tanzania.
Bank of Tanzania (1996a), Economic Bulletin, 24 (4).
Bank of Tanzania (1996b), Economic Trends in Figures, June 1966–June 1996.
Bank of Tanzania (1997), Economic Bulletin, 30 Sept.
Bank of Tanzania (1998a), Economic Bulletin, 27 (4).
Bank of Tanzania (1998b), Monthly Economic Review, May 1998.
Bank of Tanzania (various), Economic and Operations Report.
Bank of Uganda (1995), Importation of Fertilizers in Uganda (1990–1994), Agricultural
Secretariat.
Barker, C.E., M.R. Baghavan, P.V. Mitschke-Collande and D.V. Wield (1986), African
Industrialisation: Technology and Change in Tanzania, Vermont: Gower.
Barro, R.J. and Jong-Wha Lee (1993), ‘Losers and Winners in Economic Growth’,
Proceedings of the World Bank Annual Conference on Development Economics, 267–98.
Washington DC: The World Bank.
Barros, A.R. (1993), ‘Some implications of New Growth Theory for Economic
Development’, Journal of International Development, 5 (5), 532–58.
Bates, R. (1981), Markets and States in Tropical Africa, Berkeley: University of
California Press.
Beadle, L.C. and E.M. Lind (1960), ‘Research on the Swamps of Uganda’, Uganda
Journal, 24, 84–98.
Beaulac, M.N. and K.H. Reckhow (1982), ‘An Examination of Land Use – Nutrient
Export Relationships’, Water Resources Bulletin, 18, 1013–24.
Bell, M., B. Ross-Larsen and L.Westphal (1984), ‘Assessing the Performance of Infant
Industries’, Journal of Development Economics, 16, 101–28.
Benjamin, J.R. and C.A. Cornell (1970), Probability, Statistics and Decision for Civil
Engineers, New York: McGraw-Hill.
Bennett, J.T. and M.H. Johnson (1979), ‘Public versus Private Provision of Collective
Goods and Services’, Public Choice, 34:55–64.
Bertholet, C.J.L. and P.E. Lapperre (1991), The Development of Societies in Global
Perspective, Eindhoven: lecture notes, Eindhoven University of Technology, Mimeo.
Biersteker (1991), ‘Reducing the Role of the State in the Economy: A Conceptual
Exploration of IMF and World Bank Prescription’, International Studies Quarterly,
34, 477–92.
Board of External Trade (1989), Ten Years of the Board of External Trade: 1979–1989,
Dar es Salaam: Tanzania Litho Ltd.
Boardman, V.E. and Vining (1989), ‘Ownership and Performance in Competitive
Environment: A Comparison of the Performance of Private, Mixed and State
Owned Enterprises’, Journal of Law and Economics, 32 (April), 136.
Bol, D. (1995), Employment and Equity Issues in Tanzania, A paper presented at an ERB
workshop on Beyond SAP in Tanzania, Dar es Salaam.
Bongenaar, B. (1997), ‘Analyzing Technology Development, Part I: Evaluating the Role of
the Tanzania Research and Development Organization’, MSc thesis, Eindhoven:
Bongenaar, B. and A. Szirmai (1998), The Role of a Research and Development Institute
in the Development and Diffusion of Technology: the case of TIRDO, paper presented at
378 Bibliography
the conference on The Industrial Performance of Tanzania, Eindhoven, 25–27 June

1998.
Bongenaar, B. and A. Szirmai (1999), The Role of a Research and Development Institute
in the Development and Diffusion of Technology, Working Paper 99.9, Eindhoven
Centre of Innovation Studies, http://tm.tue.nl/ecis/pubs.html.
Bootsma, H.A. and R.E. Hecky (1993), ‘Conservation of the African Great Lakes: A
Limnological Perspective’, Conservation Biology, 7, 3, 644–55.
Borcherding, T.E., W.W. Pommerchne and F. Schneider (1982), ‘Comparing the
Efficiency of Private and Public Production: The Evidence from Five Countries,’ in:
Zeitschrift für National-Ökonomie, 2.
Bos, E. (1995), World Population Projections 1994–1995: Estimates and Projections with
Related Demographic Statistics, London: Longman.
Bos, O.F. (1997), Energy Conservation in Tanzanian Industry, Eindhoven: Technology
and Development Studies, Eindhoven University of Technology.
Boyco, M., A. Shleifer and R.W. Vishny (1993), A Theory of Privatization, Boston:
Harvard University.
Brabant, J.M. van (1994), ‘On the Economics of Property Rights and Privatisation in
Transitional Economics’, in: P. Cook and C. Kirkpatrick (eds), Privatisation Policy
and Performance – International Perspectives, New York: Prentice Hall.
Braun, A. (1995), Nutrient Flows in Fallow Systems and Maize in Western Kenya,
Wageningen, International Centre for Development Oriented Research in
Agriculture: ICRA.
Bromley, R.J. (1978), ‘Organization, Regulation and Exploitation in the So-Called
Urban Informal Sector: The Street Traders of Cali Columbia’, World Development, 6
(9), Sept.–Oct., 1161–86.
Bromley, R. and C. Birbeck (1984), ‘Researching Street Occupations of Cali: The
Rationale and Methods of What Many Would Call an Informal Sector Study’,
Regional Development Dialogue, 5 (2), 184–203.
BRU (1974), Minimum Requirements for Permanent Single Storey Houses, Technical
Pamphlet no. 2, Svare, TI, Dar es Salaam: Building Research Unit.
BRU (1974–95), Various Reports and Technical Guidelines, Dar es Salaam: Building
Research Unit.
BRU (1985), The Tanzanian Building Regulations, 1985, Dar es Salaam: Building
Research Unit of the Prime Minister’s Office.
Bullock, A., S.O. Keya, F.M. Muthuri, T. Bailey-Watts, R. Williams and D. Waughray
(1995), Lake Victoria Environmental Management Programme, Draft Final Report by
the Regional Consultants.
Bureau of Labour Statistics, USA (1998), Data Base Producer Price Indices, Internet:
http:\\146142.424\CGI-BIN\DSRV, Version d.d. April 26, 1998.
Bureau of Statistics (1970), Survey of Industrial Production, 1967 [1969–1974,
1979–1988, 1990], Dar es Salaam: Bureau of Statistics, Ministry of Economic Affairs
and Development Planning.
Bureau of Statistics (1971), National Accounts of Tanzania, 1966 to 1968, Sources and
Methods, Dar es Salaam: Bureau of Statistics, Ministry of Economic Affairs and
Development Planning.
Bureau of Statistics (1972), The Economic Survey 1971–1972, Dar es Salaam,
Government Printer.
Bureau of Statistics (1977), The Economic Survey 1975–1976, Dar es Salaam,
Government Printer.
Bureau of Statistics (1981), National Accounts of Tanzania, 1966–1980, Dar es Salaam:
Bureau of Statistics, Ministry of Economic Affairs and Development Planning.
Bibliography 379
Bureau of Statistics (1982), The Economic Survey 1981, Dar es Salaam, Government
Printer.
Bureau of Statistics (1983), Statistical Abstract, 1982, Dar es Salaam: Bureau of
Statistics, Ministry of Economic Affairs and Development Planning.
Bureau of Statistics (1985a), National Accounts of Tanzania, 1976–1984, Sources and
Methods, May.
Bureau of Statistics (1985b), Census of Industrial Production, vol. 3, Dar es Salaam:
Bureau of Statistics, Ministry of Planning and Economic Affairs, March.
Bureau of Statistics (1986), Input–Output Table of Tanzania, 1976, a Background on
Sources and Methods, Dar es Salaam: Bureau of Statistics, Ministry of Finance,
Economic Affairs and Planning.
Bureau of Statistics (1988a), Industrial Commodities, Quarterly Report 87–1 [1985:4,
1986:1–4, 1987:1–4, 1988:1–2, 1989:1–4, 1990:1–2, 1990:1–4, 1992: 1–2, 1991:1–4,
1992:1–4, 1993:1, 1993:3, 1993:4, 1994:1, 1995:3], Dar es Salaam: Bureau of
Statistics, Ministry of Finance, Economic Affairs and Planning.
Bureau of Statistics (1988b), Population Census: Preliminary Report of the Bureau of
Statistics, Ministry of Finance, Economic Affairs and Planning, Dar es Salaam:
Government Printer.
Bureau of Statistics, President’s Office, Planning Commission.
Bureau of Statistics (1993a), Census of Industrial Production (1989), vol. V, Methodology
Report, Dar es Salaam: Bureau of Statistics, President’s Office, Planning
Commission.
Bureau of Statistics (1993b), National Accounts of Tanzania, Dar es Salaam: Bureau of
Statistics, President’s Office, Planning Commission.
Bureau of Statistics (1994a), Census of Industrial Production (1989), vol. IV (Part 1),
Comprehensive Report 10+, Dar es Salaam: Bureau of Statistics, President’s Office,
Planning Commission.
Bureau of Statistics (1994b), Environment Statistics in Tanzania Mainland, Dar es
Salaam: Bureau of Statistics, President’s Office, Planning Commission.
Bureau of Statistics (1994c), National Accounts of Tanzania, 1976–1993, Dar es Salaam:
Bureau of Statistics, President’s Office, Planning Commission, August.
Bureau of Statistics (1994d), Statistical Abstract, 1992, Dar es Salaam: President’s
Office, Planning Commission, Government Printer, February.
Bureau of Statistics (1994e), Selected Statistical Series, 1951–1991, Dar es Salaam:
President’s Office, Planning Commission.
Bureau of Statistics (1995a), Revised National Accounts of Tanzania, 1976–1990, Dar es
Bureau of Statistics (1995b), National Accounts of Tanzania, 1976–1994, Dar es
Bureau of Statistics (1995c), Selected Statistical Series, 1951–1992, Dar es Salaam:
Bureau of Statistics (1995d), Workshop on Consumer Price Index, 6 to 8 February 1995,
Mkonge Hotel, Tanga.
Bureau of Statistics (1995e), Selected Statistical Series, 1951–1994, Dar es Salaam:
Bureau of Statistics (1996a), Producer Price Index (Manufacturing Sector), Quarterly Report
1995:3, Dar es Salaam: Bureau of Statistics, President’s Office, Planning
Commission.
380 Bibliography
Bureau of Statistics (1996b), Industrial Commodities Quarterly Report, 1996:1, Dar es

Salaam (and various issues).
Bureau of Statistics (1999), Division of National Accounts, Printout 1987–1998, Dar es
Salaam: National Bureau of Statistics, September.
Calamari, D.A., M.O. Akech. and P.B.O. Ochumba (1994), Pollution of Winam Gulf,
Lake Victoria, Kenya, CIFA Seminar on African Inland Fisheries, Aquaculture and
Environment, Harare, Zimbabwe, 5–7 December.
Callely, A.G., C.F. Forster and D.A. Stafford (1977), Treatment of Industrial Effluents,
London: Hodder & Stoughton.
Cameron, R. (1991), Economische Wereldgeschiedenis: van de Oude Steentijd tot Heden,
Utrecht: Spectrum.
Caniëls, M. (1999), Regional Growth Differentials: the Impact of Locally Bounded
Knowledge Spillovers, thesis, University of Maastricht.
Carlsson, B. (1987), ‘Reflections on “Industrial Dynamics”: The Challenges Ahead’,
International Journal of Industrial Organization, 5 (2), 135–48.
Castells, M. (1997), The Information Age: Economy, Society and Culture, vol. 1, Oxford:
Blackwell Publishers.
Caves, D.W. and L.R. Christensen (1980), ‘The Relative Efficiency of Public and
Private Firms in a Competitive Environment: The Case of Canadian Railroads’,
Journal of Political Economy, 98.
Centraal Bureau voor de Statistiek (1995), Vijfennegentig Jaren Statistiek in Tijdreeksen
1899–1994, Den Haag: Staatsuitgeverij.
Central Statistic Bureau (1964a), Census of Industrial Production in Tanganyika 1961,
Dar es Salaam: Central Statistic Bureau.
Central Statistic Bureau (1964b), The National Accounts of Tanganyika, 1960–1962,
Dar es Salaam: Central Statistic Bureau.
Central Statistic Bureau (1965), Survey of Industries, 1965, Memorandum on Definitions,
Concepts and Procedures, Dar es Salaam: Central Statistic Bureau, Ministry of
Economic Affairs and Development Planning.
Central Statistic Bureau (1966), Survey of Industries, 1965, Dar es Salaam: Central
Statistic Bureau, Ministry of Economic Affairs and Development Planning.
Central Statistic Bureau (1969), Survey of Industries, 1967, Dar es Salaam: Central
Statistic Bureau, Ministry of Economic Affairs and Development Planning.
Central Statistical Office (1976), The Measurement of Changes in Production, Studies in
Official Statistics, 25, London: Her Majesty’s Stationery Office.
Centre for Development Research (1995), Structural Adjustment in Africa: A Survey of
the Evidence.
Chambua, S.E. (1990), Technology Assessment and Acquisition in Tanzania, Institute for
Development Studies.
Chazan, N., R. Mortimer, J. Ravenhill and D. Rothchild (eds) (1988), Politics and
Society in Contemporary Africa, Boulder: Lynne Rienner.
Chenery, H.B. (1958), ‘Development Policies and Programmes’, Economic Bulletin for
Latin America, 3 (1).
Chenery, H.B. (1979), Structural Change and Development Policy, Oxford: Oxford
University Press.
CHS (1995), Urban and Housing Indicators Study for Dar es Salaam, Dar es Salaam:
Centre for Human Settlements Studies, ARDHI Institute.
Chungu et al. (1994), Assessment and Choice of Technological Alternatives – The Case of
Edible Oil Processing in Rural Tanzania, Dar es Salaam: Institute of Production
Innovation.
Bibliography 381
Clark, W.E. (1978), Socialist Development and Public Investment in Tanzania, Toronto:
University of Toronto Press.
Clerides, S.K., S. Lachs, and J.R. Tybout (1998), ‘Is Learning by Exporting Important?
Micro-Dynamic Evidence from Colombia, Mexico, and Morocco’, Quarterly Journal
of Economics, 113, 903–48.
Cody, J., H. Hughes and D. Wall (1980), Policies for Industrial Progress in Developing
Countries, Oxford: Oxford University Press.
Cody, J., R. Kitchen and J. Weiss (1990), Policy Design and Price Reform in Developing
Countries: Guidelines with Special Reference to Industry, New York: Harvester
Wheatsheaf.
Collier, P. and J. Gunning (1999), ‘Explaining African Economic Performance’,
Journal of Economic Literature, 37, 64–111.
Cooper, C. (1991), Are Innovation Studies on Industrialized Economies Relevant to
Technology Policy in Developing Countries?, UNU/INTECH, Working Paper no. 3,
Maastricht.
Cooper, C. (1995), Technology, Manufactured Exports and Competitiveness,
UNU/INTECH Discussion Paper no. 9513, December.
Cooper, C.M. (1993), Technology and Innovation in the International Economy, London:
Edward Elgar.
Cooper, C., et al. (1994), Import Liberalization, Industrialization and Technological
Capability in Africa: An Outline of a Research Project, Oxford.
COSTECH (1990), Tanzania Commission for Science and Technology, Dar Es Salaam.
Coulson, A. (ed.) (1979), African Socialism in Practice, Nottingham: Spokesman.
Coulson, A. (1982a), ‘The State and Industrialization in Tanzania’, in: M. Fransman
(ed.), Industry and Accumulation in Africa, London: Heinemann, 60–76.
Coulson, A. (1982b), Tanzania: A Political Economy, Oxford: Clarendon Press.
Crul, R.C.M. (1993), Limnology and Hydrology of Lake Victoria, Monographs of the
African Great Lakes, UNESCO/IHP.
CS I (September 1982–February 1987), Caustic Soda Project File I, collection of approx-
imately 150 notes, reports and documents, TIRDO.
CS II (March 1987–July 1993), Caustic Soda Project File II, collection of approximately
180 notes, reports and documents, TIRDO.
Dahlmann, C., B. Ross-Larsen and L.E. Westphal (1987), ‘Managing Technological
Development: Lessons from Newly Industrialising Countries’, World Development, 15.
DANIDA (1989), Country Strategy for Strengthening Environmental Considerations in
Danish Development Assistance to Tanzania.
Dao, G.C. (1989), ‘The Developing World and the Multi-Fibre Arrangement’, in: J.
Whalley (ed.), Developing Countries and the Global Trading System, vol. 2,
Basingstoke: Macmillan Press – now Palgrave.
Dar, U. (1990), ‘Management of Technological Change: An Overview’, in: UNIDO,
Management of Technological Change: Issues and Case Studies from India, chapter XIV,
Vienna, UNIDO.
Dar es Salaam Technical College (1995), unpublished internal documents.
David, P. (1975), Technological Choice, Innovation and Economic Growth, Cambridge:
Cambridge University Press.
Davies, R. (1979), ‘Informal Sector Versus Subordinate Mode of Production: A
Model’, in: R. Bromley and C. Gerry (eds), Casual Work and Poverty in Third World
Cities, Chichester: Wiley, 87–109.
DeBresson, C. (1989), ‘Breeding Innovation Clusters: A Source of Dynamic
Development’, World Development, 17 (1), 1–16.
382 Bibliography
DeBresson, C. (1994), Mapping the Sources and Application of Innovative Activities in

Developing Countries, Research Development Project for a Comparative Study,
University of Quebec at Montreal (UQAM).
De Soto, H. (1989), The Other Path: The Invisible Revolution in the Third World, New
York: Harper and Row.
Dessing, M. (1990), Support for Microenterprises: Lessons for sub-Saharan Africa, World
Bank Technical Paper Number 122, Washington DC: The World Bank.
Dijk, M.P. van (1980), ‘De Informele Sector van Ouagadougou en Dakar’, disserta-
tion, Amsterdam: Free University.
Division of Environment (1995), National Environmental Policy (final draft), Ministry
of Tourism, National Resources and Environment, Dar es Salaam.
Diyamet, B.D., R.S. Mabala and R. Mandara, Poverty and Diffusion of Technological
Innovations to Rural Women: The Role of Entrepreneurship, Research on Poverty
Alleviation, Research Report no. 98.2, Dar es Salaam.
Donges, J.B. (1976), ‘A Comparative Survey of Industrialisation Policies in Fifteen
Semi-Industrial Countries’, Weltwirtschaftliches Archiv, 112, 4.
Dore, R. (1984), ‘Technological Self Reliance’, in: M. Fransman and K. King (eds),
Technological Capability in the Third World, London: Macmillan Press – now
Palgrave, pp. 65–80.
Dosi, G. (1984), Technical Change and Industrial Transformation, New York: St Martin’s
Press – now Palgrave.
Dosi, G. (1988), ‘Sources, Procedures and Microeconomic Effects of Innovation’
Journal of Economic Literature, 26, September.
Downs, A. (1967), Inside Bureaucracy, Boston: Little Brown.
Droruga, N. (1990), Report on the Purification of Industrial Waste Water in Uganda,
Kampala: Ministry of Environmental Protection.
Due, J.M. and A.E. Temu (1996), A Pilot Study of Privatised Agro-Industrial Parastatals in
Tanzania, Morogoro: Urban.
Duijsens, R.J.H. (1995), Education in Tanzania: A Sector Review, unpublished docu-
ment, Eindhoven: Eindhoven University of Technology.
Duijsens, R.J.H. (1996), Metal Industry and Technical Education in Tanzania: Model
for Matching Demand and Supply of Knowledge and Skills, MSc thesis,
Eindhoven: Technology and Development Studies, Eindhoven University of
Technology.
Duijsens, R.J.H. and P.E. Lapperre (1998), Education, Technical Education and the
Nature and Causes of Shortcomings in Knowledge and Skills in the Metalworking
Industry in Tanzania, paper presented at the conference The Industrial Performance
of Tanzania, Eindhoven, 25–27 June 1998.
Easterly, W. (1993), Is Africa Different? Evidence From Growth Regressions, Washington
DC: World Bank Policy Research Department.
Economopoulos, A.P. (1993), Assessment of Sources of Air, Water and Land Pollution,
Part One: Rapid Inventory Techniques in Environmental Pollution,
WHO/PEP/GETNET/93.1-A, Geneva: World Health Organization.
Egmond-de Wilde de Ligny, E.L.C. van (1993), International Technology Transfer to
Developing Countries, lecture notes no. 1649, Eindhoven: Eindhoven University of
Technology.
Egmond-de Wilde de Ligny, E.L.C. van (1995), International Technology Transfer to
Developing Countries, Summary of ITT lectures, 1994–95, Eindhoven: TDS,
Eindhoven University of Technology.
Egmond-de Wilde de Ligny, E.L.C. van, (1996), ‘Summarized Findings of the Studies
of the Housing Situation in Tanzania’, chapter 4 pp 29–35, in: Report of Baseline
Bibliography 383
Surveys on the National Opportunities for the Application of New Bamboo Construction
Technologies in the Housing Sector in Tanzania, Egomond-de Wilde de Ligny, E.C.L.
van, Gaillard, H. and Janssen, J.J.A. (ed.): Centre for International Co-Operation
Activities, Eindhoven University of Technology, Eindhoven, The Netherlands and
NCC, Dar es Salaam, Tanzania, February 1996.
Egmond-de Wilde de Ligny, E.C.L. and Rijkenberg, S. (1998), The Construction
Industry in Tanzania: Challenges to Contribute to Tanzanian Socioeconomic
Development? Paper presented at the Conference on Industrial Performance of
Tanzania, Eindhoven, The Netherlands, 25–27 June, 1998.
Egmond-de Wilde de Ligny, E.L.C. van (2000), Technology Mapping for Technology
Management, PhD Thesis, Delft University of Technology.
Ehrhorn, T.H. (1982), Integrierte Energieversorgung im Brauereien, Berlin, Ehrhorn.
Elbadawi, I. (1998), ‘Real Exchange Rate Policy and Export Competitiveness in sub-
Saharan Africa’, paper presented at a UNU/WIDER Workshop on the Impact of
Liberalization on Key Markets in sub-Saharan Africa, Addis Ababa, Ethiopia,
March.
Engblom, A.C. and V. Veres (1998), Capacity Utilization and the Impact of Structural
Adjustment: The Case of the Tanzania Textile Industry unpublished MA thesis,
Gothenburg University.
Engelen, D.M. van (1996), Public Policy Making and the Rise and Fall of the
Tanzanian Manufacturing Sector, MSc thesis, Eindhoven: Technology and
ESCAP (1989), A Framework for Technology Based Development: An Overview of the
Framework for Technology-Based Development, Bangalore: United Nations.
ESRF (1997), Multi-Country Comparative Study of Private Enterprise Development (MCC-
PED): A Tanzania Country Report, (S.M. Wangwe, H.H. Semboja, J.P. Kweka and
H.B.P. Moshi), Dar es Salaam, Tanzania.
ESRF (1998), Labour Market Survey for VETA, Dar es Salaam, May.
ESRF-PSRC (1996), The Impact of Divestiture and Privatization Process on Industrial
Performance: The Case of Six Divested Companies, a report submitted to the Parastatal
Sector Reform Commission (PSRC), Dar es Salaam.
ESRF-VETA (1998), Labour Market Survey in the Dar es Salaam Region, final draft report.
Evans, J. (1985), The Flowering of the Middle Ages, London: Thames and Hudson.
Co-penhagen.
Fei, J.C.H. and G. Ranis (1964), Development of the Labour Surplus Economy, Theory and
Policy, Homewood, Ill: Richard D. Irwin.
Fei, J.C.H. and G. Ranis (1976), ‘A Model of Growth in the Open Dualistic Economy:
the Case of Korea and Taiwan’, Journal of Development Studies, May.
Fereira, M.L. (1993), Tanzania: A Poverty Profile, Washington DC: World Bank.
Fereira, M.L. and C.C. Griffin (1995), Human Resource Development Survey, Tanzania,
Washington DC: World Bank, I, Annexes I & II.
Fleisheuer, M. (1994), Road Construction in Tanzania: Problems, Constraints and
Opportunities. A Case Study on the Execution of Road Construction in Tanzania,
MSc thesis, Eindhoven: Technology and Development Studies, Eindhoven
University of Technology.
Foster, A.D. and M.R. Rosenzweig et al. (1995a), Learning by Doing and Learning From
Others: Human Capital and Technical Change in Agriculture, University of
Pennsylvania.
Foster, A.D. and M.R. Rosenzweig et al. (1995b), Technical Change and Human Capital
Returns and Investments: Evidence from the Green Revolution, University of
Pennsylvania.
384 Bibliography
Fransman, M. (1984), ‘Technological Capability in the Third World: An Overview

and Introduction to Some of the Issues Raised in this Book’, in: M. Fransman and
K. King (eds), Technological Capability in the Third World, London: Macmillan Press
– now Palgrave.
Fransman, M. (1986), Technology and Economic Development, Brighton: Wheatsheaf
Books.
Freeman, C. (1982), The Economics of Industrial Innovation, 2nd edn, London: Pinter.
Gallagher, M. (1991), Rent Seeking and Economic Growth in Africa, Boulder: Westview.
Galtung, J. (1968), Sociological Theory and Social Development, Kampala: Transition
Books.
Gerschenkron, A. (1962), Economic Development in Historical Perspective, Cambridge:
Harvard University Press.
Gnoth, M. (1994), Assessment of Condition of Trucks and Busses: Road Haulage Sector in
Tanzania, MSc thesis, Eindhoven: Technology and Development Studies,
Goodlad, J. (1979), What Schools Are for, Bloomington: Peacock.
Gophen, M., P.B.O. Ochumba and L.S. Kaufman (1995), ‘Some aspects of perturba-
tion in the structure and biodiversity of the ecosystem of Lake Victoria (East
Africa)’, Aquatic Living Resources, 8, 27–41.
Graham-Brown, S. (1991), Education in the Developing World, London: Longman.
Green, R.H., D.G. Rwegasira and B. van Arkadie (1980), Economic Shocks and National
Policy Making: Tanzania in the 1970s, The Hague: Institute of Social Studies.
Greenaway, D. and C. Milner (1993), Trade and Industrial Policy in Developing
Countries: A Manual of Policy Analysis, London: Macmillan Press – now Palgrave.
Grigg, D. (1982), The Dynamics of Agricultural Change: the Historical Experience,
London: Longman.
Grossman, G.M. and E. Helpman (1995), Technology and Trade, National Bureau of
Economic Research, Working Paper no. 4926.
Guerguil, M. (1988), ‘Some Thoughts on the Definition of the Informal Sector’,
CEPAL Review, 35, August.
Gulhati, R. (1990), ‘Who Makes Economic Policy in Africa and How?’, World
Development, 18, 8.
Gulhati, R. and U. Sekhar (1982), ‘Industrial Strategy for Late Starters: The
Experience of Kenya, Tanzania and Zambia’, World Development, 10, 11.
Haggblade, S., C. Liedholm, and D.C. Mead (1990), ‘The Effect of Policy and Policy
Reforms on Non-Agricultural Enterprises and Employment in Developing
Countries: A Review of Past Experiences’, in: F. Stewart, H. Thomas and T. de
Wilde (eds), The Other Policy, London: Intermediate Technology Publications.
Hajjar, B.A. and J. Preskey (1993), ‘Privatization in an African Context: The Case of
the United Republic of Tanzania’, Industry and Development, 32.
Hammond, R.L. (1973), ‘Evaluation at the Logical Level’, in: B.R. Worthen and J.R.
Sanders, Educational Evaluation: Theory and Practice, Belmont: Wadsworth.
Hart, J. (1971), ‘Informal Income Opportunities and Urban Employment in Ghana’,
Journal of Modern African Studies, 2, 61–89.
Havnevik, K.J., R. Skarstein and S.M. Wangwe (1985), Small Scale Industrial Sector
Study – Tanzania, Review of Experiences and Recommendations for the Future, a Report
presented to the Ministry of Industries and Trade, October 1985.
Hecky, R.E. (1993), ‘The Eutrophication of Lake Victoria’, Verhandlungen der
Internationale Vereinigung für Limnologie, 25, 856–9.
Helleiner, G.K. (1990), ‘Trade Strategy in Medium Term Adjustment’, World
Development, 18, 6.
Bibliography 385
Helleiner, G.K. (ed.) (1992), Trade Policy, Industrialization and Development, Oxford:
Clarendon Press.
Helleiner, G.K. (1994), ‘From Adjustment to Development in sub-Saharan Africa:
Consensus and Continuing Conflict’, in: G.A. Cornia and G.K. Helleiner, From
Adjustment to Development in Africa: Conflict, Controversy, Convergence, Consensus,
London: Macmillan Press – now Palgrave.
Hemming, R. and K. Miranda (1991), ‘Privatization’, in: Ke-young Chu and R.
Hemming (ed.) Public Expenditure Handbook, IMF, Washington, DC 139–45.
Henley, J.S. and G.B. Assaf (1993), ‘Privatization in an African Context: The Case of
the United Republic of Tanzania’, United Nations Industrial Development
Organisation, Industry and Development, 32.
HIID (1998), Africa Competitiveness Report, Boston: Harvard,
Hillebrand, H. and H. Jansen (1991), ‘Voorwaarden voor industrialisering’, in:
J.H. Righart (ed.), De Trage Revolutie: Over de Wording van Industriele Samenlevingen,
Meppel: Boom.
Husain (1993), ‘Comment on Barro and Wha’, Proceedings of the World Bank Annual
Conference on Development Economics, 305–310, Washington DC: The World Bank.
Hyden, G. (1982), Beyond Ujamaa in Tanzania: Underdevelopment and an Uncaptured
Peasantry, London: Heinemann.
ICRAF (International Centre for Research in Agroforestry) (1995), ‘Rain Sample
Analysis’, in: A. Braun (ed.), Nutrient Flows in Fallow Systems and Maize in Western
Kenya, 163.
ILO (1972), Employment, Incomes and Equality: A Strategy for Increasing Productive
Employment in Kenya, Geneva: International Labour Organization.
ILO (1982), Basic Needs in Danger, Addis Ababa.
ILO (1992), Increasing Productive Employment in East Africa, Geneva: ILO.
Iwaarden, A. van (1997) Production and Application of Timber in the Construction Sector,
Tanzania, MSc thesis, Eindhoven: Technology and Development Studies,
Iwugo, K.O. (1990), Basic Concepts of Environmental Quality Control in Relation to
Baseline Studies on Land-Based Sources of Pollution, lecture presented at the
IOC–SAREC–UNEP–FAO–IAEA–WHO workshop on regional aspects of marine pol-
lution, Mauritius, 20 October–9November, Mimeo.
James, J. (1995), The State, Technology and Industrialization in Africa, Basingstoke:
Macmillan Press – now Palgrave.
James, J. (1998), Public Choice, Technology and Industrialization in Tanzania,
paper presented at the conference on The Industrial Performance in Tanzania,
Eindhoven, 25–27 June.
Jansen, H. (1991), ‘De sterke en Flexibele Staat: Politiek-Institutionele Voorwaarden
Voor Industrialisatie’, in: H. Righart, (ed.), De Trage Revolutie: Over de Wording van
Industriële Samenlevingen, Meppel: Boom.
Jedruszek, J. (1979), Development in Employment and Productivity in Tanzania, 1967–77,
paper 78.5/78.6 Economic Research Bureau, University of Dar es Salaam.
Jolliffe, D. (1995), Cognitive Skills, Schooling and Income: Evidence from Ghana,
manuscript.
Jørgensen, S.E. (1980), Lake Management, Oxford: Pergamon.
Kahama C.G. (1982), Tanzania Professional Centre 2nd Inter Professional Conference on
Priorities in National Development: Executive Summary, Dar Es Salaam.
Kaplinsky, R. (1984), ‘Third World Industrialisation in the 1980s: Open Economies
in a Closing World’, Journal of Development Studies, 21 (1).
Kaplinsky, R. (1990), The Economies of Small, London: IT Publications.
386 Bibliography
Katz, J.M. (1987) (ed.), Technology Generation in Latin American Manufacturing

Industries, Basingstoke: Macmillan Press –.
Kenya Grain Growers Cooperative Union (1991), ‘Government Fertilizer
Distribution, 1990/1991’, unpublished, Nakuru, Kenya.
Kers, D. (1997), Industrial Development and Electricity Supply: Mbagala, Tanzania, as a
Case, MSc thesis, Eindhoven: Technology and Development Studies, Eindhoven
Kessels, J.W.M. and C.A. Smit (1992), ‘Opleidingsnoodzaak en Evaluatie: op zoek
naar Vaardigheden’, in: Opleiders in Organisaties Capita Selecta, Deventer: Kluwer
Bedrijfswetenschappen.
Killick, T. (1976), ‘The Possibilities of Development Planning’, Oxford Economic
Papers, 41 (3).
Killick, T. (1981), Policy Economics, London: Heinemann.
Killick, T. (1989), Principles of Policy for the Adaptive Economy, Working Paper no. 32,
London: Overseas Development Institute.
Killick, T. (1990), Markets and Governments in Agricultural and Industrial Adjustment,
Working Paper no. 34, London: Overseas Development Institute.
Kimambo, R.H. (1984), Mining and Mineral Prospects in Tanzania, Arusha: Eastern
Africa Publications Ltd.
Kimambo, R.H. (1988), Development of the Non-Metallic Minerals and the Silicate
Industry in Tanzania, a Profile of the Silicate Industry in Tanzania, Arusha, Tanzania.
King, K. (1988), Aid and the Educational Research in Developing Countries: The Role of
Donor Agencies in the Analysis of Education, Edinburgh: Edinburgh University,
Centre of African Studies.
Kiondo, A. (1991), ‘The Nature of Economic Reforms’, in: H. Campbell and H. Stein
(eds), The IMF and Tanzania: The Dynamics of Liberalisation, Harare: Southern Africa
Political Series Trusts.
Kirkpatrick, C.H., N. Lee and F.I. Nixson (1984), Industrial Structure and Policy in Less
Developed Countries, London: Allen and Unwin.
Kirkpatrick, C. and J. Maharaja (1992), ‘The Effect of Trade Liberalization on
Industrial-Sector Productivity Performance in Developing Countries’, in:
J.M. Fontaine (ed.), Foreign Trade Reforms and Development Strategy, London:
Routledge.
Kirugara, D. and N. Nevejan (1996), Identification of Pollution Sources in the Kenyan
Part of the Lake Victoria Catchment Area, Kisumu, Kenya, and Brussels, Belgium:
Kenya–Belgium Joint Project in Freshwater Ecology.
Kisanga, A.U. (1990), ‘The Challenge Faced by the Building Materials Industries in
Developing Countries: with Special Reference to Tanzania’, Habitat International,
14 (4), 119–21.
Kisimbo, A.T. (1994), Product Development and Commercialisation: TEMDO Experience,
Fourth Annual Scientific Seminar, Dar es Salaam, COSTECH.
Koenders, P. (1994), Development of a Travel Planning System at Tanzania Railways
Corporation (TRC) and Effect of Liberalization on TRC, MSc thesis, Eindhoven:
Komba (1999), Structural Change and Competitiveness in Tanzanian Manufacturing
Sector. PhD thesis. (unpublished) Washington University.
Kravis, I., A. Heston and R. Summers (1982), World Product and Income, Baltimore:
Johns Hopkins.
Krueger, A.O. (1992), Economic Policy Reform in Developing Countries, The Kuznets
Memorial Lectures at the Economic Growth Center, Yale University, Oxford:
Blackwell Publishers.
Bibliography 387
Krugman, P. (1987), ‘The Narrow Moving Band, the Dutch Disease, and the
Competitive Consequences of Mrs. Thatcher – Notes on Trade in the presence of
Dynamic Scale Economies’, Journal of Development Economics, 27, 41–55.
Kulindwa, K.A. (1994), Energy Consumption Patterns in the Manufacturing Sector of
Tanzania, PhD thesis, Dar es Salaam: University of Dar es Salaam.
Kumar, N. and M. Saqib (1994), Firm Size, Opportunities for Adaptation, and In-House
R&D Activity in Developing Countries: The Case of Indian Manufacturing, UNU-
INTECH Working Paper no. 13.
Kyessi, A.G. (1995), Overview of the Human Settlements Development and
Management in Tanzania, conference paper presented at the National Workshop on
Launching Habitat II for NGOs, CBO and other actors, Dar es Salaam: CHS,
ARDHI.
Kyessi, A., A.A. Mwaiselage and B. Seleki (1995), Shelter Situation in Tanzania, a paper
presented during the National Workshop on Prefabrication in Housing
Construction, Dar es Salaam: CHS, ARDHI.
Lake Basin Development Authority – Republic of Kenya (1984), Technical Course on
Water Quality Monitoring and Analysis, Kakamega, January, Final Report.
Lake Basin Development Authority – Republic of Kenya (1987), The Study of the
Integrated Regional Development Master Plan for the Lake Basin Development Area, 7
vols.
Lall, S. (1982), Developing Countries as Exporters of Technology: A First Look at the Indian
Experience, London: Macmillan Press – now Palgrave.
Lall, S. (1985), ‘Trade in Technology by a Slowly Industrializing Country: India’, in:
N. Rosenberg and C. Frischtak (eds), International Technology Transfer: Concepts,
Measures and Comparisons, New York: Praeger.
Lall, S. (1987), Learning to Industrialize: The Acquisition of Technological Capability by
India, London: Macmillan Press – now Palgrave.
Lall, S. (1990), Building Industrial Competitiveness in Developing Countries, Paris: OECD.
Lall, S. (1992a), ‘Technological Capabilities and Industrialization’, World
Development, 20 (2).
Lall, S. (1992b), Alternative Development Strategies in sub-Saharan Africa. New York: St.
Martin’s Press – now Palgrave.
Lall, S. (1995), ‘Structural Adjustment and African Industry’, World Development, 23,
2019–31.
Lall, S. (1996), Opening Up, and Shutting Down?: Import Liberalization and Industrial
Technology in sub-Saharan Africa, Study for the United Nations Institute for New
Technologies (INTECH), QEH, Oxford.
Lall, S. (1999), ‘Technological Change and Industrialization in the Asian NIEs:
Achievements and Challenges’, in: A. Szirmai, J. Halman, and B. Verspagen (eds),
Innovation in Theory and Practice, papers of the ECIS Opening Conference,
Eindhoven: Eindhoven University of Technology.
Lall, S. and S.M. Wangwe (1997), ‘Industrial Policy and Industrialization in sub-
Saharan Africa’, A paper prepared for discussion at the AERC Research Workshop,
25 May.
Lall, S. et al. (1997), Import Liberalisation and Technological Capabilities Building in
Tanzania, Kenya and Zimbabwe, Research Report, n.p.
Lapperre, P.E. (1992), Man, Technology, Society and Development, from the Hunters and
Gatherers to the First Industrial Revolution, 3.5 Million Years Ago until AD 1760,
Helmond: Wibro Press.
Lapperre, P.E. (1993), Problems of Developing Countries in Institutional Perspective,
lecture notes, Eindhoven: Eindhoven University of Technology.
388 Bibliography
Lapperre, P.E. (1997), Sociologie en Techniek: Lange Termijn Wisselwerking Tussen

Technologie en Maatschappij, lecture notes, Eindhoven: Eindhoven University of
Technology.
Lapperre, P.E. (1998), Industrialisation of Tanzania: Can Tanzania Learn from European
History?, Paper for the Conference on the Industrial Development of Tanzania,
Eindhoven.
Lapperre, P.E. and J. van der Ploeg (1997a), Technology Audit Methodology for Medium
and Small Scale Manufacturing Enterprises, Working Paper, Eindhoven: Eindhoven
Lapperre, P.E. and J. van der Ploeg (1998), Assessment of Six Manufacturing Enterprises
in Tanzania: Application of a Technology Audit Methodology for Medium and Small
Scale Manufacturing Enterprises, paper presented at the conference on Industrial
Performance in Tanzania, Eindhoven, 25–27 June.
Laseur, W.J.J. (1989), Technologie Investeren Overzee, Groningen.
Laseur, W.J.J. (1991), Managing Technology Transfer, Den Haag: Thieme.
Lebrun, O. and C. Gerry (1975), ‘Petty Producers and Capatalism’, Review of African
Economies, 1 (3), 20–33.
Lee, Kyu Shik and A. Anas (1990), Manufacturers’ Response to Infrastructure Deficiencies
in Nigeria: Private Alternatives and Policy Options, Policy Research and External
Affairs Working Paper, no. 325, Washington, DC: The World Bank.
Lehman, J.T and D.K. Branstrator (1994), ‘Nutrient Dynamics and Turnover Rate of
Phosphate and Sulphate in Lake Victoria, East Africa’, Limnology and Oceanography,
39 (2), 223–7.
Lemmens A.M.C. et al. (1996), Inventory of the Monitoring Activities, Needs and
Possibilities for the Key Environmental Issues in Tanzania, consultancy report for the
Royal Dutch Embassy in Dar es Salaam, Tanzania.
Lemmens A.M.C., H.A. Zanting, J.H.Y. Katima, J.S. Balwira, F.W. Bugenyi and A.G.M.
Scheren (1998a), A Rapid Assessment Methodology for Environmental Pollution
Inventory in Tanzania, Lake Victoria Catchment Area as a Case Study, paper presented
at the Engineering Research Partnership for Sustainable Development Symposium,
Dar es Salaam, Tanzania, July 27–9.
Lemmens A.M.C., H.A. Zanting, D.J.M. van Horen and A.G.M. Scheren (1998b),
Industry and Environment: Methodologies for Environmental Assessment in Data Poor
Situations, Paper presented at the International Conference on the Industrial
Performance of Tanzania, Eindhoven, the Netherlands, June 25–7.
Lensink, R. (1996), Structural Adjustment in sub-Saharan Africa, London: Longman.
Lewis, W.A. (1954), ‘Economic Development with Unlimited Supplies of Labour’, The
Manchester School of Economic and Social Studies, 22, 139–91.
Lieberman, I. (1993), ‘Industrial Restructuring: Policy and Practice’, Policy and
Research Series, 9, Washington: The World Bank.
Liedholm, C. and D. Mead (1987), Small Scale Industries in Developing Countries:
Empirical Evidence and Policy Implications, East Lansing, Michigan State University,
Department of Agricultural Economics, MSU International Development Paper,
no. 9.
Likwelile, S.B. (1997), Private Investment, Macroeconomic Environment and Economic
Growth in Tanzania, 1967–95, Dar es Salaam.
Lipumba, N. (1991), ‘Exchange Rate Policy in Tanzania’, Tanzania Economic Trends, 3,
4.
Lipumba, N.H.I. and A.V.Y. Mbelle (1993), ‘Policy Making for Economic Growth and
Poverty Elimination in a Pluralistic Political System in Tanzania: An Overview’, in:
Bibliography 389
M.S.D. Bagachwa and A.V.Y. Mbelle, Economic Policy under a Multiparty System in
Tanzania, Dar es Salaam: Dar es Salaam University Press.
Little, I.M.D., R.N. Cooper, W.M. Corden and S. Rajapatirana (1993), Boom, Crisis and
Adjustment: The Macroeconomic Experience of Developing Countries, Washington:
World Bank.
Lofchie, M. (1988), ‘Tanzania’s Economic Recovery’, Current History, May edition.
Luvanga, N. (1994), Growth of Small and Medium Size Enterprises under Liberalized
Markets: The Tanzania Experience, paper prepared for the ICGE, June 1994,
Harare.
Lynch, D. (1995), ‘Does Financial Sector Development Matter to Investment?’,
Savings and Development, 1 (19).
Maddison, A. (1995) Monitoring the World Economy, Paris: OECO Development
Centre.
Maddison, A. and B. van Ark (198), Comparisons of Real Output in Manufacturing,
Washington: World Bank, Working Paper, no. 5.
Maddison, A. and B. van Ark (1994), International Comparisons of Real Output and
Productivity, Research Memorandum GD-b. Groningen Growth and Development
Centre.
Mageni, O.S. (1994), The Role of Government on Industrial Research and Development
Institutions, TEMDO/ UNDP Regional Workshop on ‘Roles, Relevance, and
Capabilities of Industrial Research and Development Institutes: Optimising for the
21st Century’, Arusha.
Malecki, E.J. (1991), Technology and Economic Development – The Dynamics of Local,
Regional and National Change, Harlow: Longman Science & Technical.
Maliyamkono, T.L. and M.S.D. Bagachwa (1990), The Second Economy in Tanzania,
London: James Curry.
Maro, W.E. et al. (1991), Construction Cost Trends and Control Mechanisms in Tanzania,
Economic Research Bureau/University of Dar es Salaam.
Masuha, J., B. Mutagahywa and B. Nyichomba (1992), Phase I of Development of
Improved Production Systems (DIPS): the TDP Project, Dar es Salaam: University of Dar
es Salaam, Faculty of Engineering.
Mattikalli, N.M. and K.S. Richards (1996), ‘Estimation of Surface Water Quality
Changes in Response to Land Use Change: Application of the Export Coefficient
Model Using Remote Sensing and Geographical Information System’, Journal of
Environmental Management, 48, 263–82.
Maxwell, P. (1977), Learning and Technical Change in the Steel Plant of Acindar S.A. in
Rosario, Argentina, Research Program Working Paper no. 4, Buenos Aires: Economic
Commission for Latin America.
Mayer, A.F. (1992), ‘Richtzahlen und Ihr Sinnvoller Einsatz in der Brauerei’, II,
Brauwelt, 46, 2427–9.
Mbatia, K.S. (1993), ‘Constraints to the Growth of Non-Traditional Exports:
Incentive Schemes’, Tanzania Economic Trends, 5, 3–4.
Mbelle, A.V.Y. (1988), Foreign Exchange and Industrial Development: A Study of
Tanzania, PhD thesis, Sweden: Gothenburg School of Economics and Commercial
Law.
Mbelle, A.V.Y. (1989), ‘Industrial Performance under the Economic Recovery
Programme in Tanzania’, Tanzania Economic Trends, 2 (1), 7–10.
Mbelle, A.V.Y. (1993), New Technologies and Export Orientation: The Case of
Tanzanian Textile Industries, Department of Economics, University of Dar es
Salaam.
390 Bibliography
Mbelle, A.V.Y. (1994), ‘Sustainable Industrial Development in Tanzania: Can it be

Achieved?’, in: Msambichaka et al. (eds), Development Challenges and Strategies for
Tanzania: An Agenda for the 21st Century, Dar es Salaam University Press, 111–30.
Mbelle, A.V.Y. (1998), Impact of Liberalization of the Foreign Exchange and Financial
Markets in Tanzania, revised draft for the UNU/WIDER project ‘Impact of
Liberalization on Key Markets in sub-Saharan Africa’, presented at the meeting in
Addis Ababa, Ethiopia, 16–20 March.
Mbelle, A.V.Y. and M.S.D. Bagachwa (1995), ‘Technology, Trade and
Industrialization in sub-Saharan Africa: Tanzania Country Study’, in: S. Wangwe
(ed.), Exporting Africa, London and New York: Routledge, 199–245.
Mbelle, A.V.Y. and M.S.D. Semboja (1996), Exchange Rate Calculations, Economic
Research Bureau, University of Dar es Salaam, Mimeo.
Mbelle, A.V.Y. and J. Shitundu (1997), ‘The Management of Industrial Reforms in
Tanzania: Perspectives and New Challenges’, in: D. Bol et al. (eds), Economic
Management in Tanzania, Dar es Salaam: TEMA Publishers Co., 179–97.
Mbelle, A.V.Y. and T. Sterner (1991), ‘Foreign Exchange and Industrial Development:
A Frontier Production Function Analysis of Two Tanzanian Industries’, World
Development, 19 (4), 341–7.
Mbelle, A.V.Y., B.J. Ndulu and J.J. Semboja (1998a), ‘Trade Liberalization in
Tanzania: Episodes and Impacts’ in: A. Oyejide, B.J. Ndulu and J. Gunning (eds),
Integration and Trade Liberalization in sub-Saharan Africa, Vol. 2 Country Studies,
London: MacMillan.
Mbelle, A.V.Y., B.J. Ndulu and J.J. Semboja (1998b), Promoting Non-Traditional Exports
in Tanzania, revised Draft for the UNU/WIDER project ‘Growth, External Sector
and Role of Non-Traditional Exports in sub-Saharan Africa’ presented at the
meeting in Addis Ababa, Ethiopia, 16–20 March.
Meeus, M.T.H. and L.A.G. Oerlemans (1993), ‘Economic Network Research: A
Methodological State of the Art’, in: P. Beije, J. Groenewegen and O. Nuys (eds.),
Networking in Dutch Industries, Leuven: Garant.
Meier, G.M. (1990), ‘Trade Policy and Development’, in: M. Scott and D. Lal (eds),
Public Policy and Economic Development: Essays in Honour of Ian Little, Oxford:
Clarendon Press.
Meier, G.M. and W.F. Steel (1989), Industrial Adjustment in sub-Saharan Africa, New
York: Oxford University Press.
Metcalf and Eddy, Inc. (1991), Wastewater Engineering: Treatment, Disposal and Reuse,
New York: McGraw-Hill.
Mgana, S. and S. Mahongo (1997), Land-Based Sources and Activities Affecting the
Quality and Uses of the Marine, Coastal and Associated Fresh Water Environment,
workshop on the implementation of the Global Programme of Action Zanzibar,
Tanzania, 6–9 Oct.
Michalopoulus, C. (1987), ‘World Bank Programs for Adjustment and Growth’, in: V.
Corbo, M. Goldstein and M. Khan (eds), Growth-Oriented Adjustment Programs,
Washington: World Bank.
Millward, R. and D.M. Parker (1983), ‘Public and Private Enterprise – Comparative
Behaviour and Relative Efficiency’, in: R. Millward, D.M. Parker, et al. (eds), Public
Sector Economics, London.
Ministry of Agriculture, Animal Industry and Fisheries – Republic of Uganda (1992),
Report on Uganda National Census of Agriculture and Livestock (1990–1991).
Ministry of Agriculture, Animal Industry and Fisheries – Republic of Uganda (1995),
Department of Veterinary Services. Annual Report, 1994.
Bibliography 391
Ministry of Education and Culture (1988), Educational Statistics, Dar es Salaam:

Government Printer.
Ministry of Education and Culture (1993), Tanzania’s Educational System, Dar es
Salaam: Government Printer.
Ministry of Education and Culture (1995), Wizara ya Elimu na Utamaduni Takwimu za
Schule za Sekondari Tanzania, 1995, Dar es Salaam: Government Printer.
Ministry of Finance and Economic Planning – Republic of Uganda (1992), The 1991
Population and Housing Census.
Ministry of Finance and Economic Planning – Republic of Uganda (1993), Background
to the Budget, 1993–1994, Kampala.
Ministry of Health and Environmental Protection – Republic of Tanzania (1980),
Handbook of Emission Factors I: Non-Industrial Sources, The Hague: Government
Publishing Office.
Ministry of Industries (1976), Tanzania Industrial Development Strategy, Dar es Salaam:
National Printing Company Ltd.
Ministry of Industries and Trade (1996), Sustainable Industrial Development Policy
(SIDP), 1996–2020, Dar es Salaam: Printpak.
Ministry of Industries and Trade (1999), Speech by the Minister of Industries and
Trade, Hon. Iddi Simba, presenting budget estimates for the year 1999/2000 to
Parliament, June.
Ministry of Natural Resources – Republic of Uganda (1993), Rehabilitation of NWSC
Jinja Water and Sewerage, Final Report Volume 1, Directorate of Water
Development and International Development Association (IDA).
Ministry of Natural Resources – Republic of Uganda (1995), Lake Victoria
Environmental Management Program, Report of National Working Group No. 2 on
Management of Water Quality and Land Use including Wetlands, Kampala.
Ministry of Natural Resources – Republic of Uganda (1996), ‘Forestry Department:
Remote Sensing Data from the National Biomass Study’ (unpublished), PO Box
7270, Kampala, Uganda.
Ministry of Planning and National Development – Republic of Kenya (1993), Kisumu
District Development Plan, 1994–1996.
Ministry of Water Development – Republic of Kenya (1992), The Study on the National
Water Master Plan. Sectoral Report 1: Socio Economy, Japan International
Cooperation Agency (JICA).
Misanga, M.K. (1993), ‘Training Needs for Science and Technology Up to the Year
2000 and Beyond’, Uhandisi Journal, December.
Mitchell, B.R. (1995), International Historical Statistics: Africa, Asia and Oceania,
Oxford: Oxford University Press.
Mjema, G. and J. Shitundu (1995), ‘Performance of the Industrial Trade During the
Economic Recovery Programme and Some Reflections on Future Prospects’, A
paper presented at an ERB Workshop on Beyond SAP in Tanzania, Dar es Salaam.
Mlawa, H.M. and M.S. Sheya (1990), Profiles of R&D Institutions in Tanzania, a study
prepared for the Secretariat of the Future Actions Committee of Management of
Science and Technology for Development, Dar es Salaam, Tanzania.
MLYD (1991), Tanzania the Informal Sector, 1991, Ministry of Labor and Youth
Development, prepared with assistance from UNDP/ILO, URT/91/028, Dar Es
Salaam.
Moavenzadeh, F. (1987), ‘The Construction Industry’, in: L. Rodwin (ed.), Shelter,
Settlement and Development, Boston: Allen & Unwin.
392 Bibliography
Mongi, J.F.K. (1980), ‘The Development of Price Control in Tanzania’, in: K.S.P.
Rwegasira and L.A. Kanneworff (eds), Inflation in Tanzania, Dar es Salaam: Institute
of Finance Management.
More, C. (1997), The Industrial Age: Economy and Society in Britain, 1750–1995,
London: Longman.
Moshi, H.P.B. (1996), Fiscal Impact of Commercialization and Privatization of Public
Enterprises, Dar es Salaam: Mimeo.
Moshi, H.P.B. and A.A. Kilindo (1995), The Impact of Government Policy on
Macroeconomic Variables: The Case of Private Investment in Tanzania, Nairobi.
Mourik, C. van, E.M. van der Ouderaa and C.G. Hage (1991), Management van
Research en Development, Deventer: Kluwer.
Mukandala, R. (1985), ‘Trends in Civil Service Size and Income in Tanzania,
1967–1982’, Canadian Journal of African Studies, 2, 253–62.
Mukandala, R. (1988), The Political Economy of Parastatal Enterprise in Tanzania and
Botswana, unpublished doctoral dissertation, Berkeley: University of California.
Mumburi, U.P. (1998), Manufacturing Output Growth in Tanzania: An Empirical
Analysis, unpublished MA thesis, University of Dar es Salaam.
Musonda, M. (1992), Development Strategy and Manufactured Exports in Tanzania, PhD
thesis, Sweden: Department of Economics, University of Lund.
Musonda, M. and J.P. Kweka (1998), The Dynamics and Potential of Cluster Firms in the
Growth and Development of MSEs in Tanzania: The Case of Three Selected Clusters in
Tanzania, Dar es Salaam ESRF, Mimeo.
Mutagahywa, B. and P.N. Materu (1992), Academic–Industry Links, Development of a
Vital Partnership, Faculty of Engineering, University of Dar es Salaam.
Mwaiselage, A.A. (1992), Development of Local Construction Capacity in Building Works
through Small Scale Contractors in Tanzania, Dar es Salaam: Centre for Housing
Studies (CHS), Ardhi Institute.
Mwandosya, M. and M. Luhanga (1993), ‘Energy and Development in Tanzania,
Issues and Perspectives’, Energy Policy, May, 441–53.
MWEM (1992), The Energy Policy of Tanzania, Dar es Salaam: Ministry of Water,
Energy and Minerals.
MWEM (1995), Energy Balance of Tanzania in 1992, print-out, Dar es Salaam:
Ministry of Water, Energy and Minerals.
Mwihava, N. and A. Towo (1994), A Study and Assessment of Energy Projects and their
Effective Utilization in Tanzania, Dar es Salaam: COSTECH.
National Water and Sewerage Corporation – Republic of Uganda (1995), Statistics on
Sanitation in NWSC 9 Towns (unpublished), PO Box 7053, Kampala, Uganda.
NBAQS and BC (1992), List of Registered Architects, Quantity Surveyors and Building
Contractors Dec 1992, Dar Es Salaam.
NBS (National Bureau of Statistics), Statistical Publications, various issues, Dar es
Salaam.
NCC, Komba, S.C. (1992a), Identification & Promotion of Utilisation of Local Resources
and Locally Produced Materials for Construction Industry in Tanzania, Final report,
phase I, Dar es Salaam: National Construction Council.
NCC National Construction Council (1992b), Research and Development Issues in the
Construction Industry in Tanzania, Dar Es Salaam.
NCC National Construction Council (1995), Proposed National Construction Industry
development Policy, Dar es Salaam: National Construction Council, Ministry of
Works.
ND I (July 1981–May 1984), Natural Dyes Project File I, collection of approximately
100 notes, reports and documents, TIRDO.
Bibliography 393
ND II (May 1984–November 1990), Natural dyes project file II, collection of approx-
imately 70 notes, reports and documents, TIRDO.
Ndulu, B.J. (1993), Tanzania’s Economic Development: Lessons from the Experience
and Challenges for the Future, paper written for the International Conference on
Development Challenges and Strategies for Tanzania: An Agenda for the 21st
Century Dar es Salaam.
Ndulu, B.J. and J.J. Semboja (1994), ‘Trade and Industrialization in Tanzania: A
Review of Experiences and Issues’, in: G.K. Helleiner (ed.), Trade Policy and
Industrialization in Turbulent Times, London: Routledge.
Ndulu, B.J. and J.J. Semboja (1996), ‘The Development of Manufacturing for Export
in Tanzania’ in: G. Helleiner (ed.), Manufacturing for Export in the Developing World:
Problems and Possibilities, London and New York: Routledge.
Ndulu, B.J., W. Lyakurw, J.J. Semboja and A. Chaligha (1987), Import Tariff Study –
Tanzania, Dar es Salaam: unpublished report to the Ministry of Finance.
Nelson, R.R. (1990), ‘On Technological Capabilities and Their Acquisition’, in:
R.E. Evenson and G. Ranis, Science and Technology: Lessons for Development
Policy, Boulder: Westview Press.
Nelson, R.R. (1993), National Innovation Systems: A Comparative Analysis, Oxford:
Oxford University Press.
Nelson, R.R. and E. Phelps (1966), ‘Investment in Humans, Technological Diffusion,
and Economic Growth’, American Economic Review, 56, May 69–75.
Nelson, R.R. and S.C. Winter (1982), An Evolutionary Theory of Economic Change,
Cambridge MA Belknap Press of Harvard University Press.
Newberry, D.P. and N. Stern (1987), The Theory of Taxation for Developing Countries,
Oxford: Oxford University Press.
Niosi, J. and J. Rivard (1988), ‘Canadian Technology Transfer to Developing
Countries through Small and Medium-Size Enterprises’, World Development,
18, 11.
Niskanen, W. (1973), Bureaucracy: Servant or Master?, London: Institute of Economic
Affairs.
Njau, G.J. et al. (1997), Environmental Impact Assessment on the Proposed Thermal Power
Station in Tageta Area, Dar es Salaam, Prepared for Independent Power Tanzania
(IPTL) Ltd, May.
Nyerere, J. (1968a), Freedom and Socialism, London: Oxford University Press.
Nyerere, J. (1968b), Ujamaa: Essays on Socialism, Dar es Salaam: Oxford University
Press.
Ochumba, P.B.O. (1993), Measurement of Water Currents, Temperature, Dissolved
Oxygen and Winds in the Kenyan Lake Victoria, Contribution to the IDEAL
Symposium on the Limnology, Climatology and Palaeoclimatology of the East
African Lakes, Jinja, Uganda, 18–23 February.
Ochumba, P.B.O. and D.I. Kibaara (1989), ‘Observations on Blue-green Algal Blooms
in the Open Waters of Lake Victoria’, African Journal of Ecology, 27, 23–34.
OECD (February 1971), Latest Information on National Accounts of Less Developed
Countries, F/8, No. 5, Paris: OECD.
Oerlemans, A.M.H.M. (1985), Research on the Production of Sugar Cane in Kenya,
MSc thesis, Eindhoven: Technology and Development Studies, Eindhoven
Ogaram, D.A. and W.S. Kalema (1995), Lake Victoria Environmental Management
Program, Final report on regional task No. 15: Critical Sources of Industrial and
Municipal Pollution.
394 Bibliography
Orlova, A.P. and L.V. Yaroshenko (1976), ‘Contents of Pesticide and Mineral
Fertilizers in Runoff Water of Uzbekistan Irrigation’, Formirovanie Kontrol Kachestva
Poverkhnostnykh Vod, 3, 106–9.
Pack, H. and L.E. Westphal (1986), ‘Industrial Strategy and Technological Change:
Theory Versus Reality’, Journal of Development Economics, 22.
Pack, H. (1987), Productivity, Technology, and Industrial Development, New York:
Oxford University Press.
Pack, H. (1993), ‘Productivity and Industrial Development in sub-Saharan Africa’,
World Development, January, 21, 1–16.
Pack, H. and C. Paxson (1998), Is African Manufacturing Skill Constrained?, paper
presented at the conference on The Industrial Performance of Tanzania,
Eindhoven, 25–27 June.
Page Jr., J.M. (1990), ‘The Pursuit of Industrial Growth: Policy Initiatives and
Economic Consequences’, in: M. Scott and D. Lal (eds), Public Policy and Economic
Development: Essays in Honour of Ian Little, Oxford: Clarendon Press.
Parker, R.L., R. Riopelle and W.F. Steel (1995), Small Enterprises Adjusting to
Liberalization in Five Countries, World Bank Discussion Papers, no. 271,
Washington, DC: The World Bank.
Pasinetti, L.L. (1981), Structural Change and Economic Growth, Cambridge: Cambridge
University Press.
Pasinetti, L.L. and R.M. Solow (eds) (1994), Economic Growth and the Structure of Long
Term Development, proceedings of the IEA Conference held in Varenna, Italy,
Basingstoke and London: St Martin’s Press – now Palgrave – in association with the
International Economic Association.
Perkins, F. (1980), Technological Choice, Industrialization and Development: the Case
of Tanzania, unpublished doctoral dissertation, University of Sussex, England.
Perkins, F. (1983), ‘Technology Choice, Industrialization and Development
Experiences in Tanzania’, Journal of Development Studies, 19 (2), 213–43.
Pieper, U. (1998), Openness and Structural Dynamics of Productivity and Employment in
Developing Countries: A Case of De-Industrialization?, Geneva: ILO.
Pieterson, M. (1987), Het Technisch Labyrint: een Maatschappijgeschiedenis van drie
Industriele Revoluties, Meppel: Boom.
Planning Commission & Ministry of Labour and Youth Development (1991),
National Informal Sector Survey, Dar es Salaam.
Planning Commission & Ministry of Labour and Youth Development (1995), The
Dar es Salaam Informal Sector Survey 1994/5 [only estimates for manufacturing (in
Excel Worksheet format) at our disposal], Dar es Salaam.
Porter, M.E. (1985), Competitive Advantage: Creating and Sustaining Superior
Performance, London: Collier Macmillan.
Porter, M.E. (1986), Competition in Global Industries, Boston: Harvard Business School
Press.
Porter, M.E. (1990), The Competitive Advantage of Nations, New York: Macmillan –
now Palgrave.
Prins, I.M. (1997), Manufacturing Statistics, Reconstructing Tanzanian Manufacturing
Value Added, 1995–1995, MSc thesis, Eindhoven: Technology and Development
Studies, Eindhoven University of Technology.
Prins, I.M. and A. Szirmai (1998), A Reconstruction of GDP and Employment in
Tanzanian Manufacturing, 1961–1995, Report to the Tanzanian Bureau of Statistics,
Eindhoven, December.
Psacharopoulos, G. (1991), ‘Vocational Education in Theory: Why Does Vocational
Education Fail?’, International Journal of Educational Development, 11, 3.
Bibliography 395
Ranis, G. (1984), ‘Determinants and Consequences of Indigenous Technological

Activity’, in: M. Fransman and K. King (eds), Technological Capability in the Third
World, London Macmillan Press.
Ranis, G. (1988), ‘Analytics of Development: Dualism’, in: H.Chenery and
T.N. Srinivasan (eds), Handbook of Development Economics, Amsterdam North
Holland Publishers.
Reckhow, K.H. and J.T. Simpson (1980), ‘A Procedure Using Modelling and Error
Analysis for Prediction of Lake Phosphorus Concentration from Land Use
Information’, Canadian Journal of Fisheries and Aquatic Sciences, 37, 1439–48.
Ren, R. (1997), China’s Economic Performance in an International Perspective, Paris:
OECD.
Richardson, C.J. and D.S. Nichols (1985), ‘Ecological Analysis of Wastewater
Management Criteria in Wetland Ecosystems in the USA’, The Utrecht Plant Ecology
News Report, Dec. 186–219.
Riedijk, W. (ed.) (1987), Appropriate Technology for Developing Countries, Delft: Delft
University Press.
Righart, H. (1991), (ed.) De Trage Revolutie: over de Wording van Industriele
Samenlevingen, Meppel: Boom.
Ritter, W.F. (1988), ‘Reducing Impacts of Nonpoint Source Pollution from
Agriculture: A Review’, Journal of Environmental Science and Health, A23 (7), 645–67.
Rodriguez, E. (1998), Costa Rica: Policies and Conditions for Export Diversification,
revised draft for the UNU/WIDER project ‘Growth, External Sector and Role of
Non-Traditional Exports in sub-Saharan Africa’, presented at the meeting in Addis
Ababa, Ethiopia, 16–20 March, Addis Ababa.
Rogers, E.M. (1983), Diffusion of Innovations, 3rd edn, New York: The Free Press.
Romano, C.A. (1990), ‘Identifying Factors which Influence Product Innovation: A
Case Study Approach’, Journal of Management Studies, 27 (1), 75–95.
Romein, J. (1954), Aera van Europa: De Europese Geschiedenis als Afwijking van het
Algemeen Menselijk Patroon, Leiden: Brill.
Romer, P. (1986), ‘Increasing Returns and Long-Run Growth’, Journal of Political
Economy, 94 (5), 1002–37.
Romijn, H.A. (1996), Acquisition of Technological Capability in Small Firms in
Developing Countries, Tilburg: Katholieke Universiteit Brabant.
Romijn, H.A. (1999), Acquisition of Technological Capability in Small Firms in
Developing Countries, Basingstoke: Macmillan Press – now Palgrave.
Romiszowski, A.J. (1984), Designing Instructional Systems, London: Kogan Page.
Rosberg, C. and T. Callaghy (1979), Socialism in sub-Saharan Africa, Berkeley:
University of California, Institute of International Studies.
Rosenberg, N. (1986), Inside the Black Box: Technology and Economics, Cambridge, MA.
Cambridge University Press.
Rosenberg, N. (1990), ‘Science and Technology Policy for the Asian NIC’s: Lessons
from Economic History’, in: R.E. Evenson and G. Ranis (eds), Science and
Technology: Lessons for Development Policy, Boulder Westview.
Rosenberg, N. and L.E. Birdzell (1986), How the West Grew Rich: the Economic
Transformation of the Industrial World, New York: Basic Books, Inc., Publishers.
Rosenzweig, M.R. (1995), ‘Why Are There Returns to Schooling?’, American Economic
Review, Papers and Proceedings, May.
Rostow, W.W. (1960), The Stages of Economic Growth, Cambridge. Cambridge
University Press.
Rothwell (1977), ‘Characteristics of Successful Innovation and Technically
Progressive Firms’, R & D Management, 7.
396 Bibliography
Rothwell, R. (1992), ‘Successful Industrial Innovation: Critical Factors for the 1990s’,
R & D Management, 22 (3), 221–31.
Rothwell, R. and W. Zegveld (1982), Innovation and the Small and Medium Sized Firm,
London: Frances Pinter.
Rothwell, R. and W. Zegveld (1985), Re-industrialization and Technology, Longman:
Harlow.
RPED (1994), Development and Growth of Industrial Enterprise in Tanzania, Center for
International Business Research, Helsinki School of Economics, Helsinki, Finland.
RPED (1995), Dynamics of Enterprise Development in Tanzania, Final Report on Round II
Survey Data, Helsinki School of Economics.
Runciman, S. (1994), A History of the Crusades, London: The Folio Society.
Ruttan, V.W. (1998), ‘The New Growth Theory and Development Economics: A
Survey’, Journal of Development Studies, 35, (2), December.
Rweyemamu, J.F. (1973), Underdevelopment and Industrialisation in Tanzania: A Study
of Perverse Capitalist Industrial Development, London: Oxford University Press.
Sagasti, F. (1979), ‘Towards Endogenous Science and Technology for another
Development’, Development Dialogue, No 1.
Samplonius, C. (1994), Energy Efficiency in the Tanzanian Industry: The Cement Industry
as a Case Study, Eindhoven University of Technology, the Netherlands.
Santschi, P.W. and P.W. Schindler (1977), ‘Chemical and Geochemical Studies of
Lake Biel’, Schweizer Zeitschrift für Hydrologie, 39, (2), 182–200.
Scheren, P.A.G.M. (1995), A Systematic Approach to Lake Water Pollution Assessment:
Water Pollution in Lake Victoria, East Africa, Eindhoven University of Technology.
Scheren, P.A.G.M., J.C. Bosboom, K. Njau and A.M.C. Lemmens (1995), ‘Assessment
of Water Pollution in the Catchment Area of Lake Victoria, Tanzania’, Journal of
Eastern African Research and Development, 25, 129–43.
Schulte, W. (1996), Technology and Quality Care in Muhimbili Medical Center, Tanzania:
Methodology to Study the Role of Technology from the View Point of Total
Quality, MSc thesis, Eindhoven: Technology and Development Studies, Eindhoven
Schultz, T.W. (1975), ‘The Value of the Ability to Deal with Disequilibria’, Journal of
Economic Literature, Sept., 13, 827–46.
Schumpeter, J. (1912), Theorie der Wirtschaftlichen Entwicklung, published in English
as Theory of Economic Development (1961 edition by Redvers Opie), Oxford: Oxford
University Press.
Semboja, H.H. (1995), An Overview of the Industrial Sector and Selected Industries’
Performance During the 1980s–1990s, unpublished ESRF Technical Paper, Dar es
Salaam: ESRF.
Semboja, H.H. et al. (1996), Studies Comparing Small-Scale and Large-Scale Production in
LDCs.
Semboja, J.J. and I. Karamagi (1991), The Import Duty Tariff Study, a report prepared
for the Presidential Commission of Enquiry into Public Revenue and Taxation,
October 1991, Dar es Salaam.
Semboja, H.H. and J.P.Kweka (1997), Garment Industry, Dar es Salaam: Economic and
Social Research Foundation.
Semboja, H.H. and J.P. Kweka (1998), An Analysis of the Garment Sector, paper pre-
sented at the conference on Industrial Performance in Tanzania, Eindhoven, 25–27
June.
Semboja, H. and J. Kweka (forthcoming), Import Liberalization, Industrialization and
Technological Capability in sub-Saharan Africa: The Case of the Garment Industry in
Tanzania, ESRF Research Paper, Dar es Salaam.
Bibliography 397
Semboja, H.H., J. Kweka and G. Kahyarara (1997a), Firms’ Linkages in the

Development of Industry: A Theoretical Note with Policy Perspectives, a
paper presented at the International Conference on Entrepreneurship and
Small Business Development in Tanzania, 21–22 October, White Sands, Dar es
Salaam.
Semboja, H., J.P. Kweka, S.M. Wangwe and H.B.P. Moshi (1997b), Multi-Country
Comparative Study of Private Enterprise Development (MCC-PED): A Tanzania Country
Report, Dar es Salaam: ESRF.
Sen, A.K (1994). ‘Growth Economics: What and Why?’, in: L.L. Pasinetti and
R.M. Solow (eds), op. cit.
Sethuraman, S.V. (ed.) (1981), The Urban Informal Sector in Developing Countries:
Employment, Poverty and Environment, Geneva: ILO.
Shao, J.R. (1978), Some Features of Tanzanian Manufacturing Industry, MA thesis,
University of Dar es Salaam.
Shirley, M. (1989), ‘The Reform of State-Owned Enterprises: Lessons from World
Bank Lending’, World Bank PPR Series, 4, Washington DC.
Shitundu, J.L.M. and G.D. Mjema (1995), ‘Industrial Trade during the Economic
Recovery Programme and Some Reflections on Future Prospects’, in: L.H.P.B.
Msambichaka et al. (eds), Beyond Structural Adjustment Programmes in Tanzania:
Success, Failures and New Perspectives, Dar es Salaam.
Shivji, I.G. (1975), Class Struggles in Tanzania, Dar es Salaam: Tanzania Publishing
House.
Shu, G.F. (1990), ‘Betriebvergleich Energie 1990’, Brauwelt, nr. 9, 1992, p 361.
SIDO (1984), Summary of the Fourth Five Year Plan for Development of SIDO
1983/84–1987/88, Dar-Es-Salaam: SIDO.
Silver, M.S. (1984), The Growth of Manufacturing Industry in Tanzania, London:
Westview Press.
Sinda, S.H. (1991), Science and Technology Policies: Their Implementation with regard
to Science and Technology Initiatives in Tanzania, in: Proceedings of the International
Workshop on the South Commission Recommendations on S&T, COSTECH, Dar-
es Salaam.
Skarstein, R. and S.M. Wangwe (1986), Industrial Development in Tanzania: Some
Critical Issues, Dar es Salaam: Tanzanian Publishing House.
Sonali, D. and H.H. Semboja (1995), Import Liberalization, Industrialization and
Technological Capability in Tanzania, Oxford UNU-INTECH Research Project.
Souder, W.E. and A.K. Chakrabarti (1980), ‘Managing the Co-ordination of
Marketing and R&D in the Innovation Process’, TIMS Studies in the Management
Sciences, 15, 135–50.
Souza, P.R. and V.E. Tokman (1976), ‘The Urban Informal Sector in Latin America’,
International Labour Review, 114 (3), Nov.–Dec.
Stein, H. (1991), ‘Economic Policy and the IMF in Tanzania: Conditionality, Conflict
and Convergence’, in: H. Campbell and H. Stein (eds), The IMF and Tanzania: The
Dynamics of Liberalisation, Harare: Southern Africa Political Economy Series Trusts.
Steinmueller, E. and M-I. Bastos (1995), Information and Communication Technologies:
Growth, Competitiveness, and Policy for Developing Nations, Background Paper no. 2,
UNIDO Global Forum on Industry.
Stewart, F. (1977), Technology and Underdevelopment, London: Macmillan Press – now
Palgrave.
Stewart, F. (1978), ‘Technological Self Reliance of the Developing Countries: Towards
Operational Strategies’, in: UNIDO, Development and Transfer of Technology Series,
New York: UNIDO.
398 Bibliography
Stewart, F. (1984), ‘Facilitating Indigenous Technical Change in Third World

Countries’, in: M. Fransman and K. King (eds), Technological Capability in the Third
World, London: Macmillan Press – now Palgrave – 81–94.
Stewart, F. and J. James (eds) (1982), The Economics of New Technology in Developing
Countries, London: Frances Pinter.
Stewart, F., S. Lall and S.M. Wangwe (eds) (1992), Alternative Development Strategies in
African Development, Macmillan Press – now Palgrave.
Straaten, J. van der, et al. (1992), Technology Transfer and Development: A Contribution
to Policy Making, Amsterdam: Tool.
Syrimis, S. (1988), Sector Review: The Financing of Education in Tanzania, Paris:
UNESCO Educational Financing Commission.
Szirmai, A. (1997a), Economic and Social Development: Trends, Problems and Policies,
London: Prentice Hall.
Szirmai, A. (1997b), Ontwikkelingslanden in de Industriële Wedloop [Developing
Countries in the Industrial Race], inaugural lecture, Eindhoven University of
Technology.
Szirmai, A. and D. Pilat (1990), ‘Comparisons of Purchasing Power, Real Output and
Labour Productivity in Manufacturing in Japan, South Korea and the USA,
1975–1985’, Review of Income and Wealth, 26 (1), March, 1–31.
Szirmai, A. and R. Ren (1998), ‘Chinese Manufacturing in Comparative Perspective’,
in: M. Fouquin and F. Lemoine (eds), The Chinese Economy, London: Economica,
49–64.
Szirmai, A. and W. Schulte (1998), Comparative Analysis of Real Output and Labour
Productivity in Tanzanian Manufacturing: A Level Comparison between Tanzania
and the USA in 1989, paper presented at the conference on Industrial Performance
in Tanzania, Eindhoven, 25–27 June.
Tanzania Economic Trends (1990), ‘National Investment Promotion Policy’,
Tanzania Economic Trends, 3, 1–2.
Tanzania Economic Trends (1994). Statistical Appendix: Monthly Average Exchange
Rates in Tanzania 1989–1993, in: Tanzania Economic Trends, 7, 1201–2.
Tarimo, J. (1995), Experiences and Potential for Energy Conservation in Tanzania, a
Case Study of TIRDO, paper presented at Demand Side Management Seminar, Dar
es Salaam, 9 March.
Teitel, S. (1981), ‘Toward an Understanding of Technical Change in Semi-
Industrialized Countries’, in: J.M. Katz (ed.), Technology Generation in Latin
American Manufacturing Industries, London: Macmillan Press – now Palgrave.
TEXTMAT (1994a), The Impact of Practices of Fiscal Policies in Tanzania, Dar es Salaam:
Mimeo.
TEXTMAT (1994b), The Fiscal Impact of Trade Policy on Textile Industry in Tanzania,
Dar es Salaam.
Tezler, R. and Molenaar (1989), ‘In Search of New Approaches: Major Areas of
Attention’, Operations Review Unit, Small Enterprises, New Approaches: Proceedings
of the Workshop ‘Small Scale Enterprise Development, in Search of New Dutch
Approaches’, The Hague: Ministry of Foreign Affairs, Directorate General for
International Cooperation.
Thirsk, J. (1982), ‘The Rural Economy: Markets, Trade and Transport’, in: J. Blum
(ed.), Our Forgotten Past: Seven Centuries of Life on the Land, London: Thames and
Hudson.
Thomann, R.V. and J.A. Mueller (1987), Principles of Surface Water Quality Modelling
and Control, New York: Harper & Row.
Bibliography 399
Thomas, C.Y. (1974), Dependence and Transformations, New York: Monthly Review
Press.
Timmer, M.P (1996), On the Reliability of Unit Value Ratios in International
Comparisons, Groningen: Groningen Growth and Development Centre, Research
Memorandum GD 31.
Timmer, M.P. and A. Szirmai (1999), ‘Comparative Productivity Performance in
Manufacturing in South and East Asia’, Oxford Development Studies, 27, (1), 57–79.
TIRDO (1989), Ten Years of the Tanzania Industrial Research and Development
Organisation (TIRDO) 1979–1989, TIRDO.
TIRDO (1990), Semi-Detailed Energy Audit at Tanzania Breweries’ Dar es Salaam Plant,
TIRDO and AF-Energi Konsult.
TIRDO (1993a), Operational Guidelines and Procedures for the Extension Services –
Manual, TIRDO.
TIRDO (1993b), Draft TIRDO Five-Year Plan 1993/94–1997/98, TIRDO.
TIRDO (1993c), Incentive Scheme Package – 1992/93, TIRDO.
TIRDO (1995), Staff List as for 1-10-1995, TIRDO.
Todaro, M.P. (1997), Economic Development in the Third World, New York: Longman.
Toussaint, T.J.A. (1994), Energy Conservation in the Industry of Tanzania: Cost Reduction
by Energy Conservation at Tanzania Breweries Ltd, Eindhoven University of
Technology, the Netherlands.
TPC, Meetings of the Technical Programme Committee, vols 1–22, TIRDO.
TSG (1996), Roadmap Study of Four African Countries: Ghana, Namibia, Uganda and
Tanzania, Washington, DC: The Service Group.
UNESCO (1989), Statistical Yearbook 1989, Lanham: UNESCO Publishing & Bernan
Press.
Press.
Press.
UNIDO (1980), The Provision, Storage and Handling of Dyes and Chemicals for Textile
Dyeing, Printing and Finishing, UNIDO.
UNIDO (1982), Assistance to the Tanzania Industrial Research and Development
Organisation, UNIDO.
UNIDO (1988a), Industrial Restructuring Policies in Developing Asian Countries with
Particular Attention to the Textile and Garment Industry, Vienna: UNIDO, Limited
Distribution.
UNIDO (1988b), Regenerating African Manufacturing Industry: Approach and Programme,
Studies of the Rehabilitation of African Industry no. 1, PPD 101-29, December.
UNIDO (1990), Rehabilitation of African Industry, no. 12, PPD 168, 10 July.
UNIDO (1991), Manual for the Preparation of Industrial Feasibility Studies, Vienna:
UNIDO.
UNIDO (1993), Industry and Development Global Report, 1992/93, Vienna: UNIDO.
UNIDO (1995), New Materials and the Challenge for Developing Countries, Background
Paper no. 3, UNIDO Global Forum on Industry.
UNIDO (various), Industry and Development Global Report, Geneva.
United Nations (1961), Index Numbers of Industrial Production, Studies in Methods,
series F, no. 1, New York: United Nations.
United Nations (1968a), A System of National Accounts, Studies in Methods, series F,
no. 2, rev. 3, New York: United Nations.
United Nations (1968b), International Standard Industrial Classification of All Economic
Activities, Statistical Papers, series M, no. 4, rev. 2, New York.
400 Bibliography
United Nations (1979), National Accounting Practices in Seventy Countries, Studies in

Methods, series F, no. 26, vol. 3, New York: United Nations.
United Nations (1990), International Standard Industrial Classification of all Economic
Activities, 3rd revision, Statistical Papers, series M, no. 4, rev. 3, New York: United
Nations.
United Nations (1993), System of National Accounts, Studies in Methods, series F,
no. 2, rev. 4, Brussels/Luxembourg: Office for Official Publications of the European
Communities.
United Nations (1995), Statistical Yearbook,-40th issue, New York.
United Nations Conference on Trade and Development (1972), Guidelines for the
Study of Transfer of Technology to Developing Countries, New York: United Nations.
United Nations Conference on Trade and Development (1978), Handbook on the
Acquisition of Technology by Developing Countries, New York: United Nations.
United Nations Department of Economic Affairs (1951), Measures for the Economic
Development of Underdeveloped Countries, New York: United Nations Department of
Economic Affairs.
United Nations Development Programme, Development Cooperation Report Tanzania,
various issues.
United Republic of Tanzania (1969), Second Five Year Plan for Social and Economic
Development (1969–1974): Volume I: General Analysis, Dar es Salaam: Government
Printer.
United Republic of Tanzania (1972), The Economic Survey, 1971–1972, Dar es Salaam:
Government Printer.
United Republic of Tanzania (1976), Third Five Year Plan for Economic and Social
Development (1976–1981): Volume I: General Perspective, Dar es Salaam: Government
Printer.
United Republic of Tanzania (1977), The Economic Survey, 1975–1976, Dar es Salaam:
Government Printer.
United Republic of Tanzania (1980), The Regulation of Trade Act 1980: no. 18 of 1980,
Dar es Salaam: Tanzanian Gazette Supplement.
United Republic of Tanzania (1982a), Structural Adjustment Programme for Tanzania,
Dar es Salaam: Government Printer.
United Republic of Tanzania (1982b), The National Economic Survival Programme, Dar
es Salaam: Government Printer.
United Republic of Tanzania (1982c), The Economic Survey, 1981, Dar es Salaam:
Government Printer.
United Republic of Tanzania (1983), Speech by the Hon. Lt. Col. A.S. Muncho (MP):
Minister for Trade, for the session of the national assembly for 1982/3, Dar es Salaam:
Government Printer.
United Republic of Tanzania (1985a), The National Science and Technology for
Tanzania, Ministry of Planning and Economic Affairs, Dar es Salaam: United Republic
of Tanzania.
United Republic of Tanzania (1985b), Hali ya Uchumi wa Taifa katika Mwaka, 1984
[The Economic Survey 1984], Dar es Salaam: Government Printer.
United Republic of Tanzania (1986a), Programme for Economic Recovery, Dar es Salaam:
Government Printer.
United Republic of Tanzania (1986b), Hali ya Uchumi wa Taifa katika Mwaka, 1985
[The Economic Survey 1985], Dar es Salaam: Government Printer.
United Republic of Tanzania (1989), The Economic and Social Action Programme, Dar es
Salaam: Government Printer.
Bibliography 401
United Republic of Tanzania (1990a), Industrial Rehabilitation Study of the Textile

Subsector of Tanzania, Dar es Salaam: Ministry of Industries and Trade, Government
Printers.
United Republic of Tanzania (1990b), National Investment (Protection and Promotion)
Act: No. 10 of 1990, Dar es Salaam: The President’s Office – Planning Commission.
United Republic of Tanzania (1992a), National Investment (Protection and Promotion)
(Amendment) Act: No. 10 of 1992, Dar es Salaam: Tanzanian Gazette Supplement.
United Republic of Tanzania (1992b), Public Corporations Act: No. 2 of 1992, Dar es
Salaam: Tanzanian Gazette Supplement.
United Republic of Tanzania (1994), Fair Trade Practices Act: No. 4 of 1994, Dar es
Salaam: Tanzanian Gazette Supplement.
United Republic of Tanzania (1995a), Parastatal Sector Reform Commission: 1994
Review and Action Plan for 1995 and Beyond, Dar es Salaam: Parastatal Sector Reform
Commission.
United Republic of Tanzania (1995b), Economic Survey, 1994, Dar es Salaam: Bureau
of Statistics, Government Printer.
United Republic of Tanzania (1995c). Revised National Accounts of Tanzania:
1976–1990, Dar es Salaam: Bureau of Statistics.
United Republic of Tanzania (1996a), Sustainable Industrial Development Policy (SIDP)
(1996–2020), Dar es Salaam: Ministry of Industries and Trade.
United Republic of Tanzania (1996b), Industrial Development Policy (1995–2020):
Revised Zero Draft, Dar es Salaam, Technical Committee on Formulation of
Industrial Development Strategy and Policy
United Republic of Tanzania (1997), Mkakati wa Taifa wa Kusimamia Utekelezaji wa
Sera za Kuondoa Umaskini [National Strategy for Overseeing the Implementation of
Poverty Eradication in Tanzania], Office of the Vice-President.
United Republic of Tanzania (1998), The Development Vision 2025 for Tanzania,
Planning Commission (Coordinating).
United Republic of Tanzania (1999a), Economic Survey, 1998, Dar es Salaam: Planning
Commission, June.
United Republic of Tanzania (1999b), Speech by the Minister of Finance for the
1999/2000 Budget, Dar es Salaam.
United Republic of Tanzania (1999c), Speech by the Minister of Industries and Trade,
for 1999/2000 budget, Dar es Salaam, June.
United Republic of Tanzania (Various), Economic Surveys, Dar es Salaam: Planning
Commission and Bureau of Statistics.
United Republic of Tanzania (Various), Analysis of Parastatal Accounts.
UNU-INTECH, QEH-ESRF (1996), The Impact of Import Liberalisation on the
Manufacturing Sector; see also Semboja and Kweka, forthcoming.
US Department of Commerce (1986), National Income and Product Accounts of the
United States, 1929–82, Washington, DC.
US Department of Commerce (1990a), Bureau of the Census, 1989 Annual Survey of
Manufactures, Statistics for Industry Groups and Industries, M90(AS)-1, Washington,
DC.
US Department of Commerce (1990b), Bureau of the Census, 1987 Census of
Manufactures, General Summary and Industry Series.
US Department of Commerce (1992), National Income and Product Accounts, 1959–88,
vol. 2, Washington, DC, September.
US Department of Commerce, (Various) Survey of Current Business.
USAID/TANZANIA (1996), Environmental Threat Assessment, May.
402 Bibliography
USAID-CTI (1994), Industrial Policy Proposal, a report based on Joint USAID and
Confederation of Tanzania Industries Programme, Dar es Salaam.
Valk, P. de (1992), A General Framework for Evaluating the Performance of Textile
Enterprises in LDCs: With Application to Tanzania under Structural Adjustment, The
Hague: Institute of Social Studies.
Valk, P. de and A.V.Y. Mbelle (1990), Textile Industry under the Structural Adjustment
Programme in Tanzania, 1980–1988, Dar es Salaam.
Ven, Y. van der (1996), Privatization at Tanzania Harbors Authority, MSc thesis,
Eindhoven: Technology and Development Studies, Eindhoven University of
Technology.
Verlinden, P. (1996), Rwanda & Burundi, Country Series Amsterdam, The
Netherlands: Royal Tropical Institute (KIT).
Vighi, M. and G. Chiaudani (1987), ‘Eutrophication in Europe: the Role of
Agricultural Activities’, in: E. Hodgeson (ed.), Reviews in Environmental Toxicology,
3, 213–57.
Vleuten, F. van der (1995), Energy and Environment in the Global Cement Industry:
Strategies and Instruments to Promote Energy Efficiency in Developing Countries,
Working Paper 4, Report ECN-C–94-035, Petten: ECN.
Vleuten, F. van der and M. Yhdego (1996), Cleaner Production Technology for
Greenhouse Gas Reduction in Tanzanian Private Industry, Preinvestment Study, Dar es
Salaam: ERC.
Vocational Education and Training Authority (1994), unpublished internal docu-
ments, Dar es Salaam: VETA.
Vocational Education and Training Authority (1995), unpublished internal docu-
ments, Dar es Salaam: VETA.
Vries, J. de (1984), European Urbanization, 1500–1800, London: Methuen and Co.
Wangwe, S.M. (1979), Capacity Utilisation and Capacity Creation in Manufacturing
in Tanzania with Special Reference to the Engineering Sector, unpublished PhD thesis,
Wangwe, S.M. (1986), Technology Imports, Technological Learning and Self-Reliance in
Tanzania, Working Paper WEP 2–22/WP, Geneva: International Labor Office.
Wangwe, S.M. (1990), Reflections on the Basic Industry Strategy, professorial inaugural
lecture (unpublished), University of Dar es Salaam.
Wangwe, S.M. (1992a), ‘Building Indigenous Technological Capacity: a Study of
Selected Industries in Tanzania’, in: F. Stewart, S. Lall and S. Wangwe (eds),
Alternative Development Strategies in sub-Saharan Africa, Basingstoke: Macmillan
Press – now Palgrave – 265–93.
Wangwe, S.M. (1992b), Rehabilitations in the Manufacturing Sector on Tanzania, Dar es
Salaam.
Wangwe, S.M. (1993), ‘Implications of Changing External and Internal Conditions
for Industrial Restructuring in Tanzania’, in: M.S.D. Bagachwa and A.V.Y. Mbelle
(eds), Economic Policy Under A Multiparty System in Tanzania, Dar es Salaam
University Press.
Wangwe, S.M. (1994), ‘Technology Development in Tanzania: Some Challenges for
the 21st Century’, in: L.A. Msambichaka, H.P.B. Moshi and F.P. Mtatifikolo (eds)
(1994), Development Challenges and Strategies for Tanzania: An Agenda for the 21st
Century, Dar es Salaam: Economic Research Bureau/Department of Economic,
Wangwe, S.M. (ed.) (1995), Exporting Africa: Technology, Trade and Industrialization in
sub-Saharan Africa, London: Routledge.
Bibliography 403
Wangwe, S.M. (1998), The Vision of Industrialization, paper submitted to the Team of
Experts of the Development Vision, Dar es Salaam, ESRF.
Wangwe, S.M. and M.S.D. Bagachwa (1990), ‘Impact of Policies on Technological
Choice and Development in Tanzanian Industry’, in: F. Stewart, H. Thomas, and
T. de Wilde (eds), The Other Policy, London: Intermediate Technology Publications.
Wangwe, S.M. and N. Luvanga (1993), ‘Technological Transformation in the Third
World: The case of Tanzania’, in: S.J. Patel (ed.), Technological Transformation in the
Third World, vol. II, UNU/WIDER, Helsinki.
Wangwe, S.M. and H. Semboja (1997), Industrialization in Africa, paper contributed
to CODESRIA programme on Beyond Structural Adjustment, Dakar.
Wangwe, S.M., F. Musonda and J. Kweka (1997a), Policies for Manufacturing
Competitiveness: The Case of Tanzania, ESRF Discussion Paper, Dar es Salaam.
Wangwe, S.M., H.H. Semboja and J.P. Kweka (1997b), Multi-Country Comparative
Study of Private Enterprises Development in Africa (MCC-PED), A Draft Report for
Tanzania, Dar es Salaam.
Wangwe, S.M., H.H. Semboja, and H.B. Lecomte (1996), Exit Procedures and Economic
Transformation: The case of Tanzania, ESRF Discussion Paper Series no. 010, Dar es
Salaam.
Watanabe, S. (1987), ‘Technological Capability and Industrialization: Effects of Aid
and Sanctions in United Republic of Tanzania and Zimbabwe’, International Labor
Review, 125 (5).
Weaver, J.H. and A. Kronemer (1981), ‘Tanzanian and African Socialism’, World
Development, 9 (9/10).
Weiss, C., Jr. (1990), ‘Scientific and Technological Constraints to Economic Growth
and Equity’, in: R.E. Evenson. and G. Ranis (eds), Science and Technology: Lessons for
Development Policy, Boulder: Westview Press.
Weiss, J. (1988), Industry in Developing Countries: Theory, Policy and Evidence, New
York: Croon Helm.
Welch, F. (1970). ‘Education in Production’, Journal of Political Economy, 38, 35–59.
Wells, L. (1975), ‘Economic Man and Engineering Man: Choice of Technology in a
Low-Wage Country’, in: C.P. Timmer et al. (eds), The Choice of Technology in
Developing Countries, Cambridge: Harvard Studies in International Affairs, 69–74.
Westphal, L.E. et al. (1990), ‘The Development of Technological Capability in
Manufacturing: A Macroscopic Approach to Policy Research’, in: R.E. Evenson and
G. Ranis (eds), Science and Technology: Lessons for Development Policy, Boulder:
Westview Press.
Westphal, L.E., L. Kim and C. Dahlman (1985), Reflections on the Republic of Korea’s
Acquisition of Technological Capability, New York: Praeger.
Westphal, L.E. and R.E. Evenson (1993), Technological Change and Technology Strategy,
Development Economics Seminar, paper no. 93–1/4, Den Haag.
White, G. (1984), ‘Developmental States and Socialist Industrialization in the Third
World’, Journal of Development Studies, 21 (1).
Williams, D. (1975), National Planning and the Choice of Technology: the Case of Textiles
in Tanzania, Economic Research Bureau, paper no. 75.12, University of Dar Es
Salaam.
Williams, D. (1976), Choice of Technology and National Planning: the Case of
Tanzania, unpublished doctoral dissertation, Harvard University.
Winston, G. (1979), ‘The Appeal of Inappropriate Technologies: Self-Inflicted Wages,
Ethnic Pride and Corruption’, World Development, 7 (8/9), 835–45.
Wissema, J.G. and L. Euser (1988), Samenwerking bij Technologische Vernieuwing – De
Nieuwe Dimensie van het Management van Innovatie [Cooperation in Technological
404 Bibliography
Innovation – The New Dimension of Innovation Management], Deventer: Kluwer

bedrijfswetenschappen.
World Bank (1976), World Tables, 1976, Baltimore and London: Johns Hopkins
University Press.
World Bank (1980), World Tables, 1980, Baltimore and London: Johns Hopkins
University Press.
World Bank (1981), Accelerated Development in sub-Saharan Africa: An Agenda for
Action, Washington, DC.
World Bank (1984), Tanzania: Country Economic Memorandum, Report no. 5019-TA,
Washington: World Bank.
World Bank (1987a), Tanzania: An Agenda for Industrial Recovery, Report no. 6357-TA,
vols I & II, Washington: World Bank.
World Bank (1987b), World Development Report, 1987, Oxford: Oxford University
Press.
World Bank (1988), Parastatals in Tanzania: Towards a Reform Program, Report
no. 7100-TA, Washington: World Bank.
World Bank (1989), Industrial Development in Tanzania: An Agenda for Industrial
Recovery, Washington DC: World Bank.
World Bank (1991), Tanzania: Economic Report: Towards Sustainable Development in the
1990s, Report no. 9352-TA, vols I & II, Washington: World Bank.
World Bank (1993a), Zimbabwe: A Policy Agenda for Private Sector Development,
Washington, DC.
World Bank (1993b), First Report on the Kenyan Survey, Regional Programme on
Enterprise Development, Africa Technical Department.
World Bank (1993c), Economic Reform in the Manufacturing Sector of Ghana, Africa
Technical Department.
World Bank (1993d), East Asian Economic Miracle, Washington.
World Bank (1993e), Agricultural Sector Memorandum, Uganda, Report no. 10715-UG.
World Bank (1994a), Adjustment in Africa: Reforms, Results and the Road Ahead, New
York: Oxford University Press.
World Bank (1994b), First Report on the Zimbabwe Survey, Regional Programme on
Enterprise Development, Africa Technical Department.
World Bank (1994c), The Road Ahead, Washington, DC.
World Bank (1995a), Bureaucrats in Business: The Economics and Politics of Government
Ownership, New York: Oxford University Press.
World Bank (1995b), World Development Report, 1995, New York: Oxford University
Press.
World Bank (1995c), World Tables, 1995, Baltimore: Johns Hopkins University Press.
World Bank, (various), World Tables, Washington, DC: World Bank.
World Bank (1997), World Development Report, 1997, Oxford: Oxford University Press.
World Health Organization (1976), Health and Environment, Geneva.
Word Health Organisation (1982), Rapid Assessment of Sources of Air, Water and Land
Pollution, WHO Offset Publication no. 62, Geneva.
Wright, E. (ed.) (1984), History of the World: the Last Five Hundred Years, Feltham:
Hanslyn Publishing Group.
Zanting, H.A. (1996), A Systematic Assessment of Water Pollution in Lake Victoria, MSc
thesis, Eindhoven, The Netherlands: Eindhoven University of Technology, Faculty
of Technology Management.
Zijl, G. and R. Lassche (1997), Testing a Technology Audit Methodology: Rapid
Assessment of Six Small and Medium Scale Enterprises in Tanzania, MSc thesis,
Eindhoven: Eindhoven University of Technology.
Index
ABB National Transformers 157, 169 Cleaner Environment Production in

actor network 211–12 Industry in Tanzania (CEPITA)
African Socialism 136, 295 279
African Tanneries Ltd (AT) 346 climate 245, 247
Afro-Cooling 160 Commission for Science and
agriculture 12, 245, 246 Technology 164
focus on SMEs 365 Common Market for Eastern and
Ardhi Institute (University of Dar es Southern Africa (COMESA) 305
Salaam) 210 construction
Arrow, K. 127 current social-economic
Arusha declaration (1967) 17, 19, 30, performance/contribution to
41, 45, 137, 145, 148, 151 national development 200–1,
203, 206
Bagachwa, M.S.D. and Mbelle, A.V.Y. dwelling subsector: human resources
360 stock 208–10; natural resources
Bank of Tanzania 147, 310 stock 210–11; summary
basic industry strategy (BIS) 351, 352, 212; technological capabilities
358, 367 212; technology infrastructure
Bates, R. 135 (actor network) 211–12;
beer industry 265–6, 267, 268–9, 270, technology stock 206,
271 208
Bell, M.B. et al 196, 198
Bora Shoe Company 346 David, P. 123, 127
Bromley, R.J. 320 DeBresson, C. 154
Building Contractors (BC) 206, demographics 219, 220–1, 244–5
208 Development Vision 2025 368
Building Research Unit (BRU) 210 Dosi, G. 127
bureaucracy Downs, A. 144
foreign-exchange maximizing
behaviour 144–9, 151 East African Community 32, 38
political/managerial aspects 145, Economic Recovery Programme (ERP)
147–9 305–6, 342
public choice approach 144 economic and social action programme
size of public sector under structural (ESAP) 342
adjustment 149, 151 Economic and Social Research
Foundation (ESRF) 157, 160, 163,
cement industry 271–7 168–9, 360
Centre for Agricultural and Rural Multi-Country Comparative Study on
Mechanization and Appropriate Private Enterprise Development
Technology (CARMATEC) 161, 165
169 education 3–4, 6–7, 51, 291, 297–8
Chazan, N. et al 145 demand/supply of technical
Clark, W.E. 353 manpower 223–4
Clean Development Mechanism 263, demographics 219, 221
279 financial sources 233, 234
405
406 Index
education (cont.) exports

firm owners/managers/employees background 114–16
58–61 dual economy model 116–22, 129–30
importance 218–19 problems for new entrants 122–4;
metalworking industry example generic technological change
224–34 124–6; patterns of path
post-primary 222 dependency 126–9
primary 221
role 52–4 Fei-Ranis dual economy model
secondary 222–3 117–22, 129–30
structure/enrolment 221–3 commercialization point 117
teacher selection 233–4 export substitution 117, 119, 129
technical 359–61 labour-intensive export phase
TIRDO staff 173 117–19
universities 221 technological predictions 119–22
use of English 233 weaknesses 129–30
Elbadawi, I. 310 Financial Act (1973) 21
energy conservation firms
background 262 descriptive information 58–61
beer industry example 265–6, 267, estimates of cost functions 61, 63,
268–9, 270, 271 64, 65, 66–8, 69
cement industry example 271–7 planning in 50
in industrial sector 262–3 selection for R&D 185–6
in practice 263–5 skills/production-level data 56–7
summary 277–8 technology capability 69–71
what industries can do 278–80 Foreign Exchange Act (1992) 29
Energy Policy of Tanzania 262 Foreign Investment Act (1963) 18
Engblom, AC. and Veres, V. 306 Framework Convention on Climate
environment 5–6, 12 Change (FCCC) 263
background 237 Fransman, M. 155
industrial pollution in urban areas Freeman, C. 124–5
252, 253, 255, 256; air 254,
257–9; mining activities 256, G&T Company 159
259–60; water 254 Gallagher, M. 135
Lake Victoria results; BOD loading Galtung, J. 319
estimates 247–8; TN/TP loading Ghana 57, 58–61, 70
estimates 249–50 Ghanaian Living Standards
Lake Victoria statistics 243; Measurement Study 59
agriculture 245, 246; industrial government
characteristics 244; population partnership with private sector 370
244–5; role/capacity 167, 365–6
rainfall 245, 247; wetlands Grossman, G.M. and Helpman, E. 126
247
Environmental Resource Consultants in human resources 199, 208–9
Dar es Salaam 263 development 160–1, 359–61
ESCAP 175 educational backgrounds 209
ESRF-PSRC 157, 158, 159 occupational status 209–10
exchange rates 19, 20, 21, 23, 26,
357–8 import substitution industrialization
export credit guarantee scheme (ECGS) (ISI) 350–1
30 Indol International Holding 265
Index 407
industrial development 48–9 government support 365–6;

background 12–13 focus on technological learning
crisis in 44–5 364–5; priority of agro–based
1961–7 period 38–9 SMEs 365
1967–73 period 39, 41 failures: inadequate infrastructure
1973–80 period 41–3 359; inadequate institutional
1980–4 period 43–5 capacity 358; low investment in
1984–90 period 45–7 technology/technological
1990–5 period 47–8 learning 362–3; low level of
periods of analysis 35 human resource development
poor infrastructure/shortage of skilled 359–61; policy 357–8; weak
labour 13, 48 linkages between industry/related
statistics concerning 36, 37, 41 sectors 358–9
trends/explanations 34, 37–8 Great Depression 350
industrial policy 3, 6, 7–8 historical perspective 283
background 11 import substitution policies 350–1
classification of strategy 15–16 planning models 350
confinement 19, 22, 26 post-independence period 349–50
decontrol/shift from import prerequisites 285–6; economy
substitution 25–8, 40 288–90, 293–6; education 291,
development plans 17 297–8; health care 291, 298;
export incentives/promotion 23, kinship 290, 296–7; politics
26–30 286–8, 291–3; religion 290–1,
foreign aid 20–3, 25, 28–30 297
foreign exchange windows 46 present state 284
import licensing 19, 21, 22, 25, restructuring: economic reforms
26 354–5; insufficiency of
import substitution 17–18, 41, 43, macroeconomic reform 355–6;
160 recovery 355; technological
initial conditions 12–13 capabilities 356–7
joint ventures 19, 39 strategy 366
liberalization 28–30 innovation
nationalization 18–19 background 153–4
policy periods 13–14, 17, 31 efforts/achievements in
price control 21–2, 23, 26 manufacturing sector 156–60
private (foreign) entrepreneurship factors influencing productivity 160;
28–30 access to technical information
private/public sector balance 23, 165–6; competitive market
27–8, 41 turbulence 162–3; customer
rates of protection in manufacturing base 162–3; financial
24 capabilities 163–4; human
regulatory control 18–23, 25 resource development 160–1;
self-reliance 18–19 institutional framework 164–5;
synthesis of external influences/policy legal/regulatory framework 166;
periods 30, 32, 33, 34 R&D efforts 161–2; role of
tariff protection 17–18, 39 government 167
industrialization forms/role from conceptual/general
basic industry strategy 351 perspective 154–6
development of competitive industry offensive/defensive strategies 156
363–4; dynamic structure 364; policy recommendations 167–8
enhance role/ capacity of search capabilities 156
408 Index
Institute for Production Innovation (IPI) labour productivity 87, 92–8, 104–5
169 new indices of production 81, 82–5,
International Comparisons of Output 87, 88–91
and Productivity (ICOP) project new insights 98; changes in
92 structure 100, 102; level
International Labour Organisation (ILO) adjustments in nominal value
326 added 100, 101; trends in real
International Monetary Fund (IMF) 34 growth 102–4
post-independence 352–4
Jedruszek, J. 353 pre-independence 351–2
reconstruction of nominal 10+ GDP
Kenya 57, 58–61, 70 76; conceptual errors 78–80;
Kimambo, R.H. 211 coverage 76–7; summary of
knowledge transfer 176–7 adjustments 81; treatment of
Krugman, P. 126 non-response in official statistics
77–8
labour productivity 41–4, 45, 87 manufacturing sector
basic data 106–11 background 300–1
1989 benchmark 92; comparisons exports 315
98; data sources 95–6; investment 316
methodology for comparison macroeconomic policies 303–6;
92–5; unit value ratios 96–8, analysis of effects 307–14
97 recent developments 301–2;
comparative 104–5 dynamics 303; signs of recovery
Malaysian pattern 118 302–3
trends 104 Masuha 360
Lake Victoria Mbelle, A.V.Y. 306
pollution 237, 239, 241–2, 247–50 and Bagachwa, M.S.D. 303
statistics 243–7 metalwork industry
Lall, S. 198 research into education shortcomings:
Lewis, A. 115, 116–17 analysis 231, 233–4;
Lewis dual economy model 116–17 background 224; facilities/
costs 230; family/community
macroeconomic policies 228; finance 230–1;
empirical analysis: methodological methodology 224–6;
issues 307–10; time series organization, contents, methods
310–14 230; results 226, 227, 228, 229,
impact on manufacturing sector 230–1, 232; school management
305–6 230; students 228; teachers
industrial development 303–5 228, 230
insufficient reform 355–6 Ministry of Land and Urban
manufacturing performance 73–5 Development 210
capacity utilization 302 Ministry of Science and Technology and
data on 352–4 Higher Education 210
data sources for time series on 10+ Ministry of Trade and Industries (MIT)
75 347, 348
employment 302 Ministry of Works 210
exports 301–2 Morogoro Canvas Mill 169
gross value added/employment Morogoro Shoe Company 346
99 Moshi Leather Industries 159
growth/share in GDP 301 Mukanda, R. 147
Index 409
National Board of Architects and the creation/sustainability of enabling

Quantity Surveyors (NBAQS) 206, environment 371; action on
208 other related issues 372; fiscal
National Construction Council (NCC) policies 371; infrastructural
210, 211 development 372; investment
national consultative process for promotion 372; monetary
industrialization (NCPI) 373 policy 371; trade 371
National Development Corporation development 370–1
(NDC) 18, 139, 145, 147 partnership with government 370
national economic survival programme privatization
(NESP) 17, 343 conceptual framework 341–2
National Industries (Licensing and factors accounting for turnaround
Registration) Act (1967) 19 346–7
National Investment (Promotion and reforms/industrial performance
Protection) Act (NIA) (1990) 342–4; insights 344–6; summary
28–9 347–8
National Price Commission 147 production: aggregate evidence on
National Price Control Advisory Board technology 54, 55; background
19 50–1; determinants 51–4;
natural resources 199, 210–11 economic environment 52;
Ndulu, B.J. and Semboja, J.J. 312 firm-level data 56–7; indices
Nelson, R.R. 198 81, 82–5, 86–7, 88–91; purchase
newly industrialized countries (NICs) of modern equipment 52; role
54, 114–15 of education/technology 52–4;
Niskanen, W. 144, 145 technological ability 51
Nyerere, J. 136–7, 297 public choice approach 135, 144
Public Corporation Act (1992) 29
Pack, H. 176 public sector: balance with private 23,
and Paxson, C. 361 27–8, 41; political/managerial
and Westphal, L.E. 128 conflict 145, 147–9; size under
Parastatal Sector Reform Commission structural adjustment 149, 151
(PSRC) 348
path dependencies rapid assessment evaluation 250, 251,
and generic technological change 252, 260–1
122–4 rapid assessment methodology 238;
patterns 126–9 penetration factors 240–2;
Pieper, U. 304 reference nutrient load from
political system capacities modelling 242; waste loads
distributive 287 238–40; wet deposition 240
extractive 286–7 Regional Program on Enterprise
internal/external 287 Development (RPED) 56, 157,
productive 287 158, 162
regulating 287 research and development (R&D) 154,
symbolic 287 155, 160, 169
pollution see environment access to information 165
Portland Cement (Wazo) Ltd 302 efforts in 161–2
Presidential Parastatal Sector Reform institutions 169–70, 171–2
Commission 29 model 173–4
Prins, I.M. and Szirmai, A. 353 phases 174
private sector selection of firms 176–7
balance with public 23, 27–8, 41 Richardson, C.J. and Nichols, D.S. 241
410 Index
Rodriguez, E. 303 background 265, 344

Rogers, E.M. 176 energy; causes of inefficiency 269,
Romano, C.A. 154 271; conservation opportunities
Romijn, H.A. 198 268–9, 270; costs 266, 268;
Rosenberg, N. 197 efficiency 265–6; specific 267
Rostow, W.W. 320 performance 344–5
Rothwell, R. 163 Tanzania Bureau of Standards (TBS)
169
Schumpeter, J. 155 Tanzania Cigarette Company (TCC)
science and technology (S&T) activities 302, 346
164 Tanzania Engineering and
Semboja, H.H. et al 157, 158, 161, 165 Manufacturing Design Organisation
small and medium-sized enterprises (TEMDO) 169
(SMEs) 365 Tanzania Industrial Research and
South African Breweries (SAB) 265 Development Organisation (TIRDO)
State Trading Corporation (STC) 19, 169, 171, 265, 358–9
22 acquisition of technology 181–3
Steinmuller, E. and Bastos, M.-I. 125 adaptation of technology 183–5
structural adjustment plan (SAP) 17, educational levels of staff 173
343 field research 177–8
sustainable industrial development focus of 172–3
policy (SIDP) general characteristics of process
background 367 187, 189
described 367–8 projects: activated carbon 179;
development of private sector aluminium sulphate 178;
370–1; creation/sustainability of caustic soda 178; dehydrated
enabling environment 371–2 castor oil 178; natural dyes
implementation strategy 369–70 178; pectin 179; satellite
industrial sector mission/objectives receivers 180; school chalk
368–9 179; solar thermal systems
institutional framework/consultative 179; Turkey red oil 180;
mechanism for policy variables 191–2; wood adhesive
implementation 372–4 179
partnership between selection of firms 185–6
government/private sector 370 summary 189–91
success 374–5 technology identification/selection
180–1
Tanga Cement Company Ltd 302 transfer/implementation of
Tanzania technology 186–7
background 1–3 Tanzania Industrial Studies and
colonial period 12 Consulting Organisation (TISCO)
economic policy 3 355
education 3–4, 6–7 Tanzania Oxygen Ltd 302
environment 5–6 Tanzania Portland Cement Corporation
industrial policy 3, 6, 7–8 Ltd 346
industrialization 284–5; lessons background 271–2
from European history 298–9; conservation opportunities:
pre-requisites 291–8 individual power station 275–6;
technology 4–5 long-term options 276; new
Tanzania Breweries Ltd 157, 160, 302, measurement systems for kilns
347, 353 276; reduction of radiation/
Index 411
convection losses 276; replace and knowledge transfer 177

electrostatic filters 276 learning process 127–8, 196–7; focus
energy: causes of inefficiency 277; on 364–5; low investment in
costs 274–5; efficiency 362–3
272–4 little change in 54, 55, 56
Tanzanian Electrical Goods role 52–4
Manufacturing Company stock 199, 206, 208
(TANELEC) 346–7 transfer/implementation 186–7
background 345 upgrading 115
performance 345–6 textile industry 139, 142–4, 356
Tanzanian Investment Centre (TIC) Textile Manufacturing Association of
347, 348 Tanzania (TEXTMAT) 29
technological behaviour transnational corporations (TNCs) 38
public sector paradoxes 135–6;
explanations/limitations urban informal sector
140–4; planned vs actual background 318–19
136–9; project vs technological current theoretical points of
preferences 141–2; pronounced view/suggested research topics
variations 139–40; two 320–1
textile plant comparisons empirical evidence 322; constraints
142–4 for future growth 325–6;
technological capabilities 69–71 relations between
assessment 198 formal/informal 323–4;
background 194 relevance for socioeconomic
building 197–8, 213–16 development 324–5; support for
construction industry example formal/informal dichotomy 323
200–1, 202, 203, 204–5, 206, 207, interest in 319–20
208–12, 217 methodological problems 322
definitions 196–7, 199–200 statistics 327–36
levels 197
new theoretical views 194–6 Vocational Educational Training
restructuring 356–7 Authority 161
technological change 119–22
generic 124–6 Wangwe, S.M. 157, 163
and path dependencies 122–4 et al 303, 306
technology Weiss, C. Jr 198
ability 51 Wells, L. 140
acquisition 128, 175, 181–3 Westphal, L.E. 198
adaptation 176, 183–5 Williams, D. 140–1, 143
development 153 World Bank 25, 39, 44, 46, 56, 149,
general characteristics of development 157
process 187, 188, 189 World Health Organisation (WHO)
identification/selection 174, 180–1 252
implementation/diffusion 177
investment in 362–3 Zimbabwe 57, 58–61, 70

The Industrial Experience of Tanzania Adam Szirmai and Paul Lapperre

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

The Industrial Experience of Tanzania Adam Szirmai and Paul Lapperre

Hochgeladen von

Copyright:

Verfügbare Formate

The Industrial Experience

ECONOMIC AND SOCIAL DEVELOPMENT: Trends, Problems, Policies

Conference Participants viii

Notes on the Contributors x

Part I: Long-run Economic Performance 9

Part II: Innovation, Technological Capabilities and Choice of

Capabilities in the Dwelling Construction Sector for

Part III: Environmental and Energy Aspects of Industrialization 235

Part IV: Lessons from Past Experiences, Economic Reform,

This volume presents the proceedings of a conference on ‘The Industrial

Bartelt Bongenaar Herman Gaillard

T.S.A. Mbwette Wessel Schulte

Menno Prins Samuel M. Wangwe

Amber Beernink, Section of Technology and Development Studies, Faculty

Bartelt Bongenaar, Section of Technology and Development Studies, Faculty

Otto Bos, Section of Technology and Development Studies, Faculty of

Charles Cooper, The United Nations University/INTECH Maastricht, The

Raymond Duijsens, Section of Technology and Development Studies,

Emilia van Egmond-de Wilde de Ligny, Section of Technology and

Donné van Engelen, Section of Technology and Development Studies,

Herman Gaillard, Section of Technology and Development Studies,

David van Horen, Centre for Technology for Sustainable Development,

Jeffrey James, Tilburg University, The Netherlands

Josephat P. Kweka, Economic and Social Research Foundation (ESRF), Dar

Paul Lapperre, Section of Technology and Development Studies, Faculty of

Lex Lemmens, Centre for Technology for Sustainable Development (TDO),

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

Gregory Njau, Environmental Association of Tanzania (ENATA), Dar es

Howard Pack, World Bank/University of Pennsylvania, USA

Christina Paxson, Princeton University, USA

Menno Prins, Section of Technology and Development Studies, Faculty of

Caspar Samplonius, Section of Technology and Development Studies,

Peter Scheren, Centre for Technology for Sustainable Development,

Wessel Schulte, Section of Technology and Development Studies, Faculty

Haji H. Semboja, Economic and Social Research Foundation (ESRF), Dar es

Adam Szirmai, Eindhoven Centre for Innovation Studies, Eindhoven

Dick Toussaint, Section of Technology and Development Studies, Faculty

Frank van der Vleuten, Section of Technology and Development Studies,

Samuel M. Wangwe, Economic and Social Research Foundation (ESRF),

Michel Yhdego, ERC Consultants, Dar es Salaam, Tanzania

Harro Zanting, Centre for Technology for Sustainable Development,

This collection of articles focuses on an analysis of the industrial experience

of forced relocation and disruption of agricultural populations, it avoided

articles based on a wealth of information derived from ﬁrm-level surveys in

become more important. In this respect it is important to remember that

literature emphasizes that exporters are forced to innovate and to meet

The method is initially applied to Lake Victoria where, as it turns out,

On the basis of econometric analyses Mbelle concludes that manufactur-

Throughout the past decades the economic and industrial performance of

beverages), and 225 in consumer goods manufacturing (textiles, footwear,

Identifying the periods of coherent policy which have affected the

be established in two ways: by looking at objectives and plans, or by moni-

Aspect of Industrial strategy Scaling Policy instruments

b. Export promotion Tax concessions, export credits, foreign exchange retention,

II. Degree of direct regulatory

c. Labour Low/moderate/high Minimum wage legislation, immigration and emigration

d. Ownership Low/moderate/high Industrial licensing, legislation regarding foreign investment

Aspect of industrial strategy Scaling Policy instruments

b. Public sector Public enterprise investment, nationalisation, regulation

b. aid Requests for aid, restrictions for private foreign investment,

article, the beginning of a new policy period is deﬁned by the occurrence of

3.1 Import substitution through foreign investment (1961–67)