(Development Centre Studies.) OECD - Emerging Africa.-Organisation For Economic Co-Operation and Development (2002) PDF

«
Development Centre
Studies
Emerging Africa
INTERNATIONAL DEVELOPMENT
By Jean-Claude Berthélemy and

Ludvig Söderling, with
Jean-Michel Salmon and
Henri-Bernard Solignac Lecomte
© OECD, 2001.
© Software: 1987-1996, Acrobat is a trademark of ADOBE.
All rights reserved. OECD grants you the right to use one copy of this Program for your personal use only.
Unauthorised reproduction, lending, hiring, transmission or distribution of any data or software is
prohibited. You must treat the Program and associated materials and any elements thereof like any other
copyrighted material.
All requests should be made to:
Head of Publications Service,
OECD Publications Service,
2, rue André-Pascal,
75775 Paris Cedex 16, France.
Development Centre Studies
Emerging Africa
by
Jean-Claude Berthélemy and Ludvig Söderling
with
Jean-Michel Salmon and Henri-Bernard Solignac Lecomte
DEVELOPMENT CENTRE OF THE ORGANISATION

FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
ORGANISATION FOR ECONOMIC CO-OPERATION
AND DEVELOPMENT
Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and
which came into force on 30th September 1961, the Organisation for Economic Co-operation
and Development (OECD) shall promote policies designed:
– to achieve the highest sustainable economic growth and employment and a rising
standard of living in Member countries, while maintaining financial stability, and thus
to contribute to the development of the world economy;
– to contribute to sound economic expansion in Member as well as non-member
countries in the process of economic development; and
– to contribute to the expansion of world trade on a multilateral, non-discriminatory
basis in accordance with international obligations.
The original Member countries of the OECD are Austria, Belgium, Canada, Denmark,
France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway,
Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States.
The following countries became Members subsequently through accession at the dates
indicated hereafter: Japan (28th April 1964), Finland (28th January 1969), Australia
(7th June 1971), New Zealand (29th May 1973), Mexico (18th May 1994), the Czech Republic
(21st December 1995), Hungary (7th May 1996), Poland (22nd November 1996), Korea
(12th December 1996) and the Slovak Republic (14th December 2000). The Commission of
the European Communities takes part in the work of the OECD (Article 13 of the OECD
Convention).
The Development Centre of the Organisation for Economic Co-operation and Development was
established by decision of the OECD Council on 23rd October 1962 and comprises twenty-three Member
countries of the OECD: Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany,
Greece, Iceland, Ireland, Italy, Korea, Luxembourg, Mexico, the Netherlands, Norway, Poland, Portugal, Slovak
Republic, Spain, Sweden, Switzerland, as well as Argentina and Brazil from March 1994, Chile since
November 1998 and India since February 2001. The Commission of the European Communities also takes
part in the Centre’s Advisory Board.
The purpose of the Centre is to bring together the knowledge and experience available in Member
countries of both economic development and the formulation and execution of general economic policies; to adapt
such knowledge and experience to the actual needs of countries or regions in the process of development and to
put the results at the disposal of the countries by appropriate means.
THE OPINIONS EXPRESSED AND ARGUMENTS EMPLOYED IN THIS PUBLICATION ARE THE SOLE
RESPONSIBILITY OF THE AUTHORS AND DO NOT NECESSARILY REFLECT THOSE OF THE OECD OR OF
THE GOVERNMENTS OF ITS MEMBER COUNTRIES.
*
* *
Publié en français sous le titre :
L’AFRIQUE ÉMERGENTE
© OECD 2001
Permission to reproduce a portion of this work for non-commercial purposes or classroom use should be
obtained through the Centre français d’exploitation du droit de copie (CFC), 20, rue des Grands-Augustins,
75006 Paris, France, tel. (33-1) 44 07 47 70, fax (33-1) 46 34 67 19, for every country except the United States.
In the United States permission should be obtained through the Copyright Clearance Center, Customer
Service, (508)750-8400, 222 Rosewood Drive, Danvers, MA 01923 USA, or CCC Online: www.copyright.com. All
other applications for permission to reproduce or translate all or part of this book should be made to
OECD Publications, 2, rue André-Pascal, 75775 Paris Cedex 16, France.
Foreword
This study is the result of the “Emerging Africa” project which has been part of
the OECD Development Centre’s programme of work since 1997. The Development
Centre would like to express its gratitude to the Governments of Belgium and
Switzerland for their financial support to the project.
3
4
Table of Contents
Preface ....................................................................................................................... 7
PART ONE
ECONOMIC GROWTH SCENARIOS FOR AFRICA
Chapter 1 Renewed Progress in Africa in the 1990s? .................................................... 15
Chapter 2 Factors of Economic Take Off in Africa ....................................................... 33
Annex Structural Change in the Productive Sector .................................................. 61
PART TWO
ANALYSIS OF GROWTH FACTORS IN SIX AFRICAN COUNTRIES
Chapter 3 Capital Accumulation .................................................................................... 81
Chapter 4 Human Capital ............................................................................................... 99
Chapter 5 Exports ........................................................................................................... 119
PART THREE
THE POLITICAL ECONOMY OF REFORM IN SIX AFRICAN COUNTRIES
Chapter 6 Overview ....................................................................................................... 137
Chapter 7 Burkina Faso ................................................................................................. 149
Chapter 8 Côte d’Ivoire ................................................................................................. 159
Chapter 9 Ghana ............................................................................................................. 169
Chapter 10 Mali ............................................................................................................... 179
Chapter 11 Uganda ........................................................................................................... 189
Chapter 12 Tanzania ......................................................................................................... 197
Concluding Summary Policy Options for Emerging Africa .......................................... 207

Bibliography ...................................................................................................................... 225
5
6
Preface
The international strategy for combating poverty included in the Millennium

Development Goals (MDGs) at the Millennium Summit of the United Nations in
September 2000 has its origins in the adoption in May 1996 of a series of goals by the
OECD Development Assistance Committee (DAC) for 2015. These objectives
essentially address the effort to improve social development throughout the world,
while respecting the environment. In the same way, the African Growth and Opportunity
Act, a law recently passed in the United States, began as a resolution for Africa adopted
by the Congress five years ago. Both the DAC goals (published in Shaping the 21st
Century: the Contribution of Development Co–operation) and the US resolution have
influenced this OECD Development Centre project. The project, launched in 1997
and directed by Professor Jean–Claude Berthélemy, has sought to identify the conditions
for economic take off in Africa.
In the mid–1990s, Africa showed signs of both political and economic progress.
Yet the sustainability of these improvements remains unclear. In spite of such
uncertainties, we could not afford to wait for events to reveal whether the optimists or
the pessimists would carry the day. Instead, it seemed necessary to determine which
of the recent evolutions in African economies might lead to substantial economic and
social development over time. Keeping in mind that unforeseen events would
periodically influence outcomes, what was needed was a long–term outlook that might
identify the conditions for emergence in certain African countries.
The OECD Development Centre’s Emerging Africa project took as its basic
assumption that there can be no real poverty reduction in the least developed countries
without economic take off. The goal was then to determine which factors could influence
long–term growth. The experience of Japan, followed in the 1960s by Korea, Chinese
Taipei and other dynamic East Asian economies showed that, even in situations of
extreme poverty, there can be economic take off and poverty reduction if the right
economic policies are put in place.
Nonetheless, this process is not uniform across regions, nor internationally. There
are always countries that lead and others that follow. The countries farthest along in
the economic and political reform process can be considered to have the best chances
for take off in the middle term. They are also the places where development assistance
has the best chances of having a positive impact on growth and poverty reduction.
7
Assisting the countries farthest along in the reform process should serve as an incentive
to other countries to hasten the pace of reform. Examining “emerging” Africa is based
on the belief that these initial successes can create a ripple effect at the regional level
and stimulate take off in a significant number of other African countries in line with
the DAC goals.
The Millennium Goals, along with other strategies for Africa adopted by the G–8,
the largest African countries and the Bretton Woods institutions reveal that the OECD
Development Centre was right to begin this project, in spite of the difficulties it faced
five years ago. In this context, I would like to thank the Belgian and Swiss governments
who provided the intellectual and financial support that allowed the project to reach
completion. I would especially like to thank Mr. Paul Frix, Deputy Permanent
Representative, Delegation of Belgium to the OECD, and Mr. Henri–Philippe Cart,
Assistant Director General, Swiss Agency for Development and Cooperation (SDC),
Ministry of Foreign Affairs. The success of the workshop held in Geneva in October
2000, in collaboration with the Institut supérieur d’Études internationales, at which
Professor Jean–Claude Berthélemy presented the project’s main findings, leads us to
believe that this book will be of use to those interested in African development.
In addition to his other responsibilities in the French Government, Jean–Claude
Berthélemy directed this project and wrote Part One and Part Two of the book with
Ludvig Söderling, from the African section of the International Monetary Fund as co–
author, who also contributed to the Ghana case study. Jean–Michel Salmon wrote
Part Three in spring 2001. Henri–Bernard Solignac Lecomte was responsible for the
Concluding Summary and shared editorial responsibilities. I would like to thank them
all, as well as to acknowledge my personal responsibility for the final co–ordination
of the various contributions to the book.
❊ ❊ ❊
This work is organised as follows: Part One covers the growth scenarios from a
pan–African perspective. Chapter 1 deals with the question of renewed growth in the
1990s, while Chapter 2 analyses the factors behind take off in Africa, emphasising the
two principal determinants of structural change in production: the reallocation of the
rural labour force, and the diversification of production. Both of these affect the
production function and are taken up again in the Annex to Chapter 2, which constitutes
the essential core of this work. Part Two looks at the main factors of growth in six
emerging African countries. Chapters 3 and 4 address the accumulation of physical
and human capital, respectively, and Chapter 5 covers export performance.
Part Three contains the six case studies, focusing on the political economy of
the reform process. Chapter 6 presents a general overview, while Chapters 7 to 12
describe the specific situations of Burkina Faso, Côte d’Ivoire, Ghana, Mali, Uganda,
and Tanzania. These case studies, cited in the bibliography, were completed by
Arne Bigsten (Uganda), Gerard Chambas (Burkina Faso and Mali), Denis Cogneau (Côte
8
d‘Ivoire), Jean–Louis Combes (Burkina Faso et Mali), Anders Danielson (Tanzania),
Patrick Guillaumont (Burkina Faso and Mali), Sylviane Guillaumont–Jeanneney
(Burkina Faso and Mali), Clark Leith (Ghana), Sandrine Mesplé–Somps (Côte
d’Ivoire), Steve Kayizzi–Mugerwa (Uganda), and Bertrand Laporte (Burkina Faso
and Mali). The Democratic Republic of Congo was initially part of this study, but was
left out due to the ongoing conflict which cast doubt on its prospects for growth. Thus
the contributions of Joseph Maton, Koen Schoors and Annelies Van Bauwel were not
included.
The Concluding Summary addresses the major implications of political
economy and institutional change for emerging Africa. It begins by posing the
question of whether Africa differs from other continents in terms of its economic
growth after independence. When examined in the context of the three pillars of
rapid, sustainable growth obtainable through structural change — capital
accumulation, productivity gains, and institutional reform — many African countries
seem to have experienced accumulation without productivity, and therefore have
not achieved sustainable growth. This fact allows us to understand the prospects for
the upcoming decades and in particular the alternating periods of “pessimism” and
hope for an African “renaissance”. Finally, the growth scenarios for 2020 presented
in Chapter 2 are summarised before taking up the implications of economic policy
and institutional change for African countries and for donors.
What African countries can do for themselves is to stimulate investment in a
sustainable manner, improve factor productivity, promote structural change, increase
market openness and improve export performance (keeping in mind that imitation
equals innovation) and finally implement the necessary political reforms for sustainable
structural transformation. Donors, on the other hand, should consider re–calibrating
and re–directing assistance, increasing the capacity of African countries to share in
the benefits of globalisation, and to continue the process of trade liberalisation.
❊ ❊ ❊
The case studies included here were evaluated by the Project Planning Committee
(which met in July 1998 and February 1999) and finalised in the synthesis report
produced at the time of Ludvig Söderling’s departure in June 2000. The synthesis
report has been the object of several presentations; the growth scenarios for 2020
were published in the conclusions of the First International Forum on African
Perspectives, organised by the OECD Development Centre and the African
Development Bank, and held at the French Ministry of Finance in February 2000. In
addition to the Geneva workshop already mentioned, other presentations of this work
took place in October 2000 and in May 2001 at the French Ministry of Foreign Affairs;
in December 2000 in Nairobi with the African Economic Research Consortium; in
March 2001 at the IMF; and at the Centre for the Study of African Economies at
Oxford, as well as at the universities of Clermont–Ferrand and Dakar.
9
Upon my arrival, the case studies and synthesis report were put on the
Development Centre’s website for further commentary. The authors responsible for
Burkina Faso, Côte d’Ivoire, Ghana and Mali have published their studies separately.
The Democratic Republic of Congo case study by Professor Joseph Maton will appear
as Technical Paper No. 178. Various thematic studies of particular points completed
in the course of this project are listed in Section I of the Bibliography. These include
the Heavily Indebted Poor Countries Initiative (Technical Paper No.163), financial
sector reform, trade and exchange policy, the role of trade liberalisation in growth, the
impact of macroeconomic policy on growth (Technical Paper No. 150) and factors of
growth (Technical Paper No. 145).
This project benefited, at each stage of its development, from the exchange of
different points of view from a large group of people both inside and outside the
OECD, and these comments had a considerable influence on the final work. I would
especially like to thank the following for their contributions: Orlando Abreu, Patrick
Asea, Dominique Bocquet, Kwezi Botchwey, François Bourguignon, Bruno Cabrillac,
Richard Carey, Bernard Chane–Kune, Jean–Pierre Cling, Jean–Marie Cour, Jacqueline
Damon, Sébastien Dessus, Shanta Devarajan, Augustin Fosu, Andrea Goldstein, Jan
Gunning, Ulrich Hiemenz, Mohammad Hussein, Anne Joseph, Mustapha Kassé, Tony
Killick, Michael Klein, Peter Landymore, Jacques Loup, Claude Maerten, Allechi
M’Bet, Katharina Michaelowa, Christian Morrisson, David O’Connor, Temitope
Oshikoya, Ademola Oyejide, Sheila Page, Christine Richaud, Louk de la Rive Box,
Amos Tincani, Aristomène Varoudakis and Alain Viry.
This work demonstrates that, even if the optimism that characterised 1996–97
has to a degree subsided, Africa still has the potential for emergence. It is still possible
to expect that the successes seen in Botswana and Mauritius can be achieved in other
African countries.
In spite of its heterogeneity, evident in the three parts of the book, Africa aims
for a unitary development. In spite of the various strategies for Africa adopted by the
G–8 and the Bretton Woods institutions, what Africa does not do for itself cannot be
done by others. Mobilising the resources necessary for development requires the
pursuing of liberalisation reforms, and above all improvements in governance and
institutional structures. Without these, there is a substantial risk that Africa’s greatest
assets will be wasted by corruption. It is only through such reform that Africa can
attract foreign investment. Without a doubt, this requires real political leadership.
Recent strategies put forth by leaders of the three largest countries on the continent
lead us to believe that the essential ingredient for pursuing sustainable growth in Africa
— a long–term vision on the part of leadership — exists at present.
10
Economic modernisation takes one or more generations, however, and requires
long–term investments. We can never over–emphasise the need to strengthen human
capital through the implementation of far–reaching and sustainable education and
health policy. Higher levels of investment are also necessary, both for creating
production capacity and employment, as well as for modernising economic structures.
Given the present potential for savings, this implies the use of foreign financing.
However, such financing should not generate debt, considering the existing problems
of indebtedness still in need of solution. The quality of assistance and the modalities
of its usage are therefore fundamental, both for fighting poverty and for sustainable
development.
This project demonstrates, finally, that Africa is not essentially different from
other continents. It is simply less well understood. For this reason, I hope that the
present work will contribute to available knowledge of emerging Africa and thereby
restore faith in the prospects for the whole continent.
Jorge Braga de Macedo

President
OECD Development Centre
August 2001
11
12
PART ONE
ECONOMIC GROWTH SCENARIOS FOR AFRICA
13
14
Chapter 1
Renewed Progress in Africa in the 1990s?
Introduction
The 1990s were marked by improved economic growth in sub–Saharan Africa,

from a rate close to zero in 1991–92 to nearly 5 per cent in 1996. As a consequence,
the region’s per capita GDP, which had fallen at a rate of 2 per cent a year from 1990
to 1994, rose by 0.8 per cent a year from 1994 to 1998. This trend fostered hopes,
from 1995 to 1998, of seeing Africa move forward on the path of economic
development. The 1980s having been marked by negative growth of per capita GDP
(–1.2 per cent a year on average), Africa’s per capita GDP in 1994 stood at about the
same level as 30 years earlier.
The overall increase recorded was admittedly modest, but there were great
differences from one country to another. Some countries made real economic progress,
raising hopes of an economic revival. The first part of this chapter is devoted to an
examination of these varying degrees of progress, on the basis of a few simple
macroeconomic indicators. In the second, some social indicators will be examined to
determine whether the renewed growth made it possible to begin to reduce poverty.
The Improvement in Growth Performance
Diversity of Performances
Table 1.1 presents change in the average growth rate of sub–Saharan Africa over
the 1990s. The main characteristic that emerges is the extreme variety of economic
performances in the region. Every year, some countries sank into crisis while others
experienced a spectacular surge. This extreme variability is partly due to the fact that
economic conditions in African countries were particularly erratic, and that many of
them went through very strongly marked cyclic swings. It also reflects the real long–
term progress made by certain countries, while others experienced a prolonged crisis.
15
Table 1.1. Per Capita GDP Growth Rate
(constant 1995 dollars)
Year 1991 1992 1993 1994 1995 1996 1997 1998 Average
Simple average 2.0 0.4 0.9 1.7 5.1 5.6 5.4 3.3 3.1
Weighted average 0.5 -1.3 0.8 2.2 4.1 4.8 3.4 2.1 2.1
Weighted average
excluding South Africa 1.9 -0.6 0.3 1.3 5.0 5.4 4.1 3.5 2.6
Standard deviation 4.5 5.5 7.2 9.3 7.2 5.4 11.7 6.2 7.1
Lowest -8.4 -10.5 -23.8 -50.2 -10.0 -8.4 -20.2 -28.1 -19.9
Highest 10.4 10.7 13.4 16.8 34.4 29.1 76.1 21.3 26.5
Source: Authors’ calculations based on World Bank (World Development Indicators, 2000).
Evidence of this appears in the last column of the table, which lets us analyse the
distribution of African growth rates for the entire 1990–98 period. There is a gap of
25 percentage points of annual growth between the best performer (Equatorial Guinea)
and the worst (Democratic Republic of Congo). Moreover, the standard deviation of
country growth rates remains high, even when calculated on annual averages for the
entire decade.
The diversity of African growth performances calls for very cautious analysis,
in that averages can conceal many of the phenomena that interest us. For this reason,
we first calculated the simple average (i.e. without weighting by the relative economic
size of the various countries) and then the weighted average. Table 1.1 also reports the
weighted average excluding the South African economy, which is so large that it can
modify these calculations appreciably. Depending on the method of calculation used,
the results can differ by as much as a full percentage point of growth.
Most importantly, the diversity is so great that we cannot be content with
examining the average performances of Africa as a whole. For our purposes, studying
every country individually would not be a very practical solution because every case
is special, making it difficult to draw conclusions. For example, Equatorial Guinea
had the highest growth rate because it began to exploit its oil deposits, whereas the
economic disaster in the Democratic Republic of Congo is easily explained by the
country’s political situation.
We have chosen here to divide the African countries into four groups according
to their growth performances. Each group comprises 11 countries, on the basis of the
available data: the best performances (12 countries when South Africa is included),
moderately good performances, moderately poor performances and poor performances.
The list of countries in each group appears in Table 1.2, with their respective shares of
the regional population.
16
Table 1.2. Grouping of Countries by Growth Performance,
with their Shares in the Regional Population in 1995
(percentages)
Fast growers Moderately fast growers Moderately slow growers Slow growers
Benin 1.0 Burkina Faso 1.7 Gambia 0.2 Angola 2.0

Botswana 0.3 Chad 1.2 Kenya 4.8 Burundi 1.1
Cape Verde 0.1 Côte d’Ivoire 2.4 Niger 1.6 Cameroon 2.3
Equatorial Guinea 0.1 Ethiopia 10.0 Nigeria 19.6 Central Afr. Rep. 0.6
a
Eritrea 0.6 Gabon 0.2 Rwanda 1.3 Comoros 0.1
Ghana 3.0 Guinea 1.2 São Tomé + Pr. 0.0 Congo 0.5
Lesotho 0.3 Malawi 1.7 Senegal 1.5 Dem. Rep. Congo 7.8
Mauritius 0.2 Mali 1.7 South Africa 6.7 Guinea-Bissau 0.2
Mozambique 2.8 Mauritania 0.4 Swaziland 0.2 Madagascar 2.4
Sudan 4.6 Namibia 0.3 Tanzania 5.2 Sierra Leone 0.8
Uganda 3.4 Seychelles 0.0 Togo 0.7 Zambia 1.6
Zimbabwe 1.9
a. Included as from 1992.

Owing to the way they are constructed, these groups are somewhat heterogeneous.
It may be surprising to see Sudan appear in the first group, but this is due to the fact that
growth is measured here in volume terms. When the very sharp depreciation of the
Sudanese pound early in the decade is taken into account, per capita income in dollars
fell in Sudan over the period. This first group also includes countries whose growth is
the result of reconstruction after civil war (notably Eritrea, Mozambique and Uganda)
and countries whose growth was based more on improvement in economic fundamentals,
sometimes initiated long before (notably Botswana, Lesotho and Mauritius).
These groups are of comparable demographic size, except for the fact that Nigeria,
which by itself accounts for almost 20 per cent of the region’s population, necessarily
doubles the weight of its group. Each group includes countries of very different sizes.
In order to avoid giving excessive weight to very small countries, all the averages
calculated for the groups are weighted by an indicator of demographic size. The GDP
growth rates reported in Figure 1.1 are thus averages for each group, calculated by
weighting the individual countries’ average annual GDP growth rates by the previous
year’s GDP.
It may also be noted that none of the groups is marked by very strong geographical
concentration, although the worst performances tend to be located in central Africa
(with the notable exception of Equatorial Guinea, which registered the best
performances over the decade) and the good or moderately good performances are
concentrated in east Africa. Thus, while our grouping does follow geographical lines
to some extent, there were no areas of the continent that saw no progress at all.
17
Source: Authors' calculations based on World Bank (World Development Indicators, 2000).
Figure 1.1 shows clearly that the group of good performers had a high growth
rate right from the beginning of the decade, when the region as a whole was showing
very low growth. For Botswana and Mauritius (and to a certain extent Lesotho), this
surge of rapid growth began well before the 1990s; it also started in the 1980s in
Ghana and Uganda. At the same time, we will see in Chapter 2 that the countries that
grew slowly or moderately slowly in the 1990s have not always been poor performers.
The history of Africa since independence has been marked by a number of fast–growth
episodes which, unlike those of Botswana and Mauritius, have since broken off. Notable
examples include Cameroon, Kenya, Tanzania and Togo.
These observations raise two important questions: whether growth can be
sustained in the countries that are performing satisfactorily today, and whether other
countries will be able to follow the same path.
To investigate these questions, we will devote the following chapters to
comparative examination of the factors of growth in Africa. This examination gives
us the opportunity to comment in detail on the factors which can influence growth in
the six economies considered in Part Three of the book: two in the first group (Ghana
and Uganda), three in the second (Burkina Faso, Côte d’Ivoire and Mali) and one in
the third (Tanzania)1.
At this stage, some interesting conclusions can already be drawn from examination
of the average growth rates recorded by each group over the 1990s. It appears first of
all that the strong growth observed in the first group reached a peak in 1995 and
decreased thereafter. This trend calls for caution in the analysis, because growth has
slowed appreciably since 1995. On average, however, the growth rate of this group
remained over 5 per cent a year throughout the period, resulting in a substantial increase
in per capita GDP (see Figure 1.2).
18
Beginning in the middle of the decade, the growth performances of the other three
groups of countries improved to varying degrees. The second group, which until 1995
had an average growth rate two points lower than that of the leading group, caught up to
the latter (in terms of growth rate) in 1996. This trend can also be seen to some extent in
the other two groups. Consequently, the growth rates of the four groups showed some
tendency to converge towards the end of the decade. Although the economic growth
was not high enough to bring an increase in per capita income in a fair number of
countries (14 in 1998, representing 53 per cent of the regional population), this result is
encouraging in that it seems to indicate that certain countries began to imitate the fast
growers. This is the case in particular for a number of countries in the moderately fast–
growth group: Burkina Faso, Côte d’Ivoire, Ethiopia, Guinea, Mali and Mauritania.
Lastly, Figure 1.2 shows that the group of fast–growth countries, which started
from a level of per capita GDP comparable to the average, built a substantial lead over
the other groups in less than ten years: in 1998, their average per capita GDP exceeded
that of the region (excluding South Africa) by nearly 30 per cent. Conversely, the
group of countries with the poorest growth performances fell behind during the same
period: whereas in 1990 this group had reached a level more than 20 per cent higher
than the average, by 1998 it was lagging 10 per cent behind the average. These data
show that economic growth can, within a decade, have a significant impact on the
average standard of living of the population (we shall return below to the issue of
poverty reduction). On the assumption that the countries of this group achieve the
same growth rates in future as in the 1990s, they would need about 20 years to reach
a level of per capita GDP equivalent to $1 000 (1995 dollars). These figures illustrate
the extent to which real improvement in these countries’ living standards still depends
on achieving higher rates of growth.
19
The Macroeconomic Environment
To determine what kind of future growth performance can be envisaged for the
economies of the region, it is helpful to examine variables representing the
macroeconomic environment. This examination will be supplemented later in this
volume by a long–term econometric analysis emphasising certain factors of structural
change that are likely to be conducive to growth.
First of all, the growth performances of sub–Saharan African countries in the
1990s do not seem to be much affected by changes in the terms of trade, which may be
considered a good indicator of the quality of the international environment. Figure 1.3
clearly indicates that the precipitous fall in the region’s terms of trade in the 1980s
slowed considerably in the 1990s, but that the trend remained rather unfavourable. A
recovery began in 1995, however, which may explain part of the renewed growth
observed in many countries in that year.
When change in the terms of trade is compared by groups of countries, it becomes
clear that the fast–growth countries did not have a more favourable environment than
the others did in the 1990s (Figure 1.4). The international environment (where trade is
concerned, at least) does not then explain the differing growth performances of the
African economies. The causes of these differences are mainly domestic.
From a domestic standpoint, improved growth in the countries of the region was
accompanied by progress on other macroeconomic indicators. Monetary and budget
policies in particular seem to have been managed somewhat better over time.
Table 1.3 indicates that inflation was fairly high in the region early in the decade,
particularly in those rapidly growing countries which, like Ghana and Mozambique,
had experienced rather high rates of inflation in the past. These countries contained
the rise in prices fairly well, however, with the possible exception of Ghana, which
still had double–digit inflation at the end of the decade. Conversely, the slow–growth
countries made relatively little progress in this area. The two intermediate groups
already had relatively low rates of inflation in 1990 and kept them low thereafter,
partly owing to the fact that these groups include the great majority of the Franc Zone
countries, where the prevailing monetary policy precludes the risk of high inflation.
These observations suggest that initial control of inflation is neither a necessary
nor a sufficient condition for renewed growth in Africa. The case of the slow–growth
countries suggests, however, that persistent high inflation — at a time when the global
economic climate is favourable to deflation — is one among many indications of
governments’ inability to put appropriate macroeconomic policies in place.
Figure 1.5, which presents budget balances, suggests that the same observations
apply to budget policy. The first thing to strike the eye is an overall improvement
during the decade, but the link between control of the budget deficit and renewed
growth is, as in the case of inflation, far from automatic. Although the group of fast
growers had the best budget balances at the beginning of the period, from 1995 the
20
Source: World Bank, Africa Database.
21
Table 1.3. Consumer Price Inflation
(percentages)
1990 1998
Fast growers 33 9
Moderately fast growers 8 6
Moderately slow growers 9 11
Slow growers 33 24
Average 19 12
Notes: The groups are defined as above, according to their GDP growth performances over the 1990-98
period.
The group average is calculated by weighting the inflation rates of the various countries by their
1995 GDP.
Source: Authors’ calculations based on World Bank (World Development Indicators, 2000) and African
Development Bank (Selected Statistics on African Countries, 2000).
two intermediate groups performed higher on this score. Furthermore, the budget
performances of these groups were fairly comparable to that of the fast–growth countries
as early as 1990. Sound fiscal policy thus does not appear to be a sufficient condition
for a revival of growth. Once again, however, the group of slow–growth countries
posted poor results on this criterion, recording on average fiscal deficits of around
12 per cent of GDP at the beginning of the period and 7 per cent at the end.
On the whole, traditional macroeconomic policies do not afford a good
explanation of the region’s growth during the 1990s. The most “virtuous” countries
from the standpoint of controlling inflation and budget deficits were far from achieving
the best growth performances, although it is clear that the slow–growth countries were
also those with the poorest macroeconomic management. The serious macroeconomic
disturbances observed are more an indication of governments’ inability to implement
an economic policy programme than a cause of poor growth performances.
This does not mean, of course, that macroeconomic policy should be neglected.
We shall see on the contrary that the surge of growth in countries such as Ghana and
Uganda in the 1980s coincided with a return to sound macroeconomic policies. If
countries are to achieve sustainable growth, however, they will need to invest and, in
connection with this investment effort, to implement policies of structural change.
The role of policies aimed at bringing about structural change will be discussed
in Part III, where the individual countries’ achievements are examined in detail. Where
investment is concerned, however, the aggregate data in this chapter already provide
an idea of the role that this factor may have played in Africa’s growth.
Figure 1.6 clearly reveals an essential difference between the fast–growth
countries and the others: the former invested much more. The average investment
ratio of the fast growers (weighted by the GDP) was around 25 per cent, as against
less than 15 per cent for the other groups. Only the group of moderately good performers
achieved continuous progress over the period, reaching an investment ratio of 20 per
cent in 1997. This development should be associated with the faster growth which
these countries recorded beginning in the middle of the decade.
22
Notes: The groups are defined as above, according to their GDP growth performances over the 1990-98 period.
The group average is calculated by weighting the budget balance/GDP ratios of the individual countries by
their GDP for the current year (in constant 1995 dollars).
Note: South Africa is excluded from the third group
23
Uganda stands somewhat apart from the other fast growers in this respect, with
an investment rate comparable to that of the other groups. This casts some doubt on
this country’s chances of maintaining a high growth rate for long if it does not make
progress in the area of investment. At the same time, in the groups of countries with
lower growth rates, high investment rates are not always reflected in growth
performance, as for example in Congo (investment rate of the order of 30 per cent),
Swaziland and, until 1995, Tanzania. A high investment rate is thus neither a panacea
nor, at least in the short run, a necessary condition for growth. Differing investment
rates certainly cannot explain all of the differences in the growth rates of African
countries. In addition, the investment performances recorded are partly the result of
structural changes, notably the improvement of the institutional framework. Despite
all these qualifications, the data reported in Figure 1.6 provide a first clue to
understanding how growth can emerge in Africa.
The fact that these investments are not financed entirely from national resources
can weaken the economic growth process by making it dependent on contributions
from outside. This can be seen in particular in the balance of payments, through
comparison of exports and imports. Figure 1.7 shows the change in the ratio of average
exports to average imports for each group (measured in constant prices); it suggests
that little progress has been made in the balance between sources and applications of
funds. In particular, the group of fast–growth countries has the lowest such ratio
throughout the period: a large proportion of these countries’ investment was therefore
financed by outside resources. This is particularly true for Eritrea, Lesotho and
Mozambique. In contrast, Botswana and Mauritius managed more or less to balance
their absorption with their domestic resources.
Poverty
The renewal of growth in Africa, if it endures, does not necessarily mean a

reduction in poverty. Indeed, the fragmentary data available suggest that the fast–
growth countries had mixed results in terms of poverty reduction in the 1990s.
Poverty is difficult to estimate in monetary terms, for lack of statistics on changes
in income distribution over time. With a few exceptions, we have for the last two
decades no more than one observation per country of the number of persons below the
poverty line (whether this is measured according to national or international definitions).
The data thus will not allow a diagnosis of the decrease in poverty associated with
economic growth.
A comparison (Figure 1.8) of data on the proportion of the population living
below the poverty line and data on average per capita income (measured in terms of
purchasing power parity to avoid the biases introduced by cross–country differences
in price levels) shows however that economic growth does not automatically lead to a
reduction of poverty.
24
Notes: The proportion of poor people (y-axis) is based on national definitions of the poverty line.
Income per capita is measured in terms of purchasing power parity (1987 dollars) and observed in the same
year as the number of poor for each country (from 1984 to 1995, depending on the country).
Mauritius, which has a low poverty rate (approximately 10 per cent) and income per capita five times that of
the other countries on average, is not represented in the figure.
Source: Authors' calculations based on World Bank (Africa Database, 1998/99).
25
Figure 1.8 displays the decreasing relationship between average per capita income
and the percentage of the population living below the poverty line. It also shows that,
for a given average income, there can be very great differences in performance between
countries.
Some non–monetary poverty indicators allow us to verify that in the countries,
which performed well during the 1990s, growth did not always lead to poverty
reduction. First, the indicator of life expectancy at birth presented in Table 1.4 does
not establish a direct link between economic performance and increased life expectancy
of the population, although the group of fast growers registered the greatest progress
on average, with a gain of about one year of life expectancy during the past decade.
The rather modest progress achieved in this respect (on average, life expectancy
for the region as a whole did not progress) is due in particular to the AIDS/HIV
pandemic. Many of the countries have very high rates of HIV infection. In 1997, the
HIV–positive proportion of the adult population exceeded 20 per cent in two countries
(Zimbabwe and Botswana), 10 per cent in 12 countries (representing one–fourth of
the regional population) and 5 per cent in 20 countries (representing nearly half of the
regional population).
Although the long–term consequences are still difficult to predict, it is clear that
this phenomenon has a significant impact on life expectancy in sub–Saharan Africa.
Thus Table 1.5, which disaggregates the data of Table 1.4 according to the average
rate of HIV infection in the population, shows that life expectancy fell on average by
5 per cent in the hardest–hit countries, whereas it increased by 5 per cent in the other
countries, which started from a comparable initial level.
This table also shows that when the data are controlled for the average incidence
of HIV infection in the population, a more coherent relationship between economic
progress and life expectancy is found. This relation is not linear, however, as it allows
us only to distinguish the slow–growth group of countries, which performed poorly in
terms of life expectancy, from the other countries, which showed fairly uniform
improvement in life expectancy. Once again, there is no automatic link between
economic growth and poverty reduction.
Table 1.6, which presents change in infant mortality rates, allows a more
optimistic view regarding the health improvements that may be associated with
economic growth: it suggests that growth has been accompanied by a proportional
reduction in infant mortality over the past decade. The infant mortality rate fell on
average by one percentage point a year between 1990 and 1998, and by more than
1.5 points in the fast–growth countries (as against 0.7 points in the slow–growth group).
Social progress in the field of education is more difficult to analyse in connection
with economic growth because such progress — even more than with health — is the
result of a long–term process that begins with the provision of schooling for children
and the consequences of which can be seen only on the scale of one or more generations.
26
Table 1.4. Life Expectancy at Birth and HIV Infection
Countries with rate of adult HIV infection < 5 per cent
Year 1990 1998 % Increase
Fast growers 53.3 56.6 6.2

Moderately fast growers 46.8 49.4 5.5
Moderately slow growers 49.2 52.9 7.5
Slow growers 50.7 51.1 0.8
All countries 50.2 52.7 5
Countries with rate of adult HIV infection > 5 per cent
Year 1990 1998 % Increase
Fast growers 46.3 44.1 -4.7

Moderately fast growers 45.9 43.7 -4.8
Moderately slow growers 55.7 53.5 -3.9
Slow growers 47.3 43.5 -8
All countries 50.4 48.0 -4.8
Table 1.5. Life Expectancy at Birth

(average weighted by the 1995 population)
Year 1990 1998 Increase
Fast growers 50.4 51.5 2.2%

Moderately fast growers 46.1 45.0 -2.4%
Moderately slow growers 52.3 53.2 1.7%
Slow growers 50.1 49.6 -1%
All countries 50.3 50.5 0.4%
Table 1.6. Average Infant Mortality Rate

Year 1990 1998 Decrease
Fast growers 99.0 86.5 12.6%

Moderately fast growers 119.9 106.5 11.2%
Slow growers 105.0 99.5 5.2%
All countries 96.7 88.4 8.6%
27
Table 1.7 thus shows no particularly significant progress in basic literacy training
of the population in fast–growth countries compared to the other countries. In fact,
illiteracy decreased in most of the countries, suggesting that African countries’ efforts
to increase literacy have not yet brought about a steady state in this respect.
Under these conditions, the main explanatory variable for the reduction of illiteracy
rates during the period considered is the initial rate of primary school enrolment. This is
clearly apparent in Figure 1.9, where the rate of change in illiteracy from 1990 to
1998 is plotted against the average rate of primary school enrolment in 1990.
This relationship demonstrates why the group of fast growers made rather little
progress in raising literacy rates during the past decade. At the beginning of the period,
these countries had on average a rate of primary school enrolment of the order of
67 per cent (Table 1.8), well below that of the slow and moderately slow growers.
Similarly, the fact that the moderately fast growers posted the weakest performances
with respect to increasing literacy in the 1990s is explained by these countries’ low
school enrolment rates, which averaged 43 per cent.
Poverty reduction through literacy training is thus the result of a long–term
investment. From this point of view, current enrolment rates are a good leading indicator
of African countries’ potential future progress in reducing poverty and also (as we
shall show later in this study) in sustaining growth.
The data reported in Table 1.8 are somewhat discouraging where the fast–growth
countries are concerned. Their primary school enrolments increased hardly at all from
1990 to 1996. It is possible, however, that the shortness of the period considered (we
did not have the data needed to study a longer period) results in an overly pessimistic
assessment of the situation in these countries. In particular, Uganda, which accounts
for 20 per cent of the population of the fast–growth group of countries, had stagnant
primary enrolment rates during the period considered, but introduced a nearly universal
system of primary education after 1996, raising these rates from 74 per cent in 1996 to
close to 100 per cent today.
Table 1.8 also shows that the countries with moderately fast growth made
appreciable progress in primary schooling, allowing them to catch up to some extent
(they nevertheless remain considerably behind the other groups). This group includes
most of the countries having a low rate of school enrolment, notably Burkina Faso,
Ethiopia and Mali, which had enrolment rates of around 40 per cent in 1996.
Finally, this indicator once again confirms the social crisis in the slow–growth
countries, as these countries — namely Burundi, Cameroon, the Democratic Republic
of Congo and Madagascar — experienced a fall in their primary school enrolment rates.
Finally, we consider the UNDP’s composite indicator of human development,
which provides an overall view of the improvement in living standards in the countries
considered. Table 1.9, which reports the change in this indicator from 1990 to 1998,
suggests that there was appreciable improvement in the fast growers (3.1 points) and
moderately fast growers (2.5 points), very slight improvement in the moderately slow
28
Table 1.7. Illiteracy Rate
Year 1990 1998 Decrease
Fast growers 51.2 42.0 18%

Moderately fast growers 69.6 60.9 12%
Moderately slow growers 42.6 33.1 22%
Slow growers 47.6 37.6 21%
Average 50.4 41.0 19%
Table 1.8. Gross Rate of Primary School Enrolment

Fast growers 67.0 67.4 0%

Moderately fast growers 43.4 57.5 33%
Moderately slow growers 91.3 88.2 -3%
Slow growers 84.5 77.2 -9%
Average 76.0 76.3 0%
Figure 1.9. Relation betweeen Progress in Literacy and School Enrolment
29
Table 1.9. Human Development Indicator

Fast growers 41.8 44.9 7.4%
Moderately fast growers 32.4 34.9 7.7%
Slow growers 49.9 47.7 –4.4%
Overall average 43.9 44.9 2.3%
Note: Owing to lack of data for 1990, this table omits Eritrea, Equatorial Guinea, Sudan (first group), Gabon, the
Seychelles (second group), Sao Tomé and Principe, Tanzania (third group), Sierra Leone and Madagascar (fourth
group).
growers (0.6 points) and a sharp drop in the slow–growth countries. Thus there is
indeed a link, although not an automatic one, between economic growth and the
improvement in living standards recorded in a certain number of sub–Saharan African
countries during the 1990s.
It can also be seen from Table 1.9, however, that in 1998 the countries
experiencing fast and moderately fast growth still had an average level of human
development lower than that of the other two groups, even though their spurts of
growth had given them, on average, substantially higher levels of per capita GDP (as
shown by Figure 1.2). This result is due to the effect of averaging within the fast–
growth group: countries whose growth surges are relatively recent are averaged with
others whose growth episodes occurred much earlier. Countries such as Eritrea,
Mozambique and Uganda, which have managed to revive their economies after years
of crisis and conflict, are still lagging well behind in terms of human development. By
contrast, Botswana, Cape Verde, Ghana, Lesotho and Mauritius, which have enjoyed
peace and economic progress over a longer period, have above–average levels of
human development. The same holds for Namibia and the Seychelles in the group of
moderately fast growers.
Poverty in sub–Saharan Africa is thus rather persistent. Although rapid growth
does reduce poverty in the countries of the region, it takes well over ten years for the
effects to be visible enough to allow the best–performing countries to overcome their
initial handicap in human development.
These conclusions are confirmed by the UNDP poverty indicator, which is
strongly correlated with the human development indicator. On average, according to
this indicator, 41 per cent of the population was living below the poverty line in the
fast–growth countries and 52 per cent in countries with moderately fast growth, as
against 35 per cent in countries having slow and moderately slow growth (these
averages are based on a smaller sample of countries, owing to lack of data for Burundi,
Rwanda, Chad and Guinea).
30
Conclusion
This descriptive analysis of the economic and social progress achieved by the
countries of sub–Saharan Africa in the 1990s shows that the region saw a great variety
of individual performances and that a significant number of countries posted good
results during this period.
This progress is undoubtedly attributable to a host of factors. Traditional
macroeconomic policies played a role, as poor macroeconomic policies were usually
accompanied by economic decline. As a result, we may consider that the restoration
of sound macroeconomic policies in countries where the initial situation was very
poor contributed to the revival of growth, as happened in Uganda from 1987, in Ghana
from 1983, and in Côte d’Ivoire, Burkina Faso and Mali (and the other Franc Zone
countries) from 1994.
Nevertheless, the driving forces of sustained growth lie elsewhere. We showed
that using the investment rate as an indicator reveals a very sharp distinction between
the fast–growth countries and the other groups. Uganda is the only real exception in
this group, which casts some doubt on this country’s chances of seeing continued
growth. We shall also show below that investment in training and other policies aimed
at fostering structural change in the economy form part of the recipe for rapid growth.
Our findings are less clear where poverty reduction is concerned. The AIDS
epidemic is seriously undermining progress in health in a large number of countries,
which account for half of the region’s population. Efforts to increase school enrolments
have been falling off, except in the countries where the initial situation was particularly
bad and more recently in Uganda. As a consequence — apart from some noteworthy
exceptions such as Mauritius and the Seychelles, which have only small shares of the
region’s population — most of the African countries, including those that experienced
renewed growth in the 1990s, are still near the bottom of the UNDP’s human
development ranking. Among the fast–growth countries, only those that have enjoyed
rapid growth for more than ten years have a human development index above the
regional average. From the standpoint of alleviating poverty, then, it is essential to
consider the sustainability of economic progress.
Note
1. See Bibliography. This project also provided the opportunity for a case study of the
country that posted the worst performance, the Democratic Republic of Congo (see Maton,
2001).
31
32
Chapter 2
Factors of Economic Take Off in Africa
Long–term analysis shows that Africa has seen many episodes of sustained rapid
growth since the 1960s. Some of these have come to an end, and a smaller number of
others are still in progress. This chapter is mainly devoted to examining the sources of
long–term growth in Africa. It also presents a comparison of the countries used as
references for the purpose of formulating scenarios of future growth in the light of
past growth experiences.
We begin with an inventory of sustained growth experiences in Africa over the
past decades. For the purposes of this review, we have included the North African
countries which, in a long–term perspective, may constitute examples from which
sub–Saharan African countries at a substantially lower stage of development can draw
some inspiration. We also use a fairly broad definition of rapid growth, considering
the cut–off point to be an economic growth rate high enough to guarantee growth of
per capita income. In practice, however, the uncertainty of the data on population
growth leads us to take GDP growth of 3.5 per cent as our cut–off point for all countries,
instead of defining a threshold in terms of per capita GDP.
We subsequently study the characteristics of growth in these countries by
estimating a production function using panel data collected for a broader sample of
27 African countries. This estimation is strongly geared towards consideration of the
factors likely to influence total factor productivity (TFP) in the long term. We are thus
in a position to determine some of the main components of African growth, where
such growth has been observed. This analysis also allows us to compare these
experiences with the performances of the countries for which we will construct growth
scenarios, and which will be examined in greater depth in subsequent chapters.
Rapid Growth Episodes in Africa since 1960
The period studied (1960–96) has seen a number of examples of sustained rapid
growth in Africa. As mentioned above, the aim of this volume is to make a comparative
study of these extended growth periods. We define an extended period of strong growth as
33
an uninterrupted period of ten years or more during which the five–year moving average
of annual GDP growth met or exceeded 3.5 per cent. By using a sufficiently long period of
time, we exclude countries experiencing transitory surges in growth due to favourable
external factors, such as a temporary improvement in the terms of trade or increased demand
due to the economic cycles in the global economy. We use a five–year moving average as
opposed to annual growth rates for a similar reason: to avoid having to exclude strong
performers that experience a temporary slowdown in their growth records. The only
exception to this definition is that we accept countries with a strong growth record which
continues beyond 1996, even if by 1996 this growth period was one or two years short of
the ten–year cut–off requirement. This exception is applicable only to Uganda and
Mozambique. Once the countries have been selected, the start of the growth period is
considered to be the first year (included in the five–year average) for which GDP
growth met or exceeded 3.5 per cent. Similarly, for the interrupted growth
experiences, the period is considered to end with the first year of growth under
3.5 per cent within the last five–year average. Table 2.1 summarises all eligible
growth periods, selected from a large sample of 44 African countries according to
the definition given above.
A few of the countries in Table 2.1 have seen renewed growth in recent years
(Côte d’Ivoire, Namibia, Tanzania and Tunisia). Mauritius showed buoyant growth in
the early 1970s, but this came to a temporary halt towards the end of the decade,
presumably as a result of the short–term costs of structural adjustment. Other countries
(Lesotho, Kenya and Morocco) have experienced relatively strong but uneven growth
outside of the periods shown in Table 2.1. In addition, the date given for the end of
Egypt’s rapid growth period is not clear–cut, since the growth rate has been rather
volatile — and sometimes high — in the 1990s.
Some countries (Ethiopia, Gabon, Lesotho, Namibia, Togo and Tanzania) or parts
of extended growth periods in other countries (Botswana before 1970; Algeria, Egypt
and Tunisia before 1965; South Africa before 1961) are eliminated from the analysis
below, owing to the lack of other data needed to evaluate and analyse TFP performances.
However, the countries selected as benchmarks for our analysis of future emergence
prospects in subsequent chapters — Burkina Faso, Côte d’Ivoire, Mali and Tanzania —
are included in the analysis, even though they have not recently experienced any long
period of fast growth. In these four countries, rapid growth has been observed only since
1994 (for the three Franc Zone countries) or 1995 (Tanzania). The periods of observation
used for the countries in this chapter are very short (respectively two years and one
year), because our database for the econometric exercise is complete only until 1996,
but the most recent data available (WDI, 2000) show that the countries kept growing at
a rather fast pace, well above 3.5 per cent a year, until 1998. Their experience therefore
provides interesting examples of countries that might follow the route traced by early
reformers which have grown rapidly in the 1990s, such as Ghana and Uganda (the other
two countries selected for our scenario exercise).
34
Table 2.1. Sustained Strong Growth Experiences in Africa
Country Start End Length Average

of period growth
Algeria 1962 1985 23 5.2

Botswana 1965 31+ 9.3
Cameroon 1967 1986 19 7.0
Côte d'Ivoire 1960 1978 18 9.5
Egypt 1960 1990 30 6.6
Ethiopia 1960 1972 12 4.5
Gabon 1965 1976 11 13.1
Ghana 1983 13+ 4.8
Kenya 1961 1981 20 6.7
Lesotho 1970 1982 12 9.9
Malawi 1964 1979 15 6.6
Mauritius 1980 16+ 5.5
Morocco 1966 1980 14 5.9
Mozambique 1986 10+ 6.2
Namibia 1961 1979 18 6.4
South Africa 1960 1974 14 5.1
Tanzania 1961 1975 14 5.7
Togo 1960 1974 14 6.8
Tunisia 1960 1985 25 5.8
Uganda 1986 10+ 6.6
Note: Figures in the last column are logarithmic growth rates over the periods defined by start and end dates as
indicated above; + indicates that the growth period continues after 1996.
Source: Authors’ calculations from World Bank data (African Development Indicators, 2000).
Analytical Framework
Most of the recent comparative literature on African growth is based on cross–

section or panel data equations which explain growth rates in a conditional convergence
equation (see e.g. Collier and Gunning, 1999, for a synthesis). Interpreting these
equations as long–term relations is appropriate only if the economies in question can
be assumed to be close to their steady states, as in Mankiw et al. (1992). For African
economies, this is a very strong assumption. We therefore prefer to adopt an alternative
approach, which consists in estimating a level equation, in which GDP per unit of
labour is related to capital per unit of labour and to variables explaining total factor
productivity. In order to show that such an equation can be considered as a long–term
relation, co–integration tests have been performed. We also estimated an error
correction model based on this long–term equation, which is similar to a growth
equation but explains only short–term growth movements1. Lastly, our approach
differs from the most common approach in that we decided to estimate our model on
African data only, rather than on a larger sample; by doing so, we avoid the risk of
mixing countries which have very different growth behaviours and possibly
heterogeneous parameters.
35
The core of our analysis is therefore a production function explaining the long–
term relation between income on one hand and labour, capital and productivity variables
on the other. We assume constant returns to scale and thus obtain a co–integrated
production function having the following form:
where Y is GDP, L labour, K the capital stock and TFP total factor productivity, which
is determined by the following vector of variables:
— the black market premium in the exchange rate market (with CFA countries
distinguished from non–CFA countries), which is used as an index of domestic
price distortions;
— a human capital stock series, defined as the average number of years of schooling
in the working population;
— imports divided by labour;
— an index of the effect of labour reallocation on aggregate productivity;
— an index of economic diversification (see below);
— a dummy variable for revolutions and coups d’état;
— a country dummy variable, which takes account of cross–country productivity
differentials;
— a country–specific determinist trend, accounting for differences in exogenous
productivity growth among countries.
Most of our independent variables have been used in several previous studies
and therefore require only a brief theoretical explanation. In particular, the black market
premium, the human capital stock, the export ratio and the dummy for conflicts have
been used extensively in growth regressions.
The first variable is conceived of as a proxy for the implementation of
macroeconomic adjustment programmes. Market distortions, as measured by the black
market premium, can be expected to impede efficient allocation of resources and thereby
hamper productivity2.
The next four variables represent factors of structural change. The role of human
capital has been stressed by Nehru et al. (1993), Edwards (1998) and Romer (1990),
to mention only a few. Considering the very low level of human capital accumulated
to date by African countries, policies to increase and improve education should play a
major role in growth prospects through this variable.
36
Theoretical and empirical evidence of the influence of openness on productivity
has been provided by Feder (1982), de Melo and Robinson (1990), Tybout (1992),
Biggs et al. (1995), Sachs and Warner (1997) and Edwards (1998), among others. In
previous literature, TFP gains are assumed to derive from external effects such as
exposure to foreign competition, technology transfer and economies of scale, or from
faster convergence with richer countries. Some observers (see e.g. Dessus, 1998) have
argued that TFP is more dependent on imports than exports, imports playing the role
of a production factor. Accordingly, we preferred to introduce the ratio of imports to
labour, rather than exports to GDP or another measure of openness, in order to stick to
the interpretation of external trade flows as a quasi–production factor, in the framework
of a CRS production function.
Including the effect of reallocation of production factors is a common feature
in growth theory, although this effect is rarely introduced in empirical work. It was
extensively used, however, by Chenery et al. (1986), who showed that it was an
important factor in explaining growth performances. A more recent contribution
using this effect is Young’s (1995) work on East Asia. According to Young, most
TFP gains in East Asia from the 1960s to the early 1990s derived from inter–sectoral
reallocation of labour. In fact, non–agricultural and manufacturing employment
increased one–and–a–half to two times as fast as the aggregate working population.
Poirson (1998) also stresses that in most rapidly growing countries, growth is
accompanied by significant positive reallocation effects, based on labour movement
from agriculture (where labour productivity is typically low) to non–agricultural
sectors3.
One might think that structural adjustment programmes have adverse
reallocation effects in the short run. Given that agriculture is the principal
exportable good in African countries (with the exception of oil–exporting countries),
the restoration of a price system that is less distorted in favour of importables and
non–tradables should provide incentives to move factors into agriculture. This would
mean that the Syrquin effect and structural adjustment programmes could have
opposite consequences for growth. However, this line of argument, based on standard
trade theory, is not quite robust. The restoration of a price system closer to
international prices means that the economy moves along its transformation curve,
but in standard trade theory such movement is analysed under the assumption that
factors — particularly labour — are perfectly mobile across sectors. This assumption
does not fit in with the factor reallocation approach, where it is assumed that factors
are only partially mobile — the economy is characterised by dualism — which is
the very reason why large differences in factor productivity can be observed across
sectors. The labour reallocation effect considered in the calculation of the Syrquin
index therefore represents an expansion of the transformation curve, rather than a
movement along an immobile transformation curve4.
37
It would be beyond the scope of this work to introduce a complete index of
reallocation effects (including inter–sectoral movement of both labour and capital),
given the difficulty in finding the detailed time–series information on the sectoral
distribution of capital that this would require. Experience shows, however, that the
labour reallocation effect is substantially higher than the capital reallocation effect,
when the latter can be measured (see e.g. Dessus et al. 1995, on Chinese Taipei).
Moreover, within the labour reallocation effect, the most significant element is
movement from agriculture to the non–agricultural sector. In our empirical application,
we introduce only this effect, which is defined by the following equation, adapted
from Syrquin (1986):
where ρ t is the TFP gain due to labour reallocation from agriculture to non–agriculture
sectors at time t, l i,t is sector i’s share of the total labour force and υ i,t is sector i’s
contribution to GDP. A level index of the effect of sectoral labour allocation is then
computed by calculating cumulated annual increments. This index is one explanatory
variable of TFP, with a theoretical parameter equal to 1.
In principle, the reallocation effect can be tested empirically (see e.g. Poirson,
1998), but the available time–series information is somewhat sketchy. The World Bank
time series that we use are merely interpolations based on (at best) one estimate every
five years. Therefore, instead of estimating the parameter for the labour reallocation
effect, we have set it at its theoretical value of 1 (with α — the share of capital in
production — in the equation used to compute ρ t being set at its robustly estimated
value of 0.45)5.
The other main originality of our study is the inclusion of an index of diversity
(see annex). Although moving factors from a low–productivity sector to a higher–
productivity sector enhances TFP, this is not necessarily the only impact of structural
changes on TFP. In this chapter, we also attempt to test whether a diversification of
economic activity has an impact on TFP, diversification being defined as the spreading
of production to a growing number of different outputs which do not necessarily imply
different productivity levels. The reason for testing the impact of diversification on
productivity is empirical: it derives from the observation that rapid economic growth
seems to be accompanied by a higher degree of diversification (in our sample, Mauritius
provides an illustration of this, to be contrasted with the absence of progress in
diversification in South Africa’s industry after 1960).
The impact of diversification on income may be transmitted through two main
mechanisms. The first is the idea that diversification in itself may enter as a production
factor by increasing the productivity of both labour and human capital, as in the now–
standard model developed by Romer (1990). In Romer’s model, the economy is divided
into three sectors: a final goods sector, an intermediate goods sector and a research
38
sector. The research sector uses human capital and common accumulated knowledge
in order to produce new designs, which it sells (or rents) to the producers of intermediate
goods; in a developing economy, even if there is in fact no R&D sector, imitation
activities may play the same role as the innovation activities of the developed countries.
Lastly, the final goods sector acquires intermediate goods in order to produce goods
for consumption. A crucial point in this model is that the diversity of intermediate
inputs enhances productivity in the final goods sector. This technological assumption
that diversity enhances productivity may be indirectly tested by studying the impact
of production diversity within an economy 6. This approach has recently been
generalised by Feenstra et al. (1999), who show that diversification of output may be
analysed according to a similar line of argument: a more diversified output mix means
a larger transformation set, for reasons of convexity.
The second mechanism through which diversification can increase income is by
expanding the possibilities for spreading investment risks over a wider portfolio. In
other words, greater diversification will enhance average capital productivity in the
long run by providing better investment opportunities at lower risk. Acemoglu and
Zilibotti (1997) present a model in which lack of diversification leads economic agents
to invest in safe, low–return, traditional projects rather than in riskier projects with
higher growth potential. The absence of possibilities for spreading risk by investing in
a diversified high–growth portfolio will hamper capital productivity in the short run
and capital accumulation in the long run (see Box 2.1).
The ideal measure of diversification would include data on production of all
goods and services in the economy. Since GDP data are not available at a sufficiently
detailed level, we use the composition of exports to the OECD countries as a proxy
for the diversification of the economy as a whole8. This approach has the weakness of
not taking into account the diversification of non–tradables, especially services, but
there is no a priori reason to assume that this will bias our results in any particular
direction. The diversification index is calculated as:
where xi,t is exports of product i (at the three-digit level) in year t and Xt is total exports
in year t. The inverse of Div takes on a maximum value of 1 when a single product
accounts for all exports, and it tends towards 0 when there is an infinite number of
equally weighted export products. In other words, the value of Div itself starts at unity
— for the case of complete specialisation where a country concentrates all its exports
on one product — and increases with the degree of diversification. The index excludes
exports of combustibles such as petroleum products and natural gas, in order to limit
the mechanical impact of terms-of-trade shocks. For instance, if oil is included in the
index, a sharp increase in oil prices, exemplified by the oil crises of the 1970s, will
automatically lead to an increase in the relative importance of the oil sector in the
economy without necessarily reflecting any structural change9.
39
Box 2.1. The Impact of Economic Diversification on Investment Levels
In the model developed by Acemoglu and Zilibotti (1997), a low level of diversification
deters economic agents from diversifying their investments. It follows that these agents will
then invest in low–risk projects, which probably also have low productivity. In addition,
microeconomic indivisibilities do not allow for a wide dispersal of capital. Lastly, poor countries
can initiate only a limited number of projects. In this context, diversification becomes
endogenous and constitutes the principal motor of growth7.
A significantly simpler model than that of Acemoglu and Zilibotti can be used to illustrate
this point. Assume a representative agent, maximising profit subject to a certain aversion to risk:
where Ei (θ ) is the expected value of profits from project i given the level of risk, θ , Vi (θ ) is
the variance of the profits as a function of risk, β is a constant measuring the degree of risk
aversion of the agent and N is the number of projects available to the agent for investment. If
all projects are equivalent, the maximising problem becomes:
which implies:
Both the expected value and the variance of profits can be assumed to increase with
the level of risk. Hence:
It is also reasonable to assume decreasing marginal returns to risk.
Moreover, the risk itself can be defined as the variance of the outcome of a project,
which is to say . This implies, from the solution of the maximisation
problem:
Since and β is a constant, θ must increase as N increases. In other words, agents

will invest in riskier, and on average more profitable, projects if they are able to spread the
risk through a more diversified portfolio.
40
The estimation proceeds as follows. First, we estimate the production function
defined as GDP per unit of labour, as a function of the capital/labour ratio and of the
various determinants of TFP discussed above. This equation can be considered as a
long–term (co–integration) relationship, describing the determinants of potential output
of the economies, inasmuch as the dependent variable and explanatory variables are
I(1), while the regression residual is stationary.
In Chapter 3, we investigate the determinants of capital accumulation (or rather,
the variation in the capital/labour ratio). Capital accumulation will depend on factors
from two categories. The first category is that of variables influencing the capacity to
finance investments, with particular emphasis on the role of foreign aid, be it through
grants or through net debt flows. The second is factors affecting incentives to invest.
These variables include infrastructure, risk and the overall efficiency of the economy,
measured by the estimated TFP from the long–term production function. The idea is
that a low productivity level implies a low return to capital, which means low incentives
for investment. Moreover, a low productivity level implies high transaction costs,
which reduce the profitability of investments in the economy as a whole. In this way,
productivity gains or losses prove to have a double effect on the economy: a direct
effect on growth as well as an indirect effect by modifying investment incentives. The
investment function estimated therefore takes the following form:
where dll is the growth of the active population, debtaid is the flows of debt and aid,
ToT is the terms of trade, TFP is the estimated productivity level (from the production
function), risk is a variable capturing country risks and roads measures the availability
of physical infrastructure. The interaction between investment and productivity
underlines the importance of a growth path based on both capital accumulation and
productivity gains.
Estimation of the Production Function
A number of authors have already estimated production functions based on panel

data (e.g. Collier and Gunning, 1999, on African data). It is often impossible to estimate
a production function on a country time series for African economies, for lack of
sufficient information (e.g. attempts on data for Senegal in Berthélemy et al., 1996).
Using a panel data set combining cross–section and time–series information leads to
substantially better econometric results. This is the approach we will use here, with
the understanding that there are some parametric differences among countries.
41
Unit root tests have been performed, following the method proposed by Levin
and Lin (1993). This method consists in computing a Dickey–Fuller statistic aggregated
across countries. The dependent variable (GDP/labour), the capital stock divided by
labour, the diversification index and the export/labour ratio are I(1), while the black
market premium is stationary. Moreover, although they are not conclusive, our tests
appear to indicate that the human capital variable is I(1).
The results reported in Table 2.2 were obtained under the assumption that the
production function exhibits constant returns to scale, with a log–linear specification.
Testing this hypothesis was impossible due to the high degree of correlation among
labour, capital and trend series. The least–squares dummy variable method (fixed–
effects method) used here appeared to be preferable to the random–effects method,
according to the Hausman test.
Table 2.2. Panel Data Estimates of the Production Function

Dependent variable: LYLA
Variable Coefficient Standard error t-statistic
LKL 0.397 0.032 12.60

LH 0.251 0.044 5.75
LDIV 0.043 0.013 3.48
LML 0.123 0.015 8.19
LBMPCFA 0.007 0.156 0.05
LBMPNCFA -0.041 0.009 -4.76
REVCOUP -0.014 0.007 -2.13
Estimation method: within (fixed effects)

Number of observations: 760
Number of countries: 27
Adjusted R squared: 0.99
Hausman test: χ2(8) = 1 001
Note: LYLA=ln(GDP/labour)-ln(reallocation effect, see above), LKL=ln(capital stock/labour), LH=ln(average number of

years of schooling in working population), LDIV=ln(diversification index), LML=ln(imports/labour),
LBMPCFA/NCFA=ln(1+black market premium in foreign exchange market) for CFA and non-CFA countries
respectively, REVCOUP=dummy for revolutions and coups d’état. Trends and fixed effects are not reported.
We find an elasticity of GDP to capital equal to 0.4, which seems plausible given
the extremely low level of capital endowment of African economies. We also find a
positive and significant impact of human capital on GDP, with a rather high elasticity
(0.25). The effect of the black market premium proved to be negative and significant
for the non–CFA countries and non–significant for the CFA countries. The lack of
significance of the black market premium for the CFA countries is logical in view of
the guaranteed convertibility of their currency, supported by France. Conflicts have a
negative impact on GDP, as can be observed in the negative and significant coefficient
of the dummy for revolutions and coups. Finally, the impact of imports divided by
labour is positive and significant for labour productivity. All in all, the properties of
this estimated equation look reasonable.
42
Unit root tests performed on residuals of the equation reported above show that
these residuals are I(0). This equation can therefore be interpreted as a long–term co–
integration relation.
Growth Accounting
Of the 27 countries making up the database that we assembled to perform the

preceding production function estimation, 14 have experienced rather long periods of
fast growth. We add to these the 4 countries whose recent strong performances are
under review in this study: Burkina Faso, Côte d’Ivoire (henceforth called Côte
d’Ivoire II, to distinguish this recent fast–growth episode from that observed from
1960 to 1978, called Côte d’Ivoire I), Mali and Tanzania. Table 2.3 reports growth
performances for the 18 countries as well as the contributions of capital and TFP to
growth for the relevant periods.
We report the growth contribution of labour productivity (GDP/labour) rather
than GDP, because in our sample of growth episodes the increase in the working
population plays a significant role, and a role that differs from country to country.
Although not the standard way of presenting growth accounting, this method is
preferable because it cancels out the consequences of vast differences in population
growth, which would bias country comparisons.
Table 2.3 reveals a distinct difference between the current growth periods and
the earlier episodes that ended in the 1970s or the 1980s: the early growth episodes
relied much more on capital accumulation than the current growth periods. In Algeria,
Cameroon, Côte d’Ivoire I, Egypt, South Africa and Tunisia, capital deepening
explained roughly two–thirds of the growth in GDP per capita, while the corresponding
figure was around 100 per cent in Malawi and Morocco.
Although the share of capital accumulation in the total contribution to growth is
somewhat lower in Kenya, there are similarities with the other earlier growth periods
in the sense that Kenya relied on a relatively high investment ratio (24 per cent of
GDP on average), which could not be easily sustained. This high investment ratio did
not lead to the same degree of capital deepening as in the countries analysed above,
because the initial capital ratio was much higher in Kenya10 than in the other economies
considered. Moreover, Kenya’s labour force grew at a very high rate.
In the current growth periods, capital accumulation has accounted for only about
13 per cent of growth on average, as opposed to 69 per cent for the earlier periods.
Ghana and Uganda show declining or stagnant capital ratios. In other words, their
growth processes rely entirely on productivity, while capital accumulation does not
contribute at all. The situation is similar in Mozambique, although investment has
played a somewhat larger role there than in Ghana and Uganda. The only countries in
the “current growth period” group that rely on both capital accumulation and
productivity gains are Botswana and, to a much lesser extent, Mauritius.
43
Table 2.3. Growth Accounting for Selected Economies
Memo item: Contribution to
Average annual growth (%) growth (%)
GDP Labour GDP/ Capital/ Capital/
Country Period TFP TFP
growth growth labour labour labour
ratio
Failed take offs

Algeria 1965-85 5.6 3.3 2.3 3.1 0.9 60.6 39.4
Cameroon 1967-86 7.0 2.0 4.9 6.8 1.9 61.8 38.2
Côte d’Ivoire I 1961-78 9.5 3.1 6.4 7.9 2.8 55.7 44.3
Egypt 1965-90 6.4 2.3 4.1 6.3 1.3 68.9 31.1
Kenya 1961-81 6.7 3.3 3.4 2.5 2.3 33.1 66.9
Malawi 1964-79 6.6 2.5 4.1 8.6 0.2 94.5 5.5
Morocco 1966-80 5.9 4.4 1.5 3.8 -0.2 110.3 -10.3
South Africa 1961-74 5.3 2.8 2.5 3.7 0.8 66.3 33.7
Tunisia 1965-85 5.7 3.3 2.4 3.8 0.7 71.4 28.6
Average 6.5 3.0 3.5 5.1 1.2 69.2 30.8
Current long growth periods

Botswana 1970-96 10.3 3.1 7.2 8.4 3.4 52.4 47.6
Ghana 1983-96 4.8 2.9 1.9 0.0 1.9 -0.4 100.4
Mauritius 1980-96 5.5 2.1 3.4 1.4 2.8 18.0 82.0
Mozambique 1986-96 6.2 1.9 4.3 1.2 3.7 12.9 87.1
Uganda 1986-96 6.6 2.5 4.0 -1.4 4.7 -16.1 116.1
Average 6.7 2.5 4.2 1.9 3.3 13.3 86.7
Recent growth experiences

Burkina Faso 1994-96 4.7 2.1 2.6 0.2 1.7 34.6 65.4
Côte d’Ivoire II 1994-96 6.7 2.3 4.4 -0.9 4.7 -6.8 106.8
Mali 1994-96 5.0 2.7 2.3 2.6 1.4 39.1 60.9
Tanzania 1995-96 7.2 2.9 4.3 -3.0 4.2 2.3 97.7
Note: Logarithmic growth rates.
Source: Authors’ calculations.
The moderate contribution of capital accumulation in Mauritius can be partly

explained by the fact that the country invested substantially prior to its take off.
According to our data, Mauritius’s capital stock increased by nearly 5 per cent annually
on average during the decade preceding the period studied. Moreover, the trend of the
investment rate has again increased recently, from 20 per cent of GDP on average
during the first part of the period studied to 28 per cent on average over the last decade.
This is a result of increasing labour costs, which have induced many firms to use more
capital–intensive technologies. Hence, the contribution of capital accumulation to
growth may be expected to increase in the future.
44
In Burkina Faso, Côte d’Ivoire II, Mali and Tanzania, the contribution of TFP
gains in recent years is similarly high. It even exceeds 100 per cent in Côte d’Ivoire II,
where the capital/labour ratio declined between 1994 and 1996.
It is well known that emerging–country success stories, particularly in Asia,
resemble the early growth episodes in our sample in that they relied heavily on capital
accumulation, contrary to the current growth periods (see Young, 1995, and, for a
treatment comparable to the one used here, Berthélemy and Chauvin, 2000). The
question is therefore why these episodes ended much earlier in Africa than in East
Asia. Although there are substantial non–economic explanations in a number of cases
(such as the social and political unrest in South Africa), some economic factors may
be considered.
In the case of Côte d’Ivoire, the collapse of international coffee and cocoa prices
deprived the country of one of its main sources of savings, precipitating an economic
crisis. Algeria, Cameroon and Egypt had similar problems when oil prices declined
(Cameroon became an oil–exporting country in the late 1970s). In the early 1980s, the
rapid growth episode in Kenya came to an end for reasons similar to those seen in
Côte d’Ivoire: when the coffee boom ended, saving resources dried up (Azam and
Daubrée, 1996). This effect, in conjunction with increasing real interest rates, held
back investment in the late 1970s. A similar picture can be found in Malawi, which
saw substantial growth up to 1979, based on exports of tobacco and tea. Although the
immediate reason for the interruption of this strong growth performance was a severe
drought, declining terms of trade also played a role11. The notion that all of the historical
growth periods studied here — with the exception of South Africa, Morocco and perhaps
Tunisia — were induced primarily by surges in investment, fuelled by temporary
commodity booms, will be studied further in Chapter 3, where the impact of the terms
of trade on investment is analysed.
Another possible explanation lies in the lack of productivity gains. As argued
earlier, productivity is likely to have a double effect on growth. In addition to its direct
effect, a low level of productivity may create disincentives for investment by reducing
the return on capital. Productivity gains were stifled by socialist economic policies in
Algeria, Egypt, Morocco and Tunisia and by other forms of government intervention
in combination with corruption in Cameroon, Côte d’Ivoire, Kenya and Malawi. Social
and political unrest compounded the effect of heavy government intervention in South
Africa. The hypothesis that productivity level influences investment will be tested
econometrically in Chapter 4. The following section seeks to explain the contrasting
development of productivity in the recent and earlier periods, while trying to relate
this to the economic performance of the sample countries.
45
Analysing the Sources of TFP Growth
Table 2.4 illustrates the change in productivity in the countries studied during
their respective periods by dividing TFP growth into its main sources. There are some
clear differences between the earlier and the current growth periods.
Table 2.4. Sources of TFP Growth
(percentage points, annual averages)
Contribution Memo item:

Country Period Total TFP Black
growth market Human Exports/ Diversi- Reallo- GDP
capital labour fication cation Other acceleration
prem.
Failed take offs

Algeria 1965-85 0.9 -0.4 1.0 -0.2 0.4 1.0 -1.0 -0.14
Cameroon 1967-86 1.9 0.0 1.7 0.6 -0.1 1.4 -1.7 -0.15
Côte d’Ivoire I 1961-78 2.8 0.0 3.0 0.3 0.0 1.2 -1.7 0.01
Egypt 1965-90 1.3 0.1 0.6 0.2 0.4 0.4 -0.4 -0.02
Kenya 1961-81 2.3 -0.1 1.1 0.2 -0.1 0.6 0.6 -0.11
Malawi 1964-79 0.2 -0.2 0.0 0.4 -0.1 1.0 -0.8 0.05
Morocco 1966-80 -0.2 0.0 2.0 0.0 0.0 0.3 -2.5 0.02
South Africa 1961-74 0.8 0.0 0.5 0.0 0.1 0.5 -0.2 0.05
Tunisia 1965-85 0.7 0.1 1.3 0.4 0.2 0.4 -1.8 0.07
Average 1.2 0.0 1.2 0.2 0.1 0.8 -1.1 -0.02
Current growth periods

Botswana 1970-96 3.4 0.0 1.0 0.9 0.3 2.0 -0.7 0.03
Ghana 1983-96 1.9 1.6 0.5 0.6 0.4 0.1 -1.2 -0.04
Mauritius 1980-96 2.8 0.2 0.5 0.5 0.5 0.2 0.9 0.00
Mozambique 1986-96 3.7 2.5 0.4 0.7 0.1 0.3 -0.3 -0.30
Uganda 1986-96 4.7 1.3 0.8 0.6 0.2 0.3 1.5 0.23
Average 3.3 1.1 0.6 0.7 0.3 0.6 0.0 -0.01

Burkina Faso 1994-96 1.7 0.0 1.3 0.2 0.1 0.7 -0.6
Côte d’Ivoire II 1994-96 4.7 0.0 0.6 0.8 0.1 0.5 2.7
Mali 1994-96 1.4 0.0 0.4 0.7 0.0 0.6 -0.3
Tanzania 1995-96 4.2 -0.1 0.6 1.1 0.1 0.7 1.8
Note: Logarithmic rates. The “Other” column is the part of the TFP growth rate which is not accounted for by the factors
identified in previous columns.
First, reduction of distortions on the foreign exchange market has played an important
role in some of the strong performances (Ghana, Mozambique, Uganda) in the recent
period, and this is a good proxy for successful structural adjustment policies implemented
in non–CFA African economies. The black market premium has been all but eliminated in
these three economies, from levels of approximately 2 000 per cent, 4 700 per cent and
380 per cent respectively during the relevant periods (in Uganda, the black market premium
peaked at 920 per cent before the period studied, in 1978). Reduction of the black market
premium has in many cases coincided with broader structural adjustment measures. The
variable is therefore likely to catch some of the generally beneficial effects of
macroeconomic stabilisation programmes. As Botswana and Mauritius have long had
sound macroeconomic policies, this variable does not play a visible role in the dynamics
of their economies.
46
In Franc Zone countries, there is for all practical purposes no black market
premium. Nevertheless, the CFA franc devaluation in 1994 has played a similar role
in their recovery, because it corrected the macroeconomic imbalance created by an
overvaluation of their currency. The devaluation has been particularly important for
Côte d’Ivoire, which had huge unused industrial capacity until January 1994 and was
able to resume industrial activity quite rapidly thereafter. This is accounted for in the
“Other” column. In Burkina Faso and Mali, which have virtually no industrial capacity,
the impact of devaluation seems not to have been strong enough to offset other sources
of TFP decline in 1995 and 1996. Our numbers should be interpreted with caution,
however, as they are computed on a very short period and are therefore very sensitive
to the consequences of possible measurement errors.
Where structural change indicators are concerned, Table 2.4 suggests that human
capital accumulation played a more important role in the earlier periods than in recent
growth episodes. One should keep in mind, however, that nearly all sub–Saharan
countries started from extremely low levels of human capital in the 1960s, which
partly explains the high rate of growth. Nevertheless, investment in education did
decrease significantly in the 1980s, leading to slower growth in the human capital
stock. We see that human capital accumulation has played an important role in Algeria,
Botswana, Cameroon, Côte d’Ivoire I, Kenya, Tunisia and Uganda. The most impressive
case is Côte d’Ivoire I, where the improvement of human capital contributed 3.0 points
of annual average productivity growth (conversely, Côte d’Ivoire II has performed
poorly in education in the 1990s, with a decline in school enrolment rates). Mauritius’s
seemingly moderate growth in human capital is due to the fact that it started from a
relatively high level. The country is currently at the highest level of human capital in
our sample, with nearly eight years of schooling on average. Thus where human capital
accumulation is concerned, there remains much room for improvement in the years to
come, in particular for the least advanced economies in the sub–sample. If Mozambique
and Uganda were to attain Mauritius’s current level of human capital, they would
register a gain in TFP of around 25–30 per cent, while the corresponding figure for
Ghana and Botswana is about 10 per cent. These potential gains are significant, but
they will be obtained only slowly, since the accumulation of human capital through
education is a long and costly process.
Export growth has contributed significantly to increases in labour productivity
in several cases from both the earlier and the current growth periods, although its
contribution is more modest on average for the earlier periods. Export growth has
been an especially important contributor to labour productivity gains in Botswana,
Cameroon, Mozambique, Uganda and Ghana. This has also been the case recently in
Côte d’Ivoire II and Mali, through the combined effects of the CFA franc devaluation
and trade liberalisation policies. Further, there is still progress to be made in terms of
export promotion as an engine of growth. Despite substantial progress in trade
liberalisation in many cases, Africa remains relatively closed.
Reallocation of labour from agriculture to more productive sectors of the economy
has contributed significantly to growth in both the current and the earlier periods
analysed here (see annex). The most spectacular case is Botswana, where reallocation
47
away from agriculture has raised productivity by two percentage points on average
during the period studied. In this context, the results from Mauritius merit further
explanation. The modest contribution of reallocation is somewhat misleading,
considering that until the 1970s the country’s economy was dominated by sugar
production (which is both an agricultural and a manufacturing activity). In Cameroon,
the discovery of oil in the late 1970s provoked a substantial reallocation of labour
from the agricultural sector, through a sort of “Dutch disease” mechanism. Similarly,
the cocoa and coffee booms resulted in a transfer of labour mainly to the food–
processing industry in Côte d’Ivoire I and Kenya. It should be noted, however, that in
the cases of Cameroon, Côte d’Ivoire I and Kenya, this reallocation did not result in
increased diversification of the economy. In fact, the commodity booms induced
increased specialisation in these economies.
Generally speaking, diversification is a recent phenomenon in Africa, and it is
an important source of growth primarily for the countries currently in a phase of rapid
growth, with Mauritius as the most prominent example. One exception from the earlier
growth period is Algeria, where excessive investment supported by high oil income
did in fact result in a considerable increase in diversification12. However, the direction
of these investments was determined by government decree rather than economic
rationale, and the outcome was consequently highly inefficient. As a result, the country’s
dependence on capital accumulation for growth was magnified by a negative TFP
performance. Algeria’s growth process proved unsustainable after the decline in oil
prices in 1985. Parallels can be drawn between Algeria and Egypt. Egypt experienced
a significant degree of diversification during the period considered, stemming mainly
from the public sector. The overwhelming size of the public sector created severe
inefficiencies in the economy. Public investment in Egypt was financed to some extent
by oil exports, but also by substantial foreign transfers.
The analysis above provides some indication as to what caused the low levels of
productivity in countries whose growth experiences were interrupted. In several cases,
the main reason was no doubt policy–related. Algeria followed a socialist path of
command economics. Tunisia’s socialist period ended during the years under study
(in the late 1960s) but government policies remained heavily interventionist. In
Morocco, the state played a major role in the economy until an adjustment programme
was introduced in 1983. And Egypt’s public sector remains disproportionately large
despite attempts to cut back its role. Commodity booms in Cameroon and Côte d’Ivoire
led to wasteful investment and rent–seeking behaviour, as illustrated by the substantial
negative residuals for these countries in the growth accounting exercise in Table 2.3.
Another important explanation for low productivity is most likely related to the
fact that capital accumulation did not promote economic diversification. This lack of
diversification in African economies — both in countries which initially had a
manufacturing base (South Africa and Tunisia) and in primary goods producers —
stands in stark contrast to East Asian experiences13. In South Africa, as shown by
McCarthy (1998), capital accumulation was used to build increasingly capital–intensive
manufacturing industry, whereas the comparative advantage of this country was
48
presumably in (unskilled) labour–intensive manufacturing. In Tunisia, similarly,
manufacturing investment did not provide any significant diversification of the
industrial structure. The other countries considered above, Cameroon and Côte
d’Ivoire I, faced a similar lack of diversification: they remained exporters of traditional
primary products. The only countries for which diversification can be considered to
have contributed substantially to growth are all from the recent period: Mauritius,
Botswana, Ghana and, to some extent, Uganda. In the case of Mauritius, diversification
added as much as 0.5 percentage points to growth on average during the period studied.
Industrial diversification in Mauritius started with the development of textile and
clothing production (Mauritius is the largest textiles exporter in sub–Saharan Africa
and the third exporter of woollen goods in the world) and has continued recently with
electronic products. Moreover, services have been developed and diversified, in
particular tourism and financial services. This diversification has played a major role
in sustaining economic growth.
The progress in terms of diversification in Ghana and Uganda should be
interpreted with care. Given that Uganda started from an extremely low level at the
beginning of the period studied, the sustainability of its diversification — based on
new agricultural products such as flowers rather than on manufacturing — is not certain.
In the case of Ghana, diversification gains seem to be derived from a diminishing
relative importance of cocoa to the benefit of aluminium, gems and low–end wood
processing. In other words, in contrast to Mauritius, the diversification in these two
countries does not appear to imply any significant increase in the importance of high
value–added industries. This is probably due to the lack of capital deepening, without
which these economies cannot really diversify their industrial structures. As a matter
of fact, these two countries are still at a very low productivity level. It thus appears
unlikely that they will be able to pursue a growth path similar to those of Mauritius or
Botswana. The economy in Mozambique is relatively well diversified for its low level
of income, and this may improve further in the near future as substantial FDI projects
are under way in a wide range of sectors.
The rate of acceleration of GDP provides one indication as to whether the current
growth processes have reached a point where the gains from adjustment have started
tapering off. A negative acceleration indicates that GDP growth is slowing down, which
may necessitate stronger measures in order to induce more far–reaching structural
change and increased investment. Among the “current long growth period” countries,
acceleration is negative for Botswana and Ghana14. Botswana’s performance has been
somewhat less impressive on average since 1990 than in the previous two decades,
owing primarily to a cyclical effect of the diamond industry. By contrast, the relative
slowdown in Ghana is more likely to be a result of the adjustment process running out
of steam. Growth decreased gradually from over 8 per cent in 1984 to 3.6 per cent in
1994 and revived somewhat in the following two years, as in several other African
countries. GDP acceleration is zero in Mauritius and positive in Uganda. Growth in
Uganda was on an upward trend during the last few years of the period considered,
possibly indicating that the country has not yet reaped all the benefits of structural
49
adjustment. The lack of capital formation and the slow structural change in the economy
are nevertheless worrisome for the near future. Finally, useful analysis of GDP
acceleration in Mozambique is difficult, given the erratic growth record of the country.
Our tentative conclusion is that, with the exception of Botswana and Mauritius,
the economic take off in the economies currently enjoying rapid growth is not
necessarily sustainable, inasmuch as such a process requires capital accumulation and
faster structural change. At the opposite end of the spectrum, we have argued that low
levels of productivity may have undermined the sustainability of investment–driven
growth in the earlier periods.
Future Growth Scenarios for Six African Countries
The analytical framework presented above may be used to build coherent growth
scenarios. This framework is based on three relations — an aggregate production
function, an investment equation (presented in Chapter 3 through an analysis of capital
accumulation) and an accounting relation (for the balance of payments) – that enable
us to observe the required consistency between the data on national savings and foreign
capital inflows, on investment and on foreign trade (see Box 2.2).
The aggregate production function contains most of the variables which we
will be considering in order to formulate our growth scenarios: labour reallocation,
diversification of the economy (see annex), accumulation of human capital and export
expansion (detailed analysis of these factors for the six countries studied is found in
Chapters 4 and 5). Although all of these variables may be influenced by changes in
economic policy, for the next ten years at least they will necessarily be in line with
recent trends in the economies considered — since changes such as improving the
institutional framework and skills development can be brought about only in the
long term. The range of growth possibilities is thus ultimately rather narrow in the
economies studied.
Similarly, our investment equation allows economic policy variables to play a
role, through the volume and efficiency of foreign resource flows, which finance a
fairly large share of capital accumulation. However, future policy changes will be
able to alter past trends only slowly.
All in all, our analysis tends to mark out a fairly narrow path for the potential
growth of the African economies studied. Deviations from this path are possible, of
course, chiefly because nothing can guarantee that future political troubles will not
deflect certain economies from their potential growth paths, or that terms–of–trade
shocks will be neutral. What follows should therefore be interpreted less as a forecasting
exercise than as an attempt to quantify the growth potential of these economies in the
absence of unfavourable shocks. It is also possible to envisage more optimistic scenarios
in which the coming years see more progress on the reform front. For this reason, we
also construct “high” scenarios based on the hypotheses of increased structural change
50
Box 2.2. Explanation of the Model used for Long–term Predictions
The model is made up of only three fundamental equations. With all variables in
logarithmic form, the production function has the following properties:
ly=α*lk+β*ll+γ*lm+tfp,
α+β+γ=1
where ly is GDP, lk the capital stock, ll the labour force, lm imports and tfp total factor
productivity, explained as a linear combination of the human capital stock (lh), a diversification
index, a measure of reallocation of labour from agriculture to non–agricultural sectors, the
black market premium on the foreign exchange market (with CFA countries distinguished
from non–CFA countries) and a dummy for the number of revolutions and coups d’état during
the given year. The last two of these variables are assumed to stay at their optimal levels in the
future: black market premia — which are already close to zero — and the dummy for revolutions
and coups are cancelled in the equation; in other words, we assume that the economies under
study will face a relatively peaceful environment and will not be subject to strong macroeconomic
mismanagement.
The level of imports is determined according to the following function:
lm–ly=ε1*(laid–ly)+ε3*lxy+ε3*ltot
where lm–ly is the ratio of imports to GDP (lm is the level of imports), laid–ly is the flow of aid
and debt as a ratio of GDP (laid is the level of aid and debt flows), lxy is the exports–to–GDP
ratio and ltot the terms of trade. This accounting relation can be seen as a result of balance–of–
payments restrictions. Its determinants are exogenous in the long term given that access to aid,
debt and the terms of trade are considered to be exogenous, and the ratio of exports to GDP will
reach a maximum at some point, since exports cannot grow faster than GDP indefinitely. It
should be mentioned that various attempts to estimate an export function have been unsuccessful.
Prior studies (see e.g. Rodrik, 1998) have revealed difficulties in explaining exports by trade
policy variables. For these reasons, we propose scenarios for the ratio of exports to GDP, rather
than for export performances as such. Terms–of–trade movements, which also appear in this
equation, are ignored, not because we think that there will be no such movements, but because
this seems to be the most neutral hypothesis that can be adopted concerning a very uncertain
future.
Solving (1) and (2) gives the growth equation:
+ γ[ε1 dlaid + ε1 dlxy + ε 3 dltot ]+ dtfp
dly =
1 γ(1 ε1 )
where dxxx is the variation in xxx.

The investment function, discussed in chapter 3, is of the following form:
dlkt = +δ1lkprodt–1δ2*laidkt–1+δ4∗ltott–1
where lkprod is road length per capita, laidk is the ratio of aid and debt flows to the capital
stock and ltot is the terms of trade.
This equation gives a prominent role to two further variables which need to be given
careful consideration: the level of aid flows and their efficiency in the investment process, as
measured by the δ2 parameter.
51
and greater efforts to improve education. This allows us to make an order–of–magnitude
assessment of the manoeuvring room available in the long term for stimulating growth
in Africa. To evaluate the potential impact of the policy options open to African
governments and aid agencies, we subsequently examine the extent to which the various
initiatives that might be taken at the national level would contribute to any increase in
the growth rate (Chapters 3 and 5). Lastly, we examine the possible consequences of
increased foreign aid to the countries studied (Chapter 3).
Scenario Results
We will now use the model described above to build scenarios of future growth
to the year 2020 in six countries: Burkina Faso, Côte d’Ivoire, Ghana, Mali, Tanzania
and Uganda. All of these countries have recorded satisfactory recent growth. We
construct two main scenarios, one of which has two variations. The “baseline” scenario
essentially assumes a continuation of current trends in the underlying determinants of
growth. The second scenario, called the “high” scenario in the figures below, is more
optimistic and assumes improvement in certain variables, particularly those relating
to structural change, such as human capital and diversification of the economy. Two
variations of the high scenario are constructed in order to isolate the effect of
diversification in one case and that of aid efficiency in the other. In all scenarios, we
assume an absence of revolutions and coups, unchanged terms of trade and a zero
black market premium15. The results are shown below (Figure 2.1).
Before we analyse the sources of growth in more detail, a few general
characteristics of the scenario results deserve attention. Table 2.5 summarises the
scenarios by giving some key indicators for the year 2020, which should be reasonably
close to the steady state. Investment rates for the year 2000 are also given, in order to
show the impact of enhanced aid on the investment ratio16. In the baseline scenario,
Uganda achieves the highest long–term growth rate of GDP per capita, about
0.5 percentage points greater than those of Côte d’Ivoire and Mali on an annual basis,
and one percentage point greater than Ghana’s. This performance can be explained
primarily by Uganda’s strong commitment to education and the fact that it has made
more progress in diversification than the other countries in the sample, with the
exception of Ghana.
Ghana’s GDP per capita growth rate is the lowest among the countries studied in
the baseline scenario, despite its better ranking in terms of productivity growth. This
is principally attributable to the low investment rate in the country. In both Ghana and
Uganda, in fact, the capital stock has barely kept up with the growth of the labour
force since the introduction of reforms in the 1980s. Moreover, the gains obtained by
reallocating labour from agriculture to the modern sector are expected to be less in
Ghana than in the other countries. The difference in productivity between the two
sectors is the lowest among all studied countries, mainly due to the substantial
comparative advantages of Ghana’s cocoa sector.
52
Figure 2.1. Scenario Results to 2020
53
GDP per capita (1996 constant $) GDP per capita (1996 constant $)
250
300
350
400
450
500
550
600
400
450
500
550
600
650
700
350
1996 1996
1997 1997
1998 1998
1999 1999
2000 2000
2001 2001
2002 2002
2003 2003
2004 2004
2005 2005
2006 2006
2007 2007
2008 2008
2009 2009
54
2010 2010
2011
GDP per Capita: Mali

2011
2012 2012
GDP per Capita: Ghana
2013 2013
2014 2014
2015 2015
2016 2016
2017 2017
2018 2018
2019 2019
2020 2020
Less div.
Baseline
Less div.
High scen.
High scen.
Enhanced aid
Baseline scen.
Enhanced aid
55
Table 2.5. Scenario Overview
(percentages)
Baseline scenario High scenario Enhanced aid Less diversification

GDP growth per
habitant 2020
Burkina Faso 2.49 3.67 3.64 3.44

Côte d’Ivoire 2.06 3.09 3.09 2.85
Ghana 1.59 2.59 2.61 2.40
Mali 1.99 2.88 2.88 2.65
Tanzania 1.66 2.43 2.44 2.20
Uganda 2.61 3.42 3.39 3.20
Baseline scenario Enhanced aid
Investment/GDP Year 2000 Year 2020 Year 2000 Year 2020
Burkina Faso 18.4 18.1 22.0 19.6

Côte d’Ivoire 20.1 19.7 21.7 20.7
Ghana 15.4 15.5 19.1 17.4
Mali 20.7 21.3 24.7 23.2
Tanzania 24.2 23.5 27.3 25.3
Uganda 12.1 11.9 15.0 13.1
Tanzania’s growth prospects are the second lowest in the baseline scenario and
the lowest in all other scenarios, even though the country starts from a comparatively
low level17. The reason is that there has been very little structural change in Tanzania
recently, particularly where education is concerned, and this aspect is factored into
our assumptions for the future. As a result, the country’s productivity growth is the
lowest in the sample. In contrast, Tanzania has the highest investment rate, at around
24 per cent of GDP in the baseline scenario. According to our data, it has a relatively
high capital–to–GDP ratio (around 3), and hence the investment rate required to replace
depreciated equipment and sustain a given growth rate is higher than in other countries.
Burkina Faso — which also starts from a very low level, but has recently shown
better performance in terms of structural change — has the second–highest growth rate
in our sample. It must be noted, however, that Burkina Faso will still be far from its
steady state in 2020, particularly where openness and factor reallocation are concerned.
This explains to some extent the relatively high growth rate reported in Table 2.5.
A comparison of the baseline and high scenarios for all countries in Table 2.6
reveals the crucial importance of structural change for growth. This observation is
particularly relevant for the countries studied here, given the lack of structural change
in their growth processes so far. As was demonstrated above, the recent revival of a
number of African economies is mainly attributable to catch–up effects in the form
of one–off productivity gains from adjustment measures. There has been little increase
in investment and structural change. As we have already mentioned, the high scenarios
are distinguished from the baseline scenarios by an acceleration of structural change.
Apart from aspects relating to factor reallocation and economic diversification, treated
in the annex to this chapter, we will analyse in more detail the accumulation of physical
capital, the accumulation of human capital and the development of exportations in
Chapters 3, 4 and 5, in Part Two of the book.
56
Table 2.6. Baseline and High Scenarios
GDP per capita in 2020 (1996 $) Total structural Total structural

Memo: 1996 Baseline scenario High scenario gap gap (%)
Burkina Faso 237 423 566 143 34

Côte d’Ivoire 745 1 358 1 693 335 25
Ghana 362 522 635 114 22
Mali 264 424 521 97 23
Tanzania 214 356 415 59 17
Uganda 311 496 629 133 27
57
Notes
1. To save space, the error correction model is not shown.

2. CFA Franc Zone members constitute an exception, as very small black market premia do
not necessarily imply good macroeconomic management in these countries.
3. Poirson’s (1998) analysis is based on a growth regression estimated on a panel data set
of 40 countries over three five–year periods for which the required data were available.
She finds that the principal explanatory factor of the growth differential for fast–growing
countries relative to the average is the labour reallocation effect, which accounts for 43
per cent of this differential.
4. Moreover, our data are measured in constant 1987 prices, i.e. from a time before any
significant adjustment had taken place in Africa. Hence, these prices are distorted against
agriculture. To the extent that the elimination of these distortions led to a reallocation of
factors towards agriculture, the effect of adjustment on GDP measured at the distorted
prices would be negative, although the impact on welfare may be positive. In other
words, reallocation of labour as a result of modified incentive structures would empirically
be consistent with, rather than contradictory to, the Syrquin reallocation effect.
5. We applied a two–step procedure. First, we estimated the production function without
taking the Syrquin effect into account; this gave an initial value for a equal to 0.45,
which we used in our computation of the Syrquin index. Next, we re–estimated this
equation while controlling for the Syrquin effect. The issue is that the true relationship
between observable data and TFP is not log–linear. However, we have successfully tested
a log–linear approximation, which gives reasonably similar results. This estimation is
not reported.
6. In theory, the same effect could be obtained through an increase in the degree of openness.
In other words, diversified inputs could be imported rather than produced locally. One
may therefore expect the effect of diversification to decrease with the level of openness.
We attempted to test this assumption by introducing the diversification index interacted
with imports as a share of GDP, but we did not obtain any significant results.
7. With this type of model, we could attempt to prove that one of the essential determinants
of diversification (and thus growth) is the size of the economy — that is to say not only
the country’s level of development but also its population size. However, once we pose
realistic hypotheses as to the limits of capital mobility, it becomes apparent that capital
is distributed the same way in a large country as between several small, open countries.
This demonstrates that a large population does not guarantee a high level of diversification.
58
In reality, only the size of population centres is important. When a large number of
people is concentrated in one area, economies of scale can be developed via the sharing
of intermediate goods, thus liberating resources that can be invested in other projects
promoting diversification. In this respect the African continent is at a serious disadvantage,
as it counts relatively few large cities. In effect the lack of large population centres is
perhaps one of the main obstacles to the creation of externalities in general (see the
discussion on human capital above). All the same, the demographic projections of the
OECD’s Club du Sahel project the formation of large cities within the next twenty years.
8. We restrict ourselves to the export structure towards the OECD countries in order to
obtain more reliable data. The export data are taken from OECD sources on imports of
OECD countries.
9. In an attempt to overcome the impact of the terms of trade on the diversification index, the
latter was regressed on the former, leaving the residuals free from terms–of–trade influence.
However, we did not obtain significant results using these residuals in the production
function. In any case, the influence of export prices remaining after the exclusion of oil
exports from our calculation of the index does not imply any serious flaw in the index, for
two reasons: first, the impact would be felt only in the short term; second, while an
improvement in the terms of trade might induce a mechanical decrease in the diversification
index, it is associated with an increase in economic growth, and hence terms–of–trade
fluctuation could only understate the impact of diversification on productivity.
10. Using the Harberger method based on the 1960–70 period, we estimated the initial capital/
output ratio at 3.2 (calculated from World Bank data, WDI, 2000). This estimate looks
rather high, but is substantially lower than that proposed by Nehru and Dhareshwar
(1994).
11. This was further aggravated by higher transportation costs for landlocked Malawi, due
to the civil war in Mozambique.
12. We remind the reader, however, that oil exports are excluded from the calculation of the
diversification index. Including oil in the index reveals a significant specialisation during
the studied period. A similar remark can be made regarding the other oil–exporting
countries in the sample.
13. See Feenstra et al. (1999) on the role of diversification in the Republic of Korea and
Chinese Taipei.
14. This may also have contributed to the unexplained negative part of TFP growth (see
above).
15. There is some evidence that a parallel foreign exchange market is re–emerging in Ghana,
as a result of heavy intervention by the central bank to slow the depreciation of the cedi.
However, we assume any such tendencies to be transitory.
16. As Table 2.5 shows, the effect of enhanced aid on the investment rate decreases over
time. This is due to two features in our model: first, capital inflows enhance import
capacity, which has a direct effect on production, and hence on the denominator of the
investment ratio; second, the importance of aid and other capital flows for investment
gradually decreases if the capital stock grows more rapidly than these flows.
17. Note, however, that the income levels shown in Table 2.5 are not measured in PPP terms
and thus are not directly comparable.
59
60
Annex
Structural Change in the Productive Sector
Accumulation of production factors alone cannot lead to lasting growth, as neo–

classical growth theory affirms — the returns on both investment and education would
eventually diminish if the productive structure of the economy remained unchanged.
This annex examines two types of structural change: the reallocation of production
factors and diversification of the economy. To some extent, these structural changes
occur naturally as a result of factor accumulation, but they are not automatic. The
institutional framework and the incentive structure also play an important role.
In a context of suitable incentives for agricultural modernisation, training the
labour force can enable agricultural productivity gains to be made, freeing up labour
initially used in this sector. It will then be possible to employ this labour in sectors
where it is more productive, which will have a positive effect on growth. In this respect,
the economies studied here still have considerable long–term growth potential, as the
greater part of their working populations is employed in agriculture.
Capital accumulation makes it possible to develop new economic activities, if
investors encounter an institutional and incentive framework that favours such
diversification. Diversification — which at an initial stage may occur even within the
agricultural sector — then leads to productivity gains to the economy. The African
economies studied here are starting from scratch in this respect, and Mauritius’s success
in diversification suggests that this is a source of growth that they can tap.
We will address the themes of factor reallocation and economic diversification
in turn, in an attempt to highlight the factors likely to stimulate these structural changes
and thus to help increase the long–term growth of the economies studied.
Factor Reallocation
All developing countries that have taken off experienced a substantial shift in
the factors of production, mainly through a transfer of labour out of the agricultural
sector. This movement stems both from the fact that the structure of domestic demand
changes over time and the fact that productivity gains make it possible to release
labour for other activities.
61
The Distribution of Labour between Agriculture and Other Activities Plays
a Leading Role
As shown in Chapter 2, the movement of labour away from agriculture, where

labour productivity is lower than in all other sectors, automatically has a positive
effect on growth. The data presented show that, in the past, the labour reallocation
effect, or Syrquin effect, played a significant role in the episodes of rapid growth,
contributing 0.6 to 0.7 percentage point to growth on average. The same is true
elsewhere. For example, Berthélemy and Chauvin (2000) evaluated the growth
contribution of labour reallocation away from agriculture for the 1968–96 period at
1.2 percentage points of annual growth in Thailand, 0.7 points in Korea, 0.6 points in
Malaysia and 0.5 points in Indonesia.
The future growth potential represented by the Syrquin effect cannot, however,
be extrapolated directly from past data, because, assuming that labour movements
continue at an unchanged pace, this reallocation effect will contribute less and less to
growth since it diminishes automatically as the share of the agricultural working
population falls. We therefore need to study the potential for change in the sectoral
structure of the labour force in order to evaluate, using Syrquin’s formula, the effects
that such change may be expected to have on growth.
The study by the Club du Sahel (1998) on the future of West Africa to 2020
concludes that, owing to population pressure, labour movements out of agriculture
seem inevitable in the long term, since the increase in population density will lead to
a complete geographical redistribution of the population, notably in favour of cities.
According to the study, the agricultural population in this region (excluding Nigeria)
should grow by 1.6 per cent a year from 1990 to 2020, as against 2.6 per cent for the
overall population. For West Africa, including Nigeria, the reduction in the share of
the agricultural population would be 1.4 per cent annually.
Labour migration away from agriculture cannot be brought about by decree. It
can only come as the result of the economic forces at work on the labour and goods
markets. In dynamic terms, the main factor allowing labour to move to other sectors is
increased agricultural productivity. This can be seen fairly clearly from the available
African data, as we will show below.
The fact is that there are several general equilibrium mechanisms that may lead
to a transfer of labour in favour of non–agricultural activities if there is progress in
agriculture. An increase in the supply of basic food products from farmers will, other
things being equal, lead to a decrease in the relative price of these products, which
will in turn stimulate migration to other activities. Moreover, the drop in the price of
foodstuffs — the largest item in the consumption basket of urban dwellers — will
offer modern sector employers a drop in unit labour costs at a constant real wage. If
the progress in agriculture stems from export activities, it will give rise to improvement
in the balance of payments and, consequently, pressure to re–evaluate the real exchange
rate. This will work to the benefit of non–tradable activities, particularly services, and
to the detriment of the tradable goods sector (particularly export agriculture). Lastly,
62
it may be noted that advances in agriculture that allow smallholders’ families to provide
for their needs using less labour may stimulate migration to the cities. Conversely, if
there is no progress in agriculture, traditional village structures will remain in place,
population pressure will push food prices up, and to maintain the balance, a very high
proportion of the working population will have to be kept in agriculture.
As a result, there is every reason to believe that agricultural modernisation is a
pre–requisite for the reallocation of labour to non–agricultural activities. The causality
between the two phenomena may become partly circular, if the quantity of labour in
agriculture falls and, owing to decreasing marginal productivity, the productivity of
agricultural labour rises. As land is not yet a scarce resource in most of the economies
studied, however, the productivity of agricultural labour can hardly increase without
an initial effort to modernise and intensify agriculture.
Progress in Agriculture and Labour Reallocation: Scenarios of Future Trends
Progress in agriculture is a very important variable in and of itself for the

economies under study, which are still highly agrarian. Such progress has two main
consequences. First, it contributes to the overall gain in productivity, with a weighting
equal to the GDP share of the agricultural sector. Second, agricultural progress makes
it possible to release labour, which can then be employed in more productive activities.
In our macroeconomic approach, the first effect cannot be dissociated from total factor
productivity and thus is not isolated as such in our scenarios. The second, however, is
specifically identified in our model because it is the source of the increased growth
obtained through factor reallocation. We begin by studying the link between progress
in agriculture and the shift of labour out of agriculture.
Figure 2.2 enables us to compare the growth of apparent labour productivity in
agriculture with the change in the proportion of the working population employed in
agriculture, over a 25–year period (1966–90), in 46 African countries (both north and
south of the Sahara) for which these data are available. These data show that there is
indeed a positive relationship between agricultural intensification and the reduction
of agricultural employment.
Over this period, the countries which achieved the greatest productivity gains in
agriculture are also generally those which were able to release a portion of the
agricultural labour force for other activities, thus reproducing the development of
Europe in the nineteenth century and that of East Asia as from the 1950s. It can also be
seen from Figure 2.2 that when productivity gains are low or negative, the proportion
of the working population employed in agriculture tends to stabilise or to fall slightly.
The relation between the two variables studied is thus not very elastic for these countries.
In contrast, there seems to be a stronger response in terms of labour movements when
agricultural progress is more marked (agricultural productivity growth of over 0.5
percentage points per year). The relation linking these two variables is thus non–
linear, as suggested by the hyperbolic curve constructed in Figure 2.2 from the
observation data. This is fairly natural: without progress in agriculture, there is little
63
Figure 2.2. Agriculture Progress and Labour Reallocation in Africa, 1996-90
-0.06 -0.04 -0.02 0 0.02 0.04 0.06
0
Growth in share of agricultural labourr
-0.02
-0.04
-0.06
Growth of labour productivity in agriculture
Note: We end this period at 1990 because of a lack of data after this date.
Source: Authors' calculations based on data from the World Bank, African Development Indicators and
the African Development Bank, African Development Report. Growth of agricultural output is
calculated on the basis of the "Crop Production Index".
scope for change in rural conditions, and hence for a transfer of labour out of agriculture,
whereas rapid technical progress opens new horizons for people living in rural areas
by allowing them, for example, to take greater risks or to overcome the liquidity
constraints that hold back the creation of new businesses.
This hyperbolic curve implies that the elasticity of the share of agricultural labour with
respect to agricultural productivity is high when productivity gains are high, and low in the
contrary case. For example, a simulation of the preceding curve shows that a 0 to 1 per cent
increase in technical progress raises the rate of labour migration away from agriculture by
0.3 percentage points, while an increase of 3 to 4 per cent raises it by 0.7 points.
In building our growth scenarios, we will not use this relation directly, but will use
it to construct variants. The reason for this is that each country has its own characteristics,
and as a result the same degree of technical progress in agriculture will not have the
same consequences for labour from one country to another. For example, in Burkina
Faso, a country characterised by emigration, labour productivity gains in agriculture
comparable to those achieved by Côte d’Ivoire during the 1966–90 period had no visible
effect on the distribution of the working population, as more people emigrated abroad
than to the cities of Burkina Faso. This phenomenon is also related to the fact that
Burkina Faso is much poorer than Côte d’Ivoire and has fewer employment opportunities
outside of agriculture in the medium term. A similar phenomenon is observed in Mali.
64
As a result, we postulate in our baseline scenario that the recent trends observed
in each country with respect to labour movement out of agriculture will continue in
the future. To this end, we take as our starting point for calculations the years 1984 for
Ghana and 1987 for Uganda, which are the years in which the resumption of agricultural
growth began. For the other countries, where reforms came later, we do not have
sufficient historical perspective and we have chosen simply to reproduce the trends
observed since 1990. We will make use of the above relation, however, in trying to
draw up a more optimistic scenario for the future of these countries.
Table 2.7 indicates the labour reallocation effect that might result from the above
assumptions. To perform this calculation, we further assumed that the relative
productivity of agriculture and of the rest of the economy would be stable in the future,
an assumption that has little impact on our numerical results.
In our baseline scenario, these calculations constitute a cautious assessment of
the reallocation effect that may be expected in the future. The effect ranges from 0.1
to 0.5 percentage points of growth annually on average from 1996 to 2020, representing
as much as half of the effect observed over the reference period. For Ghana, the
reallocation effect is slight because labour productivity in agriculture is currently fairly
close to that in the rest of the economy. Conversely, a relatively strong effect is observed
for Burkina Faso, which is in the opposite situation. In Côte d’Ivoire, despite a sharp
drop in the relative share of agricultural labour, the reallocation effect is moderate
because this share is already rather low (close to 50 per cent).
Table 2.7. Labour Movement Away from Agriculture and the Syrquin Effect
Past trends Baseline High scenario
scenario
Country Reference Relative Growth in Growth in Growth Growth Growth in Growth
period productivity agricultural relative due to due to agricultural due to
in 1996 share of agricultural reallocation, reallocation share of reallocation
labour force productivity ref. period 1996-2020 labour force 1996-2020
Burkina Faso 1990-96 38.1 -0.2 1.4 0.8 0.5 -0.5 1.0
Côte d’Ivoire 1990-96 50.7 -1.6 0.0 0.5 0.4 -2 0.4
Ghana 1984-96 76.4 -0.3 0.0 0.1 0.1 -1.5 0.2
Mali 1990-96 57.5 -0.4 1.4 0.6 0.3 -1 0.6
Tanzania 1990-96 58.6 -0.6 0.5 0.7 0.4 -1 0.6
Uganda 1987-96 54.5 -0.2 0.0 0.3 0.3 -1 0.8
Source: Authors’ calculations from World Bank data (African Development Indicators), supplemented where the
agricultural share of the working population is concerned by the African Development Bank (African Development
Report).
65
To construct a high scenario, we examine the consequences of a 2 per cent annual
increase in the rate of growth of agricultural productivity. For countries starting from
a rather low level of labour migration, the effect should be no more than moderate,
owing to the hyperbolic form of the equation. We may therefore consider that the drop
in the share of the labour force employed in agriculture would not exceed an annual
rate of 1 per cent in these countries. Given a situation more conducive to change in the
countryside, the effect could be more pronounced: in a country like Côte d’Ivoire,
where substantial progress has been recorded in the past, the decrease in the agricultural
share of the labour force could reach an annual rate of around 3.5 per cent. This is,
moreover, the rate assumed by the government of Ghana in its projections for 2020.
Although Ghana has in the past shown very little dynamism in terms of labour
movements out of agriculture and increased agricultural productivity, the potential
exists, and this potential should be comparable to that of Côte d’Ivoire, since the
conditions under which agricultural activity is exercised are fairly similar there.
Furthermore, appreciable advances in agricultural productivity have been achieved in
Ghana since 1984 (see Table 2.8), although this has not had much impact on the
employment structure of the working population.
For our high scenarios, we therefore assume labour displacement at a rate of
3.5 per cent a year for Côte d’Ivoire and Ghana, and 1 per cent a year for the other
countries. The latter percentage is halved for Burkina Faso, however, because using a
rate of 1 per cent would, given the country’s initial structures, result in an excessively
high Syrquin effect. Under this assumption, the reallocation effect for Burkina Faso is
already greater than the others (1 per cent per year). If we assumed a 1 per cent annual
drop in the agricultural share of the working population, the reallocation effect would
correspond to annual growth of 1.6 per cent. The results reported in Table 2.7 show
that there is appreciable scope for increased growth through the Syrquin effect in the
countries studied.
Potential for Progress in Agriculture
The potential for productivity gains in agriculture is therefore an essential element

to consider in evaluating prospects for the future.
Table 2.8 reports the main data available on agricultural productivity growth
and situates the countries examined in this study with respect to the rest of Africa.
Over the 1966–90 period, only three countries (Benin, Namibia and Tunisia) saw real
growth in agricultural labour productivity at an annual rate of over 3 per cent. Six
other countries registered agricultural productivity growth of 1 to 3 per cent a year:
three of the countries studied here (Mali, Côte d’Ivoire and Burkina Faso), as well as
Mauritius, Morocco and Mauritania. These productivity gains were accompanied by a
significant drop in the share of the agricultural labour force in Côte d’Ivoire, but by an
appreciably smaller one in Mali and Burkina Faso, as indicated above. In contrast,
gains in agricultural labour productivity were low or negative in Uganda, Tanzania
and Ghana from 1966 to 1990.
66
Table 2.8. Agricultural Productivity Gains in Africa
Agricultural Agricultural Agricultural Share of Land under Implicit Cereal
output labour labour agriculture cultivation yield grains yield
productivity in working per
population agricultural
worker
Years 1966-1990
lab. prod >3% 4.3 0.8 3.5 -1.9 0.1 3.4 2.5
3%> lab. prod .>1% 2.7 1.2 1.4 -1.3 -0.4 1.9 1.7
of which:
Mali 3.5 1.7 1.9 -0.3 -0.8 2.7 -0.2
Côte d'Ivoire 3.8 2.1 1.7 -1.2 -0.8 2.5 0.2
Burkina Faso 3.0 1.8 1.2 0.0 0.2 1.0 0.8
1%> lab. Prod.>0.5% 2.1 1.3 0.7 -1.1 -1.1 1.8 1.5
of which:
Uganda 2.3 2.2 0.1 -0.3 -0.9 1.0 2.1
0.5%> lab. Prod.>0% 2.2 1.9 0.3 -0.6 -0.9 1.2 0.8
lab. Prod.<0% 0.7 1.9 -1.2 -0.3 -1.2 0.2 0.4

of which:
Tanzania 2.2 2.6 -0.4 -0.3 -1.3 0.9 2.7
Ghana 0.8 2.5 -1.7 -0.2 -1.5 -0.2 -0.4
Recent period
Burkina Faso 1990-96 5.4 1.9 3.5 -0.2 -2.5 6.0 5.7
Côte d'Ivoire 1990-96 2.1 0.7 1.4 -1.6 1.9 -0.5 4.3
Ghana 1984-96 5.3 2.6 2.7 -0.3 -1.3 4.0 4.5
Mali 1990-96 3.9 2.3 1.7 -0.4 1.1 0.6 1.3
Tanzania 1990-96 -0.5 2.3 -2.8 -0.6 -1.8 -1.0 -1.4
Uganda 1987-96 2.4 2.3 0.0 -0.2 -2.2 2.2 1.1
Source: Authors’ calculations based on data from World Bank (African Development Indicators) and African
Development Bank (African Development Report). Growth of agricultural output is calculated from the
“Crop Production Index” and that of area under cultivation from the “Land Area under Major Crops”
series.
The recent data (second part of Table 2.8) show no radical changes in this
ranking except for Ghana, which has achieved real progress in agricultural labour
productivity since 1984, although to date this has not been reflected in a shift of
labour away from agriculture.
A key question for the future, then, is whether the agricultural sectors of the
African countries studied will be able to raise labour productivity enough to allow
structural transformation of economic activity. As a first approach, we can compare
the contributions of the expansion of crop production areas and of higher yields to the
growth of agricultural output. For this comparison, we have data on the land area
devoted to major crops. The difference between the growth of total agricultural output
67
and growth of total area under crops (which we call the growth of “implicit yield” in
Table 2.8) should be the result of both an improvement in crop yields and a possible
structural change, namely increasing the area devoted to the most profitable crops.
These two effects cannot be dissociated with precision on the basis of aggregate data,
which are the only data comparable across countries (since the countries produce
different agricultural commodities). However, the growth in cereal grain yields, which
may be considered to be comparable across countries, provides some indication of the
overall growth of agricultural yields.
The data in Table 2.8 clearly indicate that very little of the increase in agricultural
labour productivity in the earlier period can be attributed to extension of cultivated
area per worker. Regardless of whether the countries experienced fast or slow growth
of agricultural labour productivity, the growth in crop area per worker is low or negative,
while the countries that recorded the most rapid productivity gains also show
improvement in their implicit agricultural yields.
In the countries that achieved productivity gains in agriculture, the growth in
implicit yields was higher than that of cereal grain yields on average, which suggests
that a structural effect was at work. This phenomenon is particularly marked in Mali
and Côte d’Ivoire, where the rise in implicit yields is not associated with a simultaneous
increase in cereal crop yields. In the case of Mali, gains were obtained through the
extension of cotton farming under the supervision of the Compagnie malienne de
développement des textiles (CMDT), which allowed the agricultural sector in southern
Mali to raise its cash–crop income considerably while simultaneously modernising.
In Côte d’Ivoire, agricultural gains were due primarily to the extension of agriculture
in forested areas, with the spectacular growth of coffee and subsequently of cocoa
crops as from the 1950s.
Past progress in agriculture was thus based only in part on modernisation and
the accompanying increase in yields. In recent years, however, the increase in yields
seems to be a decisive factor, as far as may be judged from a comparison of implicit
yields and cereal crop yields for the six countries under review. The same will probably
be true in the future.
Genuine modernisation of agriculture will be needed in the future to raise
productivity. Such modernisation has not really begun in the countries studied, apart
from some modern plantation crops in Côte d’Ivoire (see Club du Sahel, 1998) and
cotton farming in Mali. It will require increased market activity on the part of farmers
— which in any event is necessary to meet the food requirements of the ever–growing
urban population — as well as increased use of modern inputs in agriculture. In this
respect, fertiliser consumption per hectare constitutes a good indicator of the extent to
which African agriculture lags behind: in 1996, the countries of sub–Saharan Africa
consumed 14 kg of fertiliser per hectare on average, as against 34 kg in Tunisia and
Morocco and 411 in Egypt. These data should also be compared with the average
levels observed in east and south Asia (240 kg and 90 kg respectively). Excluding
Côte d’Ivoire, where fertiliser consumption amounts to 24 kg per hectare, the countries
68
studied consume considerably less fertiliser than the average for sub–Saharan Africa,
from 6 kg (Ghana and Mali) to 10 kg per hectare (Tanzania). In Uganda, fertiliser
consumption is insignificant (0.1 kg per hectare). The beginning of agricultural
modernisation should therefore be reflected in a considerable increase in consumption
of fertiliser — and other modern inputs — in the countries studied. For example, to
catch up by 2020 to Tunisia’s current level, these countries would need average annual
growth of fertiliser consumption of the order of 5 per cent.
Increased consumption of fertiliser and other inputs should bring considerable
gains in crop yields, as suggested by Figure 2.3. This figure plots the growth of implicit
crop yields (as defined above) against the growth of fertiliser consumption per hectare
over the 1966–90 period. In constructing the figure, we threw out the observations
corresponding to very low consumption in 1966 (less than 5 kg per hectare), as growth
rates have little meaning in these cases. We also eliminated Mauritius and Egypt,
whose fertiliser consumption in 1966, greater than 200 and 100 kg per hectare
respectively, cannot be compared to that of the other African countries.
Figure 2.3 Increase in Yields and Fertiliser Consumption, 1966-90
0.03
0.02
Increase in implicit crop yields
0.01
-0.01
-0.1 -0.05 0 0.05 0.1
Increase in fertiliser consumption
Source: See Table 2.3.
69
The data presented in this figure suggest that a 1 per cent increase in fertiliser
consumption could bring a 0.25 per cent increase in crop yields in the countries under
consideration. This is a very rough calculation; improving it would require case–by–
case agronomic study, since the impact of fertiliser consumption on yields depends
both on the type of crops grown and on consumption of others which have
complementary effects. Thus, the green revolution has been a success in those
developing countries where new seeds, fertilisers, pesticides and irrigation have been
used in combination. Nevertheless, taking fertiliser consumption as an indicator of
agricultural modernisation gives us an order of magnitude for the yield gains that the
sample countries could obtain through an appropriate policy to upgrade agricultural
technology. For example, if these countries caught up by the year 2020 to the fertiliser
consumption level of Morocco (which at around 35 kg per hectare is rather modest
compared to the levels reached in Asia), yields would increase at an annual rate of
1.5 per cent, which would be reflected in comparable gains in apparent labour
productivity in agriculture. The potential scope for agricultural productivity gains,
using only basic technology, is thus considerable.
In the final analysis, agricultural progress in the countries studied depends entirely
on the initiative of farmers. At this stage, enhancing yields requires only basic
techniques, which do not necessarily require large outlays on agronomic research, for
example. These advances are not easily implemented, however. The success of managed
cotton cultivation in Mali shows that in some cases the modernisation process can be
organised, but the history of African agriculture is littered with examples of the failure
of agricultural extension services. In the past, such failures have sometimes been due
to unfavourable incentive systems, notably including taxation of cash crops. The slow
pace of agricultural progress in a country such as Ghana, after two decades of decline
due in particular to over–taxation of cocoa, shows that such progress can be obtained
only in the long term — if for no other reason than the fact that it takes time to reform
the institutions involved, whose operations are subject to the play of many conflicting
interests (those of farmers, marketing channels, financial channels, the tax authorities).
The Ghanaian government’s problems in reforming the Cocoa Board, which is largely
responsible for over–taxation, are characteristic of the difficulty experienced by African
economies in making progress in this area.
Diversification
The economies considered here are not very diversified, which is both a
consequence of their low level of development and a handicap on their growth. The
diversification of an economy cannot be decided by decree: the experience of state–
led industrialisation policies that the majority of African economies went through in
the 1960s and 1970s showed that a relatively diversified industrial sector cannot be
created from scratch.
70
Relatively Undiversified Economies
The very low level of diversification of African economies, apart from South
Africa, is apparent from the data in Table 2.9. Two data sources were used to construct
this table: exports to OECD countries and — for the countries examined in this study
only — statistics that describe as consistently and completely as possible the
international trade of these countries with the entire world. The latter are structured
according to the method used for the CHELEM database of the Centre d’études
prospectives et d’informations internationales (CEPII) (see section on export
performances below). The two methods (OECD and CEPII) yield slightly different
results, both in level and in trend, due to the differing geographical coverage and
product classifications used. For Burkina Faso and Tanzania, the data calculated using
the CEPII indicators seem preferable, as there is no objective evidence indicating that
these two countries could have more diversified economies than Côte d’Ivoire and Ghana.
Table 2.9. Level and Growth of Diversification
Level of diversification in 1996 Recent growth

Country OECD data CEPII data Period OECD data CEPII data
Burkina Faso 5.1 3.1 1990-1996 1.2 -0.3

Congo (Dem. Rep.) 2.4 2.9 1990-1996 -1.2 0.1
Côte d'Ivoire 3.8 4.1 1990-1996 -1.6 -1.5
Ghana 5.5 5.8 1983-1996 6.4 2.8
Mali 2.3 1.4 1990-1996 -2.2 -4.8
Tanzania 6.4 3.7 1990-1996 -1.2 -0.1
Uganda 1.6 1.5 1987-1996 3.7 3.1
Botswana 3.5 1960-96 4.4

Mauritius 6.8 1981-96 7.0
South Africa 13.7 1960-96 1.0
Note: For more on the concepts used to define and measure diversification, see Chapter 4.
Source: OECD and CEPII calculations.
Several lessons may nonetheless be drawn from these data. First, Ghana and
Uganda have undergone real diversification since their respective changes of economic
policy in 1983 and 1987, although the levels attained are still modest, especially for
Uganda. This recent diversification process has mainly involved primary commodities
rather than manufactured goods. Figure 2.4 below shows that, in addition, the
diversification index in these two countries fluctuates rather widely around its trend
curve, suggesting the persistence of a very high degree of dependency on the state of
commodities markets. The reason is that these fluctuations are linked to terms–of–
trade movements, since the diversification index falls automatically when the price of
the main raw materials exports increases.
71
Figure 2.4. Diversification Indices for Ghana and Uganda
Ghana Uganda
1.50 1.40
CEPII data CEPII data
1.20
OECD data OECD data
1.00
1.00
0.80
0.60
0.50 0.40
0.20
0.00 0.00
-0.20
-0.40
-0.50
-0.60
-0.80
-1.00 -1.00
1980 1985 1990 1995 1980 1985 1990 1995
Note: The indices are expressed in logarithmic form.
In the Franc Zone countries and Tanzania, where economic progress is much
more recent, no significant change can be observed for the moment in the recent trends
(measured since 1990), which show stability or a slight decrease in economic
diversification (Figure 2.5). The data seem to be somewhat uncertain for these
economies, moreover, in view of the differences between our two series. The CEPII
calculations probably overestimate the increase in diversification in Côte d’Ivoire in
the 1980s and the drop in Mali in the 1990s.
The only relatively diversified African economy is that of South Africa.
Diversification of this economy began immediately after the Second World War, and
by the early 1960s it had already reached a level comparable to that of the semi–
industrialised economies. From 1960 to 1985, diversification continued at a pace that
was relatively slow, but sufficient to raise the diversification of South Africa’s exports
to a level comparable to the current level of Korea. The subsequent embargo on South
African products had a dramatic effect on the export performance of this country,
especially on exports to OECD Member countries. In 1996, the diversification index
was still far below the level it had reached in the 1980s (21.9 in 1981).
The level of diversification reached by Mauritius seems fairly modest compared
to that of South Africa, but the difference is largely due to the size difference between
these two economies. A small island economy like Mauritius cannot profitably develop
a large number of different industries, even if they are export–oriented. At any rate,
72
Figure 2.5. Diversification Indices for Burkina Faso, Côte d'Ivoire, Mali and Tanzania
Burkina Faso Côte d'Ivoire

1.00 1.00
OECD data OECD data
0.50 0.50
0.00 0.00
-0.50 -0.50
-1.00 -1.00
1980 1985 1990 1995 1980 1985 1990 1995
Mali Tanzania
1.00 1.00
OECD data OECD data
0.50 0.50
0.00 0.00
-0.50 -0.50
-1.00 -1.00
1980 1985 1990 1995 1980 1985 1990 1995
73
Mauritius’s level of diversification is roughly the same as that of Singapore. What is
most important is that Mauritius has made spectacular progress since 1981 in
diversifying its economy. Gains were made at a nearly constant rate in the 1980s, and
slowed thereafter. They can to a great extent be attributed to the establishment of the
export processing zone (EPZ), which enabled the successful development of Mauritius’s
export industry.
Lessons to be Drawn from Mauritius’s Experience
The case of Mauritius is particularly instructive regarding policy for increasing

economic diversity. This experience diverges from that of the many countries that
tried to institute state–led industrialisation policies, notably in the 1960s and 1970s,
and suggests that diversification should result mainly from private sector initiative
rather than public, although the state is responsible for establishing the proper
institutional framework.
First of all, no diversification is possible if the production factors needed for the
proper functioning of the diversified industries are absent. This requires dynamic private
investment, and hence an economic climate that favours such investment. There must
also be a local supply of skilled labour that is not diverted into rent–seeking activities.
Mauritius fulfilled both of these conditions fairly early on, owing to its early adoption
of an active education policy and to the favourable performance of private savings.
In a small economy, diversification is necessarily conditional on exports, since
local markets are too narrow in most sectors to allow profitable industrial investment.
The development policy tool most often used to boost exports is the establishment of
EPZs, as in Mauritius and, more recently, Ghana.
EPZs are “zones” (not necessarily defined in geographical terms) in which
producers enjoy tax exemption and a measure of deregulation. They spread widely
during the 1980s, especially in Asia and the Caribbean, and a few were established
in Africa. Worldwide, there were 500 EPZs in 1996, in 73 countries, compared to a
dozen in the 1970s. In some countries, EPZs have helped to create tens of thousands
of jobs.
In theoretical terms, the point of setting up an EPZ lies in the fact that clustering
effects and externalities are created within an enclave by attracting multinational firms
possessing the resources needed to export (marketing, networks etc.).
The case of Mauritius, which set up one of the first EPZs, shows that this incentive
framework can play a major role in diversification, even though the process was
somewhat slow in getting under way. It was the EPZ that allowed Mauritius to diversify
its economy considerably in the 1980s, whereas previously it had been specialised
exclusively in the sugar industry. The country thus became the leading African exporter
of clothing and the world’s third–largest producer of woollen goods. More recently, it
has further diversified into industries such as electronics.
74
Figure 2.6. Diversification Indices for Mauritius
Analysis of failures can also be instructive. The duty–free zone in Dakar,

Senegal is a well–known example showing that EPZs fail when the overall economic
environment does not function properly. The main reasons for the failure of the
Dakar free zone were: the severe restrictions imposed by the labour code, a market
for skilled labour dominated by rent–seeking behaviour, high costs for inputs like
water and electrical power, and the lack of local investors (in contrast to Mauritius,
where 60 per cent of the investments in the EPZ were of local origin). Conversely,
the EPZ in Mauritius succeeded because policymakers had grasped that it could not
be considered in isolation from the rest of the economy: in addition to creating the
EPZ, Mauritius adopted liberal policies, and more generally created an environment
favourable to investment.
The idea that an EPZ can be an effective catalyst for industrial diversification
only in a relatively liberal economic environment is consistent with our previous
analysis of rent–seeking behaviour. In a highly restrictive environment, an EPZ is
regarded primarily as an opportunity to capture rents, and thus leads to a migration of
existing export ventures into the EPZ, without creating any really new ventures. This
does not occur if the economic environment is fairly liberal, because in this case no
wealth can be derived from exploiting the differences between EPZ and non–EPZ tax
and regulatory regimes. Under these conditions, finding profitable investments in the
EPZ requires the creation of real export ventures.
75
An econometric analysis of a dozen countries with EPZs and a dozen without
EPZs (Johansson and Nilsson, 1997) confirms these observations. A significant catalytic
effect is seen when the countries that established EPZs have open economies, like
Hong Kong, Malaysia, Mauritius, Singapore and Sri Lanka. Conversely, a negative
effect is observed in economies that were relatively closed during the period considered
by the authors, such as Mexico and the Dominican Republic.
The success of a diversification policy backed by the creation of an EPZ–type
mechanism may also depend on the openness of foreign markets. Mauritius’s success
was partly due to the favourable treatment granted by the European Union in the
context of EU–ACP relations, notably with respect to the application of the GATT
Multi–fibre Agreement. Other African countries that diversify in the future might not
receive such generous preferential treatment, if only because of the dismantling of the
Multi–fibre Agreement. Generally speaking, the European Union offers less liberal
treatment to the products of ACP countries when these products compete directly with
its own industries. The proposals for reciprocal preferential agreements which the
ACP countries and the European Union will negotiate over the coming decade (under
the new convention signed in June 2000) contain no indication that the ACP countries
will be protected against measures introducing non–tariff barriers (e.g. anti–dumping
measures) in the event that they diversify into new industries. The experience of the
barriers encountered by emerging countries, which are the main targets of anti–dumping
duties, suggests that this risk should be monitored, although it is not yet visible today
given the low level of industrialisation of the countries considered.
Hypotheses for Future Trends
In our baseline scenario of future growth, we assume that the diversification

indices of Uganda and Ghana, which have shown fairly significant improvement since
the beginning of liberalising reforms in these countries, will continue to rise through
2020. Their current level is still very low, which means they have very great potential
for progress unless their trade partners place restraints on their exports. Thus in our
baseline scenarios, we postulate that the trend observed since the beginning of the
reforms will persist, which implies an annual growth rate of 2.8 per cent for Ghana
and 3.1 per cent for Uganda. Under these assumptions, and based on CEPII data
presented by Berthélemy and Chauvin (2000), Ghana would in 2020 reach the current
level of diversification of Malaysia, the Asian newly industrialising economy (NIE)
which is the least advanced in this respect, whereas Uganda would remain at a level
below Ghana’s current level.
In our high scenario, we assume that in 2020 Ghana will have reached a level of
diversification comparable to that currently observed in the Asian NIEs which are
most advanced in this respect, such as the Republic of Korea. This assumption entails
considerably faster growth in Ghana’s diversification index (see Table 2.10),
i.e. approximately 5 per cent per year. Such a development will nonetheless require a
diversification of industrial activity which, for the time being, the EPZ established in
76
Table 2.10. Hypotheses concerning Annual Growth of Diversification
Baseline scenario High scenario

Burkina Faso 0 3
Côte d’Ivoire 0 3
Ghana 2.8 5.2
Mali 0 3
Tanzania 0 3
Uganda 3.1 6
1996 has not really brought about. Our high scenario assumes that, in 2020, Uganda
(well behind Ghana in this area) will have reached Mauritius’s current level of
diversification (which lies between the current levels of Ghana and Malaysia),
corresponding to 6 per cent annual growth of its diversification index.
For Burkina Faso, Côte d’Ivoire, Mali and Tanzania, where there is no indication
that economic diversification began to rise in the 1990s, our hypothesis for the baseline
scenario is that the level of diversification will remain constant. This is consistent
with the fact that these countries have made less progress on reform than Ghana and
Uganda. In the high scenario, we assume that these countries will progress at a pace
comparable to that currently obtained by Ghana and Uganda, i.e. a rate of increase of
diversification of 3 per cent per year, which amounts to assuming that they will adjust
their incentive structures in a way comparable to what was done in these two countries.
For Côte d’Ivoire and Tanzania, the result would be a level of diversification in 2020
comparable to that observed in Malaysia in the late 1990s. Burkina Faso and Mali,
which are starting from a less diversified situation, would in 2020 reach a level of
diversification comparable to that of Malaysia in the mid–1980s, and slightly above
Ghana’s current level.
Conclusion
Developing realistic hypotheses about the structural changes that these economies
may undergo in the future is no easy matter, given the tenuous nature of progress
being achieved today. The example of success stories in other countries shows, however,
that deep–seated structural change can take hold once an economy has accumulated
the production factors needed.
The institutional framework will play a fundamental role in this process. There
will be no labour reallocation away from agriculture if the incentive structure required
for agricultural modernisation is not in place. Similarly, there will be no diversification
of the economy if potential investors cannot undertake profitable ventures in new
industries, or if the regulatory framework and the existence of rent situations afford
them larger profits in existing industries.
Although progress in diversification can be achieved only over the long run, the
institutional frameworks needed to encourage such progress should be established
immediately.
77
78
PART TWO
ANALYSIS OF GROWTH FACTORS

IN SIX AFRICAN COUNTRIES
79
80
Chapter 3
Capital Accumulation
We showed in Chapter 1 that the countries experiencing rapid growth generally

had investment rates well above the average, with the notable exception of Uganda
and, to a lesser extent, Ghana. Even so, capital accumulation does not seem to have
contributed much to the rapid growth observed recently in certain African countries,
by comparison with the fast–growth countries of Asia or the episodes of rapid growth
in Africa during the 1960s and 1970s. The investment performances of the fast–growing
African countries during the 1990s thus might well prove to be insufficient. In any
event, they are substantially lower than those seen in Asia. The sustainability of growth
in the countries that performed well in the 1990s may therefore depend on stepped–up
capital accumulation in the future, for two reasons. First, the contribution of capital to
growth will necessarily remain modest and second, the sluggishness of investment
may hold back the structural changes in economic activity that are needed to obtain
productivity gains.
The improvement already observed in the majority of the fast growers indicates
that there are several factors which drive progress in this respect. Our analysis in this
chapter shows in particular that, on the demand side, investment is sensitive to increases
in factor productivity; this ensures that progress will be made in those countries where
structural change and human capital accumulation make it possible to achieve such
gains. Where the supply of savings to finance investment is concerned, however, our
conclusions are less optimistic. National financial systems are still inefficient, and
inflows of foreign savings have little effect on growth because they are used more for
consumption than for investment. The scenarios demonstrate that inflows of foreign
capital (aid and FDI) remain essential for long–term growth, provided that these are
deployed effectively.
81
Savings and Investment
Weak Investment
The poor contribution of capital to economic growth in the countries currently

going through a strong growth period corresponds to investment rates that are low by
historical standards, although these countries perform better on average than current
slow–growth countries. This is particularly true for Ghana and Uganda, as shown in
Table 3.1. In these two countries, however, investment ratios have improved over
time, reaching about 19 per cent in Ghana and 15 per cent in Uganda in 1996.
By contrast, Mauritius has a reasonably high investment ratio, averaging over
24 per cent of GDP during the studied period, while the corresponding figure is 33 per
cent for Botswana. Mauritius’s lower investment rates as compared to Botswana can
be partially explained by the country’s much smaller endowment in natural resources.
Mozambique and, recently, Burkina Faso, Mali and Tanzania are also showing
reasonably good investment performance. In Tanzania, however, the observed figures
are misleading because the capital/output ratio is very high, which means that a 20 per
cent investment rate does not suffice to promote growth1.
In summary, the picture with respect to investment is rather mixed in fast–growing
African countries and in recent candidates to fast–growth status. Some improvement
has been achieved, but overall investment performance in most of the countries remains
below that of the leading countries such as Mauritius and Botswana, and far below
Southeast Asia, where investment rates often exceed 30 per cent.
Weak Savings
The level of local savings relative to both national savings (which include net
current transfers and factor income) and domestic savings is especially weak. In the
earlier growth periods, domestic savings were usually higher than national savings.
The principal reason for this is that these economies owe large interest payments and
other capital income to foreign creditors. As an example, in Côte d’Ivoire I and Côte
d’Ivoire II domestic savings covered over 100 per cent of investments, but the country
paid substantial amounts of factor income (interest and dividends on foreign capital)
and private transfers (notably remittances from migrant workers) to the rest of the
world. As a result, its national savings cover only a fraction of its investment. By
contrast, national savings are higher than domestic savings in Ghana, Mozambique
and Uganda — and in recent years, in Burkina Faso, Mali and Tanzania as well. This
is attributable to the fact that these countries have received significant amounts of
grants and other transfers during their respective reform periods, without paying much
abroad in terms of factor incomes. As will be demonstrated below, foreign aid has
been an important source of financing for investment.
82
Table 3.1. Savings and Investment Performance
National Domestic Domestic
National savings/
Country Period savings/ savings/ investment/GDP
GDP (%)
investment investment (%)
Failed take offs

Algeria 1977-85 1.00 1.05 36.9 37.1
Cameroon 1970-86 0.52 0.90 22.9 12.0
Côte d’Ivoire I 1968-78 0.95 1.23 22.7 21.6
Egypt 1970-90 0.65 0.58 23.4 15.2
Kenya 1967-81 0.68 0.85 23.5 16.1
Malawi 1977-79 0.66 0.67 26.6 17.6
Morocco 1968-80 0.76 0.64 21.4 16.2
South Africa 1967-74 0.88 1.01 28.6 25.3
Tunisia 1968-85 0.90 0.89 27.8 25.0
Average 0.75 0.85 25.4 19.4
Current extended growth periods

Botswana 1971-95 1.09 1.09 32.8 35.6
Ghana 1984-96 0.75 0.46 14.7 11.1
Mauritius 1981-96 0.99 0.94 24.2 23.8
Mozambique 1987-96 0.40 -0.14 26.5 10.5
Uganda 1987-96 0.37 0.15 13.7 5.1
Average 0.72 0.50 22.4 17.2

Burkina Faso 0.60 0.31 22.4 13.3
Côte d’Ivoire II 0.56 1.61 13.2 7.3
Mali 0.62 0.37 26.0 16.2
Tanzania 0.75 0.09 20.0 14.7
Note: The periods differ from previously studied periods owing to lack of data. All aggregates are measured in current
prices.
Sources: World Development Indicators CD-ROM, 1999 (Algeria, Botswana, Egypt, Ghana, Malawi, Mauritius,
Mozambique, and Uganda) and African Development Indicators CD-ROM, 1998/99 (other countries).
The two exceptions among the current high–growth episodes are once again
Botswana and Mauritius. In Botswana, investment is more than fully covered by
savings, whether measured as national or domestic savings. Botswana does not have a
heavy debt burden and is actually a creditor of the World Bank, which is rather
exceptional in an African context. In Mauritius, national savings cover close to the
totality of investments.
Among the earlier growth periods, savings rates can be considered low or modest
only for Cameroon, Egypt, Kenya, Malawi and Morocco, although savings dropped
substantially in Côte d’Ivoire after the fall in cocoa and coffee prices around 1979.
83
The Role of Foreign Financing
To a large extent, investment performance in these countries is financed by foreign

capital inflows, due to exceptionally low levels of local savings, whether these are
defined as national savings (including net current transfers and factor incomes) or as
domestic savings (see Table 3.1). Again, this is particularly obvious in the current
period. Among the countries currently enjoying strong growth, it is clear in the cases
of Ghana, Mozambique and Uganda.
Inflows of foreign capital represent approximately 50 per cent of financing for
investment in the fast–growth countries identified earlier, and 70 per cent if we include
foreign aid in the form of interest–free grants. In the countries having recently experienced
take off, inflows of foreign capital play an equal, if not more important, role2.
Determinants of Capital Accumulation
An analysis of the determinants of capital accumulation is shown in Table 3.2

below. In order to be consistent with the long–run production function above, we use
the variation in capital stock rather than the investment ratio. The equation reported in
this table will be simulated for future years to build our growth scenarios.
Table 3.2. Determinants of Capital Accumulation

Dependent variable: DLK
Variable Coefficient Standard error t-statistic
LTFLOWSK(-1) 0.170 0.066 2.58

LKPROD(-1) 0.113 0.010 11.65
LTOT(-1) 0.034 0.006 5.70
REVCOUP(-2) -0.007 0.003 -2.35
Estimation method: within (fixed effects)

Number of obs.: 566
Number of countries: 26
R squared adjusted: 0.50
Hausman Test: χ2(2)=18.1
Notes: DLK=variation in ln(capital stock), REVCOUP=dummy for revolutions and coups d’état,
LTFLOWSK=real net long-term capital flows (loans, aid and foreign investment) as a ratio
of the capital stock, LKPROD=ln(0.4*GDP/capital stock), LTOT=ln(terms of trade). The
numbers in parentheses indicate a lag operator. Fixed effects are not reported.
84
Risk
Political instability appears to influence investments negatively, as seen by the

lagged variable representing the number of revolutions and coups d’état in the recent
past. Risk is an important and complex issue in Africa. According to Collier (1998a),
Africa was ranked the riskiest continent in the Institutional Investor risk rating. Our
variable represents only one of many facets of risk. It would need to be considered in
conjunction with other variables such as macroeconomic stability, the quality of
institutions, the reversibility of policies, the risk of expropriation by the state, the
possibility of recourse to the courts, the availability of insurance and forward markets etc.
It would be beyond the scope of this study to explore these aspects, due to the lack of
long, systematic time series on African institutions. We therefore decided simply to
neutralise this variable in our growth scenario exercise, assuming that the political
environment will stay reasonably stable in the countries studied over the next 20 years.
Overall Economic Efficiency
Apparently, low savings rates can take only part of the blame for the unsustainability
of investment–driven growth in several of our selected countries. Low levels of productivity
may have discouraged continued high investment rates, a hypothesis that will be tested
econometrically below. Conversely, a high level of productivity would be an incentive for
investment, because it would improve profitability. Moreover, higher productivity — and
hence higher profitability — facilitates the financing of investments.
Rapid productivity gains in Botswana engendered strong investment performance.
This process was admittedly facilitated by the country’s rich endowment in natural
resources, but this is not a sufficient explanation. South Africa has similar endowments,
but its rapid growth episode proved to be unsustainable. Mauritius is another example
where high levels of productivity have promoted investment. The fact that investment
is still rather weak in Ghana, Mozambique and Uganda can be partially explained by
the lower TFP levels in these countries. Another reason, as we saw above, is that low
levels of national savings have made them highly dependent on external financing. In
the future, improvements in TFP, which will be obtained as a result of changes in all
the structural variables introduced in our production function equation in Chapter 2,
will increase incentives to invest.
Debt Burden
Where variables influencing financial capacity are concerned, we found that

investments were strongly dependent on debt flows. The variable LTFLOWSK
represents net long–term capital flows (loans, aid and foreign investment) as a ratio of
the capital stock. The degree of dependence on foreign aid is rather worrisome in
some cases, seeing that several African countries are struggling to reduce their debt
burden, although the Highly Indebted Poor Countries (HIPC) initiative has provided
85
some relief in the cases of Mozambique and Uganda in recent years, and will assist
several other countries including Burkina Faso, Côte d’Ivoire, Mali and Tanzania in
the near future. In addition, the level of aid flows is declining, due to budget constraints
in donor countries. Hence, African countries will not be able to rely on debt flows and
foreign aid as dynamic sources for financing investment. By contrast, debt is not an
issue in the case of either Botswana or Mauritius, making their prospects for sustainable
growth substantially brighter.
Terms of Trade
The significance of the terms–of–trade variable provides an important insight

into the earlier growth processes. With the exception of South Africa, Morocco and
Tunisia, the historical growth episodes all correspond to commodity booms. The surge
in export prices had a double effect on investment, by increasing financial resources
— from both export receipts and improved access to international capital markets —
and enhancing incentives. When export prices collapsed, however, these countries
were left with economies where essentially no positive structural change had taken
place but where distortions began to increase dramatically.
Financial Sector Reform and Domestic Policy
An analysis of recent developments in the financial sectors of the countries for

which we will build growth scenarios will suggest that there are few prospects of
improvement in this area. Yet the strength of the financial sector has an impact on
investment through the mobilisation of domestic savings (see Berthélemy and
Varoudakis, 1996). Although we did not succeed in establishing statistically significant
relationships in this respect, partly for lack of adequate data, these factors should also
be kept in mind3. African financial systems are still limited essentially to the banking
sector, since financial markets are still at an embryonic stage. As is generally the case
in developing economies, banking activity is equally undeveloped because of the
paucity of available savings.
Banking Sector Reforms
All of the banking sectors followed more or less the same course in the 1980s
and 1990s. In the early 1980s, banking activity was constrained by extensive state
intervention with controlled borrowing and lending rates, credit restrictions and sectoral
standards governing the granting of loans. Governments also intervened directly in
the allocation of loans to borrowers, which was particularly easy to do because the
main banks were wholly or partially state–owned. In addition, the legal environment
was not very favourable to risk taking by banks, owing to a lack of adequate guarantees
in case of default. Lastly, banks were generally under–capitalised.
86
As a result, banks were in a very poor position in the late 1980s. In particular
they had accumulated an alarming proportion of bad loans often exceeding 30 per
cent of their assets and sometimes much higher. This situation led beginning in the
late 1980s to banking reform and restructuring plans as well as initiatives to liberalise
interest rate policies.
All of the six countries studied here went through this process. The countries
which had made the most progress in terms of reforms and renewed growth did not
take corrective action much earlier than the others. The Franc Zone countries benefited
from the existence of a common regional central bank (the Central Bank of the West
African States, or BCEAO), which intervened in all Zone countries to ensure that
needed reforms were implemented.
The initial stage of banking reform may thus be regarded as complete or well in
hand in all of these countries, but their banking systems are not highly developed.
Their M3/GDP ratios, which are usually considered a useful indicator of the degree of
banking development, range from 20 to 25 per cent in all of the countries except
Ghana, whose M3/GDP ratio stands at about 12 per cent. Furthermore, these ratios
have not really increased since the early 1990s, except in Ghana and Uganda, where
they were initially very low.
By comparison with the other countries that have recently enjoyed strong growth,
the countries considered in this study are not really exceptional. According to the
indicator of banking development used, their situations are comparable to those of
Botswana and Mozambique. In Mauritius, by contrast, the M3/GDP ratio is nearly
80 per cent, but Maurititus has a substantially higher level of per capita income, far
above what the countries studied will be able to attain during the next two decades,
even under the most optimistic projections.
Embryonic Financial Markets
Financial markets are extremely thin. In 1998, the six countries of emerging
Africa under study all had stock exchanges. The two oldest exchanges are those of
Côte d’Ivoire (the Abidjan Bourse des valeurs, created in 1976) and Ghana (Ghana
Stock Exchange, 1990). The Abidjan exchange became the Bourse régionale des
valeurs mobilières (BRVM – Regional Stock Exchange) in September 1998 and
since then its role has been to serve all the member countries of the West African
Economic and Monetary Union (WAEMU) including Burkina Faso and Mali. The
stock exchanges of Uganda and Tanzania (Dar–es–Salaam Stock Exchange) began
to operate only in 1998.
The experience of Côte d’Ivoire and Ghana shows that stock exchanges play a
very modest role in these economies. In 1996, the market capitalisation of the Abidjan
Bourse des valeurs amounted to only 8.5 per cent of GDP (compared to nearly 200 per
cent in South Africa). The opening of the market in 1998 did not change this situation,
as the BRVM has not really succeeded in attaining a regional scope. In Ghana, stock
87
market capitalisation amounts to 23.5 per cent of GDP, which is approximately the
level reached by Malaysia 20 years earlier (in 1996, Malaysia’s market capitalisation
was nearly 100 per cent of GDP).
The thinness of financial systems prevents them from providing high–quality
services to customers, for several reasons: lack of competition, of diversification of
savings products, of financial innovation in lending and of responsiveness to demand.
The weakness of financial systems helps to explain the low level of savings.
The prospects for improvement of financial systems are uncertain. Local financial
markets are still too thin (the attempt at regionalisation within the WAEMU has revealed
its limits) to enable a dynamic financial sector to emerge. The reforms of the 1990s
stabilised banking systems, which were initially in a very poor situation, but did not
manage to give them new dynamism. It thus seems probable that the conditions under
which these financial systems collect national savings will not change greatly between
now and 2020. Where financial sectors are concerned, our scenarios do not assume any
radical change in the trends and structural parameters observed in the past. For this
reason, foreign capital inflows should continue to play an important role in the future.
Inflows of Foreign Capital
In examining the determinants of investment to measure the effect of capital

contributions, we have used a variable that takes account of grants and all net long–
term capital flows, including direct investment. Its value fluctuated fairly widely over
the 1990s, notably because of the suspension of a number of loan operations to Franc
Zone countries in the years immediately before the devaluation of the CFA franc. We
therefore used the average level of financial flows observed from 1995 to 1997 as the
baseline for our simulations of future growth. The individual countries receive different
amounts of aid, in accordance with their levels of development, policies, and geo–
strategic positioning. On the basis of recent observations (average for 1994–96), we
assume that at the end of the 1990s, Burkina Faso, Ghana and Mali were receiving aid
flows equivalent to 14–15 per cent of their respective GDPs. For Uganda and Tanzania,
these flows were equivalent to 11–12 per cent of GDP. Côte d’Ivoire, which has higher
per capita income and is indebted to the private sector (and thus has tighter schedules
for repayment of principal), has received considerably less financial aid in the last few
years. We assume in our scenarios that such aid will initially be equivalent to 5–6 per
cent of its GDP. For subsequent years, we assume that these capital flows will increase
steadily at a rate of 2.5 per cent per year in real terms, which should match the growth
rate in the OECD countries that provide these financial resources, but should be less
than the growth rate in the recipient countries. The share of external financing in the
latter’s GDPs would therefore decrease over time.
88
The Role of Aid in Investment
The results of our scenarios are not very sensitive to the level assigned to this
variable because each dollar of foreign funding contributes appreciably less to
investment than these accounting data suggest. This is shown by the estimation results
reported in Table 3.2 where the coefficient of the “net capital flows and grants” variable
(of the order of 0.17) may be interpreted (subject to linearisation) as the economy’s
propensity to transform net resource transfers from abroad, in the form of grants and
capital flows, into investment.
The value of this coefficient is fairly low because some of the resources transferred
are consumed rather than invested. In terms of the savings–investment balance in the
national accounts, this means that national savings react negatively to foreign savings,
a phenomenon that has been well known ever since the effect of aid on growth was
investigated in the 1970s.
Table 3.3. Share of Social Sectors in Foreign Aid

(percentages)
1990 1998
Burkina Faso 21.3 68.1

Côte d'Ivoire 16.6 54.8
Ghana 25.3 41.5
Mali 31.1 52.7
Tanzania 34.5 44.3
Uganda 47.6 64.5
All Sub-Saharan Africa 28.6 55.6
Note: Calculated on the basis of the share in those aid flows which can be allocated by sector.
Source: Authors’ calculations based on DAC data (Creditor Reporting System).
Other factors also help to explain the low level of this parameter. First, aid is
largely fungible4. Although it is intended to serve for investment in either physical or
human capital, aid enters into the budgets of recipient governments and consequently
frees up their own budget resources for other uses. Second, in some cases aid is purely
and simply misappropriated.
An increase in the propensity to invest foreign resource transfers could be obtained
by limiting the fungibility of such aid, and of course by limiting opportunities for
corruption and misappropriation. Donor countries have shown themselves to be
determined to make progress in this direction, notably in the context of the heavily
indebted poor countries (HIPC) debt–reduction initiative, which is by definition a
highly fungible form of aid. When the HIPC initiative was enhanced by decision of
the G–7 in Cologne in June 1999, it was noted that this debt relief would have to be
accompanied by a clear commitment from beneficiary governments to use the freed–
up resources to invest in development.
89
There is nothing abnormal in the fact that not all foreign aid goes to investment.
A part of such aid serves to finance social projects, and thus entails both the construction
of social infrastructures and payment of their current operating expenditures. Since
current social spending, notably for education and health, is classified as consumption
rather than investment, it is natural that part of these aid budgets should appear as
consumption in the national accounts, even though the function of such spending is
much more closely related to investment in human capital.
Table 3.3 provides a rough idea of the proportion of aid allocated to social sectors.
Although the data were obtained by approximation (see table note), they indicate that
the social sectors’ share of aid budgets is large and growing. In 1998, from 40 to
70 per cent of aid went to social projects, depending on the country, with the average
proportion in sub–Saharan Africa at 56 per cent. This provides a partial explanation
for the low propensity to invest foreign resource transfers, investment being defined
in the narrow sense used in the national accounts.
This objective will probably be difficult to meet, despite the determination of
donors. The fact that Uganda, which in 1998 became the first country to benefit from
debt–relief measures under the HIPC initiative, increased its military spending at the
same time –– owing to its involvement in the conflict in the Democratic Republic of
Congo –– suggests that there is still a great distance between word and deed.
We nevertheless examine, as part of an optimistic scenario, the growth impact of
an increase in the parameter representing the propensity to invest foreign resource
transfers. To take account of the fact that a large share of aid is allocated to current
social expenditures we limit this parameter to 50 per cent (which represents the average
order of magnitude of social spending in aid budgets). Even when limited in this way,
such a change in behaviour concerning the use of foreign aid would have substantial
positive consequences for growth.
What Can Donor Countries Do?
We have discussed the lack of structural change and investment in recent African
growth processes. Moreover, our scenarios clearly show the importance of structural
change for development. For the great majority of African countries, the lack of
domestic savings is obviously a major obstacle in this respect. In addition, many
investments involve imports, for which foreign exchange is required. Berthélemy and
Söderling (1999a) demonstrated that foreign aid is not a realistic alternative to export
growth for achieving a greater volume of imports. Aid and debt forgiveness nevertheless
remain important tools for assistance in Africa, as servicing the large debt accumulated
by many African countries drains a significant share of export revenue and other capital
inflows. Previous experience has shown, however, that domestic conditions must be
right if foreign aid is to have the desired impact. The following section will analyse
two aspects of the relationship between donor countries and Africa: debt forgiveness
and the HIPC initiative, and aid efficiency5.
90
Debt Relief
The heavy debt burden inherited from previous decades has been an obstacle to
economic growth for many African countries in recent years. The Cologne initiative,
designed to enhance the 1996 Highly Indebted Poor Countries (HIPC) initiative, should
help these countries to solve their debt overhang problem.
Our model is not specifically designed to analyse debt overhang, which can be
considered as the result of multiple equilibria, in which a high level of debt prevents
growth and the lack of growth in turn precludes any solution to a country’s financial
difficulties (see e.g. Berthélemy and Vourc’h, 1994; for a discussion of the Cologne
initiative, see Berthélemy, 1999).
Reducing the debt burden can have direct and indirect positive consequences
for African economies. Directly, lowering the debt — or more precisely, the net present
value of debt — frees up resources for investment in productive projects. Solving the
debt overhang should also, in principle, have an indirect positive impact on investment,
insofar as large debt stocks create disincentives for private investment. This indirect
mechanism is not incorporated in our model, as it would have been difficult to quantify,
but a quantitative assessment of the direct effect can be attempted.
The Cologne initiative will provide highly indebted poor countries with debt
relief allowing a reduction in the net present value (NPV) of the total debt stock to a
level considered sustainable. In the 1996 HIPC initiative, the principal sustainability
criterion was an NPV/export ratio less than or equivalent to 200 per cent; for the
Cologne initiative, this threshold ratio has been lowered to 150 per cent. A second
criterion, for countries that are highly open to foreign trade, is an NPV/fiscal resources
ratio less than 250 per cent (as compared with 280 per cent in the 1996 HIPC initiative).
In our sample, this second criterion is relevant only to Côte d’Ivoire.
Among our sample countries, Uganda had already received a debt relief package
based on the 1996 HIPC programme before the Cologne initiative was launched. In
order to take account of the enhancement of the previous scheme, Uganda’s current
package will be modified in accordance with the new rules adopted at the Cologne
summit. Three other countries — Burkina Faso, Côte d’Ivoire and Mali — have already
reached the point of decision, and the process should be completed in 2001 or 2002.
Finally, Ghana and Tanzania are also eligible for the initiative, and the IMF expects
them to reach the decision point in 20016. An evaluation of the amount of the NPV
debt reduction proposed in the Cologne initiative is provided below. This evaluation is
based on the Cologne debt sustainability criteria and on projections of the NPV/export
and NPV/fiscal income ratios at the completion point of the initiative (for countries
that have reached the decision point) or at the decision point (for other countries), as
they are reported in IMF assessments of foreseeable debt relief packages.
This evaluation is an upper limit of the direct impact of the Cologne scheme,
inasmuch as we compare the NPV of debt after the completion point with the previous
level of the NPV of debt, without taking account of what would have been provided
in previous debt relief schemes such as the 1996 HIPC programme. Table 3.4 reveals
91
wide variations in the amount of debt relief that would be provided to different
countries. In our sample, Tanzania would be by far the largest recipient, due to its
current high level of indebtedness. The amount received by Tanzania in this scenario
would be even higher than the largest HIPC package given so far (for Mozambique,
which received $1 442 million). Burkina Faso and Mali, which have much lower
debt levels (partly attributable to previous bilateral debt reductions granted by France,
and partly to stricter financial control in previous years), would receive much less
debt relief.
How would such a decline in the stock of debt affect the economy? In our
model, debt relief has essentially the same effect as an increase in aid flows.
Therefore, it should improve the balance of payments and to some extent enhance
the financing of investments. It is difficult, however, to compare this relief to the
flow of resources that the countries under study are expected to receive in coming
years in our scenarios.
Table 3.4. Direct Effect of the Cologne Initiative on Financial Flows

NPV of debt NPV of debt Cologne NPV Equivalent New financial % Aid flow
before relief after relief debt relief aid flow flows (1997) increase
equivalent to
Cologne
[1] [2] [3]=[1]-[2] [4]=[3]*(ρ-g)/(1+ρ) [5] 100*[4]/[5]
Burkina Faso 833 525 308 10 339 3

Côte d’Ivoire 7 837 6 689 1 148 38 567 7
Ghana 5 982 5 127 855 28 969 3
Mali 1 403 956 447 15 382 4
Tanzania 7 177 4 238 2 939 97 758 13
Uganda 1 796 1 076 720 24 665 4
Note : Columns [1] to [5] are in $ million.

ρ = discount rate (6 per cent) ; g = growth rate of foreign resource transfers (2.5 per cent).
In other words, the direct impact of the Cologne initiative will be modest for the
countries studied, and there is no justification for creating a new scenario that would
specifically take into account this new debt relief package with greater precision than
in the rough calculations presented above. However, one should not disregard the
possibility of a greater indirect impact that would facilitate future growth through
improving incentives to invest.
92
Although debt relief will be based on a stock reduction, its direct impact on the
economy will not appear up front, because the reductions in debt service will, in
principle, be stretched over a long period. In our model, the amount of debt relief
provided by the Cologne initiative needs to be compared to the flow of aid that the
economy will receive. We therefore propose to estimate the flow of grant aid as having
the same net present value as the debt relief provided by the Cologne scheme. In our
scenarios this flow will be compared with the flow of net resource transfers that each
economy will receive in the future. For the sake of comparison, we compute this
equivalent flow of grant aid by assuming that it will grow at the same rate (2.5 per
cent a year) as other net resource flows. Table 3.4 provides such a comparison, based
on an annual discount rate of 6 per cent. The result is clear: the direct impact of the
Cologne initiative is rather small in terms of equivalent supplementary resource flows.
For Tanzania, which would receive the highest debt relief in the Cologne scheme, the
impact is equivalent to 13 per cent of the flow of resources assumed in our baseline
scenario. For Burkina Faso, Ghana, Mali and Uganda, it is as low as 3–4 per cent of
the corresponding flow.
Aid Efficiency
To study the importance of aid efficiency, we assess the impact of an improvement
in the coefficient measuring the effect of aid and other capital flows on investment
(Table 3.5). This parameter can be interpreted as the share of foreign capital inflows
that is actually invested in the economy, with the rest going to consumption through
various leakages, as well as through general equilibrium effects. In moving from the
high scenario to the “enhanced aid” scenario, this parameter increases from the rather
low estimated level of 17 per cent to 50 per cent. It should be pointed out that the
estimated coefficient is to some extent an average for a relatively large number of
countries over a fairly long period and that it may vary substantially between countries
and over time.
Governments are likely to have a greater influence on this coefficient than on
other parameters estimated in the model. Hence, it may be unfair to suggest that all
countries in our sample are currently at the low level of aid efficiency represented by
the estimated coefficient. The difference in income per capita between the two scenarios
should therefore be interpreted purely as an indication of the importance of aid
efficiency, rather than as a measure of how much the countries in question would
actually gain if they improved their aid efficiency to a reasonable level.
The remarkable importance of aid efficiency can be clearly seen from Table 3.5.
For the countries under review, roughly a doubling in aid volume would be required
to obtain a result equivalent to that of increased aid efficiency. Naturally, the countries
receiving the most aid — Ghana, followed by Tanzania and Uganda — would need
the most additional aid in absolute terms in order to obtain such a result. The volume
of aid and other capital flows would become unrealistic, exceeding $1.5 billion per
year in four of the six countries in the sample. One obvious problem from the donors’
93
Table 3.5. Net Aid and Other Capital Flows
(annual average for 1996-2020, constant $ million 1996)
Assumption Equivalent to enhanced aid efficiency

in scenarios Level Difference Difference (%)
Burkina Faso 462 854 393 85

Côte d’Ivoire 761 1 675 914 120
Ghana 1 316 2 697 1 381 105
Mali 519 960 441 85
Tanzania 1 028 1 696 668 65
Uganda 906 1 767 861 95
point of view is that they can do very little to influence aid efficiency in recipient
countries, at least in the short run. They can, however, target aid to countries
which have a demonstrated record of making sensible use of aid funds (see World
Bank, 1999).
Finally, in productive sectors, the catalytic role of aid for investment can be
strengthened to promote diversification. This can be done, for example, by supporting
financial deepening, a key element for increasing investment in higher value added
activities, rather than just in new, but still low value added commodities. Deepening
can also help attract FDI from OECD countries, by giving a push to the promotion of
joint ventures in capital risk operations. An example of such a push can be found in
the higher profile and stronger role given by the EU to the European Investment Bank
within the framework of the Cotonou partnership agreement with the ACP7.
Foreign Direct Investment
Foreign direct investment (FDI) is a source of outside financing that a priori has
a maximal impact on investment. Consequently, a natural way of increasing the
propensity to invest foreign financing would be to attract this type of capital, which
has the additional merit of generating no debt.
African countries currently receive very little FDI, however. In 1997, they
received a total of some $3.5 billion in net direct investment flows, which amounted
to 1 per cent of their GDP, whereas developing countries as a whole received, relatively
speaking, over three times as much.
Much of the FDI inflow goes to investment projects in mining sectors, and very
little to investment in manufacturing industry. Investments are thus fairly concentrated
in a few countries where there are opportunities to tap new natural resources. Tanzania
is a notable example, as can be seen from Table 3.6. The same table also shows that
substantial investments have been made in Côte d’Ivoire and Uganda, whereas foreign
investment was low or negative in Burkina Faso, Ghana and Mali.
94
Table 3.6. Share of Foreign Direct Investment in GDP
(percentages)
1997 Average 1990-97

Cote d’Ivoire 3.1 1.0
Ghana 0.5 0.4
Mali -0.5 0.4
Tanzania 2.3 1.3
Uganda 2.4 0.6
All Sub-Saharan Africa 1.0 0.5
Source: Authors’ calculations based on World Bank, Africa Database.
The direct investment flows recorded in Africa are not only low but also highly
variable over time. Consideration of the averages for the 1990–97 period provides
some perspective on previous observations. In particular, these averages show that,
after Côte d’Ivoire, Ghana is the country that received the most FDI in proportion to
its GDP during the 1990s.
The ability of the countries studied to attract direct investment seems still to be
mainly due to opportunities for exploiting natural resources. Such opportunities
certainly play the leading role in Africa, as witness Côte d’Ivoire and Tanzania in our
sample, as well as in the oil–producing countries. Progress on liberalisation reforms is
often considered to be a decisive factor in theory8, yet our data do not support this
assertion. Ghana and Uganda initiated reforms fairly early on and posted strong growth
performances in the 1990s, but neither succeeded in attracting significantly more foreign
investment than the observed African average during that period. The opposite is true
for Côte d’Ivoire and Tanzania.
It should be recalled, however, that the available data are of uncertain quality,
and may underestimate the recent growth of foreign investment in Africa. For example,
UNCTAD data indicate that FDI in Africa exceeded $5 billion in 1997. In addition,
some direct investments are never accounted for as such. According to Bhinda et al.
(1999), for example, 70 to 80 per cent of the flows of private transfers recorded in the
current accounts of Uganda and Tanzania are in fact direct investments by Asian
entrepreneurs. In the case of Uganda, this would imply that direct investment flows
are underestimated by more than half. A similar phenomenon seems to exist in Côte
d’Ivoire in connection with investment by the Lebanese–Syrian community, although
it is difficult to quantify. The case of Ghana reveals another source of underestimation:
the available data clearly do not count the privatisation of the mining company Ashanti
Goldfields in 1994, although this operation alone involved the sale of $295 million in
the company’s equity to foreign investors, a sum more than ten times the annual flows
of direct investment recorded in the balance of payments.
Lastly, the privatisation process got under way belatedly in Africa, and
privatisations might stimulate FDI in coming years as they did in Latin America and
in Central and Eastern Europe.
95
There is thus some possibility, after all, that foreign direct investment will really
take off in Africa in the medium term. What is more, the fragmentary data available
concerning the return on investments in foreign subsidiaries suggest that these returns
are appreciably higher in Africa than elsewhere9. Although the risks are considerable,
policies to consolidate the progress and reforms already achieved should make it
possible to realise this potential in the future.
Conclusion
Investment in fixed capital will be a strategic parameter for Africa in the next
few decades. To date, investment behaviour has not been sufficiently dynamic to
promote lasting growth and structural change.
Obstacles have been encountered on both the demand and supply sides. Where
investment demand is concerned, investment in Africa, as elsewhere, depends on the
profitability of capital, and future productivity increases will provide the best assurance
of more dynamic capital accumulation. It also depends on political stability and more
generally on domestic and foreign investors’ assessment of the risks of investing in
Africa rather than elsewhere.
Where the supply of savings is concerned, African countries are still highly
dependent on foreign capital contributions. Saving capacity is intrinsically low, owing
to Africa’s low income levels. Moreover, since financial systems remain very thin and
vulnerable, it is not likely that they will obtain enough efficiency gains in the short
and medium term to allow them to stimulate and mobilise a more abundant supply of
local savings.
Foreign capital inflows have mainly taken the form of aid flows, including both
grants and loans. A large proportion of this aid is not allocated to investment, which
means that these aid flows are partly cancelled out — from the standpoint of financing
investment — by a direct or indirect reduction of national savings. Countries are thus
all the more dependent on foreign capital for financing growth.
An increase in foreign direct investment — and, by extension, in portfolio
investment — could probably bring about an appreciable improvement in this situation.
At the time of writing, the amount of such investment is modest, and is mainly geared
towards development of mining and oil resources rather than to investments in
manufacturing. There are some possibilities for increasing FDI, however, if only
because the available statistics probably underestimate the current growth rate of foreign
investment in a number of countries having made progress on the reform front.
96
Notes
1. In this respect, Tanzania is comparable to Kenya.

2. Côte d’Ivoire is a special case. Domestic savings are more than sufficient to cover
investment, but they are largely used to remunerate foreign production factors (capital
and labour).
3. Berthélemy and Varoudakis (1997) argue also that in a fixed–effects model estimation,
the role of financial development is to a large extent incorporated in the fixed effects for
the country: the principal impact of poor financial development is that it locks the economy
in a low equilibrium trap.
4. On this topic, see Assessing Aid, World Bank (1999) .
5. Actually, the analysis concerns the efficiency of all kinds of long–term capital flows in
terms of their impact on gross domestic fixed investment. Given the dominant proportion
of aid and concessional loans in these flows, we will nonetheless refer here to aid
efficiency.
6. Ghana had declared that it would decline debt relief under the Cologne initiative in order
to protect its financial creditworthiness, but has since changed positions. For this reason,
we also calculated the country’s potential gains from the scheduled debt relief.
7. Its role was more limited in the sucessive Lomé conventions that have preceded
“Cotonou”.
8. See for example World Development Report 1999/2000, World Bank (2000).
9. One study shows that the return on investment for subidiaries of US–based multinationals
is 25 per cent in Africa, as opposed to 12 per cent worldwide (figures cited in the 2000
Report of the Conseil des investisseurs francais en Afrique).
97
98
Chapter 4
Human Capital
Our econometric investigations have shown that human capital is an important

ingredient of long–term growth, for more than one reason. Not only does it contribute
to growth as a production factor, by making the labour force more productive, it is
also essential to the fight against poverty. Moreover, observation of economic success
stories such as Mauritius suggests that the availability of human capital is a necessary
condition for the achievement of structural change through diversification of the
economy and reallocation of labour away from agriculture.
In keeping with Chapter 2, this chapter examines human capital in its strict sense,
i.e. the training of the labour force. The health–related aspects of this issue are difficult
to incorporate in our growth scenarios because of the lack of adequate economic
evaluation and will not be considered here. In any event, as Chapter 1 showed, the
AIDS epidemic gives cause for serious concern, and no great improvement should be
expected on this front.
The African economies for which we have constructed scenarios of future growth
are fairly representative of the shortage of human capital observed in Africa generally,
as they are all at a relatively low level in this respect. Consequently, there is considerable
scope for future progress in these economies, particularly by comparison with South
Africa, Mauritius and Botswana, which are considerably further along, having already
begun their drives to increase school enrolments many years ago.
There are essentially three ways of obtaining progress in matters of human capital:
— First, investment in the training of individuals gradually increases the educational
level of the population, which leads to productivity gains for the entire economy.
This is the aspect considered in our previous econometric estimates, as represented
by the average number of years of schooling of the working–age population.
— Second, as far as the data allow us to judge, educational quality in Africa is
highly variable and generally very low. Consequently, the human capital of a
given African country cannot be directly compared to that of a more developed
99
economy. Here again, there is some scope for raising the effective amount of
available human capital, but improvements in educational quality can have a
high budgetary cost.
— Third, part of this theoretically available human capital is wasted when rent–
seeking activities pervade the economy — a characteristic feature of African
economies compared to the rest of the world. As a result, there is a largely
untapped pool of human capital that could be used in productive activities.
Drawing on this pool would be particularly favourable to growth since it would
be equivalent to an increase in human capital at no cost to the budget.
We will examine these three aspects in turn, using a detailed presentation of the
variables developed in the growth scenarios of Part One.
Developing Human Capital by Increasing School Enrolments
Table 4.1 illustrates the distance between the economies studied and the three
most advanced economies: Botswana, Mauritius and South Africa.
Table 4.1. Human Capital Indicators in 1996
Country Gross enrolment rate Average no. of Adult illiteracy rate

in primary school years of schooling
Burkina Faso 39 1.0 80

Côte d'Ivoire 71 2.8 58
Ghana 76 5.2 34
Mali 45 1.3 65
Tanzania 66 3.3 29
Uganda 74 3.3 37
Memo item:
Botswana 108 5.6 26
Mauritius 107 7.8 17
South Africa 131 7.0 16
Sources: World Bank, World Development Indicators, supplemented by African Development Bank, Selected Statistics on
African Countries, and authors’ calculations for the average number of years of schooling (based on Nehru and
Dareshwar) and for enrolment rates.
100
The average number of years of schooling, used here as the central indicator for
estimating the production function and constructing our growth scenarios, reflects the
state of progress in primary education in the countries considered. The proportion of
the population having received secondary and higher education is in fact very low.
The negative correlation between the average number of years of schooling and the
illiteracy rate is thus fairly good, except in the few countries such as Mauritius and
South Africa which have already reached an educational level at which secondary and
higher education play an important role (raising the average number of years of
schooling appreciably without reducing the illiteracy rate, which has reached a floor
level of approximately 10 percentage points; see Figure 4.1).
Note: Only those countries for which data are available are represented.
Source: See Table 4.1.
However, this correlation is not a full correlation. Two countries having the same
average number of years of schooling in the adult population may have illiteracy rates
differing by as much as 20 points. The reason is that the educational systems considered
function very differently, owing to varying degrees of effectiveness in delivery and to
the varying degrees of priority which governments give to primary as opposed to
secondary and higher education.
The growth of the human capital stock, as measured by the number of years of
schooling of the population, is constrained by a number of parameters, but it also
depends on policy choices.
101
First of all, such growth is all the more difficult to obtain because there are many
different age groups to be educated. The effort required to reach a given rate of growth
in the educational level of the population will be all the greater when the population is
growing rapidly. However, the differing rates of population growth among the countries
studied do not explain their divergent performances with respect to human capital
growth. As Table 4.2 suggests, these differences are basically due to the fact that the
efforts made are very different from country to country. This table uses the same
period of reference (1990–96) for each country, but obviously they did not all initiate
their educational drives at the same time — when they did so at all. From a forward–
looking standpoint, these data are not directly comparable across countries, because
what counts for the future is not the past slope of the human capital growth curve but
the current slope of this curve, based on today’s school enrolment policies. Differences
in initial levels also make comparison difficult.
Table 4.2. Recent Human Capital and Population Growth Rates
Growth rate (%)

Country
Human capital Population

Côte d'Ivoire 2.8 3.1
Ghana 1.4 2.7
Mali 1.3 2.9
Tanzania 2.4 3.0
Uganda 2.8 3.1
Memo item:
Botswana 2.7 2.4
Mauritius 1.9 1.0
South Africa 0.8 1.8
It is instructive, though, to compare the performances of Burkina Faso and Mali

in recent years, because these two countries started from comparable initial levels and
are experiencing similar population pressures. The difference between the two countries’
performances seems in contradiction with Table 4.1, which shows that both had
comparable primary school enrolment rates in 1996. This difference may be explained
by the fact that Burkina Faso reached enrolment rates of nearly 40 per cent in the early
1990s, whereas Mali attained this modest degree of progress only in 1996.
102
It is also interesting to compare Tanzania with Côte d’Ivoire and Uganda.
Table 4.2 suggests that Tanzania made significantly less progress than the latter two in
terms of human capital growth. In addition, a closer analysis of change over time
reveals that Uganda’s performance is clearly superior, due to the country’s past efforts
to improve education. Whereas Uganda’s primary school enrolment rate has remained
stable at around 70 per cent since 1987, those of Côte d’Ivoire and Tanzania have
steadily declined over the last two decades. In Côte d’Ivoire, the enrolment rate fell
from 79 per cent in 1980 to 67 per cent in 1995, then rose again to 71 per cent in 1996.
The downward movement in Tanzania was still more striking: the enrolment rate fell
from levels close to 100 per cent in the late 1970s to 66 per cent in 1996. If we consider
the recent progress made in Uganda, which had achieved nearly universal primary
education at the end of the 1990s (during which time Côte d’Ivoire did not progress),
we see a very clear ranking emerge among these three countries, with Uganda ahead
of Côte d’Ivoire, which is in turn ahead of Tanzania.
Lastly, Ghana stands out for its substantially higher level of human capital, which
largely explains why it recorded a much lower rate of growth of human capital than
Uganda despite its comparable rates of primary school enrolment. This country, where
the secondary school enrolment rate is much higher than in the rest of the region (the
available statistics indicate a rate of over 44 per cent, but this figure should be regarded
with caution), may thus be entering a stage comparable to that which Mauritius went
through in the 1980s and 1990s, in which human capital grew at only a moderate pace
because a high level had already been reached.
For the future, these comparative data suggest that there is room for improvement
of performances in the area of human capital growth, but at the cost of unremitting
efforts to raise enrolment rates. Ultimately, then, reducing illiteracy will require long–
term policy choices in favour of spending on education and — within education
budgets — in favour of primary education.
For our growth scenarios, we have calculated the growth of the human capital
stock (measured as the average number of years of schooling) on the basis of a projected
trend in the average number of years of primary education, obtained by incrementation
on the basis of assumptions concerning enrolment rates and population growth. The
human capital stock in the form of primary education will be supplemented by a simple
projection of the human capital stock in the form of secondary and higher education1.
The countries have followed very different trends where enrolment rates are
concerned. In recent years, enrolment rates have fallen by approximately 1 per cent
per year in Côte d’Ivoire and Tanzania (since 1990), remained stable in Ghana (since
1983) and risen by 1 per cent per year in Uganda (since 1987, but without counting
the rise resulting from the recent drive to provide universal primary education), by
2 per cent in Burkina Faso and by 10 per cent in Mali (since 1990).
103
We may assume, for the purposes of our baseline scenario, that the fall in
enrolment rates observed in the early 1990s will not continue. As a result, for this
scenario we will assume that primary school enrolment rates will be stable in Côte
d’Ivoire, Ghana and Tanzania. For Burkina Faso, which is starting from a very low
level, we will assume that the recent progress will continue at an unchanged pace,
bringing the country to a primary school enrolment rate of 69 per cent in 2020, a level
comparable to that of Côte d’Ivoire.
In the case of Mali, the sharp rise recorded recently will certainly not last, as
spending for the education sector has tended to drop over time as a proportion of GDP.
In the late 1990s, it was at a level comparable to that of Burkina Faso, at 2.2 per cent
of GDP. Consequently, although education spending increased in absolute terms, there
was a drop in both spending per pupil and the pupil/teacher ratio (over 70 pupils per
class in 1997). To avoid an excessively optimistic scenario based on the assumption
that recent trends will continue, we assume that the future progress of Mali will be
comparable to that of Burkina Faso, with an annual growth rate of 2 per cent.
Beginning in 1998, Uganda set up a nearly universal system of primary education,
which resulted in a considerable increase in enrolment rates in the late 1990s. As in
the case of Burkina Faso, however, budget appropriations did not really keep up,
which should lead to a drop in the quality of education. Once again, to avoid an overly
optimistic scenario, we shall assume that Uganda’s new system will not fully achieve
the expected results — full enrolment at the primary level with no loss of quality —
until 2010. From that date, we may also assume that more progress will be possible in
the secondary sector, and to reflect this we shall assume that the secondary–level
human capital stock will grow at a rate comparable to that of Ghana.
The consequences of these assumptions concerning the overall stock of human
capital are reported in Table 4.3, which shows projected change in this stock from
1996 to 2020. The table also presents a more optimistic version, called the high scenario,
in which we assume that Burkina Faso, Mali and Tanzania will attain full primary
enrolment in 2020, and that Ghana and Côte d’Ivoire, which are considerably further
along at the time of writing, will achieve this as early as 20102.
Table 4.3. Human Capital Growth Scenario
Burkina Faso Côte d'Ivoire Ghana Mali Tanzanie Uganda
1996 level 1.0 2.8 4.2 1.0 3.2 3.1

2020 level (baseline scenario) 2.2 3.5 5.9 2.1 3.7 5.0
1996-2020 growth 3.1% 0.9% 1.4% 3.1% 0.6% 2.0%
2020 level (high scenario) 2.5 4.4 7.0 2.5 4.2 5.0
1996-2020 growth 3.7% 1.8% 2.1% 3.6% 1.1% 2.0%
104
Quality of Human Capital
For the economies under study, the only indicators available to us are those
measuring the inputs of the educational system, which may be assumed to have an
impact on the quality of the instruction provided. The main indicator used for this
purpose is the average number of pupils per teacher in primary education. Overcrowded
classrooms usually mean that a majority of pupils do not assimilate knowledge, even
though they attend classes. In this respect, the French–speaking countries of our sample
obtain results that are clearly inferior to those of the English–speaking countries. This
may be due to the fact that education systems in the former group are wholly dependent
on the initiative of the state, which has little budgetary leeway to increase the number
of teachers, and also to the fact that the system largely provides education free of
charge. In the English–speaking countries, the private sector and local communities
play a much greater role. In Uganda, for example, families pay approximately half the
cost of primary schooling through enrolment fees and their contributions to parent–
teacher associations. In the early 1990s, they even paid as much as 75 per cent of
these costs. As a result, funding for education is less dependent on the government’s
budget resources.
Table 4.4 shows that the English–speaking countries in our sample are
approaching South Africa, but are still fairly far behind Mauritius and Botswana, which
have pupil/teacher ratios comparable to those of the developed countries.
Table 4.4. Indicators of Educational System Quality, 1996
Country No. pupils per teacher Expenditure per pupil ($) Expenditure per pupil
(primary) (primary/secondary) as a ratio of GDP/cap.
Burkina Faso 51 51 21.2

Côte d'Ivoire 41 115 16.9
Ghana 28a … …
Mali 70 39 15.4
Tanzania 36 … …
Uganda 35 … …
Botswana 25 402 12.5

Mauritius 24 365 9.8
South Africa 36 500 15.5
a. 1993.
Source: World Bank (World Development Indicators, 2000).
105
In a public, centralised educational system, the central government budget is of
course a major constraint on the improvement of educational quality. Burkina Faso
and Mali, for example, spend less than half of what Côte d’Ivoire spends on education3.
As these expenditures already account for a considerable proportion of per capita
GDP in the countries under consideration, there will be very little room for increased
spending in the future, unless we assume significant growth of per capita GDP.
The impact of the quantity of inputs used in the education system on the quality
of education is not easily quantified. We can, however, indulge in a few brief
calculations based on a result reported by Berthélemy et al. (1997). In this study,
which tests the impact of human capital on growth using panel data for a broad sample
of countries, the authors show that the elasticity of growth with respect to human
capital is an increasing function (approached linearly) of the pupil/teacher ratio. This
finding does not enable direct measurement of the possible gains to be derived from a
given decrease in this ratio — in particular because the demonstrated effect is calculated
over the long term and under the assumption that cross–country differences in quality
are constant over time — but it does allow us to give a few rough estimates. Table 4.5
presents these estimates, calculated on the assumption that the Mauritian education
system, in which the pupil/teacher ratio is comparable to that of the developed countries,
is an efficient system.
Table 4.5. Indicator of Effective Human Capital as a Function of Nominal Human Capital
and Pupil/Teacher Ratio
Average No.
of years
of schooling
Pupil/teacher ratio 1 1.3 1.6 2.0 2.4 3.1 3.8 4.8
80 1.0 1.1 1.1 1.2 1.3 1.4 1.5 1.6

64 1.0 1.1 1.2 1.4 1.6 1.7 1.9 2.2
51 1.0 1.2 1.3 1.6 1.8 2.1 2.4 2.8
41 1.0 1.2 1.4 1.7 2.0 2.4 2.9 3.4
33 1.0 1.2 1.5 1.8 2.2 2.7 3.3 4.1
26 1.0 1.2 1.6 1.9 2.4 3.0 3.7 4.7
106
The formula used to construct this table4 suggests that educational quality is of
rather secondary importance when the level of human capital is low. As a first
approximation, it may be assumed that the cost of lowering the pupil/teacher ratio is
roughly proportional to the number of pupils per teacher, and that, for a given level of
quality, the cost of the educational system is proportional to the number of pupils
enrolled and to the average number of years spent in school by a given age group.
Under these assumptions, Table 4.5 allows us to compare the economic return of
scenarios in which additional resources are allocated to improve the quality of
education, and of scenarios in which the same resources are used to raise enrolment
rates5. We have used grey shading to indicate the cells of the table where investment in
improving educational quality could earn a significant return. It is clear that none of
the countries studied is located near this shaded area. As a result, we may conclude
that major investments in educational quality do not constitute a very relevant
hypothesis for our growth projections. It should be noted, however, that this problem
will arise when enrolment rates reach 100 per cent, if these countries do not move to
raise quality before then, as Mauritius did. South Africa, for example, is facing this
problem today.
These calculations should of course be seen in perspective, since they are based
on a quantitative and mechanical assessment of the situation. Using survey data on
knowledge acquisition by pupils in five African countries, Michaelowa (2000) shows
that the number of pupils per class had no significant effect on the quality of the
education provided. This supports our conclusion that improving the pupil/teacher
ratio should remain a secondary consideration and should not be given priority over
increasing the number of children in school.
There is probably scope, on a case–by–case basis, for institutional improvements
to the education system that would help to raise educational quality (and in some
cases coverage as well). The case of Uganda, where parent–teacher associations have
come to compensate for the shortcomings of the public system6, indicates one possible
path. On the other hand, the same country’s establishment of a system of nearly universal
primary education in 1997 (each family can send four children to school at no charge)
led to a huge increase in the number of pupils per teacher at the primary level (from 30
in the early 1990s to 53 in 1997), which could compromise educational quality in
Uganda in the short and medium term.
The performance of education systems could also be improved by maintaining
sufficient budget appropriations to raise enrolment rates, especially since the number
of pupils to be enrolled should grow rapidly as a proportion of the population. Recent
developments, reported in Table 4.6, show that Uganda alone among the countries
studied made progress in this respect during the 1990s. Without a fairly radical
change in the governmental choices revealed by these data, such improvements
cannot be expected.
107
The falling GDP share of education spending does not always correspond to a
decrease in the share of public resources devoted to education. Hence these data reveal
both the difficulty of providing the public funds needed for development and the fact
that public finance choices do not always favour education. In the long run, however,
it is reasonable to assume that the unit costs of education will not fall as a proportion
of GDP, notably because teachers’ wages should follow the same trend as those of
other skilled employees, as a matter of labour market equilibrium. The data in Table 4.6
thus reveal a number of trends that could compromise the quality of education.
Michaelowa (2000) shows, however, that spending per pupil is not a good
indicator of the quality of the education provided. Côte d’Ivoire spends three times as
much per pupil as Burkina Faso (and 40 per cent more as a proportion of per capita
GDP), but knowledge acquisition by pupils is comparable in the two countries. Concern
for the quality of education should therefore not be emphasised unduly in analysing
our scenarios of macroeconomic growth. It is at the microeconomic level that the
means of improving educational quality need to be identified.
Table 4.6. Recent Trend in the GDP Share of Education Spending
1990-93 1994-98 Notes

Burkina Faso 2.5 2.2 1990-93 and 1994-96
Côte d'Ivoire 7.8 5.3 1992-93 and 1994-97
Ghana 4.4 4.4 1992-93 and 1994-96
Mali 2.8 2.2 1992-93 and 1994-97
Tanzania 3.4 2.3 1990 and 1995-98
Uganda 1.7 2.4 1990-91 and 1994-95
Source: Authors’ calculations based on World Bank (World Development Indicators, 2000), Chambas et al. (1999) and
Lambert and Sahn (2000).
Rent Seeking and Wastage of Human Capital
The availability of a better–trained workforce does not always guarantee a higher

growth rate. Many studies (see for example Pritchett, 1997) have challenged the idea
that the accumulation of human capital automatically leads to growth. In Africa, one
of the main reasons for this difficulty in transforming educational investment into
growth is that a large proportion of human capital is wasted in rent–seeking activities.
In the past, the countries studied established mechanisms that guaranteed
public sector employment to graduates of secondary and higher education
institutions, thus placing them in a rent situation. Although these systems were
dismantled in most of the countries under structural adjustment programmes, the
basic problem — a strong tendency on the part of degree holders to enter activities
in which rent situations exist — may persist.
108
In practical terms, rent–seeking behaviour is a waste of economic resources
because it leads people to work towards capturing rent situations and reaping benefits,
instead of engaging in productive activities. As a first approximation, it may be assumed
that such behaviour is mainly to be found among skilled workers. Rent–seeking activities
consist in finding or inducing the executive branch to create arrangements for
circumventing laws and regulations or manipulating them to one’s own benefit, and this
generally can be accomplished only by individuals having a certain level of education.
Corruption is a special case of rent–seeking behaviour, a narrow definition of
which is the illegal use of a government position for personal gain. The borderline
between corruption and rent–seeking behaviour is basically a legal one, between actions
that comply with the letter if not the spirit of the law and actions that are clearly
illegal. In countries where legal rules are not well enforced for lack of a modern judiciary
system, the distinction between rent–seeking behaviour and corruption is of little
importance, and for purposes of a first approximation we can combine the two.
Rent–seeking behaviours range from misappropriation of foreign aid by high
officials (for example, the members of the former governing elite of Mali were accused
of this after the overthrow of President Traoré in 1992, as was President Bédié of Côte
d’Ivoire), to bargaining over the payment of a fine to a police officer, to the bribes
extracted by customs inspectors and other officials responsible for enforcing the tax
code and other rules. A number of sources indicate that the level of corruption in
African countries is high.
Table 4.7 presents two of the main corruption indices’ ratings of the six countries
studied, along with those of Mauritius and Kenya for purposes of comparison. The
first column is based on the corruption index compiled by Transparency International
(TI) from several independent sources. The TI indicator rates countries on a scale
from 0 to 10: the more corrupt the country, the lower the rating. This indicator, which
has the merit of being based on a combination of different points of view and on
several years of experience, has the drawback of rating only a few African countries.
We therefore compare it with a second classification, presented in the second column.
The second column uses the results of the business survey conducted by the
World Bank for the World Development Report, in particular the survey question
most directly concerned with detecting corruption. Question 14 is worded as follows:
“It is common for firms in my line of business to have to pay some irregular
‘additional payments’ to get things done.” Responses to this question are classified
in six categories, from the highest (1=always) to the lowest (6=never). We use here
the average of the numerical values assigned to firms’ responses. The closer the
calculated indicator is to 1, the more corrupt the country. This indicator should thus
be interpreted in the same way as the TI index: the lower the rating, the higher the
level of corruption. It can be seen from Figure 4.2, constructed on the basis of the
data available for African countries, that this indicator and the TI corruption index
for 1998 are strongly correlated7.
109
Table 4.7. Corruption Indicators
Country TI 1998 index WDR 1997 Proportion of human capital employed

(question 14) in rent-seeking activities (%)
Burkina Faso … … …
Côte d'Ivoire 3.1 2.9 30
Mali … 3.3 …
Ghana 3.3 3.4 27
Tanzania 1.9 2.8 …
Uganda 2.6 3.3 …
Memo item:
Kenya 2.5 3.4 40
Mauritius 5 4.2 29
Source: Authors’ calculations based on World Bank data (World Development Report, 1997), Transparency
International and Berthélemy et al. (2000).
Source: World Bank and Transparency International.
110
The countries listed in Table 4.7 may thus be classed in four groups, in increasing
order of corruption: (1) Mauritius (little corruption), (2) Mali, Côte d’Ivoire, and Ghana,
(3) Uganda and Kenya, and (4) Tanzania (very corrupt). Burkina Faso, which does not
appear in the available surveys, is not easy to rank, but factual observations suggest that
corruption is not endemic to this “land of honest men”, as noted by Chambas et al.
(1999). This leads us to assign it to the second group, with Côte d’Ivoire and Mali8.
Corruption can seriously impair growth in a number of ways that fall essentially
into two categories. First, it leads to slower accumulation of the resources needed for
growth, in particular by inducing foreign enterprises to abandon their investment plans.
Second, it wastes existing resources by attracting energy away from productive
activities. We shall be concerned here with the second category.
The question to be resolved is the following: given the level of corruption and/or
the proportion of rent–seeking activities in the economy, what proportion of the skilled
population is occupied in these activities? An attempt to answer this question was put
forth by Berthélemy et al. (2000), who estimated the wastage of skilled labour in
rent–seeking activities using an indicator proportional to the share of civil service
employment9 in total formal non–agricultural employment. The data collected by
Schiavo–Campo et al. (1997) allowed this indicator to be calculated for the early
1990s for some 50 countries. When included in a standard growth regression, following
Mankiw et al. (1992), this indicator has a significant negative impact on growth from
1986 to 1996 in the economies considered. We interpret this result as being due to
rent–seeking behaviour. Since corruption requires interaction with the public employees
responsible for drafting and administering the rules, it may be considered as a first
approximation that the number of individuals engaging in rent–seeking activities is an
increasing function of the number employed in the civil service. Assuming that this
function is linear, Berthélemy et al. (1998) perform an econometric estimation yielding
a plausible coefficient of proportionality between the percentage of these public
employees in formal non–agricultural employment (defined as all activities in which
human capital is employed) and the wastage of human capital in rent–seeking
activities10.
The third column of Table 4.7 presents the results of these calculations, which
unfortunately could be carried out for only a few countries, owing to lack of data.
Given the quality of the data, the results should be considered as rough estimates for
illustrative purposes only. Côte d’Ivoire and Ghana are in fairly similar positions with
respect to this indicator, at the same level as Mauritius: they “waste” 30 per cent or
slightly less of their human capital in rent–seeking activities. Kenya (which as we
have seen may be compared with Uganda) wastes 40 per cent of its human capital. By
way of comparison, the same calculations for other continents indicate that the level
of waste is only about 20 per cent in East Asia, Latin America and the OECD countries,
a level which may be assumed to represent an irreducible minimum.
111
Measures to reduce rent seeking and corruption should bring Ghana and Côte
d’Ivoire the equivalent of an increase in human capital of roughly 14 per cent.
Our earlier comparative observations lead us to suppose that this would be true of
Mali and Burkina Faso as well. For the last two countries, this is equivalent to the
difference in human capital between our high and low scenarios (and slightly over
half of this difference for Côte d’Ivoire). For countries with a moderate corruption
index and very little educational capacity, the rewards from fighting corruption
can thus be as great as those from increased outlays on education. Assuming that
the above–mentioned comparison between Kenya and Uganda is valid, the latter
country could register an increase of 33 per cent. This estimate may also represent
the lower limit of the gains that could be obtained by Tanzania, where the gains
from fighting corruption could be appreciably greater than those from stepped–up
efforts in education.
Thus, from the standpoint of using human capital for productive ends, substantial
progress can be made in the countries examined. Moreover, these gains can be achieved
at a low opportunity cost — for everyone, that is, except the political leaders and civil
servants who are reaping the benefits of corruption. In this respect, our high scenarios
concerning the growth of human capital can be considered alternatively as scenarios
of increased efforts in education, or as scenarios of better use of existing human capital
at lower cost through the curbing of rent–seeking behaviour.
It is nonetheless impossible to predict future developments in the African countries
studied, and the lack of quantitative observations makes it difficult to identify trends.
Côte d’Ivoire’s TI index fell in 1999 (from 3.1 to 2.6), but the governments that took
power following the December 1999 coup and the subsequent elections (in 2000)
seemed determined to take action against corruption. Uganda’s rating also fell in 1999
(from 2.6 to 2.2).
Assumptions for the Future and the Role of Domestic Policies
Table 4.8 shows the development of the human capital stock and its influence
on productivity growth under the two scenarios. The substantially higher impact on
growth in Burkina Faso and Mali partly reflects the fact that both countries start from
a significantly lower level but also with a demonstrated commitment to improving
educational standards. Gross enrolment in primary school increased from around 25 per
cent in both countries in the mid–1980s to over 40 per cent in 1996. The cumulative
impact on total factor productivity from 1996 to 2020 is estimated at over 18 per cent
in the baseline scenario and around 22 per cent in the high scenario.
Ghana has the highest level of human capital stock among the sample countries,
and we assume that it will continue to make significant gains in the near future. The
main strength of Ghana’s educational system compared to those of the other countries
112
Table 4.8. The Impact of Education
Human capital stock Cumulative impact on TFP

1996 2020 2020 Baseline scenario High scenario
baseline scenario high scenario (%) (%)
Burkina Faso 1.0 2.2 2.5 18.7 22.2

Côte d’Ivoire 2.8 3.5 4.4 5.4 10.9
Ghana 4.2 5.9 7.0 8.7 12.9
Mali 1.0 2.1 2.5 18.4 21.9
Tanzania 3.2 3.7 4.2 3.4 6.5
Uganda 3.1 5.0 5.0 11.9 11.9
Note: The human capital stock is measured as the average number of years of schooling among the population aged 15-65.
studied here is the extent of secondary schooling. According to our data, the average
number of years of secondary schooling among the active population is as high as
1.2 years, and in the baseline scenario this figure is assumed to double by the year
2020. In the high scenario, we assume that the number of years of secondary schooling
will increase in 2010 — after enrolment in primary school has reached 100 per cent
— and will reach an average of three years in 2020, based on a trend comparable to
the rate of progress observed in Mauritius in the 1980s.
In the case of Uganda, we have used a different approach in constructing
scenarios, given the country’s recent efforts to provide free primary education for
up to four children per family. Although this measure had already been implemented
in 1997, we assume a gradual increase in primary school enrolment to 100 per cent
by 2010, in order to compensate for some of the loss in educational quality due to
the higher pupil–to–teacher ratio, and to allow some time for hiring more teachers.
Moreover, Uganda plans to extend the duration of primary education from seven to
eight years by 2010 (Bigsten and Kayizzi–Mugerwa, 1999). We assume that from
the moment universal primary education is achieved, efforts will be directed towards
secondary education. Therefore, from 2010 onwards, we assume that human capital
with secondary schooling will grow at a rate equal to that observed in Ghana over
the last ten years, i.e. 3 per cent annually. Accordingly, Uganda stands to gain more
under the baseline scenario than all other sample countries that have a similar or
higher level of human capital stock. We did not think it reasonable to construct a
more optimistic “high” scenario, however, in view of the substantial efforts required
under the baseline scenario.
The two laggards among the countries studied are Côte d’Ivoire and Tanzania,
which have actually seen a decline in primary school enrolment rates since the early
1980s. Under the baseline assumption that their enrolment rates remain at the current
level, we observe much more moderate gains in TFP than for the other countries.
113
Box 4.1. Investment in Education: Necessary But Not Sufficient
In the radically new model proposed by Lucas (1988), human capital is perceived to be the driver
of growth. The accumulation of human capital ultimately benefits a large number of individuals through
spillover effects. This model is based on the concept that learning and the accumulation of human capital
are social activities, in the sense that they are fostered by interactions between people of different levels
of human capital and different types of experience. Without spillover effects, one person’s human capital
would be lost the moment he or she ceased to be part of the working population. Each generation would
then have to start from zero and the volume of skills likely to be acquired by the labour force would be
limited. We can conclude that intergenerational transfer of human capital requires a minimum of initial
skills, although Lucas (who compares economic growth in the US with that of developing countries)11
does not address this point specifically. In the case of Burkina Faso or Mali for example, where average
number of years of schooling is 1, the present level of human capital can be considered negligeable in its
effect on future generations. In situations such as these — with few spillover effects — traditional
endogenous growth models cannot be used. The accumulation of skills can be accelerated or consolidated
via the creation of an environment in which practical learning generates such spillover effects. Another
of Lucas’ models can be used in this context. In his explanation of the Asian “miracle”, Lucas (1993)
points out that the Newly Industrialised Economies (NIEs) of Asia benefited from the effects of “learning
by doing”. Inasmuch as products of increasing technical sophistication are always being introduced into
the range of goods produced by these countries, the skill level continues to increase.
Certainly, efforts made by countries such as Burkina Faso, Mali or Uganda (to cite only a few) in
the area of education will ultimately amplify potential spillover effects of human capital in families and
communities. But this process is both long and costly. Moreover, the average level of human capital is
not the only determining variable. Its distribution among the population is equally important. In Africa as
in other poor regions, human captial (like income) is generally concentrated in urban areas. Certain
externalities might be present in large cities but not in rural areas, and this fact brings up several other
complex considerations. It might be assumed, for example, that investments in education are more
productive in urban areas, and that it is therefore preferable to concentrate resources there, instead of
working toward making primary education available to the largest number of children as Uganda has
done. Yet such a strategy would be highly ineffective in terms of poverty reduction. Secondly, raising the
level of instruction is not necessarily productive if the economic structure is not adapted to the use of a
relatively large qualified labour force or if skilled labour is employed primarily in rent–seeking or other
non–productive activities (see above). Thirdly, substantial productivity gains can be derived from the
reallocation of labour from the agricultural sector to more productive sectors. However, since Africa
enjoys a comparative advantage in agriculture, industrialisation should not be achieved to the detriment
of this sector. On the contrary, in order to free up labour while ensuring the sufficient agricultural production,
agricultural techniques must be modernised, a process which necessitates the development of human
capital in rural areas. Finally, reallocation of labour from rural to urban areas will be a source of productivity
gains only if migrants from rural areas have a sufficient level of education.
There is also a relationship between the potential spillover effects from human capital and the
degree of economic diversification. The low level of diversification in African economies can limit the
externalities resulting from the accumulation of human capital in at least two ways. First, and above all,
the more an economy is specialised in a limited range of products, the less the possibilities of useful and
instructive interactions between people of different backgrounds and experience. When everyone performs
the same job, the potential for sharing new skills is limited. Secondly, the lack of diversification implies
direct competition that discourages cooperation and open interaction.
114
Conclusion
The scope for increased human capital accumulation in the countries studied is
considerable, first and foremost in quantitative terms. If only to combat illiteracy,
increased enrolment rates at the primary level are needed, particularly in Burkina
Faso and Mali where these rates are still very low. The recent example of Uganda
suggests that even a poor country can implement more ambitious primary education
policies than those observed to date in sub–Saharan Africa.
Increasing school enrolment has a budgetary cost, which raises two issues. First,
the central government probably cannot bear the cost of education by itself. Once
again, the example of Uganda, which is fairly representative of the school systems
developed in English–speaking Africa, indicates the road to follow: calling on
contributions from parents, in a decentralised school system, is a possible option which
would result in giving public expenditure a great deal of leverage.
The second issue raised by the budget constraint is that of the quality of education,
since the GDP share of public spending on education is very strongly constrained by
the funds available. At the present stage, however, the main objective is to teach school–
age children to read, and the quality problem is perhaps secondary to that of meeting
basic needs.
Lastly, there is a great deal of scope for a greater contribution of education policy
to growth, by limiting the wastage of trained human capital. It is not rare in Africa for
more than half the skilled labour force to be employed in civil service departments
(excluding education and health). At the same time, the waste of skilled labour in
rent–seeking activities is evident, and appreciably greater than on other continents.
Although this problem primarily concerns people educated at the secondary and higher
levels rather than at the primary level, it needs to be taken into consideration because
of the high unit cost of secondary and higher education, which places a strain on
education budgets.
115
Notes
1. As the data on secondary and higher education have little impact on the overall indicator,
owing to the small size of the secondary and higher education sectors, it would serve no
purpose to formulate specific hypotheses in this respect for our scenarios.
2. In this high scenario, we also make the assumption that the stock of human capital having
secondary education grows more quickly as from 2010, at an annual rate of 5.5 per cent,
which corresponds to the rate observed in Mauritius over the 1977–87 period.
3. The lack of data for the English–speaking countries is partly due to the existence of
mixed public–private systems.
4. The equation used, drawn from the estimates of Berthélemy et al. (1997), is:
~
log(h )=(1−0.013(PT1−25))log(h)
where PT1 is the pupil/teacher ratio in primary school, h the average number of years of
schooling and the corresponding level of human capital measured in effective units.
5. To do so, one need only compare the increase in the total available human capital stock
obtained through an increase in staffing levels (for a given level of school enrolment,
i.e. in the same column) with the increase obtained from an equivalent increase in the
resources consumed by expanding enrolment (for a given staffing level, i.e. on the same
row). Thus, when the average number of years of schooling is 1.6, raising the teacher/
pupil ratio (and hence the number of teachers) by 20 per cent (which reduces the pupil/
teacher ratio from 51 to 41) raises the effective stock of available human capital only
from 1.3 to 1.4; by comparison, raising the enrolment rate by 20 per cent, which increases
the average number of years of study from 1.6 to 2 (for a constant teacher/pupil ratio),
has the same cost in terms of the number of teachers, but increases the total stock of
available human capital from 1.3 to 1.6.
6. See the case study by Bigsten and Kayizzi–Mugerwa (1999).
7. The correlation coefficient is 0.81. The data are not perfectly comparable, owing to the
one–year lag, but 1998 was the first year in which TI rated a fairly large number of
African countries.
8. The Davos Forum’s Africa Competitiveness Report (1998) does not indicate that the
foreign firms surveyed complain that corruption is a major obstacle in Burkina Faso.
116
9. Schiavo–Campo et al. (1997) define this category as employment in the civil service
(central and local), excluding education and health services.
10. This relation, based on a modified version of the model developed by Mankiw, Romer
and Weil (1992), is estimated for a sample of about 50 countries for which data were
available. It is written as follows:
where ln(y/y–1) is the growth rate for 1985–95, sK the investment rate, n the population
growth rate, h human capital and lG the proportion of public sector employees in the
non–agricultural working population.
11. In Lucas’ model (1988), the accumulation of human capital is determined by the following
function/formula:
where δ is a constant, (l–u) the time devoted to the accumulation of human capital and γ
a parameter that measures the amplitude of externalities that flow from existing human
capital. Lucas considers that this parameter is equal to one, but that we believe that its
value depends on the level of human capital, the degree of diversification, the degree of
economic distortion, etc. For example, if the level of human capital is below a certain
minimum level, γ can be 0, in which case the accumulation of human capital generates
no externalities.
117
118
Chapter 5
Exports
Openness to the outside world is widely considered to be necessary for growth.

In the analytical framework developed in Chapter 2, openness contributes to growth
through the availability of foreign exchange to finance imports, which for all practical
purposes play the role of an intermediate production factor. In the countries studied,
past episodes of exchange–rate controls and hard currency shortages — for example,
in Ghana, Uganda and Tanzania — have been marked by considerable economic
downturn. Conversely, the progress achieved since reforms began has brought
significant improvements, such as an increased volume of available imports and the
disappearance of the black market for foreign exchange.
The key to these positive developments is the restoration of good export
performance, the conditions for which are studied in this chapter. Export performances
may be evaluated through two complementary indicators: the ability to respond to
global demand by means of increased competitiveness, and the supply capacity of the
economy. These two approaches are not mutually incompatible, of course. The first is
constrained by growth in global demand, and thus may be regarded as a short– or
medium–term approach, while the second considers only the long term, taking as its
sole constraint an economy’s capacity for exportable goods supply. We will use each
approach in turn.
Increased Competitiveness of the Emerging Economies
African economies are specialised in low–growth market segments, because

agricultural and mining commodities dominate their exports and world demand for
these commodities is less vigorous than that for processed products. This causality
can be illustrated by breaking down the export growth of a given economy, in relation
to world export growth, into three components:
— the growth which would result automatically from growth of international demand
in the sectors in which this economy is specialised in the initial year, as a ratio of
the aggregate growth in international trade;
119
— the growth which would result automatically from growth in the demand
expressed by each trading partner of this economy, compared to aggregate world
growth, less the component resulting from the fact that the effects of sectoral
and geographical specialisation are partly correlated (the most rapidly growing
economies do not express demand for the same products as slow growers do);
— the growth resulting from neither of the first two effects, which may be considered
as being due to productivity gains1.
When applied to all of the non–oil–exporting countries of sub–Saharan Africa
(excluding the Southern African Customs Union), as well as to the countries of our
sample, this method yields the results reported in Table 5.1.
This table shows that over the 1980–96 period, African exports grew much less
rapidly than those of the rest of the world. Their lower rate of growth can be explained
in structural terms as resulting from specialisation in products with little growth
potential, which corresponds to a negative sectoral effect of 4.5 percentage points of
annual growth. Over the period considered, this sectoral effect was accompanied by
substantial losses of competitiveness, amounting on average to nearly 2.4 points of
growth per year.
Table 5.1. Breakdown of Growth in African Exports, 1980-96

(percentage points of annual growth in value terms)
Growth of world Growth Relative Factors explaining relative growth

trade of exports growth
Country/region Sectoral Geographical Competitiveness
effect effect effect
(1) (2) (2) - (1)
Sub-Saharan
Africaa 9.04 2.07 -6.97 -4.47 -0.12 -2.39
Burkina Faso 9.04 -0.02 -9.07 -5.13 0.15 -4.09
Côte d'Ivoire 9.04 1.55 -7.50 -3.45 -1.32 -2.72
Ghana 9.04 2.26 -6.79 -5.11 1.77 -3.44
Mali 9.04 -0.25 -9.29 -4.47 -1.46 -3.36
Tanzania 9.04 -0.33 -9.38 -4.12 2.68 -7.94
Uganda 9.04 0.22 -8.83 -4.27 0.33 -4.88
a. Sub-Saharan Africa excluding Democratic Republic of Congo, Gabon, Nigeria and Southern African Customs
Union.
Source: CEPII (1998).
120
Over the same period, such mediocre performances can also be observed in the
countries examined in detail for this study. However, the choice of calculation period
is not neutral for these countries. Table 5.2 presents the same calculations for the six
episodes of pre–emergence under study2, revealing in every case an improvement in
competitiveness during these periods.
Table 5.2. Breakdown of Growth in African Exports — Rapid Growth Episodes

(percentage points of annual growth in value terms)
Factors explaining relative growth

Growth of Relative
Country Period world Growth of growth Geographical Competitiveness
trade exportations (2) - (1) Sectoral effect effect effect
Burkina Faso 1994-96 12.29 15.01 2.72 -3.44 -3.40 9.41

Côte d'Ivoire 1994-96 12.29 18.61 6.32 -3.52 -0.83 10.59
Mali 1994-96 12.29 38.01 25.72 -4.21 0.54 29.39
Ghana 4983-96 14.08 13.96 -0.11 -5.04 4.06 0.87
Tanzania 1994-96 12.29 23.61 11.32 -4.10 -0.35 15.41
Uganda 1987-96 13.79 7.71 -6.07 -6.44 0.34 0.03
Source: CEPII (1998).
This table reveals that in each period identified above as a period of substantial
economic progress, export performances were better than over the 1980–96 period as
a whole. The years 1994 to 1996 were particularly favourable for the Franc Zone
countries, which took advantage of the devaluation of the CFA franc to boost their
competitiveness, thereby recording rates of export growth higher than the rate of growth
in world trade. The same spectacular progress can be seen in Tanzania following
liberalisation of the economy. Ghana and Uganda recorded less striking results over
relatively long growth periods, as their exports lagged behind growth of world trade.
However, they too posted competitiveness gains (albeit modest ones) during the period
considered, in contrast to losses for the 1980–96 period as a whole. It follows that
from 1980 to 1983 in Ghana and from 1980 to 1987 in Uganda, these countries sustained
considerable losses of competitiveness, which were subsequently reversed.
The differences in magnitude between the competitiveness gains realised in the
most recent growth episodes and those of Ghana and Uganda are due to the length of
the period under consideration and the nature of the adjustment policies followed. In
the short term, a substantial adjustment such as the January 1994 devaluation in the
Franc Zone countries makes it possible to harness previously idle production capacity,
resulting in rapid improvement of export potential. In Tanzania, the nominal devaluation
that accompanied liberalisation of the foreign exchange market in the 1990s has been
offset by relatively high inflation, undermining the gains in price competitiveness
achieved in recent years. In the same country, however, the dismantling of controls on
agricultural exports (60 per cent of which were previously controlled by the Marketing
Board) has had a positive impact on exports.
121
In Ghana and Uganda, the initial transition stage that followed liberalisation of
the economy is winding down and long–term constraints are probably beginning to
take effect, so that export growth now depends not only on price competitiveness and
liberalising policies opening up the economy to the outside world, but also on the
growth potential of the economy. Export growth (see Table 5.3) was slightly higher
than economic growth for Ghana, but lower for Uganda. In the latter, the relative
sluggishness of export growth is confirmed by national accounting data, which show
that the volume of exports grew less rapidly than the volume of GDP from 1986 to
1993 and subsequently gained ground3. This can be attributed to the fact that the earlier
years were a period of reconstruction in this country after a prolonged civil war.
Table 5.3 also reveals that the current sectoral specialisation of the economies
studied constitutes as great a handicap, from the standpoint of export dynamism, as in
our assessment for the reference period 1980–96. Nothing less than deep structural
change will allow these economies to reverse this tendency, but such change must
stem from shifts in comparative advantage rather than from a state–led industrialisation
policy. The process of endogenous economic diversification can play a decisive role
in this respect, once it is set in motion. The example of Mauritius, which in 25 years
diversified its exports and made the transition from a sugar–based economy to an
economy that exports textiles and clothing, and more recently electronic products,
indicates that such a transformation is possible.
In sum, it would seem that with a successful adjustment policy, African economies
can raise their export growth rates in the short to medium term by increasing their
competitiveness, particularly when the adjustment begins with a substantial devaluation,
and by liberalising their cash–crop sectors. These gains in competitiveness have their
limits, however. Over the long term, progress will come only from structural change
in the sectoral specialisation of these economies. Following the example of today’s
developed economies, they must diversify.
Table 5.3. Level and Recent Trend in the Exports/GDP Ratio

(accounts at constant 1995 prices)
1988 1998 Growth rate
Burkina Faso 15.8 14.1 -1.1

Côte d'Ivoire 35.5 44.5 2.3
Ghana 18.1 30.2 5.2
Mali 16.3 27.5 5.3
Tanzania 12.0 18.1 4.1
Uganda 9.6 13.7 3.6
Source: Authors’ calculations from World Bank data (World Development Indicators, 2000).
122
Supply–side Capacity and Trade Policy
In his study of trade policy and economic performance in sub–Saharan Africa,

Rodrik (1998) confirms that economic growth is the main determinant of export growth.
He shows in particular that the exports/GDP ratio in Africa is not significantly different
from what is observed elsewhere, once the data are controlled for the size, level of
development and geographical location of each country. The positive and significant
impact of the level of development (measured by per capita income) reported by Rodrik
even suggests that export elasticity is greater than one. This phenomenon can hardly
continue over the long term, however, since that would mean that the GDP share of
exports would rise indefinitely with the level of economic development.
Rodrik (1998) also shows that, other things being equal, there are great differences
among the African economies where exports are concerned. A glance at the data for
the six countries of our sample (see Table 5.3) confirms this. For example, in recent
years the GDP share of Côte d’Ivoire’s exports has been twice that of Ghana’s, although
the two countries are similar in terms of geographical location, size and comparative
advantages. This is probably due to the major investments made by Côte d’Ivoire to
develop its export crops as from the 1950s. In Ghana, by contrast, the 1960s and
1970s were years of severe crisis, marked by over–taxation of cocoa by the Cocoa
Board and strict currency controls: in 1965, Ghana exported proportionally twice as
much as Côte d’Ivoire, whereas the reverse was true in the early 1980s. Similarly,
Mali exports twice as much as Uganda as a proportion of GDP. The two countries are
both very poor and landlocked, but Uganda suffered from many years of civil war as
well as, before the reform period, very heavy taxation of export crops (50 per cent on
coffee), strict currency controls and non–tariff barriers to imports. It would therefore
seem that export performances, which obviously depend on geographical conditions,
are also deeply marked by the history of individual countries.
A partial explanation of differences in export growth rates can be found in the
fact that the export ratio is positively correlated with the degree of openness of trade
policy. According to Rodrik’s (1998) calculations, there is a significant correlation
between the export ratio and the apparent rate of taxation of foreign trade. Rodrik
finds that, other things being equal, a 10 per cent reduction in taxation of foreign trade
could raise the GDP share of trade (exports + imports) by 17 per cent. Trade
liberalisation may take different forms in different cases — examples include the
elimination of marketing board levies on cash–crop agriculture and the dismantling of
quantitative restrictions on imports — but the result will be much the same.
Trade liberalisation can undoubtedly improve export performance by boosting
incentives to develop productive activities in tradable goods sectors. Several of the
countries studied here have, to varying degrees, made progress in this respect and
have increased their exports as a result.
123
A ranking of African countries according to progress on trade liberalisation
reforms points up the changes that have occurred in trade policies, which were all
regarded as closed initially and which have gradually opened up. This ranking factors
in both tariff and non–tariff barriers and controls on imports and exports (see
Figure 5.1.) Three stages of openness are distinguished: closed policies (indicated in
dark grey shading), somewhat open policies (light grey) and open policies (blank).
Figure 5.1 clearly indicates that Ghana, followed by Uganda and Mali, were the
first to liberalise their trade policies, with Tanzania just behind. Liberalisation proceeded
in a fairly uniform fashion in all of the countries except Uganda, where a tax on coffee
exports was temporarily reintroduced in 1994 and 1995 to offset the impact of higher
international prices on the domestic market.
In contrast, Côte d’Ivoire and Burkina Faso’s early attempts at liberalisation, in

the mid–1980s for the former and the early 1990s for the latter, were followed by the
resumption of restrictive trade policies. The countries returned to more liberal policies
only in 1994, after the CFA franc devaluation.
The consequences of these different patterns of progress on liberalisation reforms
can be clearly seen in the export growth data reported in Table 5.3: the exports of
Burkina Faso and Côte d’Ivoire were sluggish in the 1990s, while those of the other
four countries grew appreciably faster.
Continued liberalisation of trade policy should lead to a further rise in these
economies’ export ratios. However, such a development can only take the form of a
transition towards a new steady state, characterised by more or less complete trade
liberalisation and by an exports/GDP ratio stabilised at a higher level. A good example
in this respect is the experience of Mauritius, where the exports/GDP ratio has begun
to level off in recent years (reaching 69 per cent in 1996).
124
Hypotheses on Future Trends
In constructing our growth scenarios, we therefore assume that export growth is

equal to GDP growth over the very long term. The level of the export ratio will influence
the path of transition towards this long–term state, and hence the level of per capita
income during each period. This effect will fade over time, however; our hypotheses
imply that the long–term impact of trade liberalisation on the growth rate will be null.
The next 20 years should be regarded as a phase of transition towards the long term, in
which liberalisation policy will have an influence on growth.
To lay out a plausible transition path, it is necessary to develop reasonable
hypotheses concerning the export ratio to which each of these economies might
converge in the long term. Even with a comparable degree of trade policy openness,
there can be substantial differences between these countries owing to their differences
in size, location and endowment in natural resources. Economies that are similar to
but more open than the economies studied are used for reference wherever possible;
in the case of economies that subsequently went through episodes of restrictive trade
policy, past export performance is used. Table 5.4 summarises these hypotheses.
Table 5.4. Hypotheses concerning the Long-term Export Ratio
Country Long-term export ratio (%) Hypothesis
Burkina Faso 20 Openness comparable to Mali today

Côte d’Ivoire 69 Openness comparable to Mauritius today
Ghana 69 Openness comparable to Mauritius today

Mali 30 Openness comparable to the African average
Tanzania 40 Openness comparable to the country’s historical maximum
Uganda 20 Openness comparable to Mali today
Mali, which has extensively liberalised its foreign trade, is used as the reference
for the poor, landlocked countries (Burkina Faso and Uganda). Mauritius, which
combined liberalisation with export promotion policies and is the only African country
to have succeeded in starting the industrialisation process, is used as the reference for
Côte d’Ivoire and Ghana. For Tanzania, the historical maximum is appreciably higher
than the current level; it was observed during the pre–Arusha period (before 1967),
during which the economy could be regarded as open. Lacking a yardstick for Mali,
we have assumed that further trade liberalisation and economic development could
lead to an export ratio close to the African average.
125
These are fairly rough assumptions, but their only purpose is to allow us to
construct a plausible central scenario of changes in export performance in a context of
continued trade liberalisation. As there is a great difference between the initial situation
and the hypothetical long–term position, our results are in fact not very sensitive to
our choice of assumptions.
The transition path of the export ratio towards its hypothetical long–term value
is assumed to converge asymptotically towards this value at a convergence rate
calculated from recent past performances. The results of these hypotheses are displayed
in Figure 5.2.
Sensitivity tests showed that variations on the baseline scenario had no major
consequences for the growth path of the economies studied. We have therefore not
drawn up a high scenario for economic openness –– its parameters would be very
close to those of the baseline scenario and would not provide any fresh insights.
In our scenarios, an increase in export volume has a much smaller impact on
growth than that obtained by enhancing diversification of exports. In the case of Côte
d’Ivoire, for instance, if the export/GDP ratio — which in the baseline scenario
converges asymptotically to Mauritius’s current level — were instead maintained at
its current level, there would be a per capita income loss of only 5 per cent by 2020. If
diversification were also maintained at its current level instead of converging to the
level of the NIEs, there would be an income loss of 11 per cent. The effect of
diversification in this example is thus more than double the effect of a pure increase in
export volume.
In light of these findings, we concentrate our analysis on diversification rather
than on export growth as such. The fact that diversification leads to greater gains than
a pure volume increase does not imply, however, that export growth should be de–
emphasised to the exclusive benefit of diversification. Indeed, openness to trade is a
pre–requisite for diversification, given the small size of the markets in the African
countries under study. We will therefore end this chapter with an analysis of the potential
role of trade agreements in the effort to increase economic openness in Africa.
Openness to Trade and Diversification: the Role of Institutions
Under these conditions, what can policymakers do to encourage diversification?

The experiences of several African countries demonstrate that diversification cannot
be created by government decree. These countries include Algeria, Côte d’Ivoire (see
Berthélemy and Bourguignon, 1996) and Senegal (see Berthélemy and Vourc’h, 1996),
where attempts to diversify the economy by government–directed investment resulted
in distortions and substantial inefficiencies. The driving force for diversification must
without a doubt be market incentives. In other words, the government’s role is to
create an economic environment conducive to private enterprise. One way it can do
this is by establishing export processing zones (EPZ; see annex to Chapter 2).
126
Figure 5.2. Hypotheses Concerning Change in the Export Ratio
127
Ghana
Mali
128
Tanzania
Uganda
129
Diversification can be viewed as the result of a combination of high–quality
production factors (skilled labour and well–functioning infrastructure in particular), a
macroeconomic climate favourable to investment, openness and access to non–
traditional export markets, and institutions favourable to risk taking. Promotion of
diversification is therefore dependent on a number of policy measures, making it
difficult to disentangle the influence of diversification from that of other sources of
growth.
The way we have designed the scenarios, however, gives us an opportunity to
study the institutional aspects of encouraging diversification. In the “less
diversification” scenario, we make the same assumptions as in the high scenario, except
that the degree of diversification in the economy follows the same trend as in the
baseline scenario. As we have indicated, an increase in diversification depends on a
number of structural factors, including the availability of skilled labour and
infrastructure, external factors such as capital inflows as well as access to export markets
and imported technology, and factors related to risk and to political and economic
stability. We include the quality of institutions in the last category. The business climate,
and thereby the incentives for entrepreneurs to invest in new areas, is negatively affected
by corruption and rent seeking, a poor judicial system, red tape and other distorting
effects of heavy government intervention. One concrete example of institutional
malfunctioning relevant to Africa is the inadequate protection of property rights, which
increases the risks faced by banks and other financial institutions in their loan
operations. Note that in the “high” and “less diversification” scenarios we have assumed
the same levels of human capital, openness to trade, and macroeconomic and political
stability. Moreover, capital flows are unchanged between the two scenarios, implying
that investment in infrastructure remains essentially the same4. In other words, the
difference between the two scenarios in terms of the growth of diversification can
stem only from institutional improvements, or from better access to export markets
for non–traditional products. It should be noted, however, that the total volume of
exports as a ratio of GDP is the same in the two scenarios.
Given the institutional character of this variation of the high scenario, we will
call the difference between the “less diversification” and the “high” scenario the gap
due to “institutional aspects of diversification” (see Table 5.5). This gap is of course
in addition to the more direct impact institutions may have on growth through misuse
of public funds, or the deterrent effect poor institutions may have on investment. In
other words, good institutions enhance not only the volume but also the diversification
of investment.
Table 5.5 clearly points to the importance of the quality of institutions for economic
growth, although we show only a fraction of the effect here. In all the countries studied,
the enhanced diversification resulting from institutional improvement has a cumulative
effect on income amounting to approximately 5 per cent by the year 2020.
130
Table 5.5. Institutional Aspects of Diversification
GDP/capita in 2020 (1996 $) Gap due to Gap due to

institutional aspects institutional aspects
Less diversification High scenario of diversification of diversification
(%)
Burkina Faso 540 566 26 5

Côte d’Ivoire 1 612 1 693 81 5
Ghana 611 635 24 4
Mali 497 521 24 5
Tanzania 396 415 19 5
Uganda 601 629 28 5
As indicated above, increasing diversification also involves some attention to

external co–operation. For instance, Uganda is a landlocked country, dependent on
Tanzania and Kenya for access to the sea. If the three countries cannot agree on joint
investment in infrastructure, Uganda will have few prospects of export growth, which
in turn will imperil its potential for diversification. Undeniably, the relationship with
potential trade partners also affects the diversification of an economy. A key factor in
the success of the EPZ in Mauritius was that the country gained preferential access to
export markets, particularly in Europe, through various international trade agreements.
However, it is doubtful that trade agreements will have any major impact unless
domestic institutions display a certain level of quality. Well–functioning domestic
institutions are thus a prerequisite for trade and regional co–operation agreements in
achieving the desired effect on diversification. The role of international trade
agreements will be discussed in some detail below.
Integrating Africa into the Global Economy: the Role of Trade Agreements
Compared to other developing countries, African countries have so far enjoyed

relatively easy access to world markets, largely attributable to the preferential trade
treatment granted to the ACP countries by the European Union under the successive
Lomé Conventions and the more recent (2000) Cotonou agreement. At present, only
Mauritius has actually cashed in on this advantage, by developing its exports of
manufactured goods (especially textiles). Overall, Africa’s share in world trade has
declined in recent decades, due to poor macroeconomic performance.
This does not mean, however, that the question of access to world markets will
not be an issue for African countries in the future. The principal policy issue under
consideration today concerns the future ACP–EU trade regime, which is due to replace
current non–reciprocal trade preferences by 2008 at the latest, and be compatible with
131
WTO rules. The previous agreement is incompatible with these rules because it implies
discrimination against other developing countries, which face less generous customs
treatment by the EU under its Generalised System of Preferences (GSP). This should
not be much of an issue for the so–called least developed countries (LDCs), since all
LDCs will likely be entitled to keep at least the current level of access to the EU
market, but a new system needs to be developed for non–LDC ACP countries (in our
sample, Côte d’Ivoire and Ghana).
Essentially, the EU proposes two solutions for non–LDC ACP countries: either
to abandon their preferential treatment and access the EU market under the GSP rules,
or to sign reciprocal free trade agreements (so–called EPAs, or Economic Partnership
Agreements) with the EU. An EPA would imply that EU products could, after an
adaptation period, enter these economies free of charge. This is considered unacceptable
by a number of countries wanting to protect their “infant” industries. In view of the
limited size of the economies under study, however, the infant industry argument cannot
really be considered pertinent. Rather, openness to trade is a prerequisite for the
development of non–traditional industries.
Under the GSP, in contrast, non–LDC ACP countries such as Côte d’Ivoire and
Ghana could face two limits to their diversification processes.
First, the GSP provides for duties on imports that compete with European
products. Production resulting from a diversification process would inevitably fall
into this category, making this provision a serious impediment to diversification.
Second, the GSP is unilateral rather than negotiated with developing countries.
If these countries negotiated EPAs instead, they would have the opportunity to include
a discussion of non–tariff barriers in the negotiations. Non–tariff barriers could become
an even greater impediment to exports of diversified manufactured products from
these countries in the coming years. In the realm of protective measures tariffs are
being replaced increasingly by various non–tariff barriers, for example anti–dumping
duties. Such measures offer high protection to UE producers and affect principally the
textile, chemical and electronic industries rather than high–tech industries. The same
applies to the United States. The countries currently most affected are Asian NIEs and
China. African countries are not affected to any great extent, but this is simply because
they are not yet competing with European manufacturers in the European market. If
countries such as Côte d’Ivoire and Ghana were to diversify and follow the path of the
Asian NIEs, they would likely face the same reactions from European companies in
the future and would in turn be affected by such non–tariff barriers, which are not
only high but also unpredictable and to some extent beyond the control of governments.
Securing preferential access to the European market under free trade agreements with
the EU may be the safest means of avoiding such measures.
132
Conclusion
Since 1980, the economies studied have posted mediocre export performances,
attributable both to specialisation in sluggish demand segments and to losses of
competitiveness. Although the first handicap was difficult to overcome early in the
reform process, as it requires true structural change, the recent period witnessed
significant progress in competitiveness, due in particular to the introduction of
appropriate macroeconomic policies. However, such competitiveness gains cannot be
reproduced indefinitely. The primary determinant of these economies’ ability to export
will be their output capacity.
Trade policy can play a negative role, however, both when it restricts exports
directly and when it protects against imports, by limiting the openness of the economy
and reducing incentives to start up export businesses. All of the countries studied had
restrictive trade policies until the mid–1980s, but they have diverged since then. Côte
d’Ivoire and Burkina Faso are still relatively closed economies. In Côte d’Ivoire, this
is not reflected in the level of the export ratio, because the country has considerable
export potential owing to its natural resource endowments and the export infrastructures
built in the past. In contrast, Burkina Faso has very little export capacity.
These two countries thus have considerable room for export growth. The same
is true of the other countries, either because they liberalised their economies relatively
recently or only partially (as in Tanzania) or because the infrastructures and services
needed for export business need to be built up or rehabilitated.
The liberal policy stances adopted by the governments of the countries studied
justify the fact that our scenarios are based on an assumption of growth in their export
ratios. As shown in Figure 5.2, we assume that export ratios will be subject to an upper
limit in the long term.
133
Notes
1. For the formulae used in this accounting method, the reader is referred to research by the
CEPII. See e.g. CEPII (1998).
2. For Tanzania, we used the 1994-96 period in our calculations, in order to have at least
two years of observation.
3. A similar initial drop in the exports/GDP ratio, of shorter duration, was observed in
Ghana.
4. In fact, according to the investment function in our model, the total capital stock will
remain unchanged between the two scenarios, with the only exception being the beneficial
impact of higher capital productivity on investment incentives. This improvement in
productivity is the direct effect of a higher degree of diversification.
134
PART THREE
THE POLITICAL ECONOMY OF REFORM

IN SIX AFRICAN COUNTRIES
135
136
Chapter 6
Overview
The present analysis takes the form of a series of chapters devoted to each of the
six countries identified in Part One of this work as belonging to the group of dynamic
African economies. This classification, though based on a very recent period, is possible
because of the progress these countries have achieved in terms of economic policy,
even though it is too early to affirm that they have real chances of sustaining economic
take off in the immediate future. These six chapters summarise the long–term
development experience of each country, reviewing its growth mechanisms and socio–
political context. We begin with a general introduction offering a short political–
economic analysis of the six countries, drawing comparisons between their experiences
and those of other African countries — either countries in difficulty (e.g. the Democratic
Republic of Congo) or the countries often pointed to as success stories or examples to
be followed (e.g. Botswana and Mauritius). A few experiences in other developing
regions of the world will be mentioned in passing.
❊ ❊ ❊
The majority of African economies are mired in an inferior economic equilibrium

characterised by a poverty trap, low institutional capacity and a lack of economic
transformation. Agriculture and the informal sector still account for a dominant share
of output and exports, which in most cases means that the economy is specialised in
activities generating low profits in the long term. Public revenues — necessary to
meet the demand for the public goods (institutional modernisation, infrastructure
network, human capital, creation of a business–friendly environment) required for
any emergence from the trap of under–development, and a fortiori for any take off
process — are chronically low. Foreign aid currently helps to offset the fiscal weakness
of African states, but it simultaneously maintains a degree of dependence as well as
the perverse effects that can characterise this situation: less saving, less local ownership
of reforms, little independent strategic thinking and so forth.
137
In most cases, the economic policies implemented during the early decades of
African development were ill–judged and carried out by governments somewhat averse
to exercising development–oriented leadership. Economic analysis of growth has shown
that the poor performances of African economies were largely due to bad economic
policies (Collier and Gunning, 1999). A necessary condition for success is therefore
the implementation of significant and lasting (irreversible) reforms that are clearly
owned by the governments that implement them (or even chosen on a wholly
endogenous basis by these governments). Substantial reforms have been undertaken
over the last decade or more, usually (but not always) following a very serious economic
crisis, and in a context of democratisation. They have played an important role in the
economic recovery observed, but still show a number of limitations. In particular,
structural reforms — undoubtedly even more important for long–term growth than
macroeconomic policy reforms (see Chapter 1 of this volume) — raise more problems
of political feasibility than macro reforms, require more time to implement and, as a
result, are far from complete. The long–term outlook is thus uncertain, since the lack
of any real sectoral diversification, combined with a persistently high level of poverty
(which is in turn tied to a worrying educational and demographic situation), makes
these economies highly vulnerable to economic and/or political shocks.
Economic Policies Leading to Unsustainable Growth
Economic policies in Africa have for many years been unfavourable to long–term
growth, failing to trigger the main mechanisms of such growth. In the best cases (apart
from the success stories of Mauritius and Botswana), resource allocation was often
rendered inefficient by the existence of an excessive number of price distortions, leading
either to ill–judged long–term gambles, as in Côte d’Ivoire1, or to the creation of various
rents that diverted economic agents from activities contributing to the general welfare.
In the worst cases, countries were given over to predatory governments (Collier and
Gunning, 1999). In all cases, economic policy showed varying degrees of anti–
development bias: massive public borrowing to finance the growing deficits (fiscal and
balance of payments); overvalued currency, financial repression and capital flight2;
expansion of a semi–public sector characterised by low productivity and negative value
added activities, often sheltered by high tariff protection. Lastly, Africa set new world
records for political instability and reversals of economic policy. These problems are
often attributed to the vicious circle of bad economic management (see Guillaumont et al.,
1999)3. Human (intrinsic quality of leaders) and cultural factors are undoubtedly important
as well. Many African countries have been caught in a vicious politico–institutional
triangle with, at its apex, a state that usually lacks effective leadership, and sometimes
acts in a divisive and predatory manner (see Box 6.1); at the next level down, an assortment
of ineffective or decaying public institutions; and at the base, individuals who have
become demoralised or turned to rent seeking as a result. Conversely, the success stories
of Mauritius and Botswana (see Boxes 6.2 and 6.3) can be attributed to solid democratic
institutions, sound and endogenously chosen economic policies, and the ability to seize
economic opportunities (diamonds in Botswana, preferential trade agreements in
Mauritius).
138
Box 6.1. A Typology of African Leaders
Chege (1999) distinguishes four types of African leaders:
– conservative and effective leaders,
– ideological and radical leaders,
– predatory leaders, and
– tyrants.
In fact, another category should be added to Chege’s four types: that of rent–
seeking leaders (who introduce economic policies that are unsustainable over
the long term). Leaders in this category are somewhat more restrained than
predatory leaders (exemplified by Mobutu). This typology is applied to each
country in the chapters below.
Box 6.2. Botswana

Botswana (Salkin et al., 1997) has enjoyed exceptional growth by African
standards: per capita GDP grew by a factor of six from 1960 to 1995
(Guillaumont et al., 1999). This growth was due to the profits from abundant
and well–managed mineral resources (mainly diamonds), combined with sound
macroeconomic policy. The countries’ mines, operated by joint ventures between
the state and multinational corporations, accounted in 1988/89 for 51 per cent
of GDP, 59 per cent of tax receipts and 90 per cent of exports. The state has
behaved like a prudent saver (the budget surplus was for many years greater
than 10 per cent of GDP, with tax revenues of over 40 per cent of GDP); the
currency has nonetheless never been overvalued and foreign exchange reserves
are ample (30 months of imports in 1994); conditions are favourable to business
and the economy is open (within the SACU). Despite the downward trend in
unit labour costs (–25 per cent from 1980 to 1992), the manufacturing sector
still accounts for only a small share of output (5 per cent of GDP) and
diversification is currently one of the government’s foremost objectives. The
political regime is a stable democracy with a strong propensity for consensus.
However, the country’s level of human development remains far below its
income level: life expectancy is only 51 years, and the rampaging AIDS epidemic
strikes one of every four adults (L’État du monde, 2001).
139
Reforms: Progress and Limits
Growth in the six countries studied under this project unquestionably revived
during the 1990s (see Table 6.1), and this upturn is related to clear improvements in
economic policy. The path followed in implementing reforms is important, in
accordance with the traditional sequence4: reduction of internal and external deficits
(including foreign donations), flotation of the national currency and adjustment of the
real exchange rate, privatisation, trade and price liberalisation, etc. As a result, rates of
domestic saving and investment have picked up, but in many cases the policy mix
could still be improved (see Table 6.1) and many structural conditions for economic
take off have not yet been met .
Similarly, the reforms owe a great deal to donor involvement, and there is not
enough local ownership of the reform process. As a result, the credibility of the reforms
is far from assured: on the whole, it is unlikely that capital has been repatriated and
the level of private investment remains modest even when FDI is included, despite the
pull exerted by privatisations. There are several reasons for this lack of credibility.
First, in the eyes of international investors Africa has a generally negative image,
which is reinforced by the “media contagion” effects of the numerous armed conflicts
and civil wars in progress on the continent. As international investors are lacking in
discrimination, they have little ability or simply little inclination to single out Africa’s
good pupils. Second, the implementation of most reforms followed a genuine economic
collapse that left the government no other choice than to turn to the donors (applying
the well–known principle “there is no alternative”). Reforms were therefore not
sufficiently endogenous.
In the period under review for the six countries studied, there is only one case of
reform initiated in the absence of a major economic crisis: the reform programme
implemented in Burkina Faso by Compaoré in 1991. The decision to undertake reform
was prompted by mere stagnation of growth, at a time when economic fundamentals
were relatively under control. Moreover, we should note Compaoré’s remarkable
political approach: the reform was implemented during the period of political
liberalisation, and Compaoré took care to negotiate at length with donors and to
undertake broad–based consultation of civil society. The reform process was thus largely
a matter of local ownership, and consequently will not be reversed. Two other cases
should be mentioned, concerning the deepening or the resumption of reforms in the
absence of a serious economic crisis. The first occurred in Mali. Konaré, democratically
elected in 1992, gave a boost to the reform process following the political (not
economic) crisis that had brought Traoré down the previous year, and in this case as
well the reforms have lasted. The second was in Tanzania: in 1995, recently elected
President Mkapa restored the dialogue with donors, which had been cut off in a climate
of widespread corruption, and stepped up the pace of reform.
140
Table 6.1. Overview of Economic and Social Conditions in 9 African Countries at the end of the 1990s
Burkina Faso Côte d’Ivoire Ghana Mali Tanzania Uganda Dem. Rep. Botswana Mauritius
Congo
GNP per capita, 1997 ($) 240 690 370 260 210 330 110 3 260 3 800
AAGR of real GDP, 1991-99 4.3 3.1 4.4 4.1 2.7 6.8 -5.8 5.0 5.3
(%)
National (domestic) saving/GDP
(%), 1998 21 (9) 14 (20) 13 25 (10) 16 18 (7) -1.9 26 31
Investment/GDP (%), 1998 27 18 17 26 19 25 ou 17* 8 28 28
Literacy rate, 1995 19 40 64 31 68 62 77 83 70
Primary/secondary school
enrolment rate, 1996 (%) 39/9 71/24 76/31 37/11 66/5 74/12 70/30 107/65 112/66
External aid/GDP (%), average 15 13 (1995) 11 15 20 15
for decade 1990s 23 (1994) 25 (1994)
Beginning of 1991 1981 (i) 1983 1982 (i) 1986 (i) 1987 No structural 1979
reforms (SAPs) 1989 (ii) Rawlings 1992 (ii) 1995 (ii) adjustment
Compaoré
141
1994 (iii)
Causes of Econ. Crises (i, ii, Econ. Macro Crisis and Crisis; Macroeconomic
reforms slowdown; iii); exogenous collapse; imbalances disastrous EP (i); disastrous EP imbalances
mediocre EP; shocks and disastrous EP 1982 (i); excessive public (excessive
consultations; unsustainable political crisis spending, break in absorption as a
endogeneity EP 1992 (ii) the dialogue with consequence of
donors and change increased public
of government (ii) spending)
Flaws in post-reform macro- No major flaw REER through Electoral No major flaw Budget through Budget n.s. n.s. n.s.
economic policy through 1997 noted 1994; falling cycle in public noted 1996; through 1992;
(as indicated in the six case education and spending; REER appreciation exchange
studies) health unstable at end of period policy
spending REER through 1997
Sources: Numerical values: ADB (1999, 2000); text: case studies, Salmon (1995) for Mauritius, Salkin et al. (1997) for Botswana.
* The 1999 and 2000 editions of the ADB’s African Development Report provide different figures for the year 1998 (Table 2.4 of the statistical appendix: the 1999 edition gives 24.7 per cent and
the 2000 edition 16.9 per cent).
Key: AAGR = average annual growth rate, EP = economic policy, REER = real effective exchange rate, SAP = adjustment structural programme, n.s.= not studied.
Box 6.3. Mauritius
The gift for compromise of Dr Ramgoolan, the father of Mauritian independence,
made it possible to distribute power in Mauritius between the whites or “Franco–
Mauritians”, who retained their monopolistic capital holdings stemming from
the colonial period, and the Hindus (two–thirds of the population) who
established themselves in politics and the civil service. Mauritius achieved
vigorous long–term growth by exploiting its comparative advantage: skilled
and initially inexpensive labour (Salmon, 1995). When economic fundamentals
deteriorated as a result of the slump in sugar prices, they were re–established in
a sustainable manner through structural adjustment (1979–85). Real wages were
adjusted downwards and the currency substantially devalued. The duty–free
zone then experienced a veritable boom (benefiting from Hong Kong Chinese
know–how and capital in the period just before Hong Kong’s reintegration
with China, but also from local capital stemming from the sugar industry) and
created nearly 100 000 jobs in five years, which was enough to eliminate
unemployment and create an impression of industrial take off in this small
island with a population of about a million. The textile and clothing industry
accounts for the greatest share of business activity in the free zone; other
manufacturing activities are surviving in the shelter of considerable tariff
protection, which the authorities have so far jealously maintained. The two
main exports (sugar and textiles) depend on preferential trade agreements that
are now endangered (exemption from the MFA and the sugar quota set in the
context of EU–ACP relations), while real wages have risen. The future of the
Mauritian economy now lies in tradable services that are intensive in information
and technology (financial services in particular). The many partisan coalitions
which have successively held power (very often led by A. Jugnauth since 1982)
have never placed the stability of the democratic regime at risk, nor really
compromised the quality of economic policy. Governance is highly satisfactory,
notably owing to a corps of brilliant high officials.
Ultimately, reforms are still too recent to confer credibility on young states which
have been independent for less than half a century and whose institutional capacities
are still weak.
Outlook
Economic take off and lasting growth depend on the consolidation of what has
already been achieved, but also on meeting all the conditions required for true structural
change, with a view to the vitally needed diversification of output and the modification
of countries’ international trade specialisations5. If these conditions are not met, the
progress achieved could prove to be ephemeral because African economies are highly
vulnerable to exogenous shocks, whether these shocks are economic in nature
142
(e.g. shocks to the terms of trade or climatic shocks affecting agricultural output) or
political (entailing the risk of a return to autocratic or predatory regimes or to political
instability, as shown by the very recent experience of Côte d’Ivoire). We will discuss
in turn domestic economic conditions, domestic political conditions and international
political–economic conditions.
Domestic Economic Conditions
Action is required at several levels:

— Domestic saving rates are too low6, which hurts investment. The investment
ratios required for take off — around 30 per cent, as shown by the experiences
of Botswana, Mauritius (see Table 6.1) and the east Asian countries — are not
yet attainable, although access to external financing has allowed Burkina Faso
and Mali to approach this level. Similarly, the upturn in factor productivity must
continue. Measures should therefore be taken concerning both the demand for
and the supply of capital including, among other things, attracting more foreign
investment capital.
— Similarly, investment in human capital is often inadequate, and the stock of
such capital is already meagre: literacy and school enrolment rates are very low
in Burkina Faso and Mali, the situation in Côte d’Ivoire is worsening, and
secondary–level enrolment rates are negligible in Uganda and Tanzania.
Moreover, some countries allocate their education spending poorly, with a bias
in favour of higher education. In this respect, donors still have a role to play in
terms of providing incentives and monitoring developments.
— The demographic situation remains explosive in many cases, owing to a delayed
demographic transition and rapid growth of urban populations. In 1998, fertility
rates in the countries studied were still greater than or equal to 5 children per
woman (and as high as 6.3 in Mali and 6.9 in Uganda, as against 4 in Botswana
and 2 in Mauritius); this is due in part to the fact that women’s educational level
is below the national average, which itself is very low. Furthermore, a dualistic
analysis is still useful. Although the labour market is no longer seen as overly
rigid (Collier and Gunning, 1999; World Bank, 2000), the urban bias long
observed in pre–adjustment economic policy7 (Bates, 1981) exerts hysteretic
effects in terms of urban population growth and unemployment rates8, which in
turn generate a propensity to frustration and political conflict. It is more necessary
than ever to promote the prosperity of the agricultural sector, giving a high degree
of priority to modernising it9, as shown by the experience of southeast Asia10.
The latter region also witnessed strongly marked family planning policies
(combined with incentives and even administrative penalties in some cases) that
very quickly led to a sharp drop in fertility rates, allowing the saving rate and
average educational level to rise. Africa is lagging far behind in this area, thus
compounding its difficulties in escaping from the poverty trap11.
143
— Sound, credible macroeconomic policy over the long term and the restoration of
price signals may not suffice to bring about significant growth in manufacturing
investment, given the highly imperfect state of international markets and the
financial market. Except in mining sectors and in the context of privatisation,
foreign direct investment is still rare in African countries, including those which
have conducted reforms successfully, and even in Burkina Faso and Mauritius.
This means that in addition to the reform issue — the way reforms are interpreted
is tarnished by the negative image of the continent, and investors continue to
fear that reforms will be reversed (whence the need to develop “restrictive
agencies”; Collier, 1998b) — the size of African markets remains a major
handicap in attracting capital. Regional integration can only reduce this difficulty,
not eliminate it: the aggregate GNP for all of Africa ($544 billion in 1998) is
only equivalent to that of Canada. Now that imports have been liberalised in
most of the countries under review (or will be increasingly liberalised through
open regionalism), some form of active industrial policy might prove necessary
in the future (see e.g. Laffont, 1998; Lall and Wangwe, 1998). This must wait
until states’ institutional capacity has been restored, however, because it is vital
not to repeat the errors of the past, and this requires safeguards, checks and
balances. For example, although in certain cases support to domestic firms is
considered advisable12, it must go hand in hand with the development of internal
and external competition, or failing this, it should be conditional on compliance
with strict performance criteria (profitability, a given proportion of turnover
earned through exports, etc.), which might be set in dynamic terms. To continue
with the theme of microeconomic efficiency, the privatisation process should be
assigned two new objectives to ensure that rates of return in reformed sectors
reach desired levels (in terms of the competitiveness of factor costs; World Bank,
2000): broadening the pool of investors and giving preference to transferring
the asset itself, instead of merely its management.
Domestic Political Conditions
The first point to be noted is that although the level of human development has
a limited effect on short–term growth (see Chapter 1), it works in favour of political
stability (other things being equal) and hence of long–term growth. Moreover, according
to Ranis et al. (2000), almost all of the countries which have achieved a virtuous
circle combining human development and high rates of economic growth initially
placed the emphasis on human development (with the exception of Botswana).
Human development in itself does not constitute a guarantee against conflict.
The Democratic Republic of Congo, for example, ranked significantly higher in 1995
in the UNDP’s HDI index than it did in terms of income level, owing to its adequate
performance in education13; yet the country suffered for many years from the predatory
regime of Mobutu, and the war which has raged since 1997 is a good illustration of
the principle that the probability of domestic conflict or civil war increases with the
existence of a substantial natural resource14. The example of the Democratic Republic
144
of Congo also shows how great the economic cost of conflict can be (see Table 6.1):
increased military spending in a context where tax revenue has a high opportunity
cost, destruction of infrastructure, falling savings and investment, a brain drain and,
in the end, economic collapse, combined with a bad reputation that will generate
hysteretic effects when the country tries to rebuild (and that has already rubbed off on
neighbouring countries).
In many African countries (Collier and Gunning, 1999) — and the six studied
here are no exception to the rule — political feasibility was less of a problem for
macroeconomic policy reform than for structural reforms, as the latter have a more
direct effect on vested interests. Although the latter inherently require more time, it is
also possible that they have been delayed owing to a delicate balance of political
forces15. In this respect, the democratisation process has a complex effect. The fact is
that, over the history of the 20th century, strong political leadership almost exclusively
concerned with national development and wealth creation seems to be a necessary
condition for economic take off, because such leadership is needed to mobilise all
surplus labour at a time when working conditions are very hard16. At the same time,
and conversely, autocracies have more often than not been disastrous for development
(Taylor, 1993), particularly in Africa and Latin America, in contrast to south–east
Asia17. In the African context, democratisation is probably the best domestic check on
predatory government and inter–ethnic conflict (Collier, 1999). The World Bank, for
example, states that in Africa “checks and balances must be exerted in order to limit
the arbitrary exercise of government. Public decision–making should be transparent
and predictable” (World Bank, 2000). This assertion is not devoid of ambiguity: it
neglects a risk and raises a contradiction. The risk is that in the context of under–
development, democratic governments will be too weak to break up rent situations
(such as those of trade unions or corporatist groups) or to regulate monopolistic markets
(especially, but not exclusively, when confronting multinational companies with
considerably greater financial power): as a result, market failures would persist,
reducing social welfare in the long term. The contradiction is that the reform process
itself, being imposed by donor conditionality, is hardly democratic, despite donors’
efforts to make it more so18. Parliaments are still too often prevented from playing
their proper role: they serve as rubber stamps for decisions taken by a few senior
national and international officials, which can undermine the take–up of reform
measures by economic agents and, in the long term, bring about further political
instability.
International Political Economy
Several improvements are advisable.

Donors should continue to target their contributions more precisely to those
countries which have demonstrated an ability to implement reforms successfully and
to follow a path leading to lasting growth, and possibly to those projects or sectors
having the greatest impact on long–term growth (once the difficulties of identifying
145
such projects or sectors have been overcome) or to poverty reduction (see below).
This is one of the conditions for aid efficiency (Burnside and Dollar, 1997), and in
addition it sends a message to countries which are more reluctant to undertake reform.
The continued implementation of the HIPC debt–reduction initiative will be very
important in order to stimulate growth — the debt burden has a negative impact on
private investment (Servén and Solimano, 1993) — and shore up the democratisation
process by freeing up public resources having a high opportunity cost. In this respect,
a special effort should be made to target foreign aid more to the direct benefit of the
poorest groups; this applies to both current and investment spending, including on
infrastructure (inland transport towards isolated areas, water supply, etc.). Donor
involvement in aid to Tanzania, which is biased in favour of higher education, is from
this point of view regrettable.
146
Notes
1. Côte d’Ivoire invested heavily in education and infrastructure, but based its development
strategy on cocoa and coffee, for which the forecasts of good long–term profitability
proved to be illusory (as for any other agricultural products, since such products are
subject to decline in the terms of trade). In the end, when all of the growth in per capita
output (1960–78) had evaporated, the post–Houphouët–Boigny period was marked by
rising corruption, and the political instability of the late 1990s completed the
destabilisation of the country.
2. According to the World Bank (2000), capital flight in 1990 was much greater in Africa
than elsewhere (39 per cent of national wealth compared to 10 per cent in Latin America
and less than 6 per cent in east Asia).
3. Guillaumont et al. (1999, p. 51) conclude that persistently high levels of poverty may
lead to political and policy instability which can reinforce the vicious circle of poor
management, doubtful economic security and slow economic growth.
4. First, the policy mix is modified to stabilise the economy (reducing macroeconomic
imbalances), then structural reforms are implemented to restore factor productivity.
5. In view of the medium– to long–term constraints on the primary commodities
specialisation of these economies, in particular the prospect of low (relative) prices for
agricultural products on international markets, we consider that sustainable take off is
possible only if it is primarily based on the expansion of non–agricultural tradable sectors.
Under no circumstances does this mean that the conditions for agricultural modernisation
should be neglected — quite the contrary.
6. Limited to 10 per cent in many cases. Côte d’Ivoire is a notable exception, but the country
cannot take advantage of its higher saving rate owing to massive capital repatriation.
7. In particular, excessive inter–sectoral differences in wage levels (sometimes reaching
1 to 4 or 1 to 5 for a given skill level, as against 1 to 1.4 in southeast Asia) that stimulated
rural emigration, even though urban labour markets were already very tight. The effect
of the urban bias is still present: access to education and health services is still much
easier in urban areas, especially capital cities.
8. Urban unemployment rates often exceed 20 per cent, even when under–employment is
not taken into account.
9. The example of Côte d’Ivoire is entirely convincing in this respect: the growth in cocoa
production was mainly achieved by extending the area under cultivation, without
productivity gains; agricultural prosperity proved short–lived when cocoa prices fell,
and the agricultural frontier has now been reached.
147
10. In southeast Asia, the elasticity of non–agricultural value added with respect to agricultural
value added was greater than unity, signifying that the faster the rate of agricultural
growth, the faster was the fall in the agricultural share of GDP (Birsdall et al., 1995).
This interaction, which is typical of the take off process, can be expected to recur in
Africa: the labour reallocation effect (or Syrquin effect) is already clearly visible in the
six countries studied under this project (see annex to Chapter 2).
11. The gloomy predictions of Malthus seem to be coming true in Africa: the inability to
introduce “preventive” checks on population growth leads to the development of what
he called “positive” checks (in the philosophical sense of the term), namely war, famine
and epidemics. This is undoubtedly a major reason for the endogeneity of conflicts and
the explosive expansion of AIDS in Africa.
12. To help them bear the costs arising from the acquisition of technical know–how, from
the need to bear higher unit costs in a situation of increasing returns to scale and low
initial share of the target market, from commercial entry barriers, etc.
13. In 1995, the literacy rate in the Democratic Republic of Congo (77 per cent) was well
above the African average (44 per cent), while the school enrolment rate in 1996 was in
line with the African average of 80 per cent in primary education and 28 per cent in
secondary education.
14. Another example is Angola. Of course, the presence of a natural resource rent is neither
a sufficient condition for the outbreak of conflict (as shown by the example of Botswana),
nor a necessary condition (the conflicts in the Horn of Africa – the internal conflicts in
Sudan and Somalia, the war between Ethiopia and Eritrea – took place in countries with
few natural resources). The human factor is still decisive in explaining political instability
and the outbreak of conflicts, as shown by the adverse political developments in Côte
d’Ivoire in the post–Houphouët–Boigny period, and in contrast, the favourable ones in
Mauritius since independence, owing to the pacifying influence of Dr Ramgoolan and
subsequently of Jugnauth, both of whom were past masters in the art of compromise
and/or forming coalitions (see Box 6.3).
15. The case of Côte d’Ivoire is telling: the government’s civil service reforms (hiring freeze
and wage adjustment) gradually alienated the urban middle class, while political support
in rural areas diminishes when there are adverse shocks to agricultural prices. As a result,
Konan Bédié’s political base eroded, and the odour of corruption surrounding the scandal
of misappropriated European aid funds left him completely discredited until his fall
from power in the coup mounted by Gueï.
16. Rodrick (1995) describes at length the obsession of Korea’s President Park with the country’s
economic take off and the way in which it was useful in mobilising productive forces. The
case of Lee Kwan Yew in Singapore is similar, not to mention the Japanese experience.
17. The authoritarian character of the governments of Singapore, the Republic of Korea and
Chinese Taipei during the take off period is now a matter of common knowledge. Governments
helped to mobilise surplus labour and reduced unit costs by various means (wage directives,
repression of trade unions). See Mazumdar (1993). Similarly, they “managed” the productive
activities of the private sector in a rather forceful manner (Rodrick, 1995).
18. Such as the principle of consulting civil society written into the Cotonou agreement by
the EU and the ACP countries, for which there is also a provision in the formulation of
strategic poverty reduction programmes with the World Bank.
148
Chapter 7
Burkina Faso 1
Background Information and Long–term Growth
Burkina Faso is a very poor, landlocked country with a population of 10.6 million.
GDP per capita was $240 in 1997 (in 1994 the country ranked 119th of 129 countries,
according to the World Bank’s atlas method). The level of human development is also
very low (172th of 175 countries in the UNDP’s HDI ranking). The economy is basically
agrarian: 88 per cent of the working population is employed in agriculture, which
accounted for 42 per cent of GDP in 1997. Agriculture is traditional and small scale,
made up mainly of small family farms where people cultivate the land without holding
title to it. Mining resources are scarce, apart from gold. Population density is high
compared to the Sahel countries (39 inhabitants per sq. km), and land that could be
converted to new farms is limited. Approximately 45 per cent of the population lives
below the poverty line, but the percentage of poor in urban areas is no more than
10 per cent. The literacy rate is only 19 per cent (as against 56 per cent for sub–Saharan
Africa), and drops to 4 per cent for women in rural areas. Life expectancy rose from
36 years in 1960 to 49 years in 1995. The rate of population growth remains high, at
3 per cent annually since 1966, and the dependency ratio2 stood at 50 per cent in the
1990s. There is a tradition of heavy emigration towards Côte d’Ivoire: estimates indicate
700 000 Burkina nationals working in the neighbouring country and some 2 million
individuals of Burkina origin living there. The gap in per capita income between the
two countries has shrunk substantially over time (from a factor of five in 1980 to a
factor of 2.5 in 1997).
The country’s main exports are cotton, gold and livestock. Although their
aggregate value does not cover the cost of imports, the current account is often in
surplus owing to external aid donations and remittances from Burkina workers abroad,
which in some years amount to 10 per cent of GDP (the economic crisis in Côte d’Ivoire
reduced this figure to under 4 per cent in the first half of the 1990s, and it may be even
lower since the recent political crisis).
149
Burkina Faso’s growth rate, though modest compared to that of many developing
countries, exceeds that of sub–Saharan Africa as a whole: per capita GNP grew on
average by 1.5 per cent annually over the 1965–96 period. Although the growth rate
has shown some short–term instability, mainly due to climatic factors (and to a lesser
extent to variation in the terms of trade), the long–term trend is one of lasting growth.
In particular, the country has never fallen into a prolonged recession as a consequence
of specific phases of ill–judged economic policy, a rare distinction in Africa. Lastly,
Burkina Faso belongs to the group of emerging African countries according to the
composite index constructed by Guillaumont et al. (1999)3 ; this index remained positive
(though not high) throughout the period studied.
Political History
Burkina Faso experienced a great deal of political instability from independence

to the late 1980s, a period which saw a succession of short–lived regimes that came to
power through military coups (five in 17 years4) or more rarely through elections.
Still, almost every government enjoyed a period of relative calm (not all of the same
duration, obviously) during which it was able to apply its economic policy without
interruption. It is also possible to discern a budgetary cycle, in which regimes that
showed concern for budgetary restraint came after regimes that sought instead to
increase the amount of public funds for purposes other than economic development.
In 1991, a referendum was approved and a new constitution proclaimed (which
was subsequently revised in 1997), providing for the establishment of a pluralistic
democracy, the institutions required for a state of law, and the election of the president
to a seven–year term. Blaise Compaoré, who had succeeded Sankara by means of a
coup in 1987, was elected president, thus gaining a political legitimacy confirmed by
his re–election in 1998.
The decade of the 1990s differs in two respects from previous decades : i) the
regime in place was both democratic and stable ; and ii) it was also a period of relatively
well–conducted economic reforms (see below). Burkina Faso is thus engaged in a
win–win process, with both political and economic reform. Municipal elections were
organised in 1995 in a bid to strengthen the democratic process and reduce the high
degree of state centralisation by encouraging local democracy. This may help to bring
about some reallocation of public resources to the most disadvantaged areas.
Lastly, despite this turbulent history, governance in Burkina Faso may be
described as increasingly good, while the degree of corruption seems lower than in
other countries in the region.
150
Development Strategy: From Import Substitution to Structural Adjustment
Burkina Faso’s development strategy after independence was based on import

substitution and the proliferation of public enterprises in a rather large number of
sectors (textiles, food, water, energy, transport, telecommunications). Before 1993,
38 state–owned enterprises accounted for 58 per cent of national output. They were
not under tight management, having a combined deficit that in some years was estimated
at over 10 per cent of fiscal revenue. For example, hiring was done with a view to
developing a political clientele, and the state tried to maximise employment in these
companies. Regulation of imports and prices grew considerably more extensive, making
Burkina Faso the most protected and the most controlled economy in west Africa.
The state also became deeply involved in managing the agricultural sector, notably
creating Sofitex to implement an integrated approach to managing the cotton industry,
the mainstay of the economy. Overall, this industry has obtained very good results:
cotton yields per hectare increased eight–fold from 1960 to 1987, and the surface area
farmed was highly elastic with respect to cotton prices. In 1987, however, it became
thoroughly disorganised as a consequence of numerous defaults of payment5. It was
subsequently restructured, with producer groups taking a growing share of the capital
of Sofitex, while the 1994 devaluation had a strong positive effect on producer prices.
Producers were paid more quickly for their crops, and a transparent mechanism for
setting producer prices, including a floor price, was established beginning in 1997.
Overall, centralised management of the cotton industry has been a success, judging
from the increases in production and yields. A lively debate is under way, however,
concerning liberalisation of the industry.
In the early 1990s, Burkina Faso undertook several successive structural
adjustment programmes, after long negotiations with the Bretton Woods institutions
that began in 1989 and after an unusual consultation of the people in 1990. It is also
unusual that the context of the reforms implemented from 1991 was not one of deep
crisis, as was often the case elsewhere, but rather one of relative economic stagnation
with no real loss of control over fundamentals. On one hand, per capita GNP stagnated
on average from 1986 to 1993 and capital productivity fell off in the 1980s (the
incremental capital–output ratio, or ICOR, rose from 4 to 11 over the 1983–91 period);
on the other, the real effective exchange rate and the debt ratio continued to fall, and
rice and cotton yields were good. In a context of democratisation and fairly good
governance, the decision to undertake reforms seems to have been motivated by
endogenous recognition of the substantial benefits to be derived from following an
orthodox adjustment policy — one of the foremost benefits being the increase in net
public aid contributions (including debt relief and restructuring).
151
Factors of Long–term Growth
In accordance with the method used in this volume to estimate growth, we review
here the main factors that influence long–term growth: investment in physical and
human capital, foreign aid, and elements affecting total factor productivity, namely
macroeconomic policy (the budget and the tax system), the exchange rate, trade policy
and lastly the business environment as viewed through the domestic regulatory
apparatus, factor costs and privatisation. A similar approach will be used in the chapters
devoted to the other five countries.
— The investment ratio averaged 17 per cent for the period as a whole (1965–97)
but fluctuated considerably6, though without a strong impact on the growth rate.
This implies that the ICOR and capital productivity underwent substantial
fluctuation as well. Public investment accounted for about 45 per cent of total
investment and over 50 per cent of public spending (twice the average percentage
observed in sub–Saharan Africa). Public expenditures are heavily dependent on
foreign aid, owing to the low saving rate, although the latter did pick up
considerably during the adjustment period (it stood at an average of 2 per cent
for the period as a whole, but rose to 6 per cent in 1994 and even 9 per cent in
1997). Even so, Burkina Faso performed less well in this respect than the other
low–income countries and sub–Saharan Africa as a whole.
— Human capital grew at a fairly fast rate, but it started from a very low level —
below a certain threshold that limits the absorption of new technologies and
increases in productivity. Human capital requirements are still huge, and well
beyond the budget resources available, even though the education sector’s share
rose from 2.6 per cent of GDP in 1980 to 3.7 per cent in 1995, which represents
over one–fourth of tax receipts. The enrolment rate in primary education is 38 per
cent, compared to 75 per cent in sub–Saharan Africa as a whole and 106 per
cent in the low–income countries. Health spending is also low at 2.2 per cent of
GDP, or only $5 per capita.
— Foreign aid is substantial and essential. At the end of the period, net official
contributions amounted to 15 per cent of GDP, nearly half of which was in the
form of grants. The amount of aid skyrocketed during the 1990s, making it
possible to finance substantial public spending (22 per cent of GDP at the end of
the decade) without excessive tax pressure. Debt increased sharply (from 30 to
60 per cent of GDP during the 1990s), but the proportion of concessional debt is
high (95 per cent), meaning that the burden represented by the external public
debt (which accounts for almost all of total debt) is only 1.5 per cent of GDP, as
against an average of 6 per cent in sub–Saharan Africa as a whole. The debt
service ratio stayed below or near 10 per cent throughout the decade, and Burkina
Faso’s eligibility for the HIPC initiative should allow it to clear more than a
sixth of its debt.
152
— Macroeconomic management has not been overly affected by the troubled
political history of the country, despite the existence of a budget cycle drawn on
political lines. To the contrary, the quality of economic policy seems to have
followed a regular rising trend, which may have encouraged the take–up of the
reform measures by the country’s citizens.
Budget policy: Morrisson and Azam (1999) have provided evidence of a
budget cycle in which periods of relaxed budget policy were followed by
periods of spending restraint, with the policy changes often coinciding with
changes of regime. For example, after independence, the Yameogo
government was marked by a generous budgetary policy that led to financial
difficulties. Following the 1966 coup that brought Lieutenant–Colonel
Lamizana to power, finance minister Garango implemented a policy of belt–
tightening and budget consolidation that was unusual for the time: tight control
over spending, stabilisation of the number of government employees and
dismissal of officials considered unproductive, elimination of free housing
for government officials, and mandatory employee savings schemes in
exchange for an easing of the tax pressure on the lowest–paid employees.
Garango was himself let go in 1975 following a series of massive strikes at a
time when the second Lamizana government was struggling with a number
of problems. Lamizana, with an eye to the 1978 elections, loosened the
budgetary purse–strings and subsequently won the election, but public
finances fell further and further into arrears. With the advent of Sankara in
1983, budgetary restraint was quickly placed back on the agenda, and a
programme of self–adjustment and public spending cuts was implemented.
Compaoré, who came to power in 1987, moved gradually towards a structural
adjustment programme, after an initial period of generous wage hikes.
It should be noted, however, that over the long term Burkina Faso has been
relatively careful in managing its public finances. Since independence, the
Burkina state has recorded a negative rate of public saving in only a very
few years7. Since 1986, the net budget deficit (including grants) has remained
modest, averaging 2.5 per cent of GDP; and since the West African Economic
and Monetary Union (WAEMU) came into force, public arrears (especially
domestic arrears) have steadily decreased (from 7 to 1 per cent of GDP),
which boosts the credibility of the state and the financial health of domestic
banks and companies. The share of the wage bill in total spending has fallen
by half since 1989. Lastly, military spending accounts for only 2 per cent of
GDP, owing to relatively good relations with neighbouring countries (apart
from occasional tension with Mali). The tax ratio is regarded as moderate
(13 per cent of GDP), although it has been somewhat erratic over the long
term. At the beginning of the revolutionary period (1984), it stood at 16 per
cent, but was brought down to 10 per cent by 1985. As in most sub–Saharan
African countries, which depend heavily on exports of primary commodities,
153
public revenues are very sensitive to exogenous factors (drought, terms of
trade); but the ratio remains lower than that which should theoretically result
from the structural features of the economy (per capita GNP, degree of
monetisation, and the ratio of mining exports to output).
The real effective exchange rate has depreciated steadily (the pace of
depreciation picked up after the 1994 devaluation, which was readily
absorbed): the decrease over the 1966–95 period was 60 per cent. This was
of course a result of membership of the WAMU (and since 1994 the WAEMU),
as well as of control over inflation (except for the 1973–78 period, which
saw excessive money creation). Thus, although interest rates long remained
under state control (until the early 1990s), the economy was not subjected to
strong financial repression, in contrast to many other developing countries.
The share of the non–tradable goods sector eventually fell in the long term
(unfortunately, the main “beneficiary” sector was agriculture).
Trade reform came late (early 1990s), but was effective, at least in part: non–
tariff barriers were eliminated, the number of tariff rates was reduced to only
three in 1993 (11, 31 and 57 per cent; the top rate was further reduced to
37 per cent in 1997). In 1997, the unweighted average tariff rate applied was
31 per cent, as against 12 per cent in the other WAEMU countries: it will
therefore continue to fall under the WAEMU’s common external tariff (CET)
(a planned maximum rate of 25 per cent in the customs union that should
come into effect in January 2000). VAT was introduced in 1993 at a rate of
15 per cent, then raised to 18 per cent in 1996. It is mainly applied in customs,
owing to the increasingly informal character of the economy. The overall level
of taxation on imports is thus still high (nearly 45 per cent of public revenue).
— The business climate improved gradually during the 1990s, but further progress
is needed, particularly where the cost of inputs is concerned.
Liberalisation of domestic trade (except for fuels) and prices has been carried
out; recently, all of the jurisdictions required for economic activity in a
liberalised framework were established (harmonisation of commercial law
within the Organisation for the Harmonisation of Business Law in Africa, or
OHADA, labour code, insurance code).
Privatisation programmes were implemented as part of the first structural
adjustment programme (1991). They have been delayed for a number of
reasons (political resistance; lack of investors, in particular national investors;
persistent uncertainty concerning business law), and as a result, in 1997 only
19 privatisations of the 41 planned in 1991 had been carried out. Moreover,
strategic sectors (water, electricity, telecoms) are excluded from the
privatisation process. Public monopolies in these sectors give rise to extra
costs that substantially increase the cost of factors for local firms8. This
problem will be felt increasingly as the trade liberalisation process continues
(tariff reduction under the structural adjustment programme, customs union
in the WAEMU, partnership agreement with the EU under the Cotonou
154
agreement). Lastly, a programme of bank restructuring and privatisation
has put the banking system on a more sound footing, although the supply of
loans is still not well suited to demand: in a context of imperfect information,
private banks have a strong preference for short–term loans, particularly to
those commercial activities which are regarded as less risky. The financial
depth of the economy is increasing, however, as the M3/GDP ratio has
remained over 20 per cent since the early 1990s.
Overview and Outlook
Burkina Faso’s long–term macroeconomic performance has been among the best
in sub–Saharan Africa, though not nearly good enough to meet the needs of this very
low–income country with strong educational and social imbalances. The acceleration
in the rate of per capita GDP growth is still very recent (3.2 per cent in 1996 and 2.2 per
cent in 1997), while earlier years were well below past performance (+1.8 per cent
annually over the 1966–86 period, but 0 per cent for 1986–93 and 1 per cent in 1995).
This performance is due to governance of acceptable quality, reflected in a
measure of respect for macroeconomic fundamentals and a good image in the eyes of
donors and lenders. In addition to relatively good governance — which could still be
improved — a democratisation process has been under way for a decade. All in all,
this is a good illustration of the kind of evolution donors hope to see accompany
growth in Africa, characterised by the democratisation–good governance “ticket”
(World Bank, 1991).
This evolution raises two important questions: What are the reasons for Burkina
Faso’s good governance? Will it suffice to bring about long–term economic take off
and lift the country out of extreme poverty? These questions are not easily answered,
of course, but a few explanatory factors are worth mentioning.
Apart from human factors and co–operation with donors (whose contributions
have had a great impact, as we have seen), the quality of Burkina Faso’s economic
policy is probably due to a propitious mix of politico–institutional and socio–economic
conditions. Admittedly, the country suffers from massive, endemic poverty and
illiteracy, and both of these factors are capable of engendering socio–political crises
in many cases. In Burkina Faso, however, they are probably offset by a number of
advantages. As there is little poverty in urban areas and the rate of urbanisation is low
(18 per cent9), the probability of popular uprisings (which rarely occur in the
countryside) is lower, especially since the high rate of emigration to Côte d’Ivoire has
served as a safety valve, reducing discontent over rationing of the labour supply in
urban areas. Moreover, urban population growth was certainly held in check10 by
advances in agriculture, which were considerable and perhaps less inequitable than in
other countries. Progress in agriculture was due both to good public management of
the sector (and a decent policy concerning producer prices) and to appropriate
management of the currency and exchange rate: the currency was never overvalued
despite the adoption of an import–substitution strategy, whereas in other countries
155
this strategy was often characterised by an unsustainable exchange rate that was
unfavourable to tradable activities, especially agricultural exports. The benefits of
agricultural growth, moreover, were certainly distributed in a relatively equitable way,
given the agrarian structure. Major social and economic advances were realised, and
the people certainly felt their effects, particularly through the reduction of the income
gap with Côte d’Ivoire. This probably helped to bring about a measure of popular
acceptance of the economic policies followed. Lastly, the absence of ethnic conflict
and of substantial mineral resources in Burkina Faso made it less probable that factions
would be formed under “predatory leaders” to seize power and reap the spoils, as
happened in other countries.
Following the typology of Chege (1999), it can be stated that Burkina Faso has
fortunately not experienced the third and fourth categories of leaders (predatory leaders
and tyrants), but rather the first category of “conservative and effective” leaders
(e.g. Compaoré) and the second, ideological and radical leaders (e.g. Sankara).
Lamizana’s long term of office is harder to classify: it was more a hybrid, taking on
some features from both type 1 and type 3 in different periods (type 1 with Garango
from 1966 to 1975, and a moderate type 3 following the dismissal of Garango, 1975–80).
In fact, another category should be added to Chege’s four types: that of rent–seeking
leaders (who introduce economic policies that are unsustainable over the long term).
Leaders in this category are somewhat more restrained than predatory leaders
(exemplified by Mobutu). This type would correspond to Lamizana’s second term and
to J–B. Ouadraogo’s two years in power (1980–82).
The question of Burkina Faso’s long–term growth outlook is equally complex.
Given the current state of economic policy and the weight of foreign aid (even though
the latter is well managed), further productivity gains will be limited until the stock of
human capital increases, allowing for significant diversification of the economy.
Diversification will require decades despite very substantial outlays in this area. The
fact is that nearly half of public expenditure is already devoted to investment, and
public spending levels are heavily dependent on aid flows11, which probably limits the
simulated effect of increased aid efficiency (this consists in raising the average share
of aid allocated to investment from 17 per cent to 50 per cent).
Without development of manufacturing, notably for export, economic growth
will remain unstable in the short term and dependent on exogenous factors (climate,
terms of trade, public and private capital contributions). Economic take off could also
be held back by ripple effects from the political crisis in Côte d’Ivoire (as well as
elsewhere in Africa), which makes the region appear more risky in the eyes of investors.
Economic diversification is still essential in the long term. At this stage, there
are two crucial issues: it is very clearly necessary to target education to young women
(in rural areas, the literacy rate for this group is 4 per cent), which is the surest method
of reducing the fertility rate and hence of improving, for a given level of budget
resources, the ratios of education spending and health care per capita (i.e. human
capital) in the medium to long term. In addition, trade policy and the effectiveness of
156
market incentives will play a major role. Will economic liberalisation and the restoration
of prices and institutions be enough — even when accompanied by the steady but
very slow increase in the stock of human capital — to trigger a take off in
manufacturing, or will market imperfections (increasing returns to scale, learning costs,
entry barriers on international markets, etc ) be such that a more aggressive policy will
once again come under consideration?
Notes
1. Unless otherwise indicated, the information in this paragraph is drawn from the report
by Chambas et al. (1999) for the OECD’s emerging Africa project.
2. Defined as the ratio of the working–age population (15–64 years old) to the rest of the
population.
3. This composite indicator is based on 14 variables representing economic policy choices
(structural and macroeconomic policy), economic performance, and any domestic and
foreign conflicts.
4. Including the well–known coup led by Captain Sankara in August 1983, which opened
the so–called revolutionary period in which Sankara defied the Bretton Woods institutions
and changed Burkina Faso’s name to the “country of honest men”. Many accusations of
abuse of power accompanied Sankara’s restoration of financial order.
5. For example, producers experienced problems related to supplies of inputs and payment
for crops.
6. The investment ratio rose from 10 per cent in the 1965–70 period to 21 per cent in 1973–
74, fell back to 13 per cent in 1982, then rose to 24 per cent in 1985 and has since
remained in the vicinity of 20 per cent, apart from the devaluation period 1993–95, when
it fell sharply.
7. Defined as the difference between current revenues and current expenditures.
8. Water, diesel fuel, telephone and rail transport prices in Burkina Faso are much higher
than those prevailing in the region.
9. L’État du monde, 2001.
10. From a Harris–Todaro perspective.
11. Public spending amounts to 22 per cent of GDP, while the tax take is only 13 per cent
(see above).
157
158
Chapter 8
Côte d’Ivoire 1
Côte d’Ivoire, a country of 15 million inhabitants, experienced a long period of

vigorous growth followed by a crisis of exceptional severity. Per capita GNP at
purchasing power parity, which stood at approximately $1 000 (international dollars)
in 1960, reached a peak of $2 237 in 1978 (placing Côte d’Ivoire in the middle–
income group of countries, after a long period during which the real per capita GDP
growth rate stood at about 4 per cent despite a very high rate of population growth),
then fell back to $1 113 in 1995 (annual average growth of –3.7 per cent in real per
capita GDP). The growth rate picked up sharply in the second half of the 1990s, with
a simultaneous increase in the pace of structural reforms, but Côte d’Ivoire will make
up for this heavy loss of income only in the long term. The economy is still mainly
based on exports of cash crops: the main exports for 1997 were cocoa (35 per cent of
total exports), refined petroleum products (11 per cent), coffee (8 per cent) and wood
(6 per cent). The level of human development is low (Côte d’Ivoire ranks in 148th
position in the UNDP’s HDI ranking for 1998), particularly the literacy rate (40 per
cent, compared to an average of 56 per cent for sub–Saharan Africa). Life expectancy
was 52 years in 1995, according to the UNDP. Nevertheless, Côte d’Ivoire belongs to
the category of emerging African countries according to the composite index
constructed by Guillaumont et al. (1999)2, as this index is recovering after having
turned negative in the early 1990s.
Political and Social History
Côte d’Ivoire enjoyed a long period of political stability under the authority of
Houphouët–Boigny, who died in late 1993. According to Chege’s (1999) typology of
African leaders, it is reasonable to conclude that the father of Côte d’Ivoire’s
independence belonged to the category of conservative and (relatively) effective leaders.
159
Henri Konan Bédié, the president of the National Assembly and a former World Bank
officer, succeeded him just before the CFA franc devaluation of January 1994 and
held power until late 1999, in an atmosphere of serious political crisis and runaway
corruption. The IMF and the World Bank suspended their aid in December 1998, as
did the European Union after discovering that its FF 180 million aid package for health
and decentralisation had been misappropriated. A military coup on 24 December 1999
brought General Gueï to power. Dialogue with donors was re–established shortly
thereafter, in early 2000, while Gueï eventually relinquished power and Gbagbo won
a highly controversial election3. Côte d’Ivoire now needs to demonstrate that it is once
again capable of democratisation and good political governance. Immediately before
this grave political crisis, the stability, continuity and competence of government
services were recognised by a large number of companies working in Côte d’Ivoire,
although they criticised a certain number of barriers to competition and a lack of
transparency on goods markets (which are highly oligopolistic), and what is more
important, they deplored the uncertain rule of law, lack of security for property and
people, and the poor functioning of the justice system.
As the crisis deepened in the late 1980s, the poverty rate tripled from 10 to
30 per cent between 1987 and 1993. The recent recovery has helped to reduce it, but
mainly in Abidjan4 and in cocoa–producing areas; elsewhere, it is still very prominent,
and the poverty rate is close to 35 per cent. Trickle–down effects have been too slow
in coming, because modern–sector activities are still overly concentrated and economic
activity has not diversified.
Development Experience: Agricultural Growth, Lack of Diversification, Crises

and Structural Adjustment
Côte d’Ivoire’s mode of development is unique in that it has combined three

characteristics: exploitation of a large agricultural rent set up even before independence,
openness to foreign capital and especially to foreign labour (both skilled labour from
Europe and unskilled labour from the African sub–region), and a strongly marked
effort to develop human capital (which slackened toward the end of the period,
however). The period since independence may be divided into three long phases:
growth and boom (1960–78), failed reforms and a lasting crisis (1979–93), further
reforms and economic recovery (since 1994).
The first phase (1960–78) includes the so–called miracle period, which lasted
until 1974 and was characterised by a high growth rate (8 per cent annually), driven
by extensive cash–crop agriculture (coffee, cocoa and wood). Much of the agricultural
surplus was appropriated by the state5, which in return invested heavily in the economy
(public enterprises and infrastructure). The immigration rate was therefore high at the
time, accounting for one percentage point of the 4.1 per cent rate of population growth
from 1960 to 1980 (Berthélemy and Bourguignon, 1996). Beginning in 1975 there
was a veritable explosion of international coffee and cocoa prices6 which the state
160
exploited along the lines described above, but with even greater intensity: while
producer prices eventually fell over the period7, the resources of the Caistab rose to as
high as 16 per cent of GDP in 1977, and the state undertook massive increases in
public spending (particularly for investment), which accounted for 42 per cent of GDP
in 1978. Dutch disease effects were soon felt: appreciation of the real exchange rate
(by 50 per cent between 1971 and 1978), reduction of exports except for coffee and
cocoa, expansion of non–tradable goods sectors, and parallel increases in the fiscal
and current account deficits.
The second phase (1979–93) started with a reversal in the terms of trade. The
state initially refused to cut expenditures, notably those incurred through the
continuation of investment programmes which proved to have low productivity and
even negative profitability. The balance–of–payments crisis came in 1980, when the
fiscal and current account deficits amounted respectively to 8.5 and 18 per cent of
GDP. Structural adjustment began in 1981. An initial stabilisation plan (1981–83) was
thwarted by a series of exogenous shocks (appreciation of the dollar and rising real
interest rates, deterioration of the terms of trade and a drought leading to a deep
recession in 1983). It was followed by a second adjustment plan (1984–86), which, in
contrast to the preceding one, managed to restore the fundamental balances, though
this success was mainly due to a coffee and cocoa mini–boom that proved very short–
lived: the prices of these commodities dropped suddenly in 1986, plunging the country
once again into a severe economic and financial crisis, to the surprise of the authorities,
who thought they had stabilised the economy for the long term. The reforms —
particularly trade reform — were halted or even reversed, as the government sought
to forestall the social and political turbulence that would be entailed by the drastic
budget cuts required. Similarly, it refused to reduce the price paid to coffee and cocoa
producers, and the dialogue with donors was broken off for two years (1987–89).
The end of the decade is a pivotal period from a political and economic standpoint,
as it saw not only the authorisation of a multi–party system but also the resumption of
structural reforms and liberalisation (stabilisation, structural adjustment programmes
for finance — PASFI, and competitiveness — PASCO). In a climate of social unrest,
producer prices for coffee and cocoa dropped by about 40 per cent, and overall public
spending by 25 per cent. But these reforms failed to restore the fundamentals: the
public deficit reached 12 per cent of GDP in 1993, as receipts fell by over 8 per cent
of GDP; the current account deficit rose to 11 per cent of GDP; and debt swelled
rapidly (77 per cent of GDP in 1985, 196 per cent in 1990 and 243 per cent in 1993).
The third phase (1994–98) began with the devaluation of the CFA franc. An
Enhanced Structural Adjustment Facility (ESAF) agreement was concluded, comprising
a structural adjustment programme for the 1994–97 period and considerable debt
rescheduling. In April 1997 Côte d’Ivoire became eligible for the HIPC initiative, and
a second ESAF agreement was signed for the 1998–2000 period. This period showed
signs of a recovery, helped by a massive return of outside aid, which rose from 8 per
cent of GNP on average in the early 1990s to 23 per cent in 1994 and was still at
161
13 per cent in 1995, while at the same time commodities prices picked up and cocoa
production hit a new high. The budget deficit was rapidly brought under control (by
1996), the successful devaluation made possible a 30 per cent depreciation of the real
effective exchange rate and external balances were restored. Economic recovery is
clearly evident, but does not yet seem to be robust.
Market Reform and Factors of Long–term Growth
macroeconomic policy, trade policy and lastly the business environment, the banking
system and privatisation.
— The investment ratio varied widely with the long–term phases of Côte d’Ivoire’s
economy. For example, it rose to 30 per cent of GDP at its peak in 1978, owing
to the rapid rise in public investment (13 per cent of GDP in the same year). The
shocks of the early 1980s and the 1990s led to massive deflation and the
investment ratio fell to 11 per cent as early as 1986 and 6 per cent in 1993.
Private investment picked up as from 1994, but public investment did not follow
suit (despite some restructuring of public spending): in 1997, they stood
respectively at 10 and 5 per cent of GDP. The investment ratio thus remained
well below the average for the low–income countries (25 per cent of GDP in
1997) and slightly below that of sub–Saharan Africa as a whole (18 per cent the
same year). Foreign direct investment in particular was still insufficient, though
clearly rising; part of this rise was due to exceptional investments related to
privatisation and BOT programmes (see below). Lastly, capital intensity, which
had long been rising at a rapid pace (physical capital per worker grew at an
annual average rate of 8 per cent over the 1960–78 period), fell steadily from
the beginning of the 1980s (by approximately 1 per cent per year), reflecting
adjustment towards a growth path more in line with comparative advantage, but
negatively affecting growth in the short to medium term.
— The domestic saving rate has bounced back well, to 20 per cent of GDP in the
second half of the 1990s, after a severe decline during the crisis, from nearly
30 per cent (in the 1970s) to 10 per cent of GDP (1990–93). However, national
saving is still clearly insufficient, and amounts to less than half of domestic
saving, due to capital outflows related to the repatriation of profits by foreign
firms, repatriation of the earnings of immigrant workers (European and African),
repayment of foreign debt and various forms of capital flight.
— Despite substantial budget outlays (40 per cent of the budget, or 6–8 per cent of
GDP in the 1960s and 1970s), accumulation of human capital has been held
back by exceptionally fast population growth, over 4 per cent a year during the
1960s and 1970s. A demographic transition is now getting under way, but the
fertility rate is still high, at over five children per woman. The level of human
162
capital, estimated via the average number of years of schooling, is one of the
lowest in the world (2.1), and there is a large gender gap where access to literacy
training is concerned. Furthermore, investment in human capital is very low in
rural areas (only half of boys and a third of girls attend school). Overall, the
primary school enrolment rate regressed during the 1990s (68 per cent, as against
73 per cent in the 1980s). Côte d’Ivoire’s level of human development is low in
relation to its level of per capita income: in 1998 (using 1995 data) the country
ranked 15 places lower in the UNDP’s HDI ranking than in the per capita GDP
ranking. The educational share of the budget nevertheless continued to shrink
over the recent period, to 19 per cent of GDP for 1994–98. Within the education
budget, moreover, the share devoted to higher education — a characteristic feature
of the Côte d’Ivoire “model” — has been growing. Thus, in its current state, the
country’s education system is inequitable and needs to be reoriented towards
primary education. Much the same can be said of the health system, which is
costly, tilted against primary care and, in a word, inegalitarian.
— Foreign aid has played an important and even a decisive role in the economic
recovery observed since 1994, in terms of both direct aid and of debt restructuring
and debt forgiveness. A series of agreements (the so–called Naples treatment,
the 1997 agreement under the Brady plan, an agreement with the Paris Club in
1998) brought the public debt/GDP ratio down from 210 per cent in 1995 to
126 per cent in 1997 and to 72 per cent in 1998 — even before any
implementation of the HIPC initiative. Côte d’Ivoire is thus now on a sustainable
path where external debt is concerned.
— Macroeconomic management has been erratic at best: with its long–standing
bias against exports, due to tariff protection and especially the over–valuation of
the real exchange rate, it was not able to anticipate shocks or to adopt a cautious
interpretation of them. The internal and external deficits often swelled, in an
admittedly highly unstable economic climate. Similarly, distortions in cocoa and
coffee producer prices were particularly large, as the state refused to pass on
both the major price increase during the boom and, for a long time, the price
drop of the late 1980s. Macroeconomic management improved greatly in the
second half of the 1990s, but it remains to be shown that this progress in economic
governance will last. The regional integration process should contribute to this,
but the political events at the very end of the decade could have substantial
negative effects in this respect.
— Trade reform, which was begun in the early 1980s, regressed during the crisis of
1987–89. In 1990, the average apparent rate of taxation on imports was 37 per
cent, the same level as in 1983. Actual trade reform started again in 1991 (levelling
of tariff rates, elimination of non–tariff barriers and export taxes and licences,
except for coffee, cocoa and wood, the three cash crops) and the pace of reform
has picked up since 1994. Further dismantling of customs barriers came with
the CFA franc devaluation: in 1997, only two products were still subject to NTBs
(refined petroleum products and cotton), and the average tariff rate fell to 29 per
cent in 1995. According to the annual IMF report, however, Côte d’Ivoire was
163
still one of the least open African countries in 1996. In this policy area where the
government has shown a degree of instability in its choices, the implementation
of the WAEMU’s effective exchange rate should bring faster and more lasting
progress, as well as confer credibility on the reform process.
— Liberalisation of domestic trade has moved much faster since 1994, with the
near–total abolition of price controls, liberalisation of trade in foodstuffs and
liberalisation of the transport sector. New legal codes have been introduced for
investment (1995), mining, labour (1995) and telecommunications. Similarly,
the privatisation process made great strides, starting from a high level of state
involvement in the productive economy8: the 1994–97 period saw the partial or
total privatisation of some 30 firms, in productive sectors, public utilities
(telephone, rail) and public management of specific industries (the CIDT for the
textile industry). The coffee marketing entity was sold off to the private sector
during the 1998–99 campaign, and that for cocoa was slated for privatisation
under the 1999–2000 campaign. Restructuring of the banking sector has in theory
been completed; however, as a consequence of the highly oligopolistic structure
of the banking market (four major banks hold between 70 per cent and 80 per
cent of the market ) and its vulnerability due to the recent emergence from a
crisis, the proportion of loans allocated to medium– to long–term investment is
insufficient (only a third of total loans). Most investment operations are therefore
financed by foreign banks that are not established in Côte d’Ivoire, and there is
no financing mechanism for SMEs–SMIs. The transformation of the Abidjan
stock exchange into a regional stock exchange should, however, support the
development of the private sector and the inflow of foreign capital. The financial
depth of the economy is rather favourable, with an M2/GDP ratio of 27 per cent
since the beginning of the 1990s (although it has fallen with respect to the
“miracle” years, when it stood at over 30 per cent).
— Partnership between the public and private sectors is also on the rise, with the
launch of the heavy infrastructure programme called the “12 labours of the
African elephant”. The actual construction is to be handled by private contractors
on a BOT basis (“build–operate–transfer” operations, in which the facilities
concerned must be transferred to state ownership after being operated on a
concession basis for 15 to 30 years). At this stage, however, despite the large
investment outlays during the “miracle” period, Côte d’Ivoire does not look like
a country that has made or is making greater efforts than other developing countries
where basic investment is concerned. In particular, transport costs to and from
Côte d’Ivoire remain high compared to those of Asia and Central America.
The economic recovery in Côte d’Ivoire, though quite noticeable, seems still to
be somewhat shaky, and there are a number of factors threatening long–term growth.
164
First, the factor accumulation process is not very encouraging. At the end of the
1990s, investment in physical capital was still inadequate. There was little FDI apart
from that related to the privatisation programme, and public investment spending was
only modest. Similarly, investment in human capital is now clearly insufficient — it
is decreasing in relative terms — and remains disproportionately geared towards higher
education, to the detriment of primary education.
Second, the state has not managed to create the conditions needed for true
structural change. Côte d’Ivoire remains an economy based on cocoa, which accounts
for 35 per cent of export revenues, and an economy that lives on primary commodity
rents, with traditional export industries bringing in 80 per cent of revenues. For far
too long, the authorities have placed their hopes for economic take off almost
exclusively on agricultural industries which are insufficiently profitable in the long
term. Moreover, the strategy of extensive development of cocoa production by gradually
putting large forested areas under cultivation is hitting its limits (the forest frontier has
been reached), while the various policies for agricultural intensification seem to be
viable only in periods of steady prices for agricultural products9. Overall, Côte d’Ivoire’s
agriculture has seen almost no increase in productivity in 40 years.
Côte d’Ivoire’s recent economic growth is more a catching–up process than a
real take off; factor productivity has grown perceptibly in the recent period (+4.8 per
cent for 1994–96), but according to Verdoorn’s law this could be mainly due to the
currently dynamic state of the economy.
The development of labour–intensive manufactured exports must remain a leading
objective. The reforms already carried out and under way are helping to establish
incentives in line with this objective, but achieving it will require the fulfilment of
other conditions, having to do with financing of activities, the skills of the labour
force and the quality of infrastructure. Production costs are still too high, despite the
reforms; for example, the devaluation did not bring about any really significant
reduction of real wage costs10.
Lastly, the observed improvement in governance needs to be maintained, and
extended to the judicial system. Yet the doubt surrounding Côte d’Ivoire’s political
stability — apparent since the end of the 1990s — is hurting its international image
and holding back the emergence of its economy. This adverse socio–political
development had been in the making for a number of years. Cogneau and Mesplé–
Somps (1999), for example, point out that although Côte d’Ivoire may have been
described as the “republic of good pupils”, it seems to have punished its children, i.e.
educated employees, and particularly government employees. Their real wages were
adjusted downwards by 30 per cent from 1994 to 1996 through the money illusion
and political fatigue (after the failed attempt to reduce public–sector wages in nominal
terms, which killed the 1989–90 adjustment programme). Moreover, public
employment has been frozen and private modern–sector wage employment fell until
1994, causing a high unemployment rate among higher education graduates: wage
employment in the modern sector (public and private) was no higher at the end of the
1990s than in 1985. An even more fundamental problem is that the years of crisis have
165
eroded the prosperity of the educated urban middle class and its trust in the ruling
elite, which has lost legitimacy as further cases of corruption have come to light. The
difference in income and living conditions between the Abidjan lower middle class
and the ruling elite has probably widened, while the new class of educated young
entrepreneurs, now deprived of access to the civil service, is not yet very large.
Furthermore, the crisis provoked a rise in xenophobia and conflict between natives of
Côte d’Ivoire and immigrants from other countries in the region. Lastly, the reforms
in progress, particularly the privatisation of the cocoa and coffee industries, have
limited the government’s options in terms of re–distributive policies and maintenance
of the traditional balance of political forces between planters and government
officials. The state will have to demonstrate its ability to identify and implement
new instruments of political and social management essential for democratisation
and long–term growth.
166
Notes
by Cogneau and Mesplé–Somps (1999) for the OECD Development Centre’s Emerging
Africa project.
foreign conflicts.
3. The election was compromised by the inegibility of former prime minister Ouattara,
who precipitated the crisis of 1999 and the rise of xenophobia and discord between the
Christian and animist southern part of the country and the Muslim north.
4. Abidjan was sheltered to some extent — less than 5 per cent of poor people — by the
maintenance of public employment and wage levels, at least until 1993; in 1995, the
poverty rate had risen to 20 per cent, then fell again to 11 per cent in 1998.
5. Through the Caistab and export taxes on the products concerned.
6. The prices of these products increased three–fold or more from 1975 to 1977, and Côte
d’Ivoire’s terms of trade thus rose by 80 per cent.
7. Using index numbers (1975 = 100), the real producer price fell to 72 for cocoa and 84 for
coffee in 1977.
8. Public capital accounted for 55 per cent of the capital in the modern sector in 1982, and
still 43 per cent in 1992.
9. They should also be accompanied by a good system of agricultural credit, which did not
yet exist in 1999.
10. Over the 1993–96 period, the average real wage in the private sector fell by 15 to 20 per
cent, while the real minimum wage increased by 6 per cent.
167
168
Chapter 9
Ghana 1
Ghana is a poor country with 19 million inhabitants, 38 per cent of whom live in
urban areas2. The population grew at an annual rate of 2.7 per cent in 1997, owing to
a slight drop in the fertility rate to 5.1 children per woman (ADB, 1999). Per capita
GNP was $370 in 1997 (ADB, 1999). The country’s level of human development is
slightly below its income level: according to UNDP data (Human Development Report,
1998), Ghana came in 133th place in the HDI ranking, or eight places lower than in
the per capita GDP ranking (at PPP). A substantial proportion of the population lives
in poverty (30 per cent in 1999, according to the government), mainly in rural areas.
In 1997, life expectancy was 49 years and the literacy rate 64 per cent (ADB, 1999).
Political History, Development Strategy and Economic Performance
Ghana became independent in March 1957. With the support of the urban working
classes, Nkrumah, leader of the independence movement, took over the reins of power
and became president in 1960, determined to pursue his grand vision of Ghana.
He remained in power — various elections were thoroughly rigged within the single–
party framework — until 1966, when he was overthrown by Brigadier Afrifa and his
National Liberation Council. Nkrumah’s economic policy was increasingly based on
import substitution and the development of a large public enterprise sector; public
spending shot up rapidly (up to 25 per cent of GDP), leading to fiscal deficits (between
5 and 10 per cent of GDP), current account deficits and sustained money creation.
The currency became over–valued (real appreciation of 50 per cent from independence
to 1965) and the export ratio fell from 30 to 18 per cent. Prices paid to cocoa producers
dropped by over 50 per cent. Real per capita GDP stagnated overall, and it even declined
as from 1964.
169
The military government of Afrifa (1966–69) had to cope with the critical situation
left by its predecessor, notably in terms of foreign exchange reserves. It cut spending
somewhat, particularly investment spending, and introduced a slightly tighter monetary
policy. It was forced to devaluate the currency by 43 per cent in 1967, which rectified
the over–valuation only partially (from 50 percentage points to 30), but as a whole its
policy mix helped to stabilise the economy, helped along by the gradual doubling of
the international price of cocoa from 1965 to 1969. In the latter year, the military
government handed over power to Busia, who was elected democratically in August.
The foreign debt inherited from Nkrumah was also reduced and restructured. Per capita
GDP picked up slightly.
The Busia government held power for only three years (1969–72); its political
base (mainly comprising the cocoa–producing areas where the Akan language is
spoken) was shaky, as witness the many parliamentary conflicts provoked by
accusations of illicit financial gain on the part of government administrators and elected
officials. A few economic reforms were initiated, including the gradual elimination of
import licences and a rise in the refinancing rate, with mixed results. Import
liberalisation, combined with a 50 per cent fall in cocoa prices from 1969 to 1971, put
further pressure on the balance of payments. In December 1971, the government
devalued the currency by over 80 per cent, a measure that lopped off 20 percentage
points of over–evaluation, but also reduced the real cash reserves of many actors,
including the army.
In January 1972, another coup brought Acheampong and the National Redemption
Council to power, inaugurating 30 years of military government in Ghana. The new
government knew nothing whatsoever of the basics of macroeconomic policy, thinking
that it could run the economy in the same way as an army (through orders rather than
incentives). It decided to re–value the currency (wiping out three–quarters of the
previous year’s devaluation), quadruple the minimum wage and suspend debt service
payments. The illusion lasted for two years owing to an exceptional positive shock
(the surge in international cocoa and gold prices), which enabled output per capita to
rise by 15 per cent (in 1973 and 1974 with respect to 1972) and reach its highest level
in the history of the Ghanaian economy3. Over the longer term, the government’s
refusal to devalue4 and the rising inflation rate led to serious over–valuation of the
currency (80 per cent in 1978). The government then conceded some very timid parity
adjustments in the summer of 1978 (17, 8 and 15 per cent). The increased inflation
rate is easily explained by a loose budget policy (a deficit amounting to 11 per cent of
GDP in 1976, and inflation of 100 per cent in 1977). Nominal interest rates were low,
which meant that real rates were very negative (–50 per cent in 1977). The military
government also nationalised the foreign companies operating in the mining sector
and reduced foreign equity holdings in the banking sector to minority stakes. This
created substantial rents which were appropriated by the military leadership, and
corruption pervaded Ghanaian society in a context of growing discontent. Most
economic agents responded by fleeing the monetary system. Tax receipts dropped
(less than 5 per cent of GDP), the investment rate collapsed (5 per cent of GDP in
1979) and infrastructures deteriorated. The economy had completely degenerated.
170
The army responded to criticism of its economic policy by excluding all civilians
from the government. It formed the Supreme Military Council, and in June 1978
replaced Acheampong with Akuffo, who freed a few political prisoners and readjusted
the exchange rate and interest rates, but much less than was required. Indeed, no
fundamental change in economic policy was made; the military leadership continued
its predatory behaviour.
In 1979, Rawlings’ attempted coup failed to come off, but a mutiny in the army
led to the replacement of the SMC by the Armed Forces Revolutionary Council, with
Rawlings at its head. Afrifa, Acheampong and Akuffo were executed. Elections were
held in July, and Rawlings did not run for office, proclaiming his desire to see the
country purged of fraudulent practices.
The Limann government which came to power in September 1979 did virtually
nothing, however, at a time when there was a tremendous need for reform of the
policy mix, as well as liberalisation of economic activities and the restoration of state
institutions. In 1981, the indicators were once again blinking red. Per capita GDP
continued its fall, which had begun in 1975 (The index level which was 100 at
independence fell to under 90 in 1984).
In December 1981, Rawlings took back the reins of power, formed the
Provisional National Defence Council (PNDC) and tried to regain control over
corruption through a number of alternative administrative structures (including ad hoc
courts and local councils), conducting an often violent “revolution”. No serious
analysis of the causes of the economic collapse5 was undertaken, and economic
policy remained unchanged until 1983.
Rawlings then decided to reverse the course of policy completely, regarding a
market–friendly development strategy as a third way, i.e. an alternative to kleptocracy
and populism, and the only path capable of stopping the downward spiral of poverty.
At the time, the index of the real exchange rate stood at only 7 per cent of its value
at independence (which was therefore over–valued by 93 per cent in real terms. In
the period following independence, the black market premium for foreign exchange
reached 2 000 per cent of the official exchange rate.) The export ratio had fallen to
5 per cent (cocoa and mineral production were at half their level under Busia).
Attempted coups by populist radicals failed. The economic recovery programme
was launched in 1983 and the reforms began. The PNDC was ready to use force to
hold onto power, which was an asset in terms of providing a measure of isolation
from the various corporatist and pressure groups — and hence in terms of the
feasibility of the reforms. The year 1992 was a turning point in this respect, as
democratisation led to electoral struggles (and possibly to cycles in public spending
as well) and a shorter time horizon for decision makers. Rawlings won the 1992 and
1996 elections, however, at the head of the National Democratic Congress (successor
to the PNDC), surrendering power only in 2000 in compliance with the constitution,
which forbade him from seeking a third term.
171
This long, troubled political and economic history plunged the country into a state
of extreme under–development and undermined the recovery process for years — the
time required to build the credibility of the new regime. Using Chege’s (1999) typology,
Ghana’s successive political regimes may be classified as “ideological and radical”
(Nkrumah, Rawlings 1981–83), “conservative and effective” (Rawlings from 1983 on,
Brigadier Afrifa, President Busia) or “predatory” (Acheampong, Akuffo, Limann).
Market Reform and Factors of Long–term Growth
— The investment ratio fell to under 5 per cent during the crisis of the early 1980s,
then gradually rose as from 1983; it stood at only 15 per cent of GDP in the early
1990s, and from 15 to 20 per cent in the second half of the decade (Fosu, 1999).
Domestic private investment is still insufficient. The domestic saving rate was
long held down by the climate of financial repression; it rose to 8 per cent of
GDP at the end of the period. External saving was slightly higher than domestic,
mainly because of international aid. FDI amounted to only 0.4 per cent of GDP
and 2.5 per cent of total investment, even though privatisation operations were
implemented.
— Despite the lack of significant data6, there is no doubt that investment in human
capital picked up considerably when public finances did, although the enrolment
rates in primary and secondary education hardly changed over the long term
(even with respect to the crisis years), standing on average at 70 and 35 per cent
respectively. The quality of education certainly improved greatly, as more teachers
were hired (the total number of teachers rose by half from 1985 to 1994), the
“brain drain” and the exodus of educated people in general came to an end, and
teaching materials became available again. The share of GDP devoted to the
health sector remained low (under 3 per cent of GDP in 1996).
— Macroeconomic management was truly disastrous through 1983 but has
subsequently improved over time, particularly where exchange rate management
is concerned. On the other hand, absorption has remained in excess of GDP,
sustained by foreign aid and accentuated by the electoral cycle that emerged as
part of the democratisation process of the 1990s.
— The state of public finances rapidly improved: as early as 1986, tax receipts had
climbed back to 13.6 per cent of GDP (from 5.6 per cent in 1983), allowing the
PNDC government to spend an equivalent amount. The increased tax take was
172
due to the recovery in declared foreign trade: the trade tax base increased by a
factor of 30 and customs tariffs were simplified, with the overall result that
customs revenues tripled as a percentage of GDP. The central government budget
showed a small surplus until 1991, but democratisation brought electoral cycles
in public spending that had the potential to undermine the solvency of the state.
In 1992, spending rose by 4 per cent of GDP, while revenue dropped (a fall in
privatisation receipts); as a result, the deficit increased to 5 per cent of GDP.
Spending followed a rising trend until 1996 and then accelerated further in the
second electoral period of the decade (1996), reaching one–third of GDP. The
introduction of VAT, which required two attempts7, did not bring in revenues to
match this rise in spending (although they did exceed 20 per cent of GDP in the
second half of the 1990s). The gap between revenue and expenditure swelled to
10 per cent of GDP, and was financed from a number of sources: issuing treasury
bonds (at the risk of crowding the private sector out of a narrow financial market),
borrowing from the central bank (money creation), foreign debt, receipts from
privatisation (ranging from 1 to 5 per cent of GDP per year over the 1993–96
period) and accumulation of payment arrears. The ex post balance between
revenue and expenditure remains extremely delicate, since receipts have not
followed the same trend as spending.
Reform of the exchange rate regime became a priority as early as 1983. A dual
exchange rate system was immediately set up8, but it was hard to manage and met
with the disapproval of donors (the IMF). The dual system was replaced at year–end
1983 with a single exchange rate, at 30 cedis per US dollar (wiping out 20 percentage
points of over–valuation), and from 1984 to 1986 the government carried out several
other de–valuations, eliminating 20 further percentage points of over–valuation. A
weekly currency auction system was introduced in September 1986 and unified in
February 1987, wiping out 30 more points of over–valuation; at this point the currency
was still over–valued, but only by about 20 per cent. An inter–bank foreign exchange
market was created in 1990 for large transactions only; this market was unified in
1992, a year which marked the completion of the transition to a transparent and market–
determined exchange rate. This institutional modernisation gradually won popular
acceptance in Ghana, even though the real effective exchange rate proved to be very
unstable over the 1990s. From 1991 to 1994, the currency depreciated by 60 per cent
in real terms, owing to the looser budget policy and increased inflation during the
electoral year 1992. While this stimulated exports, access to external financing meant
that there was no slowdown in imports. As from 1995, the currency began to appreciate
quickly in real terms, as the government put pressure on foreign exchange dealers in
order to reduce the rate of nominal depreciation (to below the inflation rate). The
balance of payments deteriorated significantly and a nominal adjustment became
unavoidable; it was implemented in 1999 (–40 per cent), bringing the real effective
exchange rate back to its 1995 level. This high degree of variability is obviously harmful
from the standpoint of attracting investors; perhaps it was the price to be paid for
exchange rate maturity.
173
The current account balance has followed the trade balance fairly closely. It has
been consistently negative since the reforms began, with the deficit reaching a peak in
1993 (approximately 13 per cent of GDP) and subsequently stabilising at 10 per cent
or less in the middle of the decade. Overall balance is maintained by net positive
capital inflows, mainly consisting of foreign aid.
— Foreign aid reached a peak in 1991, at the equivalent of 13 per cent of GDP, and
then stabilised in the first half of the 1990s at 11 per cent of GDP. It accounts for
70 to 80 per cent of total investment depending on the year. The face value of
the debt, which was modest initially for historical reasons, quadrupled during
the period of market reform (1983–97); as a percentage of GNP, it rose from
41 to 102 per cent in 1994 (88 per cent in 1997). The debt service ratio increased
to over 50 per cent in the late 1980s, then fell back to 25–30 per cent in the
1990s, thanks to several debt restructuring packages. Ghana was declared eligible
for the HIPC initiative at the Cologne summit9.
— Reform of foreign trade has been extensive and rapid. Many special taxes and
non–tariff barriers were eliminated in the second half of the 1980s. Tariffs were
somewhat unstable until 1990, but in 1993 the tariff structure was simplified
through the establishment of three rates (0, 25 and 30 per cent).
— Privatisation and the reform of domestic trade. The privatisation of a very
extensive quasi–public sector10 was another crucial feature of the reform effort.
A great deal of time was required, as the government wanted to manage the
process in transparent fashion, and also because the subject was politically
sensitive, even within the ruling party. From 1991 to 1996, however, of the
183 firms slated for privatisation (Fosu, 1999), 91 were sold, partial equity stakes
were sold in 27, 18 were transformed into joint ventures, 5 were placed under
private management and 42 were liquidated. Examples include the privatisation
of Ghana Telecoms and the sale of a 41 per cent stake in Ghana Commercial
Bank (with a promise to sell a further 40 per cent later on), while the country’s
largest mining company, Ashanti Goldfields, was floated on the national stock
exchange in an operation that transferred a huge proportion of its capitalisation
(90 per cent).
Most price controls were eliminated in 1983. The agricultural and mining sectors
recovered. With the 1987/88 harvest, the producer price for cocoa was tripled just at
the moment when the international cocoa price fell by half, and this forced the Cocoa
Marketing Board to cut the producer price again, though only moderately. Production
increased through 1989 and 1990 (while illicit cross–border trade fell), and exports
subsequently stabilised around this higher level. The share of cocoa in exports dropped
from 60 to 25 per cent owing to the relative buoyancy of other sectors, particularly
mining, where institutional reform (privatisation and realistic regulation) was a success,
leading to rapid growth of output. The revenues from gold exports have exceeded
those from cocoa since 1992. Liberalisation and further promotion of the cocoa industry
were approved in 1999, in a bid to double the amount produced.
174
— Reform of the financial system, which was in a deplorable state in the early
1980s11, was a very important feature of the adjustment. The reform was applied
in three stages: liberalisation of interest rates (1987–88), restructuring of banks’
balance sheets and the introduction of prudential regulations (1989–91)12, and
lastly structural change via privatisation and promotion of the private sector
(from 1992). The reforms have not been completed; the supply of medium– and
long–term credit remains small, since banks still regard this as a high–risk area.
In addition, the M2/GDP ratio stands between 15 to 20 per cent (of which M1
accounts for 10 to 15 percentage points), as against 25 per cent on average in
sub–Saharan Africa as a whole. This reflects the reluctance of economic agents
to hold money for any purpose other than that of the transaction, owing to the
monetary erosion they have experienced in the past and the still–significant level
of recent inflation13.
— The infrastructure sector (road, rail, telephone), which was in a state of near–
total neglect, has been substantially rehabilitated since the reforms began14.
In the aftermath of this highly volatile economic and political history, output per
capita in the second half of the 1990s was barely higher than at independence, and
below the level which prevailed in the early 1970s. Overall, Ghana’s long–term growth
pattern is easily explained by the fact that capital intensity fell during the crisis years
and has picked up slowly since the beginning of the reform period, while the collapse
in factor productivity lasted until 1983. Factor productivity has recovered significantly
since then, but remains below its level in the early 1960s. The quality of economic
policy is now fairly good. According to the composite index constructed by
Guillaumont et al. (1999)15, Ghana belongs to the category of emerging African
countries: this index returned to a clearly positive value in 1990 and has continued to
improve. Economic governance is not sufficiently stable, however, and this maintains
the wait–and–see attitude of observers, who tend to associate such instability with the
bad habits of the past. They also continue to be concerned about the possibility of
renewed political instability, or even a return to military rule following Rawlings’
withdrawal from politics16.
Further progress is still needed on a number of fronts: tight budget policy,
controlling inflation and the real exchange rate, modernising the financial sector and,
correlatively, domestic saving and investment.
Change in the structure of output is a vital long–term goal. In this respect, an
encouraging sign — though a somewhat ambiguous one — can be seen in the steady
rise of the share of manufactured goods in total exports (18 per cent for the 1990–94
period), although the share of manufacturing value added in GDP fell to 8 per cent in
1996 owing to a slowdown of growth across the whole manufacturing sector. This
175
may be interpreted as reflecting increased export competitiveness and the emergence
of the sub–sector in which the country has a comparative advantage, while a de–
industrialisation process is at work in formerly protected sub–sectors.
If Ghana continues to improve its economic policy, in order to restore a significant
level of confidence in the country, and if the political environment remains democratic,
the Ghanaian government can hope that the economy will continue to advance on a
regular growth path that could eventually lead it to economic take off.
176
Notes
by Leith and Söderling (2000) for the OECD’s Emerging Africa project.
3. Using index numbers (level at independence = 100), the index of real per capita GDP
reached 130.
4. The currency, or rather the nominal exchange rate, was regarded as a matter of national
pride.
5. The collapse was accentuated by negative exogenous shocks in 1982, such as droughts
and fires in the plantations.
6. Fosu (1999) shows, using his own calculations, that since 1983 the share of spending
devoted to education has been roughly 20–25 per cent of the total.
7. VAT was introduced initially in 1995 at a rate of 17.5 per cent, then withdrawn two
months later as a consequence of riots; it was then re–introduced successfully in 1998, at
the more modest rate of 10 per cent.
8. An export bonus was established, to be financed by a surcharge on imports.
9. Reducing the ratio of the NPV of debt to exports from 200–250 per cent to 150 per cent;
this ratio stood at 175 per cent in 1999 in Ghana.
10. In 1987, public enterprises accounted for over half of value added and manufacturing
employment. They were also deeply involved in agricultural sectors and services,
particularly banking.
11. In addition to the financial repression linked to state intervention in the setting of interest
rates, banks (which belonged to the quasi–public sector) were long forced to base their
lending decisions on political considerations rather than financial assessment of projects.
Financial services did not generate value added; rather, they destroyed it by lending
systematically to defaulting customers.
12. In 1989, bad loans still amounted to 41 per cent of the stock of loans. Banks were therefore
supposed to restrict their loans to the new sectors (tradable goods) being promoted as
part of the reform. A loan from the IDA will be used to constitute a bank restructuring
fund.
177
13. The inflation rate hit a peak of 122 per cent in 1983 and has remained above 20 per cent
in each year since then, except for 1985, 1991 and 1992.
14. Unfortunately, the available data do not suffice for a quantified assessment.
foreign conflicts.
16. Rawlings was able to use his influence with the army to convince the latter to withdraw
from management of public affairs (Fosu, 1999). It is to be hoped that his successors
will be able to do the same.
178
Chapter 10
Mali 1
Mali is a very poor, landlocked country with a population of 9.3 million, covering
over a million square kilometres. Its population density is thus very low (7 people/sq.
km), but the population is very unevenly distributed. It is growing at an average annual
pace of 2.3 per cent (which has risen to 2.8 per cent in the period since 1990), as a
consequence of a high fertility rate (nearly 7 children per woman), whose effects are
partly offset by a high rate of net emigration (towards France but also within the sub–
region). The urban population is only 28 per cent of the total population, but it is
increasing very quickly. Per capita GNP was evaluated at $250 in 1997, and Mali
belongs to both the group of low–income countries and that of the least advanced
countries. The very low level of human development (Mali is in 171st place in the
UNDP’s HDI ranking) reflects a high poverty rate (71 per cent of the population in
1996) as well as widespread malnutrition (affecting half of all children under five
years of age). Only half of the population has access to drinking water. Life expectancy
was only 47 years in 1995.
Per capita GDP growth over the 1960–92 period was modest and unstable, but
nonetheless positive. The economy went through four major phases:
i) in the 1960–68 period it stagnated under the Marxist regime of Keita;
ii) in the 1968–74 period a few reforms inspired by the African socialism model
(including the elimination of collective fields) were introduced in a bid to establish
a mixed economy; the results were again disappointing, as per capita output
continued to stagnate;
iii) the latter increased strongly in the 1975–79 period (+5.8 per cent a year) as a
result of excellent conditions on the cotton market, but cotton prices turned down
from 1980 to 1982, wiping out the previous income gains (–6.5 per cent a year);
179
iv) the fourth phase began in 1982 with a structural adjustment programme that
eliminated the country’s socialist policies and resulted in a slightly rising long–
term trend in per capita GDP that lasted until 1997 (average annual growth of
+0.8 per cent), with a surge of faster growth resulting from the devaluation of
1994 (per capita GDP increased by 3.6 per cent in 1995, 0.9 per cent in 1996
and 3.5 per cent in 1997).
The Malian economy is basically agrarian, with agriculture and livestock raising
accounting for approximately 48 per cent of GDP. Although only 2 per cent of the
land is arable (and 25 per cent is used for grazing), it is still possible to extend agriculture
because there is considerable potential in the form of irrigable land (1 million hectares),
of which only 5 per cent has been developed to date. Exports consist mainly of cotton
(nearly 50 per cent of total exports), livestock (15 per cent) and gold (25 per cent).
The economy is thus highly vulnerable to random exogenous shocks (rainfall) and
external shocks (terms of trade2) which render short–term growth very unstable and
erratic: annual growth rates were negative for half of the years from 1968 to 1997.
It should also be noted that faster growth in Mali regularly brings a slowdown in
the rate of increase of poverty (both in extent and in depth), but no reduction of poverty.
Trickle–down effects are thus virtually null.
Lastly, according to the composite index constructed by Guillaumont et al.
(1999)3, Mali belongs to the category of emerging African countries, as this index
remained steadily positive throughout the study period (1970–95).
Political History and Development Strategy
Since 1960, Mali has seen five political regimes, which to some extent coincide
with the various phases of the economy’s long–term growth, and which were marked
by diametrically opposed choices of development strategy.
The Marxist government of Keita4, which took power when Mali was created in
5
1960 , gave way in the coup of November 1968 to the socialist regime of the Comité
Militaire de Libération Nationale, headed by Traoré. This regime was notable for a
major increase in corruption and clientelist behaviour, which fuelled growing social
discontent: five attempts were made to topple Traoré by coup d’état, and he had to
cope with five major strikes. A system of high tariff protection was instituted as part of
an import–substitution strategy, while education, health and infrastructure were
neglected. The state of public finances gradually worsened, leading as early as 1982
to the adoption of stabilisation and structural adjustment programmes which included
rejoining the WAMU in 1984. In the early 1990s, nearly 25 years after his rise to
power, popular distrust of Traoré came to a head and violent rioting broke out in
Bamako. Traoré was overthrown in March 1991 and arrested a year later. He was
condemned to death, but was pardoned a few years later by his successor, Konaré.
The latter, democratically elected in 1992, pursued and extended the reform effort.
180
Apart from this, beginning in 1990, the country experienced serious civil unrest, with
government forces confronting northern rebellions provoked by discriminatory
allocation of public expenditures. The conflict came to an end in 1995 with the
conclusion of agreements that should be laid to the credit of the democratic government.
Although Mali has had only three presidents in its 40 years of existence, it has
undergone many political crises and has lived under democratic rule only since 1992.
Following the typology developed by Chege (1999), these successive leaders may be
classified as follows: Keita and Traoré from 1968 to 1982 can be described as
ideological and radical leaders, while Konaré’s leadership may be regarded as
conservative (in that his economic policy was in line with the Washington consensus)
and relatively effective. Traoré’s leadership during the reform period (1982–91) was,
from a strictly economic standpoint, somewhat similar to that of Konaré6, but this did
not prevent Traoré from getting bogged down nor his subsequent political collapse.
The latter was due in particular to the fact that the country’s wealth had been to some
extent captured by the minority in power, a phenomenon that originated in the
bureaucratic mode of the Marxist and socialist periods. Today, governance is still
compromised by these acquisitive reflexes. Attempts to combat corruption have lacked
credibility or have been perceived as lacking continuity, in contrast to the case of
Burkina Faso.
macroeconomic policy, trade policy and lastly the business environment as viewed
through the domestic regulatory apparatus, bank restructuring and privatisation.
— The investment ratio over the 1967–97 period followed a clearly rising trend,
punctuated with a few marked dips during periods of recession: it rose from
under 15 per cent in the 1970s to 25 per cent in the mid–1990s, with a long–
term average of 18.5 per cent. The increase in investment spending was largely
financed through access to official foreign funding (see below). The productivity
of investment fell until 1992 (the ICOR rose in irregular fashion from 2 to 12)
and improved very quickly thereafter (the ICOR had fallen back to 6 in 1994).
From 1982 on, the public investment ratio amounted to 10 per cent of GDP and
held relatively steady, despite the instability of the short–term growth rate. The
overall saving rate, which had been zero or negative during the 1980s, rose to
nearly 10 per cent of GDP in the second half of the 1990s. This increase was
partly due to financial development, which began only recently in Mali, but also
to the contribution made by public savings, which have turned positive again
since 1995 (5 per cent of GDP in 1996) after a sharp decline from 1990 to 1994.
181
The overall saving rate is still too low to meet the demand for capital, however,
and also in comparison with the average for the low–income countries (18 per
cent) and for sub–Saharan Africa as a whole (12 per cent).
— The level of human capital is very low: the adult literacy rate in 1996 was
estimated at only 21 per cent (compared to 57 per cent for sub–Saharan Africa
as a whole). The enrolment rate in primary education was 34 per cent in 1995,
up ten percentage points over the previous decade, but less than half of the
average for sub–Saharan Africa (75 per cent), and even further from the average
for the low–income countries (106 per cent). The secondary school enrolment
rates for Mali, sub–Saharan Africa and the low–income countries in 1995 were
9, 27 and 56 per cent respectively. Mali is thus clearly lagging behind, especially
since school enrolments are concentrated in Bamako, where the rates are four
times the national averages. Health performance is poor. The shares of health
and education spending in total spending, which were already much too low at
0.8 per cent and 2.2 per cent of GDP respectively, nevertheless fell further in the
1990s.
— Over the long term, macroeconomic management may be described as prudent;
among other things, it featured tight control over inflation (which is mainly
imported) and substantial, regular depreciation of the real effective exchange
rate (by 61 per cent from 1967 to 1996). Trade balance deficits were structurally
significant (amounting to 6 per cent of GDP on average, with the exception of
1997, an exceptionally good year owing to the buoyancy of cotton and livestock
exports resulting from the devaluation) and were made worse by the deficits in
the services–revenue balance, but were offset by substantial unrequited transfers,
both public (ODA) and private (remittances from emigrants, amounting to 2 to
5 per cent of GDP depending on the year). The current account still remained
constantly in deficit, but more often than not this deficit was offset by large
foreign public contributions on favourable terms; moreover, the current account
balance has improved regularly since the mid–1980s (from –16 per cent of GDP
in 1985 to –2 per cent in 1997). Mali has also managed to improve its financial
solvency since 1985, by gradually moderating its net borrowing requirement
(tax revenue plus donations less public spending) from 8 per cent of GDP to
2 per cent in 1997. The country’s budget policy, in contrast to that of many
economies specialised in primary products, seems to have been reasonable and
not pro–cyclical (public expenditures have stabilised), while the rate of public
saving (current revenue less current expenditure) has been constantly positive
except for the brief 1978–80 period.
— Foreign aid is crucial to Mali’s economy: net public contributions have trended
upwards, rising to 25 per cent of GDP or more during certain periods (the droughts
of 1974 and 1985, the 1994 devaluation), and have not fallen below 15 per cent
since 1980. A large proportion of aid is in the form of donations (ranging from
about 5 per cent to 9 per cent of GDP since 1990, or, on average, over half of the
tax take), and the intensity of the aid received corresponds in part to cycles of
182
structural reform. Despite various rescheduling agreements, Mali’s level of
indebtedness remains one of the highest in Africa in relative terms (116 per cent
of GNP in 1996, which guarantees the country’s eligibility for the HIPC initiative).
However, almost all of the debt is concessional (approximately 97 per cent in
1997) and is owed entirely to official creditors. What is most important, the
level of debt is sustainable: interest payments on the debt reached a peak of
about 2.3 per cent of GDP in 1994 (as against 6.1 per cent on average for the
countries of sub–Saharan Africa) and fell to 0.9 per cent in 1997. The debt service/
exports ratio was under 20 per cent in 1996. Like some other Sahel countries,
Mali is thus in a position, with the help of foreign aid, to finance major public
expenditures (23 per cent of GDP at the end of the period) without raising the
tax ratio to levels that discourage activity (since 1960, taxes have varied around
an average value of 15 per cent of GDP, a level close to the average for the
African low–income countries and consistent with the structural characteristics
of the country, namely per capita GNP, the degree of monetisation and the ratio
of mining exports to output). The greater part of public investment — which
accounts for over half of total public spending, or twice the share observed
elsewhere in sub–Saharan Africa — is financed by ODA. The level of aid should
be maintained in the future owing to democratisation (Azam and Morrisson,
1999) and to fairly satisfactory economic results.
— A few trade reform measures were taken in the early 1980s: for example, export
taxes, which accounted for 10 per cent of customs receipts in 1978, were gradually
eliminated; in particular, the tax on cotton staple was removed in 1981. However,
the level of tariff protection was still very high in the second half of the 1980s
(the rate of taxation on imports that competed with local products was of the
order of 100 per cent in 1988). True reform was supposed to be undertaken
beginning in 1986, with the elimination of administrative prices and import
monopolies through the liquidation of SOMIEX, as well as simplification of the
tariff structure; in fact, it got under way somewhat later, and picked up steam
under Konaré, especially in connection with the WAEMU agenda (the customs
union scheduled for January 2000). In the late 1990s, the weighted tariff rate
was 14 per cent, which is close to the average value for the WAEMU but lower
than that of Mali’s neighbours in the Sahel, such as Burkina Faso, which still
had a high level of protection at the end of the period. VAT was introduced
beginning in 1991 to make tax resources less dependent on taxation of foreign
trade7, while the system of flat–rate direct taxes formerly applying to the entire
population was eliminated.
— Domestic trade has been liberalised since 1992: prices and competition are free
in most sectors; the exceptions are water, electricity, fuels and
telecommunications, which remain under public monopoly. In addition, public
management of agricultural sectors, which was initially highly regulated and the
source of many inefficiencies (inadequate producer prices, mediocre yields etc.),
has changed a great deal since the mid–1980s as a result of a sectoral adjustment
programme for agriculture. The state withdrew from the rice sector, though it
183
has retained a few supporting roles; both yields and production quadrupled over
the 1984–97 period. The cotton sector, which is the mainstay of Mali’s economy,
was successfully restructured in 1989 under a planning agreement between the
state and the CDMT (Compagnie malienne de développement des textiles). The
restructuring was based on three principles: making the sector autonomous
through co–management by producers, making the CDMT accountable in its
role as principal trader and setting up an internal stabilisation/regulation
mechanism8. In this way, the producer price was stabilised at a time when the
international cotton price remained highly volatile. The results have been
satisfactory: production has clearly picked up, largely as a result of the extension
of the land area under cultivation. Cotton income is nevertheless taxed, under
ordinary tax law applying to the CMDT and through the introduction of a
“cyclical” export tax applied when the sale price exceeds a given threshold. In
view of the success observed, liberalisation of the sector is being widely debated.
— The process of state withdrawal from productive activities began in 1987–88,
with among other things the implementation of PASEP (Programme d’ajustement
sectoriel des entreprises publiques), a sectoral adjustment programme for public
enterprises. The public enterprises formed as part of the Keita regime’s import–
substitution strategy were active in all manufacturing sub–sectors, and the
economy was highly regulated. In 1980, for example, 57 non–financial public
enterprises accounted for 70 per cent of industrial output. Under PASEP, nearly
50 companies were privatised or liquidated within ten years. The state has not
withdrawn entirely from the economy, however, owing to its determination to
keep a strategic sector (electricity, water, telephone). There are still some 20
public enterprises, but they have been restructured under the new system of
“performance contracts”: their aggregate net earnings moved back into the black
in 1994, allowing them to record a surplus of savings over investment. Lastly, a
new legal environment for companies (business law with the OHADA, labour
code, insurance code) has been created within the WAEMU framework.
— Restructuring of the banking system — most Malian banks had long been public
corporations, then semi–public companies — began in 1988. The traditional lending
institutions (of which there are seven) are characterised by inadequate networks of
branches and by services that are poorly suited to the needs of the economy. Loans
had fallen off sharply in the first half of the 1990s as a result of the now–completed
restructuring process, and today two–thirds of them are short–term loans (financing
of commercial activities and seasonal loans), while long–term loans mainly serve
to finance investment in the energy sector and the agri–food industry. SMEs–
SMIs and micro–enterprises are thus largely excluded from the credit market,
though fortunately they do have access to a fairly dynamic, decentralised system
of savings and loan co–operatives. The financial depth of the economy increased
in the second half of the 1980s (M3/GDP ratio of 20 per cent or more) and has
held steady since then, while the proportion of financial instruments properly
speaking (M3–M1) increased to around 5 per cent of GDP.
184
— Mali’s landlocked state and its extremely poor supply of transport infrastructure
(road and rail) raise production costs considerably and are holding back Mali’s
growth, even though factor costs are in line with the average of those in
neighbouring countries. Mali thus suffers from two forms of isolation: it lacks
access to the sea, and has isolated areas in its interior 9. Furthermore, the
distribution of electricity is inadequate (only 8 per cent of the population has
direct access to electricity).
Although the growth of per capita output has been only moderate (+0.7 per cent
for the 1965–96 period), it contrasts with the situation of regression observed in many
countries of sub–Saharan Africa, and places Mali in the upper quartile of the countries
in the region. It should also be evaluated in the light of the structural handicaps facing
the country (lack of infrastructure and human capital). In this sense, Mali’s growth
performance has been adequate. Economic policy has greatly improved over time, as
shown by the fact that, even though the economy is specialised in primary commodities,
there has been no “Dutch disease” crisis, for reasons that are both domestic (no over–
valuation of the currency) and external (no large, sustained drop in the terms of trade).
The reforms have certainly favoured long–term growth. Indeed, it is possible to
connect the fairly long growth period observed since 1983 to the gradual implementation
(in successive levels) of the reforms. For example, the relative decline of the services
sector since 1982, which picked up steam in 1994, may be interpreted as reflecting the
decline of non–tradable sectors following the drop in the real effective exchange rate.
According to Chambas et al. (2000), however, “it is impossible to identify
econometrically the impact of the liberalisation measures on Mali’s growth”. Moreover,
the relative growth in tradable sectors is due more to the growth of the primary sector:
whereas agricultural growth has been noteworthy, the industrial sector has grown only
very erratically since 1983. From 1985 to 1996, the proportion of GDP generated by
the primary sector in fact grew more (about +8 percentage points, from 40 to 48 per
cent) than that of the secondary sector (+2 points, from 15 to 17 per cent). Thus,
instead of the long–term diversification needed, what we are seeing is the reinforcement
of the Malian economy’s specialisation in primary products.
For this reason, the economic recovery that is evident in the second half of the
1990s appears to be very delicate, and a rise in the long–term growth rate is far from
certain. A stronger and longer–lasting shock to the price of cotton would have a
disastrous impact on Mali’s economy.
The latest stage in the reform process relates to institutional modernisation, with
the establishment of clear legal rules — progress is being made on this point in
conjunction with the WAEMU — and in particular more reliable application of those
185
rules. The fact is that the application of the law is still a matter of uncertainty in Mali,
a situation which stifles private initiative, particularly FDI. Reform of the judicial
system is needed, and governance must continue to improve.
More fundamentally, the Malian economy will have to demonstrate the capacity
to encourage real sectoral diversification towards non–primary tradable activities. To
this end, Mali will have to mobilise all the internal and external resources available to
finance needed investment in infrastructure and human capital. Development of human
capital is urgently needed, particularly in rural areas, in order to bring about the
demographic transition, without which it seems that the poverty trap will tighten
inexorably. In fact, a far–reaching ten–year development programme for education
has recently been formulated, providing for the doubling of the literacy rate and the
primary school enrolment rate. In addition, the quality of wage and trade policy will
be crucial if the country is to exploit its comparative advantage.
186
Notes
by Chambas et al. (2000) for the OECD’s Emerging Africa project.
2. The terms of trade were however much less variable in Mali than elsewhere (particularly
Côte d’Ivoire): year–to–year variation, expressed in index numbers (1970 = 100),
remained within a band ranging from 80 to 120 over the entire 1960–97 period; moreover,
since 1960 there has been no obvious downward trend in the terms of trade, according to
UNCTAD data.
foreign conflicts.
4. Keita withdrew Mali from the Franc Zone in 1962, created many public enterprises and
at the same time nationalised many private ones. Staffing levels in these enterprises
were excessive. The semi–public company SOMIEX was given the monopoly on foreign
trade. Faced with unsustainable payment imbalances, Keita was obliged to rejoin the
Franc Zone in 1967.
5. Following the break–up of the federation with Senegal.
6. In terms of growth performance and, to a lesser extent, his economic policy of initiating
reform.
7. In vain. The share of trade taxes in total tax resources has been over 45 per cent since
1993, as compared to 35 per cent over the long term, a result of the narrowed range of
tariff exemptions and the replacement of quotas by tariffs in certain cases.
8. The system is based on a floor producer price and on the definition of so–called stabilised
costs (calculated on the basis of an implicit scale and of targets for productivity gains).
The difference between a producer’s receipts and stabilised costs is either transferred to
(if it is positive) or is compensated by (if it is negative) the stabilisation fund, which can
be replenished through a Stabex mechanism. When the fund exceeds a given ceiling, the
additional margins are distributed between the CMDT (65 per cent) and producers (35 per
cent).
9. A glance at Mali’s geography shows that Bamako is at the extreme edge of the country’s
territory, and is particularly far removed from the Timbuktu area in the north.
187
188
Chapter 11
Uganda 1
Uganda is a very poor, landlocked country with 21 million inhabitants, 14 per

cent of whom live in urban areas. The population is growing at an average annual rate
of 2.8 per cent (1995–2000) owing to a high fertility rate (seven children per woman)2.
GNP per capita was evaluated at $330 in 1997 (ADB, 1999), a level which in real
terms is lower than that of 1972 and even that of 1960. This reduction is due to a
prolonged political and economic crisis. The country’s level of human development is
very low compared to its income level: according to UNDP data (Human Development
Report, 1998), Uganda came in 160th in the HDI ranking, or 23 places lower than in
the GDP per capita ranking (at purchasing power parity). Much of the population of
course lives in poverty (45 per cent in 1996), particularly in rural areas (where four–
fifths of the poor live), but poverty has fallen by ten percentage points since 1992 and
its impact is somewhat limited by a fairly acceptable income distribution (the Gini
coefficient stayed under 0.38 throughout the 1993–96 period). Life expectancy was
only 43 years and the literacy rate 62 per cent in 1997 (ADB, 1999).
At independence the first Obote regime, drawing inspiration from the import–
substitution model, followed a mixed–economy strategy with a highly interventionist
spirit that was frustrated by a decentralised rural economy. In 1966, Uganda left the
EAC East African Currency Board (the other members were Kenya and Tanzania)
in order to found its own central bank and be able to print money at will, most
particularly with a view to financing the fiscal deficit. During the first decade
following independence, substantial efforts were made in the areas of education,
health and infrastructure, and these outlays were supported by a satisfactory rate of
economic growth.
189
This progress was to be wiped out in the following decade, however, as Amin
took power in early 1971 and implemented policies that were to have a heavy impact
on the economy and the well–being of the population for decades thereafter. He proved
to be a leader of the tyrannical type (Chege, 1999). In particular, he expelled the
Ugandan–Asian trading families in 1972, in a move that cost the country a large portion
of its human capital and earned it a lasting reputation as an anarchic country where
confiscation of property was an everyday occurrence. The drop–off in private–sector
activity (investment and exports), in particular the collapse of the formal sector, went
hand in hand with the expansion of the public sector, which was managed by cronies
of the regime who were completely devoid of the managerial skills required. Faced
with gradually deteriorating macroeconomic imbalances, the government imposed
stronger controls on prices, payments and the distribution of goods. Rebel troops backed
by Tanzania finally overthrew Amin in 1979, by which time his domestic popularity
had mostly disappeared.
Three governments followed in rapid succession in 1979 and 1980, in a general
climate of extreme insecurity. Obote returned to power in 1981 and sought support
from the Bretton Woods institutions, his declared economic objective being to restore
the efficiency of the productive sectors. Several adjustment measures were announced
in this respect (letting the shilling float, increased producer prices for cash crops,
elimination of price controls, rationalisation of the tax system, reining in public
spending), but they were imperfectly implemented, and then were stopped altogether
in 1984. Faced with intense guerrilla warfare, the government increased its military
expenditures and lost control over public expenditures, which had been hugely increased
in anticipation of the upcoming elections. Civil service wages quadrupled in the year
1984 alone, and the money supply rose by 127 per cent. The economy went into a
series of crises and Obote’s second regime fell in 1985.
In January 1986 the National Resistance Movement (NRM) came to power. The
NRM’s economic policy was initially marked by indecision. Once again, an
unsustainable macroeconomic policy was implemented, with an artificial revaluation
of the shilling which led to an acute shortage of foreign exchange and various forms
of rationing. The NRM thoroughly overhauled its economic policy as from 1987, with
a new donor–backed reform plan aimed first and foremost at stabilising the economy.
The latter was handicapped by adverse terms–of–trade shocks (the terms–of–trade
index — base value 100 in 1960 — fell from 122 in 1986 to 72 in 1987 and 47 in
1990), which the regime initially managed in a piecemeal and radical manner3. But
this did not prevent the economy from recording a solid recovery that has continued
very regularly ever since, with low inflation and consequently an improvement in per
capita income. This growth is more in the nature of a catching–up effect, however,
since the rapid rise in productivity was due to the reactivation of production capacity
that had been unused during the years of crisis and bad management.
Uganda’s growth dynamic is handicapped, however, by civilian and military
instability on several fronts.
190
First, the Ugandan army has for over ten years been forced to confront the ARD,
an armed wing of the northern resistance movement; the rebels have attacked schools
and hospitals, and destroyed the region’s infrastructure and rural economy. President
Museveni (who has led the country as head of the NRM ever since his party took
power in January 1986) has tried to provide a positive response to this persistent
problem by setting up a regional reconstruction programme for the north, with the
support of donors, and by promising amnesty for rebel leaders in exchange for an end
to the fighting.
Second, the Ugandan army’s intervention in the conflict in the Democratic
Republic of Congo has had still greater economic and political consequences. Although
the government justified taking action against the regime of L.D. Kabila on national
security grounds, this intervention has been perceived by economic agents as increasing
the political risk associated with Uganda (whose risk rating was already poor). The
intervention is increasingly contested by public opinion, and its costs (direct and
indirect, i.e. opportunity cost) are high: the defence budget accounts officially for
15 to 20 per cent of current expenditures, as against 5 per cent for the health sector,
and social projects have had to be cut as a result. Moreover, the border situation has
become increasingly confused with the proliferation of armed factions, originating in
Uganda and Rwanda, that are trying to grab part of the wealth of Congo and which
fought among themselves using heavy weapons in 1999.
Politically, there has thus been no real democratisation whatsoever, although
there has been progress on civil liberties and freedom of the press. Museveni was re–
elected in 1996, and the June 2000 referendum on multi–party government was
boycotted by the partisans of a fully multi–party system, who were not allowed to
speak out freely against the hybrid system proposed by the government (the
“Movement” system, which the government regards as a sort of grassroots democracy
without political parties ). This political lockdown is a source of concern to donors,
particularly the EU, but Uganda’s economic policy ensures that it is still regarded as a
good pupil by the World Bank.
— The investment ratio is still low in comparison to that of sub–Saharan Africa as
a whole: only 16 per cent in 1996. Private–sector investment accounted for
11.5 percentage points of this total, and a huge proportion (13 percentage points)
went to the construction sector. Investments were concentrated in the region of
191
Kampala. Capital productivity was high, with an average ICOR of 2.2 in the
1990s. The domestic saving rate has picked up: it stood at about 7 per cent of
GDP in the mid–1990s, after being negative throughout the 1980s. Domestic
consumption still exceeds the GDP, however, owing to current transfers from
abroad.
— The level of human capital is too low (although the literacy rate is reasonable
given the country’s income level4) and biased in favour of the capital city and
against girls. Only 12 per cent of boys had no education whatsoever on average
in 1995, as against 31 per cent for girls; in urban areas these ratios fall to 5 per
cent and 11 per cent respectively. The crisis of the 1970s and the increase in
military spending long dissuaded the state from attempting to extend education,
and the sector sank into deep crisis for more than two decades. The introduction
of universal primary education in 1997 led to a doubling of enrolments (now at
95 per cent), and implies fresh increases in the education budget, which has
already risen from 10–12 per cent of the central government’s current spending
to over 20 per cent since 1996. The country’s very poor performance in health
reflects the effects of the civil war, the crisis of previous decades and, more
recently, AIDS. In the mid–1990s, the government was spending only $4 per
head on health care (as against $30 in Botswana for example), and access to care
was disgraceful in rural areas.
— Over the long term, macroeconomic management has greatly improved. Sound
budgetary and monetary policies have been adopted and inflation is under control.
The financial solvency of the state is assured. The overall tax ratio rose from
7 per cent of GDP in 1991/92 to approximately 12 per cent in the second half of
the 1990s, but it remains too low owing to a lack of administrative capacity and
difficulties in extending the tax base. As a result, marginal tax rates on formal–
sector firms are at levels considered too high by the managers of such firms.
Thirty–five per cent of fiscal revenue was derived from tariffs, particularly on
energy products (petrol in particular5), which raises local transport costs and
undermines the competitiveness of this landlocked country. VAT at a rate of
17 per cent was introduced in 1996 and accounted for 43 per cent of tax revenue
for the fiscal year 1996/97, but this figure fell to 31 per cent for 1997/98 as a
consequence of poor collection methods and the effects of corruption. The
government has managed, however, to balance current spending and receipts
since the mid–1990s. Several partial reforms of the exchange rate regime had
been attempted by successive governments, but without any lasting success; the
black market in foreign exchange persisted for many years (premiums were
sometimes enormous: in 1986–87, for example, the spread was several
hundred per cent ). Genuine change began with the controlled float of the currency
in 1993, combined with the formation of an inter–bank foreign exchange market.
Capital movements were fully liberalised in 1997. The real effective exchange
rate therefore converged towards its equilibrium level at the end of the period,
whereas the currency had been greatly overvalued in the 1970s (by over 60 per
cent) and to a lesser extent through 1994 (approximately 20 per cent).
192
— Foreign aid has been substantial since peace was restored and lasting reforms
introduced. Since 1990, aid flows have been greater than or equal to 15 per cent
of GNP. Aid earmarked for the multilateral debt–reduction fund, as well as debt
forgiveness and restructuring (including the implementation of the HIPC
initiative: Uganda was the first country to benefit from the initiative, which has
already erased 20 per cent of its stock of debt ), reduced the debt burden from
30 per cent of current government expenditures in 1992/93 to 16 per cent in
1997/98. Debt as a proportion of GNP fell from 105 per cent in 1992 to 60 per
cent in 1996, and the debt service ratio has been stabilised at 20 per cent of
exports (73 per cent in 1991). The sums released are to be used for the education
and social sectors as a matter of priority.
— Trade reform and liberalisation of domestic trade made rapid progress in the
1990s. Early in the decade, the elimination of the marketing boards’ monopoly
on coffee6, tea and cotton led to a rise in producer prices and a surge in production.
Uganda regained its position as Africa’s leading coffee producer. Quantitative
restrictions were eliminated, except for tobacco. The number of tariffs was
reduced, and the highest tariff rate fell to 30 per cent in 1996. In 1998, there
were only three basic tariff rates (0, 7 and 15 per cent, excluding fuels), making
Uganda one of the most open economies of the Common Market for Eastern and
Southern Africa (COMESA). Tariff revenues increased as a percentage of imports,
however (from 13 per cent in 1990 to 25 per cent in 1996), owing to the
elimination of discretionary exemptions and better collection methods.
Liberalisation of foreign trade did not lead to an export boom for several reasons,
the two most important being excessive domestic transport costs and the very
high domestic price of petrol, which obviously adds to transport costs. The
government has introduced a system for reimbursement of the petrol tax that
seems to work mainly in favour of large firms (which have the administrative
capacity needed to work the system). For the future, the government is confronted
with the following dilemma: it can reduce the petrol tax to boost export
competitiveness, in which case it would forgo a major source of public revenue;
or it can maintain the petrol tax and forgo export competitiveness.
— The privatisation of the quasi–public sector, which has become very extensive,
began with the Act of 1993. At year–end 1997, 78 public enterprises had been
privatised, 20–odd had been liquidated and 12 remained the property of the
state. The process (including the uses of privatisation proceeds) was strongly
criticised for its lack of transparency (the interaction between the various political
forces is intense) and several scandals broke out in late 1998. Privatisation resulted
in considerable savings for the state, which no longer had to subsidise enterprises
with a soft budget constraint in order to keep them in business. A fair number of
the privatised companies were able to rationalise and increase their output.
— Restructuring of the banking system began in 1992 with a sectoral reform plan
initiated by the World Bank. The years of crisis and economic decline had induced
economic agents to shun banks and conduct their transactions in cash or by
193
barter. Financial institutions, including the UCB (the Uganda Commercial Bank
is owned by the state and carries out the government’s transactions ), were all in
a situation of prolonged insolvency throughout the 1980s. The reform had three
objectives: an efficient banking sector, financial deepening and better satisfaction
of customer needs. However, the results thus far have been mediocre: real interest
rates for borrowers rose to excessive levels at the end of the period (over 10 per
cent), while the lending rate remained low (4 per cent in nominal terms). Many
small banks with links to specific business communities entered the market;
they are poorly managed due to their inexperience and/or the patronage of group
managers and politicians (who are often businessmen as well). They are in serious
risk of default, the credit market functions poorly and this jeopardises the progress
achieved in other sectors of the economy. The regulatory framework and the
transparency of the system will have to be strengthened considerably.
— The supply of infrastructure is insufficient given the recovery of growth in the
1990s, despite substantial efforts in recent years. According to a business survey
conducted in 1994, the main obstacles in this area were, in descending order,
power outages, unstable voltage, telecommunications and the road system.
Per capita GDP grew at an average rate of 3.3 per cent during the 1990s. This
growth was obtained through the (partial) restoration of peace and the implementation
of reforms. Economic policy improved greatly, and in fact, according to the composite
index constructed by Guillaumont et al. (1999)7, Uganda belongs to the category of
emerging African countries: after fluctuating erratically and then collapsing in the
1980s (–2.5 in 1985–89), this indicator recovered quickly in the 1990s, although it
was still negative (–0.60 in 1993–95).
In the long run, however, growth will depend on the mobilisation of production
factors and on their productivity, and from this point of view the progress made by
Uganda, though undeniable, is not yet sufficient to guarantee self–sustaining growth
in the future.
It is vital that the investment ratio be increased, especially in tradables sectors,
but this will be difficult as long as the perceived risk of investing in Uganda remains
high (owing to domestic and foreign political conflicts)8.
Similarly, progress is urgently needed in the financial sector, and very large
investments are needed in physical infrastructure. At the same time, investment in
human capital must continue to rise. This will make it possible to quicken the process
of demographic transition, which has been too long delayed. The need for financing
will thus remain very large, and foreign aid will remain indispensable.
194
Furthermore, the country lacks political maturity. It still suffers from widespread
corruption and political interference in business matters, as well as frequent border
troubles. Much remains to be done to establish the public institutions and the incentives
needed for smooth market development (including a better judicial system and tax
reform) and to provide the desired type of governance.
The very first step to be taken is to develop a professional, well–managed civil
service. Administrative and parliamentary surveys were recently initiated to gauge
the intensity of corruption in government (including the military and privatisation
authorities). Sanctions either do not work or do not exist, although the situation in this
respect seems less endemic in Uganda than in some of its neighbours. The quest for
good governance also entails raising civil servants’ wages, which are so low (below
the minimum level of subsistence) that they discourage performance. In the health
sector, for example, employees’ lack of motivation leads to negligence, absenteeism
and even theft. The government has become aware of this priority: between 1993/94
and 1997/98, the public–sector wage bill rose from 21 per cent to 40 per cent of current
expenditures excluding debt, even though many employees were made redundant over
this period9.
Many programmes aimed at building institutional capacity are being conducted
in collaboration with donors, and these should be pursued and extended. If lasting
improvement is obtained in public–sector efficiency, it will once again be possible to
establish a dialogue and even partnership with the private sector, with a view to raising
the growth rate (instead of maximising various rents, as was previously the case).
Similarly, the decentralisation process, which will especially affect education and social
programmes, will have to be kept under full control so that it does not become counter–
productive; in particular, the success of the process will depend on local capacities for
taxation and efficient use of funds.
195
Notes
by Bigsten and Kayizzi–Mugerwa (1999) for the OECD’s Emerging Africa project.
3. The government went through phases when it resorted to printing money, with strong
inflationary impact (230 per cent in the first half of 1992); it subsequently decided to
take drastic measures such as limiting public expenditures strictly to monthly cash flow
(the “cash budget system”). In the last fiscal quarter of 1991/92, for example, expenditures
were cut by 70 per cent. This helped to restore donor confidence, and the cash budget
system was retained throughout the 1990s, though with some degree of flexibility.
4. As in Tanzania and Ghana, and in contrast to Burkina Faso, Côte d’Ivoire and Mali.
5. In 1997, the tax rates on petrol and diesel were still at 215 per cent and 160 per cent
respectively.
6. As well as the monopoly on rail transport of coffee to Mombasa.
foreign conflicts.
8. The appendix to the ADB’s African Development Report provides two contradictory
figures for Uganda’s investment rate in 1998: the 1999 report gave this rate as 25 per
cent, but the 2000 report adjusted this to 17 per cent.
9. Four thousand “drunkards and inefficient employees” were dismissed as early as 1992,
and 150 000 people through 1996.
196
Chapter 12
Tanzania 1
Tanzania is a very poor country with a population of 33 million, 32 per cent of

which lives in urban areas2 . The population grew at an annual rate of 2.7 per cent in
1997 as a result of a slight drop in the fertility rate, to 5.3 children per woman (ADB,
1999). Per capita GDP in 1997 was evaluated at $210 (ADB, 1999). The level of
human development is fairly high in comparison to the country’s income level:
according to UNDP data (Human Development Report, 1998), Tanzania ranked 150th
in the HDI classification, or 20 places higher than in the per capita GDP ranking (at
purchasing power parity). The poverty rate is high (50 per cent of the population),
mainly in rural areas (poverty rates in rural and urban areas were 58 per cent and
29 per cent respectively). Life expectancy was 52 years and the literacy rate 68 per
cent in 1997 (ADB, 1999). Over 80 per cent of the working population is employed in
agriculture, which accounts for approximately 50 per cent of GDP. In 1998, 64 per
cent of exports were primary products (of which agricultural products accounted for
42 percentage points and mining products, including gold, for 22 percentage points).
The economic experience of Tanzania may be divided into four periods: (1) the
pre–Arusha period, 1961–67; (2) the pre–crisis period, 1968–78; (3) the crisis period,
1979–85; and (4) the reform period, since 1986.
197
Pre–Arusha Period, 1961–67
Following independence in 1961, few major policy changes were made with
respect to the colonial period, despite the government’s desire to Africanise the public
sector and stimulate growth. Industrial development was pursued through a moderate
import–substitution policy, based on private initiative, and small farmers were
encouraged. The results were positive — macroeconomic stability and 2 per cent annual
average growth of per capita income — but did not satisfy the government of Nyerere,
the country’s president since 1962.
Pre–crisis Period, 1968–78
After the Arusha Declaration (1967), which constitutes a major turning point in
the country’s history, Nyerere did an about–face. To use the typology of Chege (1999),
Nyerere’s leadership then became clearly ideological and radical, as he sought to launch
Tanzania on the path of African socialism, based on efforts to achieve self–reliance.
The state took control of the economy, notably by creating a great many public
enterprises, and virtually every sector came under its grip. It placed controls on the
prices of a large number of goods, and proceeded to assemble the rural population
(initially on a voluntary basis, then by force) in Ujaama villages (the reason for this
was that the highly dispersed state of the population was considered incompatible
with efficient rural development; Oman and Wignaraja, 1991). Abundant external
financing3 allowed it to undertake many investments as part of an import–substitution
industrialisation strategy: the investment ratio rose to 30 per cent and more during the
second half of the 1970s. Capital productivity was lacklustre, as shown by the drop in
the growth rate of per capita GDP to 0.7 per cent a year over the period. Public
enterprises were poorly managed (running large deficits) and enjoyed a soft budget
constraint that eventually placed a heavy burden on public finances.
Crisis Period, 1979–85
The period of crisis began in 1979 with the swelling of the fiscal deficit. Nyerere
rejected an agreement with the IMF that provided for a 15 per cent devaluation; he
strengthened the controls on the economy and implemented his own reform programmes,
which proved to be rather tentative and failed to tackle crucial issues such as the over–
valuation of the currency (the exchange rate premium on the black market reached a
peak of 800 per cent in late 1985) and reform of the agricultural marketing system. Part
of the economy went underground, including a large proportion of exports4. It became
difficult to measure GDP accurately, but according to official data the per capita GDP
growth rate turned negative (–1.5 per cent a year over the period). Output fell owing to
insufficient imports of inputs. The capacity utilisation rate in the manufacturing sector
fell to under 30 per cent. The crisis reached such proportions that Nyerere resigned. The
new government agreed to introduce real market reforms, but did so grudgingly, as the
Arusha Declaration continued to be a strong influence on attitudes.
198
Reform Period, 1986–present
The reform period started in 1986 with a huge devaluation. The successive reform
programmes emphasised stabilisation of the economy and subsequently structural
reform (see below). Foreign aid flowed back into the country in massive amounts. In
1993–94, however, government relaxed its budget policy and turned once again to
borrowing from the central bank; it also seemed indifferent to creeping corruption and
tax avoidance. This attitude resulted in a small crisis with donors, which suspended
their disbursements of funds (from 1994 to 1998, disbursements were cut in half with
respect to commitments). The government which has been in power since November
1995, led by President Mkapa, restored relations with donors and stepped up the pace
of reform. Mkapa’s leadership seems to be fairly conservative and effective (Chege,
1999). Per capita GDP has risen by 0.6 per cent a year over the entire reform period.
Although the quality of public services is still mediocre, Tanzania has come a
long way since 1985. Few countries have undertaken such sweeping socio–economic
change in such a short time.
The investment ratio, which rose to over 30 per cent of GDP during the early
crisis years (1979 and 1980), subsequently underwent a rapid downward adjustment
to under 20 per cent, where it stayed until the early 1990s. A strong upturn was observed
from 1991 to 1994: the investment ratio climbed back to over 25 per cent, owing in
particular to a private investment boom (17 per cent of GDP), which unfortunately
was short–lived. Investments were mainly made in infrastructure, gold mines, trade,
finance and real estate. The level of manufacturing investment remained low, reflecting
both the lack of business opportunities and the wait–and–see attitude adopted by
investors owing to the insufficient credibility of the reforms. The domestic saving rate
continued to fluctuate around 0 from 1988 to 1995; this is understandable, as real
interest rates on savings deposits were constantly negative, from –20 to –30 per cent,
up to 1995. They have risen since then, but are still negative.
Investment in human capital was a declared priority of the vision inspired by the
Arusha Declaration. Before the crisis period, substantial donor–backed investment
brought excellent results in terms of literacy and women’s participation in the education
system. The crisis of the early 1980s and the withdrawal of donors brought about a
rapid decline in both the quantity and the quality of education and health services.
This decline continued during the reform period, owing to the strict limits on public
199
spending. The education budget amounted to only 4 per cent of GDP on average during
the 1990s. An increasing share of spending was allocated to primary education (60 per
cent) to the detriment of the secondary level (8 per cent; the enrolment rate in secondary
education was under 6 per cent, as against 17 per cent in Uganda); while higher education
retained a large share (20 per cent), particularly from the standpoint of spending per
head (a higher education student “costs” 100 times as much as a primary school pupil),
which is not justified from the standpoint of the general welfare5. In the end, the
primary school enrolment rate fell from 93 per cent in 1980 to 67 per cent in 1995.
Macroeconomic management improved over the long term, but it remained
somewhat erratic in some respects.
Substantial progress in balancing public finances was achieved only after two
attempts. The deficit was reined in from 9 to 2 per cent of GDP over the 1987–92
period, but fiscal profligacy returned in 1993–95, with a rise in current expenditures
and a deficit of 11 per cent of GDP. The introduction of a cash budget system in 1996
quickly restored order, and the deficit was held to 2.4 per cent of GDP in 1997. The
overall tax ratio has varied around 15 per cent since the beginning of the decade6; this
level is not high enough to meet the demand for expenditures (18 to 24 per cent of
GDP depending on the year since the early 1990s), a fair proportion of which are
needed for development. Attempts to broaden the tax base — including the introduction
of VAT in 1998 — and improve collection have met with little success to date, despite
the establishment of the Tanzanian Revenue Authority in 19967.
A good deal of progress has been made in exchange rate management, but some
problems have persisted. From 1986 to 1992, the exchange rate was adjusted on a
discretionary basis. The black market premium fell to 30 per cent in 1992 (compared
to 800 per cent in 1985). In 1994 an inter–bank foreign exchange market was created,
and the currency allowed to float virtually freely. It therefore collapsed in nominal
terms: in mid–1997, one US dollar was worth over 600 shillings (as against 51 in
1986 and 8 in 1981). The real effective exchange rate appreciated by 40 per cent from
1993 to 1997, however, owing to a considerable inflation differential, particularly in
the second half of the 1990s8, and to a temporary hike in the real interest rate9.
The trade balance remained deep in deficit over the entire period, despite the
very recent (1994–98) and highly unstable spurt of growth in exports of manufactured
products, which helped to reduce the agricultural share of total exports by nearly
20 percentage points. The current account deficit increased in the mid–1990s, reaching
25 per cent of GDP; it has been offset by massive external aid.
Foreign aid reached a peak in the early 1990s, when it amounted to nearly 30 per
cent of GDP and $50 per capita. The amount of aid in proportion to GDP subsequently
declined to under 20 per cent in the second half of the decade. At the same time, the
debt burden increased during the first half of the 1990s, hitting a peak of 179 per cent
of GNP in 1994. Beginning in 1995, however, it was managed differently. First of all,
Tanzania decided to resort to external borrowing only in case of necessity, and to
borrow exclusively on concessional terms, and in particular from the IDA. Next, the
200
government decided to give priority to servicing the debt, and interest payments made
up a growing share of total public expenditure (20 per cent in the late 1990s). Payment
arrears were reduced by over $500 million from 1996 to 1997, while total debt fell to
114 per cent and 97 per cent of GNP respectively in the same two years. Much more
substantial progress in debt reduction will probably be achieved through the HIPC
initiative, and this progress will be very helpful in stimulating Tanzania’s growth if
the debt relief allows the country devote more resources to the public goods required
for economic take off.
Trade reform — has been rapid and has affected many aspects of trade, not only
tariffs but also the many non–tariff barriers (such as the obligation to turn in foreign
exchange earned through exports, numerous quantitative restrictions, import licences).
This liberalisation was not co–ordinated with policy concerning the exchange rate,
which appreciated in real terms. As a result, many firms in the manufacturing sector,
which could not remain competitive on prices, lost all or part of their local markets
(textiles and clothing, shoes, leather, sugar and edible oils). Lastly, the tradable goods
sector has not managed to raise its share of national output (only about a third, excluding
subsistence agriculture and services).
Although it was recognised as early as 1985 that reform of public enterprises
was urgently needed to preserve the solvency of the state, the process was long delayed,
finally getting under way from 1992. A commission for the reform of the semi–public
enterprise sector has been operational since mid–1993. Progress on privatisation has
been slow, as the commission tackled the small enterprises first. By 1996, however,
private ownership was being considered even for strategic sectors such as the railway
system, telecommunications, ports and Air Tanzania. Several obstacles to privatisation
need to be addressed: the debt burden on enterprises, the future of their employees
and problems related to political feasibility10. Of the 383 companies slated for
privatisation, the commission had processed 230 cases in June 1998. It was supposed
to get to the others in the following years, but in the meantime these enterprises continue
to encumber the central government budget. Another important step will be the
introduction of the regulatory structure needed for the privatised sectors, many of
which operate on monopolistic markets. For the moment, no regulations have been
issued in this respect, so the respective ministers continue to serve as regulators.
The government has also implemented a policy of promoting private investment
based on the issuance of “incentive certificates” granting a number of advantages,
notably in tax matters. Investors, however, still regard Tanzania as a high–risk country
where reforms can quickly be reversed; it is true that the real effective exchange rate
and inflation have been much more volatile there than in other sub–Saharan African
countries.
The years of state control of the economy under Nyerere were characterised by
financial repression, coupled with extreme concentration of the financial system in
favour of the National Bank of Commerce, which accounted for 80 per cent of deposits.
Reform of the financial system was thus unavoidable, but it did not begin until 1991,
201
when the new law on financial and banking institutions was passed. The sector
underwent a thorough restructuring (institutions were recapitalised or liquidated as
appropriate) and extensive liberalisation. The authorities now need to tackle a new
stage in financial modernisation: that of fostering competition and efficiency in the
system. The degree of financial development in Tanzania follows no clear pattern: the
M2/GDP ratio trended down from 24 per cent in 1994 to 19 per cent in 1997, owing to
a relative drop in the M1 money supply (to 10 per cent of GDP), which was not offset
by the supply of quasi–money (M2–M1). This may have been due to the tighter
monetary policy adopted in connection with the stabilisation programme. Today, credit
is both strongly geared to the short term and limited by inappropriate regulations and
banking practices (small producers are virtually excluded from credit by the requirement
that all loans be fully secured by collateral). As a consequence in particular of the
tightening of budget constraints on semi–public enterprises, the ratio of loans to
deposits has fallen to an absurdly low level (under 30 per cent in 1998, as against
141 per cent in 1990).
The supply of infrastructure is very low. The old socialist regime left a
catastrophic situation in this respect, and several years of reform have brought only a
partial recovery. The government lacks the human and financial resources needed to
bring the process to completion. Roads are of poor quality, and electricity supply is
irregular and very expensive. As a result, factor costs remain prohibitively high.
Tanzania has succeeded in greatly improving its economic policy, deregulating

its economy and raising the investment rate, but this has not led to a strong surge of
growth. Over the entire period under review, the fundamental problem of the Tanzanian
economy was not lack of investment, but the poor quality of investment. The low
level of capital productivity is attributable to the incomplete state of the reforms, or of
their results. Tanzania is the only one of the six countries studied in the Development
Centre’s research programme on “Emerging Africa” which, according to the composite
index of Guillaumont et al. (1999)11, does not belong to the category of emerging
African countries: the value of this indicator was negative throughout the 1980–95
period, and even started to drop again from 1993. The country’s most important (and
most lasting?) advances are certainly those achieved most recently, i.e. since 1996,
and which are not covered by the period studied by Guillaumont et al.
GDP per capita at the end of the decade was scarcely higher (+6 per cent) than in
1967, at the time of the Arusha Declaration; income distribution seems to have worsened;
and poverty may be even more widespread. As the recent official revisions have shown
(the official figure for GDP was revised upwards by 68 per cent for 1992 and 263 per
cent for 1998). However, the accuracy of Tanzania’s national accounts remains very
uncertain, which introduces a degree of uncertainty in all previous analyses.
202
In any case, the economy is moving in the right direction, but this movement
needs to be reinforced. Private investors are currently taking a wait–and–see attitude,
while public investment fell at the end of the period (from 27 per cent of total public
expenditures for fiscal year 1995 to only 12 per cent for fiscal year 1998). Change in
the structure of output is highly desirable, but no such change occurred in the first half
of the 1990s, when manufacturing value added fell from 9 per cent of GDP in 1990 to
7.3 per cent in 1995. Growth in the stock of human capital will be crucial, but prospects
are somewhat gloomy: public resources devoted to human capital are diminishing in
relative terms, despite the contribution of international aid, which unfortunately is not
well targeted within this sector. The primary school enrolment rate has dropped; and
enrolment in secondary education is extremely low. This may change as the
demographic transition picks up steam; although the fertility rate is still high, at 5.2
children per woman in 1998 (ADB, 2000), it is slightly below that of the neighbouring
countries.
Local ownership of the reform process, which did not exist at all under Nyerere,
is undoubtedly still insufficient. The weight of foreign aid probably undermines the
endogeneity of strategic decision–making, and governance is still poor — in fact, it is
often considered the crucial constraint on Tanzania’s development prospects. Most
public institutions are incapable of producing the documentation needed for any kind
of audit or other assessment. Building institutional capacity will be a key factor for the
future. As in Uganda, improvement of governance will require in particular a higher
wage for civil servants: 94 per cent of them are paid a wage that leaves them below
the poverty line. Moreover, middle–grade officials often receive compensation
equivalent to only 35 per cent of what their private–sector counterparts are paid, whereas
unskilled jobs are still in many cases better paid in the public sector than in the private12.
Efforts to raise public–sector wages are under way, in the wake of a civil service
reform programme that has been operational since 199313. The declared objective is to
make public employees more disciplined and accountable through the application of
modern management principles.
Lastly, corruption is an endemic problem. As soon as President Mkapa took office,
he created a Presidential Commission on Corruption, which has published a report
describing the extent of fraud among politicians, the government and the judiciary, but
no penalties have been imposed to date. The Transparency International corruption
index for 1998 ranked Tanzania 4th from the bottom. It is less a matter of large–scale
corruption, however, than of medium–scale, creeping corruption that is well established
at all levels of government. More generally, the bureaucracy is still very large and
devoted to red tape14, and the business climate is regarded with great concern by foreign
investors, which partly explains the meagre FDI flows into the country.
Tanzania will have to make progress on a number of fronts before it can advance
on a path of lasting and self–sustained growth.
203
Notes
by Bigsten and Danielson (1999) for the OECD’s Emerging Africa project.
3. The Tanzanian experience won the sympathy of many intellectuals and donors (Nyerere
was widely acknowledged to be an excellent orator) until the early 1980s. Only the
Bretton Woods institutions opposed his policy stance in the early days.
4. According to some estimates, undeclared exports amounted to nearly half of official
exports at that time.
5. Although this is also due to a bias in the aid supplied by donors.
6. And perhaps even less if one observes the upward–revised figures for national output, as
provided by the new national accounts published in 1997 (see “Overview and Prospects”).
7. This is an independent public agency which pays its employees above the regular
administrative wage schedule; however, combating tax avoidance and corruption is
difficult because of the size of the informal economy and the undeveloped state of the
financial system, which reduces the means of verification available for tax auditing
purposes.
8. This rise in inflation was mainly due to exogenous factors, i.e. the climatic shocks in
1996 and 1997 (drought followed by heavy rains), which reduced agricultural output
and raised prices owing to the imbalance between supply and demand. Foodstuffs make
up 71 per cent of the basket used to calculate the consumer price index; the latter needs
to be modified.
9. The refinancing rate exceeded 20 per cent in real terms over a discontinuous period of
several quarters from 1994 to 1997.
10. The political old guard is not enthralled with the idea of privatisation, while there is
growing resentment towards foreigners and Tanzanians of Asian origin, who might be
the main beneficiaries of the process.
foreign conflicts.
204
12. In fact, the egalitarianism advocated by the Arusha Declaration had induced the
government to reduce the wage differential between senior officials and unskilled public
employees from 30 to 1 in 1969 to 6 to 1 in the mid–1980s. The recent reforms have
expanded the wage structure again: the ratio now stands at 16 to 1.
13. The share of total public spending allocated to the public wage bill rose from 20 to
30 per cent between fiscal years 1994 and 1998, despite a 30 per cent reduction in public
employment. Part of this rise was due to the partial monetisation of certain benefits in
kind, particularly petrol. From 1992 to 1997, real wages rose by 75 per cent, but they
started from very low levels (in the late 1980s their value was only a fifth of that in the
early 1970s).
14. It still takes 18 to 36 months to obtain the many official permissions required to start up
a business.
205
206
Concluding Summary
Policy Options for Emerging Africa
In the period following independence, most African economies enjoyed more

than a decade of steady growth. In contrast to the forecasts of an “Asian drama”, the
economic future of Africa looked particularly bright. From the 1970s to the late 1980s,
however, most African economies actually slowed down or, more often than not,
regressed dramatically. The tide had turned, and grim predictions of an African drama
— “Afro–pessimism” — became the norm. Several of these economies bounced back
in the 1990s, feeding hopes of an “African renaissance”, but today this concept has
lost its fugitive glamour, mostly because of the resurgence of devastating conflicts (in
the Horn of Africa, central Africa and Guinea), civil strife (Zimbabwe, Côte d’Ivoire)
and the HIV/AIDS pandemic (particularly in eastern and southern Africa), which have
overshadowed very positive events such as the peaceful, democratic political transitions
in Senegal (2000) and Ghana (2001).
The present study has delved beneath these dramatic images to re–assess the
economic factors underlying the successive episodes of irregular growth in Africa,
and has taken a long–term perspective to evaluate Africa’s chances of achieving
sustained high growth rates over the next 20 years. It has done so by taking an original
approach with respect to both the sample of countries examined and the analytical
framework used.
First, it deliberately focuses on six recent “success stories” which point to the
factors of a possible African economic emergence. These are Burkina Faso, Côte
d’Ivoire, Mali and Tanzania, which have all experienced an acceleration in growth in
the second half of the 1990s, as well as Ghana and Uganda, which have enjoyed high
growth rates over longer periods. Although the growth episodes are quite short, such
cases are encouraging, not just because these African economies started to grow faster
again over the last decade, but also because their new growth paths differed from
those that had prevailed in the 1960s. New factors at work in the 1990s could eventually
make growth more lasting and allow sustainable poverty reduction.
207
The second original feature of this study is that, in analysing these factors, it
takes into account elements that are traditionally overlooked. For instance, while
diversification or the lack thereof is usually presented solely as an indicator of the
development of an economy, here the impact of diversification on growth is considered,
and shown to be substantial. In addition, by focusing on African economies only, the
study allows for the inclusion of parameters that are common to African economies
but would make less sense in a cross–regional analysis, such as the important roles of
foreign aid and debt. Moreover, we reject the traditional assumption that the factors of
production, including labour, are perfectly mobile across sectors, analysing instead
the implications for growth of the dualistic structure of African economies and the
imperfect mobility of factors.
Our analysis yields the following conclusions:
— African countries are not different from other developing countries. The sustained
growth essential to alleviating poverty can be achieved.
— However, there will be no Asian–style boom in African economies over the next
20 years. Africa’s emergence will be slow.
— African economies are not doomed to fail because of external factors: what matters
most is the quality of domestic policies and the sustainability of this quality over
time. Long–term efforts of investment in human capital are especially important.
— A “Marshall plan” for Africa is unlikely to turn the tide, but donors can make a
difference with sustained and more efficient aid to support these policies.
The bottom line is that, if Africa is to reduce poverty in a sustainable manner,
faster growth is needed. Otherwise, the targets set by the OECD Development
Assistance Committee (DAC) will be unreachable. The 1990s have shown that a decade
of growth can indeed have a substantial impact on living standards, but even the
economies which grew most rapidly over that decade would require another two decades
at comparable growth rates to reach a per capita GDP exceeding $500 or $600. In
other words, sustained growth may allow people to live significantly better over the
next generation, but no economic miracle should be expected.
Just as there is no single cause for slow growth, so there is no miracle solution,
no single button to press that would automatically trigger a virtuous circle of growth
and poverty reduction in African economies. Increased trade openness, accelerated
debt relief, sound macroeconomic management, modernisation of agriculture, and
more efficient institutions for financial intermediation are all necessary conditions for
growth, yet none is by itself sufficient. The need for these improvements in Africa
points to both weaknesses in governance and impediments to the institutional change
needed to strengthen property rights and civil rights. Braga de Macedo (2001) has
shown how experiences in democracy building in Latin America and elsewhere
208
demonstrate the possibility of achieving political freedoms without sacrificing financial
ones. What is necessary in Africa as in other developing countries is to ensure that
institutional reforms are not blocked by bad governance and, in particular, by corruption.
Those conclusions should be taken neither as forecasts of a continent–wide economic
tragedy nor as excessively optimistic. The point of this study is of a different nature:
by taking a close, fresh look at the actual factors of growth in African economies, in
the light of recent positive experiences, it aims to help African policy makers as well
as DAC donors identify those areas where action can best make a difference.
The Story So Far: Is Africa Different?
Early Growth Episodes: Trying to Walk on One Leg
African countries are not different from other developing countries. In economic
terms, indeed, the failure of African countries to take off can be explained by the same
factors that made several East Asian economies successful (see Box 1). The continent
has actually witnessed many sustained episodes of fast growth. Relatively rapid growth
rates in the 1960s reflected massive programmes of public investment in physical
infrastructure, in productive sectors and in education. Growth rates were all the higher
because the African economies started from very low levels. To that extent, the early
growth episodes in Africa show some similarity to those in East Asia.
Box 1. There Is No Single African Economy: Evidence from the 1990s

For all their largely historical common features, African economies cannot be
pictured as one. The diversity of experiences and performances over the 1990s
shows that any generalisation would necessarily be misleading:
– Over the 1990s, there was a 25 percentage point gap in growth rates between
the best performer (Equatorial Guinea) and the worst (Democratic Republic
of Congo).
– Botswana shares borders with some of the most heavily indebted countries
in the world, but is itself a creditor of the World Bank.
– Côte d’Ivoire’s exports expressed as a share of GDP are twice as large as
Ghana’s, despite the two countries’ marked similarities in terms of
geography and comparative advantages.
– Mauritius has pupil/teacher ratios comparable to those of developed
countries (around 25), while a Malian teacher has classes of 70 pupils on
average; etc.
209
What brought these growth episodes to an end? As the analysis clearly shows,
the sustainability of growth depends on capital accumulation and productivity gains,
which together promote structural change (Figure 1). In a nutshell, the East Asian
countries were able to do both of these, while also implementing important institutional
reforms1 — whereas African countries accomplished only the former. Eventually, low
levels of productivity undermined the sustainability of investment–driven growth. The
low quality of investment — concentrated in rent–seeking activities — led to
widespread inefficiencies and eventually to declining returns. Capital accumulation
failed to promote economic diversification, which also proved to be an obstacle to
growth in that it blocked necessary structural transformation. In particular, the failure
to modernise agricultural sectors impeded the reallocation of labour away from
agricultural activities, a decisive factor in the structural transformation of the economies
analysed in Chapter 2.
Figure 1. Three Pillars of Sustained Rapid Growth
Capital TFP
Accumulation Gains
Institutional
Reforms
Structural
change
These characteristics are largely policy related. Whereas resource allocation was
strongly geared towards public investment in the years of growth, inefficiencies resulted
in low or negative growth of total factor productivity (TFP), partially offsetting the
positive impact of capital accumulation. Diversification, for its part, depends on the
attractiveness of the economic climate for savings and investment, but most African
economies lack the administrative institutions and financial systems — shaped by
policy — which are needed to strengthen legal protection of property rights. Lack of
financial diversification impedes investment diversification, which in turn prevents
mitigation of economic risk. In the end, profits from rent–seeking activities exceed
those derived from production through innovation (or imitation); rent seeking becomes
the favourite activity; and structural change comes to a halt. The associated losses in
terms of growth are substantial: the case of Mauritius shows that effective diversification
can add up to half a percentage point of growth per year on average.
210
Growth in the 1990s and Beyond: The Challenge of Sustainability
In the 1990s, solid growth performances were achieved by a number of African

countries — those in our sample of six and others, among which the best performers
are Mauritius and Botswana. These performances have had a visible impact on income
levels, although their impact on poverty is less clear (see Box 2). The main difference
between the earlier growth episodes and those of the 1990s is the role of TFP. Capital
accumulation played a lesser role in growth in the 1990s, while productivity gains
played a greater role. Among the several factors of TFP progress at work, efforts to
correct macroeconomic imbalances — e.g. the CFA franc devaluation and structural
adjustment policies — have paid off, and export growth has contributed more to labour
productivity over recent years than in the past. It should nevertheless be stressed that,
although capital accumulation played a less important role in relative terms, it still
plays an essential role in absolute terms: the fast growers are countries that, in addition
to improvements in TFP, have maintained relatively high investment rates.
Box 2. Growth and Poverty in Africa in the 1990s

– Growth has not systematically led to poverty reduction, whether measured
in monetary or non–monetary terms.
– Faster–growing countries have nevertheless seen larger reductions in infant
mortality, in addition to visible social progress as measured by human
development indicators.
– Increases in literacy rates stem necessarily from a long–term effort in
education, so the link with growth is tenuous.
– The impact of HIV/AIDS, especially in southern African countries, makes
the picture unclear because the epidemic reduces life expectancy (a
component of poverty) regardless of growth levels. Even when one controls
for HIV incidence, the link between growth and life expectancy is not
straightforward.
This observation should serve as a warning to countries where improvements

were mainly due to catch–up effects. One–off productivity gains from adjustment
measures, if not accompanied by an increase in investment and structural change, will
not deliver sustained higher growth rates. Even the better performers need to raise
their levels of investment, for the contribution of capital to growth will necessarily
remain modest in African economies, and failure to increase investment will hold
back structural change. The good news about these recent productivity gains is that
they have a key role to play in attracting investors. However, the picture remains less
encouraging on the savings side, which is still handicapped by inefficient national
financial systems and non–dynamic remittances, mostly used for consumption rather
than investment.
211
Prospects for the Coming Decades: From “Pessimism” to “Renaissance”
… to Emergence
Projections based on the analytical framework used in this study outline a possible
performance path for these six economies over the next two decades2. Under the
assumption that current trends in the underlying determinants of growth will continue
(our “baseline scenario”), the following points emerge:
— Economic take off is not necessarily sustainable in those economies which grew
relatively fast in the 1990s, since capital accumulation and faster structural change
are still needed in such a process. Table 1 indicates that the countries that seem
set to achieve high, sustainable levels of investment, structural change and
productivity gains in 2020 should see the highest rates of per capita GDP growth.
— The projected impact of change in demographic dependency ratios on growth
can hardly be over–emphasised, as such change accounts for one–third of the
change in per capita GDP.
— Thanks to investment in education and progress in diversification, Uganda
achieves the highest long–term growth of per capita GDP under the baseline
scenario, about 0.5 percentage point greater than Côte d’Ivoire and Mali on an
annual basis, and a full percentage point greater than Ghana.
— Despite its better ranking in terms of productivity growth, Ghana has the lowest
per capita GDP growth rate, mostly because of low investment rates. The country’s
lack of capital accumulation and slowness in making structural change are sources
of concern for the future.
— Tanzania’s growth prospects are the second lowest, even though it starts from a
comparatively low level. With very little structural change recently, especially
in education, the country has the lowest productivity growth in the sample.
— Conversely, Burkina Faso — which also started from a very low level, but which
has recently performed better in terms of structural change — is projected to have
the second–fastest growth rate in our sample. It will still be far from its steady
state in 2020, however, particularly in terms of openness and factor reallocation.
The hypotheses on which the above projections are based cannot be described as
excessively optimistic or pessimistic. For the most part, they are simple extrapolations of
current trends (progress in agriculture, financial deepening, etc.). The study also provides
a more optimistic “high” scenario, however, which assumes improvement in structural
change, such as human capital and diversification of the economy. The differences between
these are substantial in the long run. Whereas the baseline scenario predicts annual growth
rates of per capita GDP ranging from around 1.6 per cent (Ghana, Tanzania) to around
2.5 per cent (Uganda, Burkina Faso), in the high scenario these rates vary from 2.4 per
cent (Tanzania) to over 3.6 per cent (Burkina Faso). This would mean that by 2020, Tanzania
would reach a level of per capita GDP of around $4153 ($350 in the baseline scenario),
Mali and Burkina Faso between $520 and $560 ($400 and $450 respectively), Uganda
$630 (slightly under $500), Ghana $630 (over $500) and Côte d’Ivoire $1 700 ($1 300).
212
Table 1. Projections for 2020: Results of the Baseline Scenario
(growth rates in 2020, percentages)
Per capita GDP Effect of change GDP/labour Capital deepening Total factor
in demographic productivity
dependency ratio
Burkina Faso 2.5 0.7 1.8 0.7 1.1

Uganda 2.6 1.0 1.6 0.5 1.0
Mali 2.0 0.6 1.4 0.6 0.8
Ghana 1.6 0.5 1.1 0.5 0.7
Côte d’Ivoire 2.1 0.9 1.1 0.6 0.5
Tanzania 1.7 0.7 1.0 0.5 0.4
Average 2.1 0.7 1.3 0.6 0.8
Total factor of which

productivity
Education Reallocation Openness Diversification
(Exp/GDP)
Burkina Faso 1.1 0.7 0.4 0.0 0.0

Uganda 1.0 0.6 0.3 0.0 0.1
Mali 0.8 0.7 0.2 0.0 0.0
Ghana 0.7 0.4 0.1 0.1 0.1
Côte d’Ivoire 0.5 0.1 0.3 0.0 0.0
Tanzania 0.4 0.1 0.3 0.0 0.0
Average 0.8 0.4 0.3 0.0 0.0
Two variants of the high scenario isolate the effects of diversification and aid
efficiency respectively. The former clearly points to the importance of institutional
quality for economic growth: safer and more efficient financial intermediation,
protection of property rights, less corruption and rent seeking, a functioning judicial
system, moderate government intervention. Together with better market access for
non–traditional exports, these improvements would enhance not only the volume but
also the diversification of investment. In the projections for the six countries, such
institutional upgrading would raise the level of per capita GDP in 2020 by 5 per cent,
through enhanced diversification. The second variant suggests that the potential
contribution of improved aid efficiency to growth is considerable, and may have a
greater impact than the more “visible” debt relief schemes (see Box 3)4.
The analysis clearly shows that there is no single cause of slow growth in Africa,
and similarly no single way of accelerating growth. The question, then, is where to
begin. The study points to several areas where African governments can make a
difference and where donors can help. These points do not form a sequence of distinct
actions. Rather, they are a set of interwoven pre–requisites for faster, sustainable growth.
213
Box 3. Debt Relief: How Much Will It Help?
Debt relief can act in the same way as an increase in aid flows, improving the
balance of payments and to some extent contributing to the financing of
investment.
The study estimates, however, that the direct impact of the Cologne initiative
will be rather small in terms of equivalent supplementary resource flows. It
would procure the equivalent of a 13 per cent increase in aid flows for Tanzania
(the largest beneficiary) and merely a few percentage points for the other
countries.
It is possible, however, that the Cologne initiative may have a greater indirect
impact, facilitating future growth through better incentives to invest.
What Africans Can Do for Themselves
The study confirms that African countries’ performances have less to do with
external conditions than is sometimes thought: terms of trade have a very small impact
on growth rates. This implies that the scope for domestic policies to foster change is
also greater than some may fear. The above analysis points to three main objectives:
enhancing the overall level of investment, improving productivity and fostering
diversification ––goals that can be met namely, though not exclusively, by enabling
the managed opening of these economies. Both Braga de Macedo (2001) and Bonaglia
et al. (2001) have pointed out that these objectives can be reached only if governments
implement the necessary political reforms.
Enhancing Investment in a Sustainable Manner
Sustained capital accumulation is of strategic importance for African economies.

For prospects of higher returns to stimulate the demand for investment, progress in
factor productivity is required (see below), as well as lower levels of both economic
and political risk. As for the supply side, national savings cover only a fraction of
investment, partly due to high debt repayments to foreign creditors. Given the structural
weakness of local saving capacity (due to low incomes), and since improvements in
the efficiency of financial intermediation institutions may be expected only in the
medium to long term, there can be no alternative to attracting more foreign direct
investment. Foreign aid has been an important source of financing, but it cannot make
up for the lack of investment capital in a sustainable manner, as it tends to substitute
for national savings.
214
From this point of view, stronger activity in the continent’s regional integration
processes is a positive trend for the future. These are no longer considered to be an
alternative to unilateral trade liberalisation, nor are they understood as mainly serving
to enable growth in the trade of goods and services between member countries. It is
thus now widely accepted that co–ordinated, flexible approaches between countries
sharing dynamic economic regions at their borders can make a considerable contribution
to investment growth and infrastructure development5. Numerous opportunities of
this sort remain to be explored, for example in central Africa and in the Great Lakes
region (where the DRC could become a pivotal force for regional integration once
peace has returned) as well as in North and sub–Saharan Africa where borders do not
necessarily reflect “real ” separations.
Improving Factor Productivity
Human capital is key for long–term growth because it makes labour more
productive, enhances the assets of poor communities and individuals, and is a necessary
ingredient of structural change in an economy. Margins of manoeuvre to enhance the
contribution of human capital include: i) increasing the number of years of schooling
of the labour force (which means that school enrolments must grow faster than the
population); ii) raising the quality of education (inevitably at high cost); or iii) reducing
the wastage of existing human capital in rent–seeking activities (at zero budgetary cost).
In the poorest countries, significant TFP gains can be realised through i) and ii),
i.e. by investing in education, but this will necessarily be a slow and costly process
requiring larger budgets for education and a higher budget share for primary education.
An increase in enrolment rates is necessary, but has a budgetary cost which central
governments may not be able to sustain alone6. Human capital formation can also be
improved by reforming education systems, as the case of Uganda shows: contributions
by parents, in a decentralised system, can give public expenditure a good deal of
leverage.
Point iii) above entails opening up opportunities in productive activities and
curbing corruption, in order to increase the social returns on investment in education.
Indeed, the problem is not only that the current level of such investment is lower than
it was during the take off of East Asian countries (and even lower than in post–
independence Africa), but also that such investment brings a smaller return in African
than in East Asian countries. The actual contribution of education efforts to growth is
far less than optimal (especially at the secondary and higher levels), because these
efforts are to a large extent wasted in unproductive rent–seeking activities and
corruption. According to our estimates, the African countries considered here waste
up to twice as much of their human capital as East Asian, Latin American and OECD
countries do. Here again, the magnitude varies: Mauritius, Côte d’Ivoire and Ghana
have relatively low levels of wastage, Kenya and Uganda higher ones. In countries
215
where corruption is relatively moderate (Mali, Burkina Faso) and education supply
capacity is low, fighting corruption could pay off as much as increasing education
efforts. In those where corruption levels are higher (Tanzania, Democratic Republic
of Congo7), it may bring an even higher return. In other words, curbing corruption and
rent–seeking activities would be an inexpensive way for African economies to make
better use of their existing human resources.
Promoting Structural Change
Macroeconomic stability is necessary, but it will not do the trick alone. Changes
in economic structures are needed, in the form of higher shares of employment in
non–agricultural sectors and a more diversified economy. Structural adjustment
programmes have indeed brought progress in terms of macroeconomic balances,
although in some cases reforms arguably need to be deepened or consolidated (e.g. in
Ghana, where inflation remains a serious obstacle to growth). The experience of the
1990s shows, however, that sound monetary and budgetary policies are neither
necessary nor sufficient to trigger growth. Imbalances do not cause bad performance;
rather, they reflect the inability of governments to (re)start growth. Progress remains
to be made in establishing the right set of incentives, with an institutional framework
that favours the modernisation of agriculture and profitable private investment in a
wider range of activities.
The modernisation and intensification of the agricultural sector are pre–requisites
to an increase in labour productivity, and eventually to the reallocation of labour to
non–agricultural activities. The developing countries in Africa and elsewhere that have
experienced the highest productivity gains in agriculture are those in which a part of
the agricultural labour force has been reallocated to other, more productive activities8.
Increasing productivity in agriculture requires only basic techniques and the potential
increase in yields is considerable, as the success of cotton growers in Mali shows.
However, extension services can succeed in this only if adequate incentives are in
place, which requires a difficult — and thus almost certainly slow — reform of
production, marketing, financial and fiscal institutions9.
As for diversification, experience clearly shows that it cannot be dictated. Rather,
the issue here is to facilitate private sector development by introducing the right
incentives (availability of skilled labour, infrastructure, macroeconomic policy
favourable to investment, openness, obtaining market access for non–traditional
exports, institutions favourable to risk taking, etc.). As Uganda and Ghana have shown,
the countries that have progressed the most in terms of diversification are those where
the change towards more market–friendly policies and the establishment of efficient
enabling institutions (trade promotion agencies, public/private sector dialogue, etc.)
have been the most marked. International trade obviously has a key role to play in
promoting the diversification of economic activities, among other things by encouraging
imitation, a possible substitute for innovation.
216
Openness to Trade and Export Promotion
Trade is not the answer to the need for faster growth, and the next 20 years will
probably not witness a generation of “African tigers” following in the footsteps of
Southeast Asia. The study does show, however, that trade is an important element of
the equation. On the import side, there is scope for more trade liberalisation, which
should further stimulate exports and which is necessary to encourage acquisition of
technology (see below). This will work only if trade policies are formulated in a manner
consistent with the country’s overall economic policies, which themselves should be
part and parcel of its development strategy. For instance, companion policies should
aim at enhancing the capacity of the poor to respond to the economic change induced
by trade liberalisation.
On the export side, the challenge is mostly one of supply. African exports grew
more slowly than those of other countries over the 1980s and early 1990s, mostly
because African economies continued to specialise in products for which demand has
been growing slowly or declining, but also — to a lesser extent — because they have
lost competitiveness in each of these sectors. Macroeconomic adjustment allowed
significant competitiveness gains in the mid–1990s, which in turn triggered a rise in
exports. It is true that in the short to medium run, radical adjustment policies (e.g. the
CFA franc devaluation, liberalisation in Tanzania) have made it possible to mobilise
under–utilised resources almost immediately, and have suddenly increased the price
competitiveness of traditional (mainly agricultural) exports, resulting in significant
improvements in export performance (e.g. cotton in Mali). Such gains cannot be made
indefinitely, however, and in the long run export performance will primarily depend
on growth in the production capacity of African economies. This requires nothing
less than deep structural transformation and diversification of these economies, as
we saw in Chapter 2.
Innovation versus Imitation
Technologically as well as commercially, African economies are perceived as

“marginalised” in the world economy. Many observers speak of a growing
technological gap between Africa and the rest of the world (developed and
developing) that can never be bridged. Endogenous growth theories, which emphasise
the key role of research and development — activities that require large investments
far beyond the reach of African economies — may seem to give an academic
justification to these fears. This study argues, however, that what African economies
cannot achieve by innovation, they can achieve, at lower cost, by imitation. This
will require greater openness to imported goods and services, foreign direct
investment, and the technology embedded in them; a minimum level of human capital
so that imitation can take place; and possibly the establishment of proactive “imitation
strategies” involving businesses.
217
Political Reform for Sustainable Structural Transformation
In most cases, the macroeconomic political reforms outlined in Chapters 2 to 5

are easier to implement than structural reforms. Since the latter are more threatening
to vested interests, implementing them is more difficult and takes more time. But this
reform process is also hindered by the precarious political situations found in many
African countries. On the other hand, the experiences of economic take off in Japan,
and later in Korea and Singapore, would seem to indicate that a strong national political
leadership which makes national development and growth a priority is a necessary
condition for take off and for mobilising the forces of production. African autocracies,
though, have often met with disastrous results. These regimes have allowed for the
enrichment of minorities to the detriment of national development and the improvement
of the population’s living conditions. Such practices also stir up conflicts (both overt
and latent) which effectively exacerbate the poverty trap, as can be seen in the tragic
example of the Democratic Republic of Congo10, only one example among many. In
young nation states where institutions are fragile and where there is often considerable
cultural and linguistic diversity, democracy and good governance are largely perceived
as essential to good leadership11.
Ongoing democratisation efforts in a large number of countries reveal the
complexity of this process. These efforts highlight the risk that weak democratic
governments may be unable to bring down rent–seeking activities that benefit trade
unions or specific economic sectors, or to regulate monopolistic markets, in particular,
but not exclusively, when faced with financially superior multinationals. Market failures
can thus persist, and decrease social well–being in the long term. Moreover,
democratisation efforts in countries dependent on foreign aid reveal an important
contradiction, e.g. that the reform process itself, because of conditionality, does not
always adapt well to democratic principles. This is in spite of efforts on the part of
donors and investors to foster the democratisation process, primarily through
decentralised co–operation and dialogue with civil society. In this respect, the difficulty
faced by both donors and recipients in implementing a truly participative and efficient
development demonstrates the need to refine this process –– even to create new ways
of working together. The role of parliaments is also too often neglected because in
many cases these bodies act as mere rubber stamps for decisions made by high level
officials, both national and international. This fact undermines citizen participation,
and ultimately contributes to political instability. Finally, African entrepreneurs should
be able to fully contribute to the economic development process. The need to encourage
constructive dialogue between the private and public sectors in the aim of building
confidence between them (which is often lacking) cannot be overestimated.
Thus besides being ends unto themselves, democracy and good governance are
prerequisites for the emergence of dynamic economies in Africa. The case studies
presented in Part Three examine several examples. However, in the absence of role
models, reaching these goals will require these countries and their partners to develop
approaches that are both cautious and innovative .
218
What Donors Can Do to Help African Countries Help Themselves
Rebalancing and Focusing Aid
Donors’ possibilities of influencing aid efficiency in recipient countries may

seem limited, at least in the short run. There has been a lively debate, largely instigated
by the World Bank, on the efficiency of aid, and particularly on whether to focus it on
“good policy” recipient countries12. This study would argue that aid can usefully be
targeted to sectors where it is more likely to support rapid growth, and to countries
which have a demonstrated record of making sensible use of aid funds.
As we saw in Chapter 4, investment in human capital is key, but in almost all
countries, national resources will not be able to keep up. The scope for enhancing the
share, and the amount, of aid dedicated to primary schooling and basic health services
is still very large: the study shows that these areas can bring high returns in terms of
social equity, but also in terms of growth. A call for curbing corruption and rent–
seeking behaviour as a way of enhancing the impact of human capital (see above)
should also explicitly remain at the core of the partnership between donors and
recipients. In productive sectors, the catalytic role of aid for investment can be
strengthened:
i) to promote regional investment by supporting financial deepening, a key
requirement for diversification into higher value added activities instead of merely
into new but still low value added commodities;
ii) to help attract FDI from OECD countries by giving a boost to the promotion of
joint ventures and venture capital operations.
An example of the latter is the European Union’s granting of a higher profile and
broader role to the European Investment Bank under the Cotonou agreement on ACP–
EU partnership, compared with the EIB’s more limited role under the successive Lomé
conventions.
One message to the donor community is to focus on those countries which have
been showing these encouraging signs of success, as well as of making good use of
aid funds, and which can become models in their own regions. While regional
champions have played an important part in technology diffusion and growth in Asia,
most African countries suffer from a lack of regional leadership. South Africa arguably
plays the role of a locomotive to some extent in the southern part of the continent13
(although the potential for regional co–operation within the Southern African
Development Community remains largely untapped), but Nigeria in the west, under
successive military regimes, has remained highly oil–dependent and, until recently,
inward–looking. Apart from these two heavyweights, the countries studied in this
report could become examples for their neighbours to emulate, as could others whose
economic performance has improved lately. Together with a greater focus on these
good performers, donor support to regional co–operation and integration strategies
can help stimulate such emulation.
219
Enhancing the Capacity of African Economies to Reap the Benefits
of Globalisation
The benefits of greater openness and participation in the multilateral trade system
will not materialise automatically, for two reasons. Owing to changes that are largely
beyond their control (see Box 4), African countries’ capacity to participate is extremely
weak, and African economies arguably have less room for manoeuvre than did other
WTO members which opened their economies earlier. Donors have a role to play in
supporting the trade policy capacity of African countries and the competitiveness of
their exporters. Mainstreaming trade in development and poverty alleviation strategies
requires renewed efforts to build the capacity of the national and regional institutions
in charge of the trade policy process. Each multilateral and bilateral donor has its own
comparative advantages in this sector. Recent and ongoing co–ordination efforts —
e.g. within the Integrated Framework14 — as well as the Development Assistance
Committee’s work focusing on developing best practices15 should help improve the
impact of trade–related aid. Donors can also provide substantial help by targeting
the poor, who are likely to be the group most affected by changes in international
trade policies.
Box 4. Timing Matters: African Economies,

Late Starters in Trade Liberalisation
African countries started opening their economies later than most other trading
nations (developed and developing). As a result, they face a narrower range of
choices and may have to expect lower returns from opening than earlier
reformers have enjoyed.
– Stronger competition on world markets. Many more countries are
competing on world markets than there were a couple of decades ago,
which lowers the potential returns on outward–oriented strategies.
– An increasing technological bias against unskilled labour may have made
trade–led growth progressively less beneficial for poor countries where
this factor is relatively abundant.
– Less scope for trade policy. Selective protectionism may have been easier
to implement — and potentially more beneficial — for East Asian nations
in the 1960s than for African nations today, as multilateral and regional
disciplines were less constraining then.
– Less scope for domestic support. WTO rules froze support to agriculture
before African economies even started to set up their own support systems.
– Less scope for bargaining. Not only are African economies small players
in world trade negotiations because of their weak supply capacity, but the
multiplicity of trade liberalisation fronts (structural adjustment programmes,
regional trade liberalisation schemes — often donor–sponsored), if not
carefully managed, may diminish their potential bargaining power in
multilateral negotiations.
220
From Trade Preferences to Trade Liberalisation
By contrast with Asia, short–lived growth periods in Africa were induced

primarily by surges in investment, fuelled by temporary commodity booms. These
booms have not triggered a process of capital accumulation, however, and have arguably
hindered diversification. Instead, in some cases (Cameroon, Côte d’Ivoire) they led to
wasteful investment and rent–seeking behaviour. When export prices fell, economies
were left with unchanged structures and increasing distortions. While domestic policies
are mostly to blame, one cannot fail to notice that the focus on product–specific trade
preferences may have provided the wrong incentives. Moreover, in terms of developed
countries’ trade policies towards developing countries, such preferences may have
substituted for enhanced market access across the board, which would have been a
better incentive for diversification. While preferential schemes may have helped a
few countries or exporters to enter world markets, they can hardly be expected to
continue playing such a role in the coming decades. A new round of multilateral trade
negotiations, which would include substantial opening of OECD markets to LDCs
and other poor countries, as well as lower barriers to large emerging economies, may
be the best way of inviting emerging African economies to participate actively in the
international trading system. Finally, apart from the issues related to border measures,
future multilateral negotiations should also ensure that the global rules of the trade
game — especially on intellectual property — do not hamper African economies’
potential to use imitation strategies to engage in world trade.
❊ ❊ ❊
In spite of its heterogeneity – a dimension underscored by the findings of this

work — Africa aims for unity in its development. The various programmes for Africa
adopted by the G–8 and the Bretton Woods institutions will not change the fact that
nothing and no one can take the place of African countries in their own development.
Mobilising development efforts assumes the pursuit of reforms such as liberalisation,
and above all improvements in governance and in institutional structures. Without
these, there is a great risk that Africa’s potential will be wasted through corruption.
Such reforms are also a prerequisite for attracting foreign investment. Without a doubt,
this requires the high quality leadership — requiring both credibility and legitimacy –
that characterises successful development efforts. Conversely, periods of uncertainty
and political instability lead much too often to a slowing, or even to a reversal of
development.
221
Notes
1. The extent of institutional reforms, however, varied from country to country. The Asian
financial crisis of the 1990s revealed the limits of some of these.
2. In the long–term scenarios, we control for factors of political and social instability — by
definition impossible to predict — by assuming that the risk remains constant from the
past decades onwards.
3. $US 1996.
4. See also Joseph (2000) and Cohen (2000).
5. These were the conclusions found by the Second International Forum on African
Perspectives organised by the African Development Bank and the OECD Development
Centre in Paris, March 2001, published in ADB–OECD (forthcoming). See also Braga
de Macedo et al. (2001).
6. It is worrying that, in all the countries studied but one (Uganda), outlays for this sector
have not been sufficient to maintain the share of education in GDP. Countries that adopted
ambitious plans in the 1990s (such as Burkina Faso) have not been able to sustain both
their initial levels of investment and the quality of education, and in some cases (Côte
d’Ivoire, Tanzania) investment in human capital has even decreased.
7. Subject, of course, to the end of the conflict and substantial improvements in the political
situation in the Democratic Republic of Congo.
8. As in 19th century Europe or East Asia in the 1950s.
9. Here too, regional integration can increase the influence of “peer pressure” as it did in
Europe. See Braga de Macedo et al. (2001).
10. See also Conflict and Growth in Africa, OECD Development Centre Studies (Goudie
and Neyapti, 1999; Klugman et al., 1999; Azam et al., 1999).
11. Collier (1999) finds that the democratisation process probably provides the best protection
from predatory behaviour and ethnic conflict, while the World Bank observes that in
Africa “checks and balances must be exerted in order to limit the arbitrary exercise of
government. Public decision–making should be transparent and predictable” (World Bank,
2001).
12. For an overview with additional analysis, see Beynon (2001).
222
13. Although the potential for regional cooperation in the SADC remains largely untapped.
See Jenkins (2001), Braga de Macedo (2001), and ADB–OECD (forthcoming).
14. The IF is a joint endeavour by six multilateral agencies (IMF, ITC, UNCTAD, UNDP,
World Bank, WTO) and bilateral agencies to co–ordinate trade–related technical assistance
to developing countries, and to help mainstream trade in national strategies and policy
frameworks such as the Poverty Reduction Strategy Papers.
15. The OECD Development Assistance Committee was to issue its policy guidelines on
trade capacity development in 2001. See OECD–DAC (2001).
223
224
Bibliography
1. DOCUMENTS PREPARED FOR “EMERGING AFRICA” PROJECT

(available on http://oecd.org/dev)
1.1. National Case Studies
BIGSTEN, A. AND A. D ANIELSON (1999), “Is Tanzania an Emerging Economy?”, OECD

Development Centre, mimeo, May.
BIGSTEN A. AND S. KAYIZZI–MUGERWA (1999), “Is Uganda an Emerging Economy?”, OECD
Development Centre, mimeo, May.
CHAMBAS, G., J.L. COMBES, P. GUILLAUMONT, S. GUILLAUMONT–JEANNENEY AND B. LAPORTE (2000),
« Mali : les facteurs de croissance à long terme », OECD Development Centre, mimeo,
February.
CHAMBAS, G., J.L. COMBES, P. GUILLAUMONT, S. GUILLAUMONT–JEANNENEY AND B. LAPORTE (1999),
« Burkina Faso : les facteurs de croissance à long terme », OECD Development Centre,
mimeo, October.
COGNEAU, D. AND S. MESPLÉ–SOMPS (1999), « La Côte d’Ivoire peut–elle devenir un pays
émergent ? », OECD Development Centre, mimeo, June.
FOSU, A.K. (1999), “Emerging Africa: The Case of Ghana”, OECD Development Centre, mimeo,
May.
LEITH, J.C. AND L. SÖDERLING (2000), “Ghana: Long Term Growth, Atrophy, and Recovery”,
OECD Development Centre, mimeo, June.
MATON, J. WITH H.–B. SOLIGNAC LECOMTE (2001), Congo 1965–1999 : les espoirs déçus du
« Brésil africain », Technical Paper No. 178, OECD Development Centre, September.
MATON, J., K. SCHOORS AND A. VAN BAUWEL (1999), « Congo 1965–1997 », OECD Development
Centre, mimeo, April.
225
1.2. Thematic Studies
BERTHÉLEMY, J.–C. AND L. SÖDERLING (1999a), The Role of Capital Accumulation, Adjustment
and Structural Change for Economic Take–off: Empirical Evidence from African Growth
Episodes, Technical Paper No. 150, OECD Development Centre, July, World Development,
Vol. 29, No. 2, February 2001.
BERTHÉLEMY, J.–C. AND L. SÖDERLING (1999b), “Trade and Development: A European View on
African Issues”, article prepared for workshop on “Regionalism and a Post–Lomé Trade
Regime: Implications for Southern Africa”, Institute for Global Dialogue, Johannesburg,
October.
COHEN, D. (2000), The HIPC Initiative: True and False Promises, Technical Paper No. 166,
OECD Development Centre, Paris, October.
GUILLAUMONT, P., S. GUILLAUMONT–JEANNENEY AND A. VAROUDAKIS (1999), Economic Policy
Reform and Growth Prospects in Emerging African Economies, Technical Paper No. 145,
OECD Development Centre, March.
JOSEPH, A. (2000), Résoudre le problème de la dette, la contribution décisive de l’initiative
PPTE, Technical Paper No. 163, OECD Development Centre, August.
JOSEPH, A. (1999), « La réforme du secteur financier », OECD Development Centre, mimeo,
April.
RICHAUD, C. AND A. VAROUDAKIS (1999), “Securing Benefits from Globalisation: Progress Report
on the Consistency between Exchange Rate and Trade Policies in Africa”, OECD
Development Centre, mimeo, January.
2. OTHER REFERENCES
— L’État du Monde (2001), Éditions La Découverte, Paris.

A CEMOGLU, D. AND F. ZILIBOTTI (1997), “Was Prometheus Unbound by Chance? Risk,
Diversification and Growth”, Journal of Political Economy, Vol. 105, No. 4.
ADB (1999, 2000), African Development Report, African Development Bank, Abidjan.
ADB/OECD (2002), Regional Integration in Africa, Proceedings of Second International Forum
on African Perspectives, OECD Development Centre Seminars, African Development
Bank and OECD Development Centre, Paris
ALTER, R. (1990), “Export Processing Zones for Growth and Development: The Mauritian
Example”, IMF Working Paper WP/90/122.
AZAM, J.–P. AND C. DAUBRÉE (1996), Bypassing the State. Economic Growth In Kenya, 1964–90,
OECD Development Centre Studies (Long–term Growth Series), Paris.
AZAM J.–P. AND C. MORRISSON (1994), The Political Feasability of Adjustment in Côte d’Ivoire
and Morocco, OECD Development Centre Studies, Paris.
226
AZAM, J.–P. AND C. MORRISSON WITH S. CHAUVIN AND S. ROSPABÉ (1999), Conflict and Growth in
Africa, Vol. 1 The Sahel, OECD Development Centre Studies, Paris.
BATES, R.H. (1981), Market and States in Tropical Africa, University of California Press,
Berkeley.
BAUMOL, W.J. (1990), “Entrepreneurship: Productive, Unproductive, and Destructive”, Journal
of Political Economy, Vol. 98(5), pp. 893–921.
BERTHÉLEMY, J.–C. (1999), “Comments on Gunning and Mash”, ABCDE Conference in Europe,
World Bank, Paris, June.
BERTHÉLEMY, J.–C. AND F. BOURGUIGNON (1996), Growth and Crisis in Côte d’Ivoire, World
Bank, Washington, D.C.
BERTHÉLEMY, J.–C. AND S. CHAUVIN (2000), Structural Changes in Asia and Growth Prospects
after the Crisis, CEPII Working Paper No. 2000–09, Paris. Disponible sur http://
www.cepii.fr/
BERTHÉLEMY, J.–C. AND S. DESSUS (2000), “Why Doesn’t Human Capital Accumulation Always
Contribute to Growth?”, Social Science Research Network Working Paper Series. http:/
/www.ssrn.com
BERTHÉLEMY, J.–C., S. DESSUS AND A. VAROUDAKIS (1997), Capital humain, ouverture extérieure
et croissance : estimation sur données de panel d’un modèle à coefficients variables,
Technical Paper No. 121, OECD Development Centre, January.
BERTHÉLEMY, J.–C., C. PISSARIDES AND A. VAROUDAKIS (1998), “Human Capital and Growth :
The Cost of Rent–seeking Activities”, mimeo, OECD Development Centre, in DE RUYTER
VAN S TEVENINCK (ed.), Economics Growth and its Determinants, Kluwer Academic
Publishers, Dordrecht.
BERTHÉLEMY, J.–C., A. SECK AND A. VOURC’H (1996), Growth in Senegal – A Lost Opportunity?,
OECD Development Centre Studies (Long-term Growth Series), Paris.
BERTHÉLEMY, J.–C. AND L. SÖDERLING (2000), “Simulating the Next 20 Years”, in BRAGA DE
MACEDO AND KABBAJ (Preface), Reform and Growth in Africa, Proceedings of First
International Forum on African Perspectives, OECD Development Centre Seminars,
African Development Bank and OECD Development Centre, Paris, pp. 163–184.
BERTHÉLEMY, J.–C. AND A. VAROUDAKIS (1997), “Financial Development, Savings and Growth
Convergence: A Panel Data Approach”, in R. HAUSMANN AND H. REISEN (eds.) Promoting
Savings in Latin America, IDB–OECD, Paris.
BERTHÉLEMY, J.–C. AND A. VAROUDAKIS (1996), Financial Development Policy and Growth
croissance, OECD Development Centre Studies (Long-term Growth Series), Paris.
BERTHÉLEMY, J.–C. AND A. VOURC’H (1994), Debt Relief and Growth, OECD Development Centre
Studies, Paris.
BEYNON, J. (2001), Policy Implications for Aid Allocations of Recent Research on Aid
Effectiveness and Selectivity, DFID.
BHINDA, N., S. GRIFFITH–JONES, J. LEAPE AND M. MARTIN (1999), Private Capital Flows to Africa
– Perception and Reality, Fondad, The Hague.
227
BIGGS, T., M. SHAH AND P. SRIVASTAVA (1995), Training and Productivity in African Manufacturing
Enterprises, RPED Discussion Paper, World Bank, Washington, D.C.
BIRDSALL, N., D. ROSS AND R. SABOT (1995), “Inequality and Growth Reconsidered: Lessons
from East Asia”, The World Bank Economic Review, Vol. 9 No. 3, pp. 477–508.
BONAGLIA, F., J. BRAGA DE MACEDO AND M. BUSSOLO (2001), “How Globalisation Improves
Governance?”, mimeo, OECD Development Centre, July.
BOURGUIGNON, F. AND T. VERDIER (2000), “Is Financial Openness Bad For Education?”,
International Colloquium “Ouverture économique et développement”, GDR Économie
et Finance Internationale, Tunis, June.
BRAGA DE MACEDO, J., D. COHEN AND H. REISEN (eds.) (2001), Don’t Fix, Don’t Float, OECD
Development Centre Studies, Paris.
BRAGA DE MACEDO , J. (2001), “Globalisation and Institutional Change : A Development
Perspective”, paper presented at the General Assembly of the Pontifical Academy of
Social Sciences on the theme Globalisation and Common Humanity : Ethical and
Institutional Concerns, 25–28 April.
BRAGA DE MACEDO, J. AND O. KABBAJ (Preface) (2000), Reform and Growth in Africa, Proceedings
of First International Forum on African Perspectives, OECD Development Centre
Seminars, African Development Bank and OECD Development Centre
BURNSIDE, C. AND D. DOLLAR (1997), Aid, Policy and Growth, Policy Research Working Paper
No. 1777, World Bank, Washington, D.C.
CEPII (1998), Compétitivé des nations, Economica, Paris.
CHEGE, M. (1999), Politics of Development : Institutions and Governance, Global Coalition
for Africa, Washington, D.C.
CHENERY, H., S. ROBINSON AND M. SYRQUIN (eds.) (1986), Industrialization and Growth : A
Comparative Study, Oxford University Press, Oxford.
COLLIER, P. (1999), “On the Economic Consequences of Civil War”, Oxford Economic Papers, 51.
COLLIER, P. (1998a) “Globalization, Implications for Africa”, in Z. IQBAL AND M. KHAN, (eds.)
Trade Reform and Regional Integration in Africa, International Monetary Fund,
Washington, D.C.
COLLIER, P. (1998b), “The Role of the State in Economic Development: Cross–regional
Experiences”, Journal of African Economies, Vol. 7, supplément 2, December.
COLLIER, P. AND J.W. GUNNING (1999), “Explaining African Economic Performance, Journal of
Economic Literature, Vol. 37, March.
DE MELO, J. AND S. ROBINSON (1990), Productivity and Externalities. Models of Export–Led
Growth, Policy Research Working Paper, World Bank, Washington, D.C.
DESSUS, S. (1998), “Total Factor Productivity and Outward Orientation in Taiwan: What is the
Nature of the Relationship?”, in T.T. FU, C.J. HUANG AND A. LOVELL (eds.) Economic
Efficiency and Productivity Growth in the Asia Pacific Region, Edward Elgar Publishing
Limited, Aldershot.
228
DESSUS, S., J.–D. SHEA AND M.–S. SHI (1995), Chinese Taipei : The Origins of the Economic
Miracle, OECD Development Centre Studies (Long-term Growth Series), Paris.
EDWARDS, S. (1998), “Openness, Productivity And Growth: What Do We Really Know?”
Economic Journal, Vol. 108, No. 447, March.
FEDER, G. (1983), “On Exports and Economic Growth”, Journal of Development Economics,
Vol. 12, No.1/2, February–April.
FEENSTRA, R.C., D. MADANI, T.H. YANG AND C.–Y. LIANG (1999), “Testing Endogeneous Growth
in South Korea and Taiwan”, Journal of Development Economics, Vol. 60, No.2.
FISCHER, S., E. HERNÁNDEZ–CATÁ AND M. KHAN (1998), “Africa: Is this the Turning Point?”,
IMF Paper on Policy Analysis and Assessment 98/6, Washington, D.C.
GOUDIE, A. AND B. NEYAPTI (1999), Conflict and Growth in Africa, Vol. 3 Southern Africa,
OECD Development Centre Studies, Paris.
GUILLAUMONT, P. AND S. (eds.) (1994), Ajustement et Développement. L’expérience des pays
ACP, Economica, Paris.
HARVEY, C. (1998), “Botswana: The Impact of the Ending of Apartheid”, in L. PETERSSON (ed.),
Post Apartheid Southern Africa, Routledge, London and New York.
JENKINS, C. (2001), Integration and Co-operation in Southern Africa, Technical Paper No. 172,
OECD Development Centre, March.
JOHANSSON , H. AND L. N ILSSON (1997), “Export Processing Zones as Catalysts”, World
Development, Vol. 25, No. 12, December.
KLUGMAN, J., B. NEYAPTI AND F. STEWART (1999), Conflict and Growth in Africa, Vol. 2 Kenya,
Tanzania and Uganda, OECD Development Centre Studies, OCDE, Paris.
LAFFONT, J.–J. (1998), “Competition, Information and Development”, paper presented at the
Annual Bank Conference on Development Economics, World Bank, Washington, D.C.,
April.
LALL, S. AND S. WANGWE (1998), “Industrial Policy and Industrialisation in Sub–Saharan Africa”,
Journal of African Economies, Vol. 7, Supplement 1, June.
LAMBERT, S. AND D. SAHN (2000), “Incidence of Public Spending in the Health and Education
Sectors in Tanzania”, OECD Development Centre, mimeo, March.
LEITH, J.C. (1997), “Growth and Structural Transformation in Botswana”, in J.S. SALKIN et al.
(eds.), Aspects of the Botswana Economy: Selected Papers, Lentswe La Lesedi, Gaborone,
and James Currey, Oxford.
LEVIN, A. AND LIN, C. (1993), “Unit Root Tests in Panel Data: New Results”, UC San Diego
Discussion Paper 93–56.
LUCAS, R.E. (1988), “On the Mechanics of Economic Development ”, Journal of Monetary
Economics, 22, 3–42.
LUCAS, R.E. (1993), “Making a Miracle”, Econometrica, Vol. 61, No.2.
229
MANKIW, G., D. ROMER AND D. WEIL (1992), “A Contribution to the Empirics of Economic
Growth”, The Quarterly Journal of Economics, Vol. 107, No. 2, May.
MCCARTHY, C. (1998), “South African Trade and Industrial Policy in a Regional Context”, in
L. PETERSSON (ed.), Post Apartheid Southern Africa, Routledge, London and New York.
MAZUMDAR, D. (1993), “Labor Markets and Adjustment in Open Asian Economies: The Republic
of Korea and Malaysia”, World Bank Economic Review, Vol. 7 No. 3.
MICHAELOWA, K. (2000), “Dépenses d’éducation, qualité de l’éducation et pauvreté : l’exemple
de cinq pays d’Afrique francophone”, Technical Paper No. 157, OECD Development
Centre, April.
MORRISSON C. AND J.–P. AZAM (1999), Conflict and Growth in Africa, Vol. 1 The Sahel, OECD
Development Centre Studies, Paris.
NEHRU, V. AND A. DHARESHWAR (1994), “New Estimates of Total Factor Productivity Growth
for Developing and Industrial Countries”, International Economics Department, World
Bank, Washington, D.C.
NEHRU, V., E. SWANSON AND A. DUBEY (1993), “A New Database on Human Capital Stock:
Sources, Methodology and Results”, Policy Research Working Paper, World Bank,
Washington, D.C.
OECD–DAC (2001), The DAC Guidelines: Strengthening Trade Capacity for Development,
OECD Development Assistance Committee, October.
OECD–DAC (1996), Shaping the 21st Century: The Contribution of Development
Co–operation, OECD Development Assistance Committee, May.
OLSON, M. (2000), Power and Prosperity, Basic Books, New York
OMAN, C. AND G. WIGNARAJA (1991), The Postwar Evolution of Development Thinking, OECD
Development Centre, Paris.
OYEJIDE, A., B. NDULU AND D. GREENAWAY (1999), Regional Integration and Trade Liberalization
in Sub–Saharan Africa, Vol. 4, MacMillan, London.
POIRSON, H. (1998), « Essais en économie de la croissance et du développement », Unpublished
thesis, EHESS, Paris.
PRITCHETT, L. (1997). Where has all the Education Gone?, Policy Research Working Paper
RANIS G., F. STEWART AND A. RAMIREZ (2000), “Economic Growth and Development”, World
Development, Vol. 28 No. 2, February, 197–219.
RODRIK, D. (1998), “Trade Policy and Economic Performance in Sub–Saharan Africa”, NBER
Working Paper No. 6562, May.
RODRIK, D. (1995), “Getting Interventions Right : How South Korea and Taiwan Grew Rich”,
Economic Policy, No. 20, April.
ROMER, P. (1990), “Endogenous Technological Change”, Journal of Political Economy, Vol. 98,
No. 5.
230
SACHS, J. AND A. WARNER (1997), “Sources of Slow Growth in African Economies”, Journal of
African Economies, Vol. 6, No. 3.
SAHEL CLUB (1998), Preparing for the Future – A Vision of West Africa in the Year, OCDE,
Paris.
SALKIN, J.S., D. MPABANGA, D. COWAN, J. SELWE AND M. WRIGHT (1997) (eds.), “Aspects of the
Botswana Economy, Selected Papers”, Published for the Bank of Botswana by Lentswe
La Lesedi Ltd.
SALMON, J.–M. (1995), Marché du travail et développement économique dans les petites
économies insulaires : Théorie et application, L’Harmattan.
SCHIAVO–CAMPO, S., G. DE TOMMASO AND A. MUKHERJEE (1997), “An International Statistical
Survey of Government Employment and Wages”, Policy Research Working Paper
SERVÉN, L. AND A. SOLIMANO (1993), “Debt Crisis, Adjustment Policies and Capital Formation
in Developing Countries : Where Do We Stand?”, World Development, Vol. 21, No. 1.
SYRQUIN, M. (1986), “Productivity Growth and Factor Reallocation”, in H. CHENERY, S. ROBINSON
AND M. SYRQUIN (eds.), Industrialization and Growth: A Comparative Study, Oxford
University Press, Oxford.
TAYLOR, L. (1993), The Rocky Road to Reform. Adjustment, Income Distribution and Growth
in the Developing World, MIT Press, Cambridge, Mass.
TYBOUT, J.R. (1992), “Linking Trade And Productivity: New Research Directions”, World Bank
Economic Review, Vol. 6, No. 2.
UNITED NATIONS DEVELOPMENT PROGRAMME (UNDP) – (1998), Human Development Report
(HDR), UNDP, New York.
WORLD BANK (2001), Can Africa Claim the 21st Century?, World Bank, Washington, D.C.
WORLD BANK (2000), World Development Indicators (WDI), World Bank, Washington, D.C.
WORLD BANK (1999), Assessing Aid, World Bank, Washington, D.C.
WORLD BANK (several years) World Development Report (WDR), World Bank, Washington, D.C.
YOUNG, A. (1995), “The Tyranny of Numbers: Confronting the Statistical Realities of the East
Asian Growth Experience”, Quarterly Journal of Economics, Vol. 110, No. 3, August.
231
OECD PUBLICATIONS, 2, rue André-Pascal, 75775 PARIS CEDEX 16
PRINTED IN FRANCE
(41 2001 08 1 P) ISBN 92-64-19655-2 – No. 52143 2001

(Development Centre Studies.) OECD - Emerging Africa.-Organisation For Economic Co-Operation and Development (2002) PDF

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

(Development Centre Studies.) OECD - Emerging Africa.-Organisation For Economic Co-Operation and Development (2002) PDF

Enviado por

Direitos autorais:

Formatos disponíveis

«

By Jean-Claude Berthélemy and

DEVELOPMENT CENTRE OF THE ORGANISATION

Concluding Summary Policy Options for Emerging Africa .......................................... 207

The international strategy for combating poverty included in the Millennium

Jorge Braga de Macedo

ECONOMIC GROWTH SCENARIOS FOR AFRICA

Renewed Progress in Africa in the 1990s?

The 1990s were marked by improved economic growth in sub–Saharan Africa,

The Improvement in Growth Performance

Benin 1.0 Burkina Faso 1.7 Gambia 0.2 Angola 2.0

a. Included as from 1992.

Note: South Africa is excluded from the third group

The renewal of growth in Africa, if it endures, does not necessarily mean a

Source: Authors' calculations based on World Bank (Africa Database, 1998/99).

Countries with rate of adult HIV infection < 5 per cent

Year 1990 1998 % Increase

Fast growers 53.3 56.6 6.2

Year 1990 1998 % Increase

Fast growers 46.3 44.1 -4.7

Table 1.5. Life Expectancy at Birth

Year 1990 1998 Increase

Fast growers 50.4 51.5 2.2%

Table 1.6. Average Infant Mortality Rate

Year 1990 1998 Decrease

Fast growers 99.0 86.5 12.6%

Year 1990 1998 Decrease

Fast growers 51.2 42.0 18%

Table 1.8. Gross Rate of Primary School Enrolment

Year 1990 1996 Increase

Fast growers 67.0 67.4 0%

Figure 1.9. Relation betweeen Progress in Literacy and School Enrolment

Year 1990 1998 Increase

Factors of Economic Take Off in Africa

Rapid Growth Episodes in Africa since 1960

Country Start End Length Average

Algeria 1962 1985 23 5.2

Most of the recent comparative literature on African growth is based on cross–

It is also reasonable to assume decreasing marginal returns to risk.

which is to say . This implies, from the solution of the maximisation

Since and β is a constant, θ must increase as N increases. In other words, agents

Estimation of the Production Function

A number of authors have already estimated production functions based on panel

Table 2.2. Panel Data Estimates of the Production Function

Variable Coefficient Standard error t-statistic

LKL 0.397 0.032 12.60

Estimation method: within (fixed effects)

Note: LYLA=ln(GDP/labour)-ln(reallocation effect, see above), LKL=ln(capital stock/labour), LH=ln(average number of

Of the 27 countries making up the database that we assembled to perform the

Failed take offs

Average 6.5 3.0 3.5 5.1 1.2 69.2 30.8

Current long growth periods

Average 6.7 2.5 4.2 1.9 3.3 13.3 86.7

Recent growth experiences

Note: Logarithmic growth rates.

Source: Authors’ calculations.

The moderate contribution of capital accumulation in Mauritius can be partly

Contribution Memo item:

Failed take offs

Current growth periods

Recent growth experiences

Source: Authors’ calculations.

Future Growth Scenarios for Six African Countries